diff options
Diffstat (limited to 'src')
871 files changed, 75958 insertions, 27425 deletions
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index e40e9b0a5..dbda528ce 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -32,7 +32,6 @@ if (MSVC) # /Zc:inline - Let codegen omit inline functions in object files # /Zc:throwingNew - Let codegen assume `operator new` (without std::nothrow) will never return null add_compile_options( - /W3 /MP /Zi /Zo @@ -43,6 +42,13 @@ if (MSVC) /Zc:externConstexpr /Zc:inline /Zc:throwingNew + + # Warnings + /W3 + /we4547 # 'operator' : operator before comma has no effect; expected operator with side-effect + /we4549 # 'operator1': operator before comma has no effect; did you intend 'operator2'? + /we4555 # Expression has no effect; expected expression with side-effect + /we4834 # Discarding return value of function with 'nodiscard' attribute ) # /GS- - No stack buffer overflow checks @@ -53,10 +59,26 @@ if (MSVC) else() add_compile_options( -Wall + -Werror=implicit-fallthrough + -Werror=missing-declarations -Werror=reorder + -Werror=unused-result + -Wextra + -Wmissing-declarations -Wno-attributes + -Wno-unused-parameter ) + # TODO: Remove when we update to a GCC compiler that enables this + # by default (i.e. GCC 10 or newer). + if (CMAKE_CXX_COMPILER_ID STREQUAL GNU) + add_compile_options(-fconcepts) + endif() + + if (ARCHITECTURE_x86_64) + add_compile_options("-mcx16") + endif() + if (APPLE AND CMAKE_CXX_COMPILER_ID STREQUAL Clang) add_compile_options("-stdlib=libc++") endif() diff --git a/src/audio_core/CMakeLists.txt b/src/audio_core/CMakeLists.txt index c381dbe1d..68c67507b 100644 --- a/src/audio_core/CMakeLists.txt +++ b/src/audio_core/CMakeLists.txt @@ -7,24 +7,59 @@ add_library(audio_core STATIC audio_out.h audio_renderer.cpp audio_renderer.h + behavior_info.cpp + behavior_info.h buffer.h codec.cpp codec.h + command_generator.cpp + command_generator.h + common.h + effect_context.cpp + effect_context.h + info_updater.cpp + info_updater.h + memory_pool.cpp + memory_pool.h + mix_context.cpp + mix_context.h null_sink.h sink.h + sink_context.cpp + sink_context.h sink_details.cpp sink_details.h sink_stream.h + splitter_context.cpp + splitter_context.h stream.cpp stream.h time_stretch.cpp time_stretch.h + voice_context.cpp + voice_context.h $<$<BOOL:${ENABLE_CUBEB}>:cubeb_sink.cpp cubeb_sink.h> ) create_target_directory_groups(audio_core) +if (NOT MSVC) + target_compile_options(audio_core PRIVATE + -Werror=conversion + -Werror=ignored-qualifiers + -Werror=implicit-fallthrough + -Werror=reorder + -Werror=sign-compare + -Werror=unused-variable + + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + + -Wno-sign-conversion + ) +endif() + target_link_libraries(audio_core PUBLIC common core) target_link_libraries(audio_core PRIVATE SoundTouch) diff --git a/src/audio_core/algorithm/filter.cpp b/src/audio_core/algorithm/filter.cpp index f65bf64f7..f34a5b9f3 100644 --- a/src/audio_core/algorithm/filter.cpp +++ b/src/audio_core/algorithm/filter.cpp @@ -55,7 +55,8 @@ void Filter::Process(std::vector<s16>& signal) { /// @param total_count The total number of biquads to be cascaded. /// @param index 0-index of the biquad to calculate the Q value for. static double CascadingBiquadQ(std::size_t total_count, std::size_t index) { - const double pole = M_PI * (2 * index + 1) / (4.0 * total_count); + const auto pole = + M_PI * static_cast<double>(2 * index + 1) / (4.0 * static_cast<double>(total_count)); return 1.0 / (2.0 * std::cos(pole)); } diff --git a/src/audio_core/algorithm/interpolate.cpp b/src/audio_core/algorithm/interpolate.cpp index 49ab9d3e1..699fcb84c 100644 --- a/src/audio_core/algorithm/interpolate.cpp +++ b/src/audio_core/algorithm/interpolate.cpp @@ -146,7 +146,7 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, return {}; if (ratio <= 0) { - LOG_CRITICAL(Audio, "Nonsensical interpolation ratio {}", ratio); + LOG_ERROR(Audio, "Nonsensical interpolation ratio {}", ratio); return input; } @@ -164,7 +164,8 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, const std::size_t num_frames{input.size() / 2}; std::vector<s16> output; - output.reserve(static_cast<std::size_t>(input.size() / ratio + InterpolationState::taps)); + output.reserve(static_cast<std::size_t>(static_cast<double>(input.size()) / ratio + + InterpolationState::taps)); for (std::size_t frame{}; frame < num_frames; ++frame) { const std::size_t lut_index{(state.fraction >> 8) * InterpolationState::taps}; @@ -197,4 +198,36 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, return output; } +void Resample(s32* output, const s32* input, s32 pitch, s32& fraction, std::size_t sample_count) { + const std::array<s16, 512>& lut = [pitch] { + if (pitch > 0xaaaa) { + return curve_lut0; + } + if (pitch <= 0x8000) { + return curve_lut1; + } + return curve_lut2; + }(); + + std::size_t index{}; + + for (std::size_t i = 0; i < sample_count; i++) { + const std::size_t lut_index{(static_cast<std::size_t>(fraction) >> 8) * 4}; + const auto l0 = lut[lut_index + 0]; + const auto l1 = lut[lut_index + 1]; + const auto l2 = lut[lut_index + 2]; + const auto l3 = lut[lut_index + 3]; + + const auto s0 = static_cast<s32>(input[index]); + const auto s1 = static_cast<s32>(input[index + 1]); + const auto s2 = static_cast<s32>(input[index + 2]); + const auto s3 = static_cast<s32>(input[index + 3]); + + output[i] = (l0 * s0 + l1 * s1 + l2 * s2 + l3 * s3) >> 15; + fraction += pitch; + index += (fraction >> 15); + fraction &= 0x7fff; + } +} + } // namespace AudioCore diff --git a/src/audio_core/algorithm/interpolate.h b/src/audio_core/algorithm/interpolate.h index ab1a31754..d534077af 100644 --- a/src/audio_core/algorithm/interpolate.h +++ b/src/audio_core/algorithm/interpolate.h @@ -38,4 +38,7 @@ inline std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> return Interpolate(state, std::move(input), ratio); } +/// Nintendo Switchs DSP resampling algorithm. Based on a single channel +void Resample(s32* output, const s32* input, s32 pitch, s32& fraction, std::size_t sample_count); + } // namespace AudioCore diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp index c187d8ac5..a7e851bb8 100644 --- a/src/audio_core/audio_renderer.cpp +++ b/src/audio_core/audio_renderer.cpp @@ -2,92 +2,46 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "audio_core/algorithm/interpolate.h" +#include <vector> + #include "audio_core/audio_out.h" #include "audio_core/audio_renderer.h" -#include "audio_core/codec.h" -#include "common/assert.h" +#include "audio_core/common.h" +#include "audio_core/info_updater.h" +#include "audio_core/voice_context.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/hle/kernel/writable_event.h" #include "core/memory.h" +#include "core/settings.h" namespace AudioCore { - -constexpr u32 STREAM_SAMPLE_RATE{48000}; -constexpr u32 STREAM_NUM_CHANNELS{2}; - -class AudioRenderer::VoiceState { -public: - bool IsPlaying() const { - return is_in_use && info.play_state == PlayState::Started; - } - - const VoiceOutStatus& GetOutStatus() const { - return out_status; - } - - const VoiceInfo& GetInfo() const { - return info; - } - - VoiceInfo& GetInfo() { - return info; - } - - void SetWaveIndex(std::size_t index); - std::vector<s16> DequeueSamples(std::size_t sample_count, Memory::Memory& memory); - void UpdateState(); - void RefreshBuffer(Memory::Memory& memory); - -private: - bool is_in_use{}; - bool is_refresh_pending{}; - std::size_t wave_index{}; - std::size_t offset{}; - Codec::ADPCMState adpcm_state{}; - InterpolationState interp_state{}; - std::vector<s16> samples; - VoiceOutStatus out_status{}; - VoiceInfo info{}; -}; - -class AudioRenderer::EffectState { -public: - const EffectOutStatus& GetOutStatus() const { - return out_status; - } - - const EffectInStatus& GetInfo() const { - return info; - } - - EffectInStatus& GetInfo() { - return info; - } - - void UpdateState(Memory::Memory& memory); - -private: - EffectOutStatus out_status{}; - EffectInStatus info{}; -}; -AudioRenderer::AudioRenderer(Core::Timing::CoreTiming& core_timing, Memory::Memory& memory_, - AudioRendererParameter params, +AudioRenderer::AudioRenderer(Core::Timing::CoreTiming& core_timing, Core::Memory::Memory& memory_, + AudioCommon::AudioRendererParameter params, std::shared_ptr<Kernel::WritableEvent> buffer_event, std::size_t instance_number) - : worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count), - effects(params.effect_count), memory{memory_} { - + : worker_params{params}, buffer_event{buffer_event}, + memory_pool_info(params.effect_count + params.voice_count * 4), + voice_context(params.voice_count), effect_context(params.effect_count), mix_context(), + sink_context(params.sink_count), splitter_context(), + voices(params.voice_count), memory{memory_}, + command_generator(worker_params, voice_context, mix_context, splitter_context, effect_context, + memory), + temp_mix_buffer(AudioCommon::TOTAL_TEMP_MIX_SIZE) { + behavior_info.SetUserRevision(params.revision); + splitter_context.Initialize(behavior_info, params.splitter_count, + params.num_splitter_send_channels); + mix_context.Initialize(behavior_info, params.submix_count + 1, params.effect_count); audio_out = std::make_unique<AudioCore::AudioOut>(); - stream = audio_out->OpenStream(core_timing, STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS, - fmt::format("AudioRenderer-Instance{}", instance_number), - [=]() { buffer_event->Signal(); }); + stream = + audio_out->OpenStream(core_timing, params.sample_rate, AudioCommon::STREAM_NUM_CHANNELS, + fmt::format("AudioRenderer-Instance{}", instance_number), + [=]() { buffer_event->Signal(); }); audio_out->StartStream(stream); QueueMixedBuffer(0); QueueMixedBuffer(1); QueueMixedBuffer(2); + QueueMixedBuffer(3); } AudioRenderer::~AudioRenderer() = default; @@ -108,257 +62,200 @@ Stream::State AudioRenderer::GetStreamState() const { return stream->GetState(); } -static constexpr u32 VersionFromRevision(u32_le rev) { - // "REV7" -> 7 - return ((rev >> 24) & 0xff) - 0x30; +static constexpr s16 ClampToS16(s32 value) { + return static_cast<s16>(std::clamp(value, -32768, 32767)); } -std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params) { - // Copy UpdateDataHeader struct - UpdateDataHeader config{}; - std::memcpy(&config, input_params.data(), sizeof(UpdateDataHeader)); - u32 memory_pool_count = worker_params.effect_count + (worker_params.voice_count * 4); - - // Copy MemoryPoolInfo structs - std::vector<MemoryPoolInfo> mem_pool_info(memory_pool_count); - std::memcpy(mem_pool_info.data(), - input_params.data() + sizeof(UpdateDataHeader) + config.behavior_size, - memory_pool_count * sizeof(MemoryPoolInfo)); - - // Copy VoiceInfo structs - std::size_t voice_offset{sizeof(UpdateDataHeader) + config.behavior_size + - config.memory_pools_size + config.voice_resource_size}; - for (auto& voice : voices) { - std::memcpy(&voice.GetInfo(), input_params.data() + voice_offset, sizeof(VoiceInfo)); - voice_offset += sizeof(VoiceInfo); - } +ResultCode AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params, + std::vector<u8>& output_params) { - std::size_t effect_offset{sizeof(UpdateDataHeader) + config.behavior_size + - config.memory_pools_size + config.voice_resource_size + - config.voices_size}; - for (auto& effect : effects) { - std::memcpy(&effect.GetInfo(), input_params.data() + effect_offset, sizeof(EffectInStatus)); - effect_offset += sizeof(EffectInStatus); - } + InfoUpdater info_updater{input_params, output_params, behavior_info}; - // Update memory pool state - std::vector<MemoryPoolEntry> memory_pool(memory_pool_count); - for (std::size_t index = 0; index < memory_pool.size(); ++index) { - if (mem_pool_info[index].pool_state == MemoryPoolStates::RequestAttach) { - memory_pool[index].state = MemoryPoolStates::Attached; - } else if (mem_pool_info[index].pool_state == MemoryPoolStates::RequestDetach) { - memory_pool[index].state = MemoryPoolStates::Detached; - } + if (!info_updater.UpdateBehaviorInfo(behavior_info)) { + LOG_ERROR(Audio, "Failed to update behavior info input parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - // Update voices - for (auto& voice : voices) { - voice.UpdateState(); - if (!voice.GetInfo().is_in_use) { - continue; - } - if (voice.GetInfo().is_new) { - voice.SetWaveIndex(voice.GetInfo().wave_buffer_head); - } + if (!info_updater.UpdateMemoryPools(memory_pool_info)) { + LOG_ERROR(Audio, "Failed to update memory pool parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - for (auto& effect : effects) { - effect.UpdateState(memory); + if (!info_updater.UpdateVoiceChannelResources(voice_context)) { + LOG_ERROR(Audio, "Failed to update voice channel resource parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - // Release previous buffers and queue next ones for playback - ReleaseAndQueueBuffers(); - - // Copy output header - UpdateDataHeader response_data{worker_params}; - std::vector<u8> output_params(response_data.total_size); - const auto audren_revision = VersionFromRevision(config.revision); - if (audren_revision >= 5) { - response_data.frame_count = 0x10; - response_data.total_size += 0x10; - } - std::memcpy(output_params.data(), &response_data, sizeof(UpdateDataHeader)); - - // Copy output memory pool entries - std::memcpy(output_params.data() + sizeof(UpdateDataHeader), memory_pool.data(), - response_data.memory_pools_size); - - // Copy output voice status - std::size_t voice_out_status_offset{sizeof(UpdateDataHeader) + response_data.memory_pools_size}; - for (const auto& voice : voices) { - std::memcpy(output_params.data() + voice_out_status_offset, &voice.GetOutStatus(), - sizeof(VoiceOutStatus)); - voice_out_status_offset += sizeof(VoiceOutStatus); + if (!info_updater.UpdateVoices(voice_context, memory_pool_info, 0)) { + LOG_ERROR(Audio, "Failed to update voice parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - std::size_t effect_out_status_offset{ - sizeof(UpdateDataHeader) + response_data.memory_pools_size + response_data.voices_size + - response_data.voice_resource_size}; - for (const auto& effect : effects) { - std::memcpy(output_params.data() + effect_out_status_offset, &effect.GetOutStatus(), - sizeof(EffectOutStatus)); - effect_out_status_offset += sizeof(EffectOutStatus); + // TODO(ogniK): Deal with stopped audio renderer but updates still taking place + if (!info_updater.UpdateEffects(effect_context, true)) { + LOG_ERROR(Audio, "Failed to update effect parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - return output_params; -} - -void AudioRenderer::VoiceState::SetWaveIndex(std::size_t index) { - wave_index = index & 3; - is_refresh_pending = true; -} -std::vector<s16> AudioRenderer::VoiceState::DequeueSamples(std::size_t sample_count, - Memory::Memory& memory) { - if (!IsPlaying()) { - return {}; + if (behavior_info.IsSplitterSupported()) { + if (!info_updater.UpdateSplitterInfo(splitter_context)) { + LOG_ERROR(Audio, "Failed to update splitter parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } } - if (is_refresh_pending) { - RefreshBuffer(memory); - } + auto mix_result = info_updater.UpdateMixes(mix_context, worker_params.mix_buffer_count, + splitter_context, effect_context); - const std::size_t max_size{samples.size() - offset}; - const std::size_t dequeue_offset{offset}; - std::size_t size{sample_count * STREAM_NUM_CHANNELS}; - if (size > max_size) { - size = max_size; + if (mix_result.IsError()) { + LOG_ERROR(Audio, "Failed to update mix parameters"); + return mix_result; } - out_status.played_sample_count += size / STREAM_NUM_CHANNELS; - offset += size; - - const auto& wave_buffer{info.wave_buffer[wave_index]}; - if (offset == samples.size()) { - offset = 0; - - if (!wave_buffer.is_looping && wave_buffer.buffer_sz) { - SetWaveIndex(wave_index + 1); - } - - if (wave_buffer.buffer_sz) { - out_status.wave_buffer_consumed++; - } - - if (wave_buffer.end_of_stream || wave_buffer.buffer_sz == 0) { - info.play_state = PlayState::Paused; - } + // TODO(ogniK): Sinks + if (!info_updater.UpdateSinks(sink_context)) { + LOG_ERROR(Audio, "Failed to update sink parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - return {samples.begin() + dequeue_offset, samples.begin() + dequeue_offset + size}; -} - -void AudioRenderer::VoiceState::UpdateState() { - if (is_in_use && !info.is_in_use) { - // No longer in use, reset state - is_refresh_pending = true; - wave_index = 0; - offset = 0; - out_status = {}; + // TODO(ogniK): Performance buffer + if (!info_updater.UpdatePerformanceBuffer()) { + LOG_ERROR(Audio, "Failed to update performance buffer parameters"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - is_in_use = info.is_in_use; -} -void AudioRenderer::VoiceState::RefreshBuffer(Memory::Memory& memory) { - const auto wave_buffer_address = info.wave_buffer[wave_index].buffer_addr; - const auto wave_buffer_size = info.wave_buffer[wave_index].buffer_sz; - std::vector<s16> new_samples(wave_buffer_size / sizeof(s16)); - memory.ReadBlock(wave_buffer_address, new_samples.data(), wave_buffer_size); - - switch (static_cast<Codec::PcmFormat>(info.sample_format)) { - case Codec::PcmFormat::Int16: { - // PCM16 is played as-is - break; - } - case Codec::PcmFormat::Adpcm: { - // Decode ADPCM to PCM16 - Codec::ADPCM_Coeff coeffs; - memory.ReadBlock(info.additional_params_addr, coeffs.data(), sizeof(Codec::ADPCM_Coeff)); - new_samples = Codec::DecodeADPCM(reinterpret_cast<u8*>(new_samples.data()), - new_samples.size() * sizeof(s16), coeffs, adpcm_state); - break; - } - default: - UNIMPLEMENTED_MSG("Unimplemented sample_format={}", info.sample_format); - break; + if (!info_updater.UpdateErrorInfo(behavior_info)) { + LOG_ERROR(Audio, "Failed to update error info"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - switch (info.channel_count) { - case 1: - // 1 channel is upsampled to 2 channel - samples.resize(new_samples.size() * 2); - for (std::size_t index = 0; index < new_samples.size(); ++index) { - samples[index * 2] = new_samples[index]; - samples[index * 2 + 1] = new_samples[index]; + if (behavior_info.IsElapsedFrameCountSupported()) { + if (!info_updater.UpdateRendererInfo(elapsed_frame_count)) { + LOG_ERROR(Audio, "Failed to update renderer info"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - break; - case 2: { - // 2 channel is played as is - samples = std::move(new_samples); - break; - } - default: - UNIMPLEMENTED_MSG("Unimplemented channel_count={}", info.channel_count); - break; } + // TODO(ogniK): Statistics - // Only interpolate when necessary, expensive. - if (GetInfo().sample_rate != STREAM_SAMPLE_RATE) { - samples = Interpolate(interp_state, std::move(samples), GetInfo().sample_rate, - STREAM_SAMPLE_RATE); + if (!info_updater.WriteOutputHeader()) { + LOG_ERROR(Audio, "Failed to write output header"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } - is_refresh_pending = false; -} + // TODO(ogniK): Check when all sections are implemented -void AudioRenderer::EffectState::UpdateState(Memory::Memory& memory) { - if (info.is_new) { - out_status.state = EffectStatus::New; - } else { - if (info.type == Effect::Aux) { - ASSERT_MSG(memory.Read32(info.aux_info.return_buffer_info) == 0, - "Aux buffers tried to update"); - ASSERT_MSG(memory.Read32(info.aux_info.send_buffer_info) == 0, - "Aux buffers tried to update"); - ASSERT_MSG(memory.Read32(info.aux_info.return_buffer_base) == 0, - "Aux buffers tried to update"); - ASSERT_MSG(memory.Read32(info.aux_info.send_buffer_base) == 0, - "Aux buffers tried to update"); - } + if (!info_updater.CheckConsumedSize()) { + LOG_ERROR(Audio, "Audio buffers were not consumed!"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; } -} -static constexpr s16 ClampToS16(s32 value) { - return static_cast<s16>(std::clamp(value, -32768, 32767)); + ReleaseAndQueueBuffers(); + + return RESULT_SUCCESS; } void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) { - constexpr std::size_t BUFFER_SIZE{512}; - std::vector<s16> buffer(BUFFER_SIZE * stream->GetNumChannels()); + command_generator.PreCommand(); + // Clear mix buffers before our next operation + command_generator.ClearMixBuffers(); - for (auto& voice : voices) { - if (!voice.IsPlaying()) { - continue; + // If the splitter is not in use, sort our mixes + if (!splitter_context.UsingSplitter()) { + mix_context.SortInfo(); + } + // Sort our voices + voice_context.SortInfo(); + + // Handle samples + command_generator.GenerateVoiceCommands(); + command_generator.GenerateSubMixCommands(); + command_generator.GenerateFinalMixCommands(); + + command_generator.PostCommand(); + // Base sample size + std::size_t BUFFER_SIZE{worker_params.sample_count}; + // Samples + std::vector<s16> buffer(BUFFER_SIZE * stream->GetNumChannels()); + // Make sure to clear our samples + std::memset(buffer.data(), 0, buffer.size() * sizeof(s16)); + + if (sink_context.InUse()) { + const auto stream_channel_count = stream->GetNumChannels(); + const auto buffer_offsets = sink_context.OutputBuffers(); + const auto channel_count = buffer_offsets.size(); + const auto& final_mix = mix_context.GetFinalMixInfo(); + const auto& in_params = final_mix.GetInParams(); + std::vector<s32*> mix_buffers(channel_count); + for (std::size_t i = 0; i < channel_count; i++) { + mix_buffers[i] = + command_generator.GetMixBuffer(in_params.buffer_offset + buffer_offsets[i]); } - std::size_t offset{}; - s64 samples_remaining{BUFFER_SIZE}; - while (samples_remaining > 0) { - const std::vector<s16> samples{voice.DequeueSamples(samples_remaining, memory)}; - - if (samples.empty()) { - break; - } - - samples_remaining -= samples.size() / stream->GetNumChannels(); - - for (const auto& sample : samples) { - const s32 buffer_sample{buffer[offset]}; - buffer[offset++] = - ClampToS16(buffer_sample + static_cast<s32>(sample * voice.GetInfo().volume)); + for (std::size_t i = 0; i < BUFFER_SIZE; i++) { + if (channel_count == 1) { + const auto sample = ClampToS16(mix_buffers[0][i]); + buffer[i * stream_channel_count + 0] = sample; + if (stream_channel_count > 1) { + buffer[i * stream_channel_count + 1] = sample; + } + if (stream_channel_count == 6) { + buffer[i * stream_channel_count + 2] = sample; + buffer[i * stream_channel_count + 4] = sample; + buffer[i * stream_channel_count + 5] = sample; + } + } else if (channel_count == 2) { + const auto l_sample = ClampToS16(mix_buffers[0][i]); + const auto r_sample = ClampToS16(mix_buffers[1][i]); + if (stream_channel_count == 1) { + buffer[i * stream_channel_count + 0] = l_sample; + } else if (stream_channel_count == 2) { + buffer[i * stream_channel_count + 0] = l_sample; + buffer[i * stream_channel_count + 1] = r_sample; + } else if (stream_channel_count == 6) { + buffer[i * stream_channel_count + 0] = l_sample; + buffer[i * stream_channel_count + 1] = r_sample; + + buffer[i * stream_channel_count + 2] = + ClampToS16((static_cast<s32>(l_sample) + static_cast<s32>(r_sample)) / 2); + + buffer[i * stream_channel_count + 4] = l_sample; + buffer[i * stream_channel_count + 5] = r_sample; + } + + } else if (channel_count == 6) { + const auto fl_sample = ClampToS16(mix_buffers[0][i]); + const auto fr_sample = ClampToS16(mix_buffers[1][i]); + const auto fc_sample = ClampToS16(mix_buffers[2][i]); + const auto lf_sample = ClampToS16(mix_buffers[3][i]); + const auto bl_sample = ClampToS16(mix_buffers[4][i]); + const auto br_sample = ClampToS16(mix_buffers[5][i]); + + if (stream_channel_count == 1) { + buffer[i * stream_channel_count + 0] = fc_sample; + } else if (stream_channel_count == 2) { + buffer[i * stream_channel_count + 0] = + static_cast<s16>(0.3694f * static_cast<float>(fl_sample) + + 0.2612f * static_cast<float>(fc_sample) + + 0.3694f * static_cast<float>(bl_sample)); + buffer[i * stream_channel_count + 1] = + static_cast<s16>(0.3694f * static_cast<float>(fr_sample) + + 0.2612f * static_cast<float>(fc_sample) + + 0.3694f * static_cast<float>(br_sample)); + } else if (stream_channel_count == 6) { + buffer[i * stream_channel_count + 0] = fl_sample; + buffer[i * stream_channel_count + 1] = fr_sample; + buffer[i * stream_channel_count + 2] = fc_sample; + buffer[i * stream_channel_count + 3] = lf_sample; + buffer[i * stream_channel_count + 4] = bl_sample; + buffer[i * stream_channel_count + 5] = br_sample; + } } } } + audio_out->QueueBuffer(stream, tag, std::move(buffer)); + elapsed_frame_count++; + voice_context.UpdateStateByDspShared(); } void AudioRenderer::ReleaseAndQueueBuffers() { diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h index c0fae669e..2fd93e058 100644 --- a/src/audio_core/audio_renderer.h +++ b/src/audio_core/audio_renderer.h @@ -8,11 +8,20 @@ #include <memory> #include <vector> +#include "audio_core/behavior_info.h" +#include "audio_core/command_generator.h" +#include "audio_core/common.h" +#include "audio_core/effect_context.h" +#include "audio_core/memory_pool.h" +#include "audio_core/mix_context.h" +#include "audio_core/sink_context.h" +#include "audio_core/splitter_context.h" #include "audio_core/stream.h" +#include "audio_core/voice_context.h" #include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" -#include "core/hle/kernel/object.h" +#include "core/hle/result.h" namespace Core::Timing { class CoreTiming; @@ -22,211 +31,30 @@ namespace Kernel { class WritableEvent; } -namespace Memory { +namespace Core::Memory { class Memory; } namespace AudioCore { +using DSPStateHolder = std::array<VoiceState*, 6>; class AudioOut; -enum class PlayState : u8 { - Started = 0, - Stopped = 1, - Paused = 2, +struct RendererInfo { + u64_le elasped_frame_count{}; + INSERT_PADDING_WORDS(2); }; - -enum class Effect : u8 { - None = 0, - Aux = 2, -}; - -enum class EffectStatus : u8 { - None = 0, - New = 1, -}; - -struct AudioRendererParameter { - u32_le sample_rate; - u32_le sample_count; - u32_le mix_buffer_count; - u32_le submix_count; - u32_le voice_count; - u32_le sink_count; - u32_le effect_count; - u32_le performance_frame_count; - u8 is_voice_drop_enabled; - u8 unknown_21; - u8 unknown_22; - u8 execution_mode; - u32_le splitter_count; - u32_le num_splitter_send_channels; - u32_le unknown_30; - u32_le revision; -}; -static_assert(sizeof(AudioRendererParameter) == 52, "AudioRendererParameter is an invalid size"); - -enum class MemoryPoolStates : u32 { // Should be LE - Invalid = 0x0, - Unknown = 0x1, - RequestDetach = 0x2, - Detached = 0x3, - RequestAttach = 0x4, - Attached = 0x5, - Released = 0x6, -}; - -struct MemoryPoolEntry { - MemoryPoolStates state; - u32_le unknown_4; - u32_le unknown_8; - u32_le unknown_c; -}; -static_assert(sizeof(MemoryPoolEntry) == 0x10, "MemoryPoolEntry has wrong size"); - -struct MemoryPoolInfo { - u64_le pool_address; - u64_le pool_size; - MemoryPoolStates pool_state; - INSERT_PADDING_WORDS(3); // Unknown -}; -static_assert(sizeof(MemoryPoolInfo) == 0x20, "MemoryPoolInfo has wrong size"); -struct BiquadFilter { - u8 enable; - INSERT_PADDING_BYTES(1); - std::array<s16_le, 3> numerator; - std::array<s16_le, 2> denominator; -}; -static_assert(sizeof(BiquadFilter) == 0xc, "BiquadFilter has wrong size"); - -struct WaveBuffer { - u64_le buffer_addr; - u64_le buffer_sz; - s32_le start_sample_offset; - s32_le end_sample_offset; - u8 is_looping; - u8 end_of_stream; - u8 sent_to_server; - INSERT_PADDING_BYTES(5); - u64 context_addr; - u64 context_sz; - INSERT_PADDING_BYTES(8); -}; -static_assert(sizeof(WaveBuffer) == 0x38, "WaveBuffer has wrong size"); - -struct VoiceInfo { - u32_le id; - u32_le node_id; - u8 is_new; - u8 is_in_use; - PlayState play_state; - u8 sample_format; - u32_le sample_rate; - u32_le priority; - u32_le sorting_order; - u32_le channel_count; - float_le pitch; - float_le volume; - std::array<BiquadFilter, 2> biquad_filter; - u32_le wave_buffer_count; - u32_le wave_buffer_head; - INSERT_PADDING_WORDS(1); - u64_le additional_params_addr; - u64_le additional_params_sz; - u32_le mix_id; - u32_le splitter_info_id; - std::array<WaveBuffer, 4> wave_buffer; - std::array<u32_le, 6> voice_channel_resource_ids; - INSERT_PADDING_BYTES(24); -}; -static_assert(sizeof(VoiceInfo) == 0x170, "VoiceInfo is wrong size"); - -struct VoiceOutStatus { - u64_le played_sample_count; - u32_le wave_buffer_consumed; - u32_le voice_drops_count; -}; -static_assert(sizeof(VoiceOutStatus) == 0x10, "VoiceOutStatus has wrong size"); - -struct AuxInfo { - std::array<u8, 24> input_mix_buffers; - std::array<u8, 24> output_mix_buffers; - u32_le mix_buffer_count; - u32_le sample_rate; // Stored in the aux buffer currently - u32_le sample_count; - u64_le send_buffer_info; - u64_le send_buffer_base; - - u64_le return_buffer_info; - u64_le return_buffer_base; -}; -static_assert(sizeof(AuxInfo) == 0x60, "AuxInfo is an invalid size"); - -struct EffectInStatus { - Effect type; - u8 is_new; - u8 is_enabled; - INSERT_PADDING_BYTES(1); - u32_le mix_id; - u64_le buffer_base; - u64_le buffer_sz; - s32_le priority; - INSERT_PADDING_BYTES(4); - union { - std::array<u8, 0xa0> raw; - AuxInfo aux_info; - }; -}; -static_assert(sizeof(EffectInStatus) == 0xc0, "EffectInStatus is an invalid size"); - -struct EffectOutStatus { - EffectStatus state; - INSERT_PADDING_BYTES(0xf); -}; -static_assert(sizeof(EffectOutStatus) == 0x10, "EffectOutStatus is an invalid size"); - -struct UpdateDataHeader { - UpdateDataHeader() {} - - explicit UpdateDataHeader(const AudioRendererParameter& config) { - revision = Common::MakeMagic('R', 'E', 'V', '8'); // 9.2.0 Revision - behavior_size = 0xb0; - memory_pools_size = (config.effect_count + (config.voice_count * 4)) * 0x10; - voices_size = config.voice_count * 0x10; - voice_resource_size = 0x0; - effects_size = config.effect_count * 0x10; - mixes_size = 0x0; - sinks_size = config.sink_count * 0x20; - performance_manager_size = 0x10; - frame_count = 0; - total_size = sizeof(UpdateDataHeader) + behavior_size + memory_pools_size + voices_size + - effects_size + sinks_size + performance_manager_size; - } - - u32_le revision{}; - u32_le behavior_size{}; - u32_le memory_pools_size{}; - u32_le voices_size{}; - u32_le voice_resource_size{}; - u32_le effects_size{}; - u32_le mixes_size{}; - u32_le sinks_size{}; - u32_le performance_manager_size{}; - INSERT_PADDING_WORDS(1); - u32_le frame_count{}; - INSERT_PADDING_WORDS(4); - u32_le total_size{}; -}; -static_assert(sizeof(UpdateDataHeader) == 0x40, "UpdateDataHeader has wrong size"); +static_assert(sizeof(RendererInfo) == 0x10, "RendererInfo is an invalid size"); class AudioRenderer { public: - AudioRenderer(Core::Timing::CoreTiming& core_timing, Memory::Memory& memory_, - AudioRendererParameter params, + AudioRenderer(Core::Timing::CoreTiming& core_timing, Core::Memory::Memory& memory_, + AudioCommon::AudioRendererParameter params, std::shared_ptr<Kernel::WritableEvent> buffer_event, std::size_t instance_number); ~AudioRenderer(); - std::vector<u8> UpdateAudioRenderer(const std::vector<u8>& input_params); + ResultCode UpdateAudioRenderer(const std::vector<u8>& input_params, + std::vector<u8>& output_params); void QueueMixedBuffer(Buffer::Tag tag); void ReleaseAndQueueBuffers(); u32 GetSampleRate() const; @@ -235,16 +63,23 @@ public: Stream::State GetStreamState() const; private: - class EffectState; - class VoiceState; + BehaviorInfo behavior_info{}; - AudioRendererParameter worker_params; + AudioCommon::AudioRendererParameter worker_params; std::shared_ptr<Kernel::WritableEvent> buffer_event; + std::vector<ServerMemoryPoolInfo> memory_pool_info; + VoiceContext voice_context; + EffectContext effect_context; + MixContext mix_context; + SinkContext sink_context; + SplitterContext splitter_context; std::vector<VoiceState> voices; - std::vector<EffectState> effects; std::unique_ptr<AudioOut> audio_out; StreamPtr stream; - Memory::Memory& memory; + Core::Memory::Memory& memory; + CommandGenerator command_generator; + std::size_t elapsed_frame_count{}; + std::vector<s32> temp_mix_buffer{}; }; } // namespace AudioCore diff --git a/src/audio_core/behavior_info.cpp b/src/audio_core/behavior_info.cpp new file mode 100644 index 000000000..3c2e3e6f1 --- /dev/null +++ b/src/audio_core/behavior_info.cpp @@ -0,0 +1,105 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> +#include "audio_core/behavior_info.h" +#include "audio_core/common.h" +#include "common/logging/log.h" + +namespace AudioCore { + +BehaviorInfo::BehaviorInfo() : process_revision(AudioCommon::CURRENT_PROCESS_REVISION) {} +BehaviorInfo::~BehaviorInfo() = default; + +bool BehaviorInfo::UpdateOutput(std::vector<u8>& buffer, std::size_t offset) { + if (!AudioCommon::CanConsumeBuffer(buffer.size(), offset, sizeof(OutParams))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + OutParams params{}; + std::memcpy(params.errors.data(), errors.data(), sizeof(ErrorInfo) * errors.size()); + params.error_count = static_cast<u32_le>(error_count); + std::memcpy(buffer.data() + offset, ¶ms, sizeof(OutParams)); + return true; +} + +void BehaviorInfo::ClearError() { + error_count = 0; +} + +void BehaviorInfo::UpdateFlags(u64_le dest_flags) { + flags = dest_flags; +} + +void BehaviorInfo::SetUserRevision(u32_le revision) { + user_revision = revision; +} + +u32_le BehaviorInfo::GetUserRevision() const { + return user_revision; +} + +u32_le BehaviorInfo::GetProcessRevision() const { + return process_revision; +} + +bool BehaviorInfo::IsAdpcmLoopContextBugFixed() const { + return AudioCommon::IsRevisionSupported(2, user_revision); +} + +bool BehaviorInfo::IsSplitterSupported() const { + return AudioCommon::IsRevisionSupported(2, user_revision); +} + +bool BehaviorInfo::IsLongSizePreDelaySupported() const { + return AudioCommon::IsRevisionSupported(3, user_revision); +} + +bool BehaviorInfo::IsAudioRendererProcessingTimeLimit80PercentSupported() const { + return AudioCommon::IsRevisionSupported(5, user_revision); +} + +bool BehaviorInfo::IsAudioRendererProcessingTimeLimit75PercentSupported() const { + return AudioCommon::IsRevisionSupported(4, user_revision); +} + +bool BehaviorInfo::IsAudioRendererProcessingTimeLimit70PercentSupported() const { + return AudioCommon::IsRevisionSupported(1, user_revision); +} + +bool BehaviorInfo::IsElapsedFrameCountSupported() const { + return AudioCommon::IsRevisionSupported(5, user_revision); +} + +bool BehaviorInfo::IsMemoryPoolForceMappingEnabled() const { + return (flags & 1) != 0; +} + +bool BehaviorInfo::IsFlushVoiceWaveBuffersSupported() const { + return AudioCommon::IsRevisionSupported(5, user_revision); +} + +bool BehaviorInfo::IsVoicePlayedSampleCountResetAtLoopPointSupported() const { + return AudioCommon::IsRevisionSupported(5, user_revision); +} + +bool BehaviorInfo::IsVoicePitchAndSrcSkippedSupported() const { + return AudioCommon::IsRevisionSupported(5, user_revision); +} + +bool BehaviorInfo::IsMixInParameterDirtyOnlyUpdateSupported() const { + return AudioCommon::IsRevisionSupported(7, user_revision); +} + +bool BehaviorInfo::IsSplitterBugFixed() const { + return AudioCommon::IsRevisionSupported(5, user_revision); +} + +void BehaviorInfo::CopyErrorInfo(BehaviorInfo::OutParams& dst) { + dst.error_count = static_cast<u32>(error_count); + std::copy(errors.begin(), errors.begin() + error_count, dst.errors.begin()); +} + +} // namespace AudioCore diff --git a/src/audio_core/behavior_info.h b/src/audio_core/behavior_info.h new file mode 100644 index 000000000..512a4ebe3 --- /dev/null +++ b/src/audio_core/behavior_info.h @@ -0,0 +1,72 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> + +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" + +namespace AudioCore { +class BehaviorInfo { +public: + struct ErrorInfo { + u32_le result{}; + INSERT_PADDING_WORDS(1); + u64_le result_info{}; + }; + static_assert(sizeof(ErrorInfo) == 0x10, "ErrorInfo is an invalid size"); + + struct InParams { + u32_le revision{}; + u32_le padding{}; + u64_le flags{}; + }; + static_assert(sizeof(InParams) == 0x10, "InParams is an invalid size"); + + struct OutParams { + std::array<ErrorInfo, 10> errors{}; + u32_le error_count{}; + INSERT_PADDING_BYTES(12); + }; + static_assert(sizeof(OutParams) == 0xb0, "OutParams is an invalid size"); + + explicit BehaviorInfo(); + ~BehaviorInfo(); + + bool UpdateOutput(std::vector<u8>& buffer, std::size_t offset); + + void ClearError(); + void UpdateFlags(u64_le dest_flags); + void SetUserRevision(u32_le revision); + u32_le GetUserRevision() const; + u32_le GetProcessRevision() const; + + bool IsAdpcmLoopContextBugFixed() const; + bool IsSplitterSupported() const; + bool IsLongSizePreDelaySupported() const; + bool IsAudioRendererProcessingTimeLimit80PercentSupported() const; + bool IsAudioRendererProcessingTimeLimit75PercentSupported() const; + bool IsAudioRendererProcessingTimeLimit70PercentSupported() const; + bool IsElapsedFrameCountSupported() const; + bool IsMemoryPoolForceMappingEnabled() const; + bool IsFlushVoiceWaveBuffersSupported() const; + bool IsVoicePlayedSampleCountResetAtLoopPointSupported() const; + bool IsVoicePitchAndSrcSkippedSupported() const; + bool IsMixInParameterDirtyOnlyUpdateSupported() const; + bool IsSplitterBugFixed() const; + void CopyErrorInfo(OutParams& dst); + +private: + u32_le process_revision{}; + u32_le user_revision{}; + u64_le flags{}; + std::array<ErrorInfo, 10> errors{}; + std::size_t error_count{}; +}; + +} // namespace AudioCore diff --git a/src/audio_core/codec.cpp b/src/audio_core/codec.cpp index c5a0d98ce..2fb91c13a 100644 --- a/src/audio_core/codec.cpp +++ b/src/audio_core/codec.cpp @@ -16,8 +16,9 @@ std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM constexpr std::size_t FRAME_LEN = 8; constexpr std::size_t SAMPLES_PER_FRAME = 14; - constexpr std::array<int, 16> SIGNED_NIBBLES = { - {0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1}}; + static constexpr std::array<int, 16> SIGNED_NIBBLES{ + 0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1, + }; const std::size_t sample_count = (size / FRAME_LEN) * SAMPLES_PER_FRAME; const std::size_t ret_size = diff --git a/src/audio_core/codec.h b/src/audio_core/codec.h index ef2ce01a8..9507abb1b 100644 --- a/src/audio_core/codec.h +++ b/src/audio_core/codec.h @@ -38,7 +38,7 @@ using ADPCM_Coeff = std::array<s16, 16>; * @param state ADPCM state, this is updated with new state * @return Decoded stereo signed PCM16 data, sample_count in length */ -std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM_Coeff& coeff, +std::vector<s16> DecodeADPCM(const u8* data, std::size_t size, const ADPCM_Coeff& coeff, ADPCMState& state); }; // namespace AudioCore::Codec diff --git a/src/audio_core/command_generator.cpp b/src/audio_core/command_generator.cpp new file mode 100644 index 000000000..fb8700ccf --- /dev/null +++ b/src/audio_core/command_generator.cpp @@ -0,0 +1,977 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/algorithm/interpolate.h" +#include "audio_core/command_generator.h" +#include "audio_core/effect_context.h" +#include "audio_core/mix_context.h" +#include "audio_core/voice_context.h" +#include "core/memory.h" + +namespace AudioCore { +namespace { +constexpr std::size_t MIX_BUFFER_SIZE = 0x3f00; +constexpr std::size_t SCALED_MIX_BUFFER_SIZE = MIX_BUFFER_SIZE << 15ULL; + +template <std::size_t N> +void ApplyMix(s32* output, const s32* input, s32 gain, s32 sample_count) { + for (std::size_t i = 0; i < static_cast<std::size_t>(sample_count); i += N) { + for (std::size_t j = 0; j < N; j++) { + output[i + j] += + static_cast<s32>((static_cast<s64>(input[i + j]) * gain + 0x4000) >> 15); + } + } +} + +s32 ApplyMixRamp(s32* output, const s32* input, float gain, float delta, s32 sample_count) { + s32 x = 0; + for (s32 i = 0; i < sample_count; i++) { + x = static_cast<s32>(static_cast<float>(input[i]) * gain); + output[i] += x; + gain += delta; + } + return x; +} + +void ApplyGain(s32* output, const s32* input, s32 gain, s32 delta, s32 sample_count) { + for (s32 i = 0; i < sample_count; i++) { + output[i] = static_cast<s32>((static_cast<s64>(input[i]) * gain + 0x4000) >> 15); + gain += delta; + } +} + +void ApplyGainWithoutDelta(s32* output, const s32* input, s32 gain, s32 sample_count) { + for (s32 i = 0; i < sample_count; i++) { + output[i] = static_cast<s32>((static_cast<s64>(input[i]) * gain + 0x4000) >> 15); + } +} + +s32 ApplyMixDepop(s32* output, s32 first_sample, s32 delta, s32 sample_count) { + const bool positive = first_sample > 0; + auto final_sample = std::abs(first_sample); + for (s32 i = 0; i < sample_count; i++) { + final_sample = static_cast<s32>((static_cast<s64>(final_sample) * delta) >> 15); + if (positive) { + output[i] += final_sample; + } else { + output[i] -= final_sample; + } + } + if (positive) { + return final_sample; + } else { + return -final_sample; + } +} + +} // namespace + +CommandGenerator::CommandGenerator(AudioCommon::AudioRendererParameter& worker_params, + VoiceContext& voice_context, MixContext& mix_context, + SplitterContext& splitter_context, EffectContext& effect_context, + Core::Memory::Memory& memory) + : worker_params(worker_params), voice_context(voice_context), mix_context(mix_context), + splitter_context(splitter_context), effect_context(effect_context), memory(memory), + mix_buffer((worker_params.mix_buffer_count + AudioCommon::MAX_CHANNEL_COUNT) * + worker_params.sample_count), + sample_buffer(MIX_BUFFER_SIZE), + depop_buffer((worker_params.mix_buffer_count + AudioCommon::MAX_CHANNEL_COUNT) * + worker_params.sample_count) {} +CommandGenerator::~CommandGenerator() = default; + +void CommandGenerator::ClearMixBuffers() { + std::fill(mix_buffer.begin(), mix_buffer.end(), 0); + std::fill(sample_buffer.begin(), sample_buffer.end(), 0); + // std::fill(depop_buffer.begin(), depop_buffer.end(), 0); +} + +void CommandGenerator::GenerateVoiceCommands() { + if (dumping_frame) { + LOG_DEBUG(Audio, "(DSP_TRACE) GenerateVoiceCommands"); + } + // Grab all our voices + const auto voice_count = voice_context.GetVoiceCount(); + for (std::size_t i = 0; i < voice_count; i++) { + auto& voice_info = voice_context.GetSortedInfo(i); + // Update voices and check if we should queue them + if (voice_info.ShouldSkip() || !voice_info.UpdateForCommandGeneration(voice_context)) { + continue; + } + + // Queue our voice + GenerateVoiceCommand(voice_info); + } + // Update our splitters + splitter_context.UpdateInternalState(); +} + +void CommandGenerator::GenerateVoiceCommand(ServerVoiceInfo& voice_info) { + auto& in_params = voice_info.GetInParams(); + const auto channel_count = in_params.channel_count; + + for (s32 channel = 0; channel < channel_count; channel++) { + const auto resource_id = in_params.voice_channel_resource_id[channel]; + auto& dsp_state = voice_context.GetDspSharedState(resource_id); + auto& channel_resource = voice_context.GetChannelResource(resource_id); + + // Decode our samples for our channel + GenerateDataSourceCommand(voice_info, dsp_state, channel); + + if (in_params.should_depop) { + in_params.last_volume = 0.0f; + } else if (in_params.splitter_info_id != AudioCommon::NO_SPLITTER || + in_params.mix_id != AudioCommon::NO_MIX) { + // Apply a biquad filter if needed + GenerateBiquadFilterCommandForVoice(voice_info, dsp_state, + worker_params.mix_buffer_count, channel); + // Base voice volume ramping + GenerateVolumeRampCommand(in_params.last_volume, in_params.volume, channel, + in_params.node_id); + in_params.last_volume = in_params.volume; + + if (in_params.mix_id != AudioCommon::NO_MIX) { + // If we're using a mix id + auto& mix_info = mix_context.GetInfo(in_params.mix_id); + const auto& dest_mix_params = mix_info.GetInParams(); + + // Voice Mixing + GenerateVoiceMixCommand( + channel_resource.GetCurrentMixVolume(), channel_resource.GetLastMixVolume(), + dsp_state, dest_mix_params.buffer_offset, dest_mix_params.buffer_count, + worker_params.mix_buffer_count + channel, in_params.node_id); + + // Update last mix volumes + channel_resource.UpdateLastMixVolumes(); + } else if (in_params.splitter_info_id != AudioCommon::NO_SPLITTER) { + s32 base = channel; + while (auto* destination_data = + GetDestinationData(in_params.splitter_info_id, base)) { + base += channel_count; + + if (!destination_data->IsConfigured()) { + continue; + } + if (destination_data->GetMixId() >= static_cast<int>(mix_context.GetCount())) { + continue; + } + + const auto& mix_info = mix_context.GetInfo(destination_data->GetMixId()); + const auto& dest_mix_params = mix_info.GetInParams(); + GenerateVoiceMixCommand( + destination_data->CurrentMixVolumes(), destination_data->LastMixVolumes(), + dsp_state, dest_mix_params.buffer_offset, dest_mix_params.buffer_count, + worker_params.mix_buffer_count + channel, in_params.node_id); + destination_data->MarkDirty(); + } + } + // Update biquad filter enabled states + for (std::size_t i = 0; i < AudioCommon::MAX_BIQUAD_FILTERS; i++) { + in_params.was_biquad_filter_enabled[i] = in_params.biquad_filter[i].enabled; + } + } + } +} + +void CommandGenerator::GenerateSubMixCommands() { + const auto mix_count = mix_context.GetCount(); + for (std::size_t i = 0; i < mix_count; i++) { + auto& mix_info = mix_context.GetSortedInfo(i); + const auto& in_params = mix_info.GetInParams(); + if (!in_params.in_use || in_params.mix_id == AudioCommon::FINAL_MIX) { + continue; + } + GenerateSubMixCommand(mix_info); + } +} + +void CommandGenerator::GenerateFinalMixCommands() { + GenerateFinalMixCommand(); +} + +void CommandGenerator::PreCommand() { + if (!dumping_frame) { + return; + } + for (std::size_t i = 0; i < splitter_context.GetInfoCount(); i++) { + const auto& base = splitter_context.GetInfo(i); + std::string graph = fmt::format("b[{}]", i); + const auto* head = base.GetHead(); + while (head != nullptr) { + graph += fmt::format("->{}", head->GetMixId()); + head = head->GetNextDestination(); + } + LOG_DEBUG(Audio, "(DSP_TRACE) SplitterGraph splitter_info={}, {}", i, graph); + } +} + +void CommandGenerator::PostCommand() { + if (!dumping_frame) { + return; + } + dumping_frame = false; +} + +void CommandGenerator::GenerateDataSourceCommand(ServerVoiceInfo& voice_info, VoiceState& dsp_state, + s32 channel) { + const auto& in_params = voice_info.GetInParams(); + const auto depop = in_params.should_depop; + + if (depop) { + if (in_params.mix_id != AudioCommon::NO_MIX) { + auto& mix_info = mix_context.GetInfo(in_params.mix_id); + const auto& mix_in = mix_info.GetInParams(); + GenerateDepopPrepareCommand(dsp_state, mix_in.buffer_count, mix_in.buffer_offset); + } else if (in_params.splitter_info_id != AudioCommon::NO_SPLITTER) { + s32 index{}; + while (const auto* destination = + GetDestinationData(in_params.splitter_info_id, index++)) { + if (!destination->IsConfigured()) { + continue; + } + auto& mix_info = mix_context.GetInfo(destination->GetMixId()); + const auto& mix_in = mix_info.GetInParams(); + GenerateDepopPrepareCommand(dsp_state, mix_in.buffer_count, mix_in.buffer_offset); + } + } + } else { + switch (in_params.sample_format) { + case SampleFormat::Pcm16: + DecodeFromWaveBuffers(voice_info, GetChannelMixBuffer(channel), dsp_state, channel, + worker_params.sample_rate, worker_params.sample_count, + in_params.node_id); + break; + case SampleFormat::Adpcm: + ASSERT(channel == 0 && in_params.channel_count == 1); + DecodeFromWaveBuffers(voice_info, GetChannelMixBuffer(0), dsp_state, 0, + worker_params.sample_rate, worker_params.sample_count, + in_params.node_id); + break; + default: + UNREACHABLE_MSG("Unimplemented sample format={}", in_params.sample_format); + } + } +} + +void CommandGenerator::GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voice_info, + VoiceState& dsp_state, + s32 mix_buffer_count, s32 channel) { + for (std::size_t i = 0; i < AudioCommon::MAX_BIQUAD_FILTERS; i++) { + const auto& in_params = voice_info.GetInParams(); + auto& biquad_filter = in_params.biquad_filter[i]; + // Check if biquad filter is actually used + if (!biquad_filter.enabled) { + continue; + } + + // Reinitialize our biquad filter state if it was enabled previously + if (!in_params.was_biquad_filter_enabled[i]) { + dsp_state.biquad_filter_state.fill(0); + } + + // Generate biquad filter + // GenerateBiquadFilterCommand(mix_buffer_count, biquad_filter, + // dsp_state.biquad_filter_state, + // mix_buffer_count + channel, mix_buffer_count + + // channel, worker_params.sample_count, + // voice_info.GetInParams().node_id); + } +} + +void AudioCore::CommandGenerator::GenerateBiquadFilterCommand( + s32 mix_buffer, const BiquadFilterParameter& params, std::array<s64, 2>& state, + std::size_t input_offset, std::size_t output_offset, s32 sample_count, s32 node_id) { + if (dumping_frame) { + LOG_DEBUG(Audio, + "(DSP_TRACE) GenerateBiquadFilterCommand node_id={}, " + "input_mix_buffer={}, output_mix_buffer={}", + node_id, input_offset, output_offset); + } + const auto* input = GetMixBuffer(input_offset); + auto* output = GetMixBuffer(output_offset); + + // Biquad filter parameters + const auto [n0, n1, n2] = params.numerator; + const auto [d0, d1] = params.denominator; + + // Biquad filter states + auto [s0, s1] = state; + + constexpr s64 int32_min = std::numeric_limits<s32>::min(); + constexpr s64 int32_max = std::numeric_limits<s32>::max(); + + for (int i = 0; i < sample_count; ++i) { + const auto sample = static_cast<s64>(input[i]); + const auto f = (sample * n0 + s0 + 0x4000) >> 15; + const auto y = std::clamp(f, int32_min, int32_max); + s0 = sample * n1 + y * d0 + s1; + s1 = sample * n2 + y * d1; + output[i] = static_cast<s32>(y); + } + + state = {s0, s1}; +} + +void CommandGenerator::GenerateDepopPrepareCommand(VoiceState& dsp_state, + std::size_t mix_buffer_count, + std::size_t mix_buffer_offset) { + for (std::size_t i = 0; i < mix_buffer_count; i++) { + auto& sample = dsp_state.previous_samples[i]; + if (sample != 0) { + depop_buffer[mix_buffer_offset + i] += sample; + sample = 0; + } + } +} + +void CommandGenerator::GenerateDepopForMixBuffersCommand(std::size_t mix_buffer_count, + std::size_t mix_buffer_offset, + s32 sample_rate) { + const std::size_t end_offset = + std::min(mix_buffer_offset + mix_buffer_count, GetTotalMixBufferCount()); + const s32 delta = sample_rate == 48000 ? 0x7B29 : 0x78CB; + for (std::size_t i = mix_buffer_offset; i < end_offset; i++) { + if (depop_buffer[i] == 0) { + continue; + } + + depop_buffer[i] = + ApplyMixDepop(GetMixBuffer(i), depop_buffer[i], delta, worker_params.sample_count); + } +} + +void CommandGenerator::GenerateEffectCommand(ServerMixInfo& mix_info) { + const std::size_t effect_count = effect_context.GetCount(); + const auto buffer_offset = mix_info.GetInParams().buffer_offset; + for (std::size_t i = 0; i < effect_count; i++) { + const auto index = mix_info.GetEffectOrder(i); + if (index == AudioCommon::NO_EFFECT_ORDER) { + break; + } + auto* info = effect_context.GetInfo(index); + const auto type = info->GetType(); + + // TODO(ogniK): Finish remaining effects + switch (type) { + case EffectType::Aux: + GenerateAuxCommand(buffer_offset, info, info->IsEnabled()); + break; + case EffectType::I3dl2Reverb: + GenerateI3dl2ReverbEffectCommand(buffer_offset, info, info->IsEnabled()); + break; + case EffectType::BiquadFilter: + GenerateBiquadFilterEffectCommand(buffer_offset, info, info->IsEnabled()); + break; + default: + break; + } + + info->UpdateForCommandGeneration(); + } +} + +void CommandGenerator::GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, EffectBase* info, + bool enabled) { + if (!enabled) { + return; + } + const auto& params = dynamic_cast<EffectI3dl2Reverb*>(info)->GetParams(); + const auto channel_count = params.channel_count; + for (s32 i = 0; i < channel_count; i++) { + // TODO(ogniK): Actually implement reverb + if (params.input[i] != params.output[i]) { + const auto* input = GetMixBuffer(mix_buffer_offset + params.input[i]); + auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]); + ApplyMix<1>(output, input, 32768, worker_params.sample_count); + } + } +} + +void CommandGenerator::GenerateBiquadFilterEffectCommand(s32 mix_buffer_offset, EffectBase* info, + bool enabled) { + if (!enabled) { + return; + } + const auto& params = dynamic_cast<EffectBiquadFilter*>(info)->GetParams(); + const auto channel_count = params.channel_count; + for (s32 i = 0; i < channel_count; i++) { + // TODO(ogniK): Actually implement biquad filter + if (params.input[i] != params.output[i]) { + const auto* input = GetMixBuffer(mix_buffer_offset + params.input[i]); + auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]); + ApplyMix<1>(output, input, 32768, worker_params.sample_count); + } + } +} + +void CommandGenerator::GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled) { + auto* aux = dynamic_cast<EffectAuxInfo*>(info); + const auto& params = aux->GetParams(); + if (aux->GetSendBuffer() != 0 && aux->GetRecvBuffer() != 0) { + const auto max_channels = params.count; + u32 offset{}; + for (u32 channel = 0; channel < max_channels; channel++) { + u32 write_count = 0; + if (channel == (max_channels - 1)) { + write_count = offset + worker_params.sample_count; + } + + const auto input_index = params.input_mix_buffers[channel] + mix_buffer_offset; + const auto output_index = params.output_mix_buffers[channel] + mix_buffer_offset; + + if (enabled) { + AuxInfoDSP send_info{}; + AuxInfoDSP recv_info{}; + memory.ReadBlock(aux->GetSendInfo(), &send_info, sizeof(AuxInfoDSP)); + memory.ReadBlock(aux->GetRecvInfo(), &recv_info, sizeof(AuxInfoDSP)); + + WriteAuxBuffer(send_info, aux->GetSendBuffer(), params.sample_count, + GetMixBuffer(input_index), worker_params.sample_count, offset, + write_count); + memory.WriteBlock(aux->GetSendInfo(), &send_info, sizeof(AuxInfoDSP)); + + const auto samples_read = ReadAuxBuffer( + recv_info, aux->GetRecvBuffer(), params.sample_count, + GetMixBuffer(output_index), worker_params.sample_count, offset, write_count); + memory.WriteBlock(aux->GetRecvInfo(), &recv_info, sizeof(AuxInfoDSP)); + + if (samples_read != static_cast<int>(worker_params.sample_count) && + samples_read <= params.sample_count) { + std::memset(GetMixBuffer(output_index), 0, params.sample_count - samples_read); + } + } else { + AuxInfoDSP empty{}; + memory.WriteBlock(aux->GetSendInfo(), &empty, sizeof(AuxInfoDSP)); + memory.WriteBlock(aux->GetRecvInfo(), &empty, sizeof(AuxInfoDSP)); + if (output_index != input_index) { + std::memcpy(GetMixBuffer(output_index), GetMixBuffer(input_index), + worker_params.sample_count * sizeof(s32)); + } + } + + offset += worker_params.sample_count; + } + } +} + +ServerSplitterDestinationData* CommandGenerator::GetDestinationData(s32 splitter_id, s32 index) { + if (splitter_id == AudioCommon::NO_SPLITTER) { + return nullptr; + } + return splitter_context.GetDestinationData(splitter_id, index); +} + +s32 CommandGenerator::WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples, + const s32* data, u32 sample_count, u32 write_offset, + u32 write_count) { + if (max_samples == 0) { + return 0; + } + u32 offset = dsp_info.write_offset + write_offset; + if (send_buffer == 0 || offset > max_samples) { + return 0; + } + + std::size_t data_offset{}; + u32 remaining = sample_count; + while (remaining > 0) { + // Get position in buffer + const auto base = send_buffer + (offset * sizeof(u32)); + const auto samples_to_grab = std::min(max_samples - offset, remaining); + // Write to output + memory.WriteBlock(base, (data + data_offset), samples_to_grab * sizeof(u32)); + offset = (offset + samples_to_grab) % max_samples; + remaining -= samples_to_grab; + data_offset += samples_to_grab; + } + + if (write_count != 0) { + dsp_info.write_offset = (dsp_info.write_offset + write_count) % max_samples; + } + return sample_count; +} + +s32 CommandGenerator::ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples, + s32* out_data, u32 sample_count, u32 read_offset, + u32 read_count) { + if (max_samples == 0) { + return 0; + } + + u32 offset = recv_info.read_offset + read_offset; + if (recv_buffer == 0 || offset > max_samples) { + return 0; + } + + u32 remaining = sample_count; + while (remaining > 0) { + const auto base = recv_buffer + (offset * sizeof(u32)); + const auto samples_to_grab = std::min(max_samples - offset, remaining); + std::vector<s32> buffer(samples_to_grab); + memory.ReadBlock(base, buffer.data(), buffer.size() * sizeof(u32)); + std::memcpy(out_data, buffer.data(), buffer.size() * sizeof(u32)); + out_data += samples_to_grab; + offset = (offset + samples_to_grab) % max_samples; + remaining -= samples_to_grab; + } + + if (read_count != 0) { + recv_info.read_offset = (recv_info.read_offset + read_count) % max_samples; + } + return sample_count; +} + +void CommandGenerator::GenerateVolumeRampCommand(float last_volume, float current_volume, + s32 channel, s32 node_id) { + const auto last = static_cast<s32>(last_volume * 32768.0f); + const auto current = static_cast<s32>(current_volume * 32768.0f); + const auto delta = static_cast<s32>((static_cast<float>(current) - static_cast<float>(last)) / + static_cast<float>(worker_params.sample_count)); + + if (dumping_frame) { + LOG_DEBUG(Audio, + "(DSP_TRACE) GenerateVolumeRampCommand node_id={}, input={}, output={}, " + "last_volume={}, current_volume={}", + node_id, GetMixChannelBufferOffset(channel), GetMixChannelBufferOffset(channel), + last_volume, current_volume); + } + // Apply generic gain on samples + ApplyGain(GetChannelMixBuffer(channel), GetChannelMixBuffer(channel), last, delta, + worker_params.sample_count); +} + +void CommandGenerator::GenerateVoiceMixCommand(const MixVolumeBuffer& mix_volumes, + const MixVolumeBuffer& last_mix_volumes, + VoiceState& dsp_state, s32 mix_buffer_offset, + s32 mix_buffer_count, s32 voice_index, s32 node_id) { + // Loop all our mix buffers + for (s32 i = 0; i < mix_buffer_count; i++) { + if (last_mix_volumes[i] != 0.0f || mix_volumes[i] != 0.0f) { + const auto delta = static_cast<float>((mix_volumes[i] - last_mix_volumes[i])) / + static_cast<float>(worker_params.sample_count); + + if (dumping_frame) { + LOG_DEBUG(Audio, + "(DSP_TRACE) GenerateVoiceMixCommand node_id={}, input={}, " + "output={}, last_volume={}, current_volume={}", + node_id, voice_index, mix_buffer_offset + i, last_mix_volumes[i], + mix_volumes[i]); + } + + dsp_state.previous_samples[i] = + ApplyMixRamp(GetMixBuffer(mix_buffer_offset + i), GetMixBuffer(voice_index), + last_mix_volumes[i], delta, worker_params.sample_count); + } else { + dsp_state.previous_samples[i] = 0; + } + } +} + +void CommandGenerator::GenerateSubMixCommand(ServerMixInfo& mix_info) { + if (dumping_frame) { + LOG_DEBUG(Audio, "(DSP_TRACE) GenerateSubMixCommand"); + } + const auto& in_params = mix_info.GetInParams(); + GenerateDepopForMixBuffersCommand(in_params.buffer_count, in_params.buffer_offset, + in_params.sample_rate); + + GenerateEffectCommand(mix_info); + + GenerateMixCommands(mix_info); +} + +void CommandGenerator::GenerateMixCommands(ServerMixInfo& mix_info) { + if (!mix_info.HasAnyConnection()) { + return; + } + const auto& in_params = mix_info.GetInParams(); + if (in_params.dest_mix_id != AudioCommon::NO_MIX) { + const auto& dest_mix = mix_context.GetInfo(in_params.dest_mix_id); + const auto& dest_in_params = dest_mix.GetInParams(); + + const auto buffer_count = in_params.buffer_count; + + for (s32 i = 0; i < buffer_count; i++) { + for (s32 j = 0; j < dest_in_params.buffer_count; j++) { + const auto mixed_volume = in_params.volume * in_params.mix_volume[i][j]; + if (mixed_volume != 0.0f) { + GenerateMixCommand(dest_in_params.buffer_offset + j, + in_params.buffer_offset + i, mixed_volume, + in_params.node_id); + } + } + } + } else if (in_params.splitter_id != AudioCommon::NO_SPLITTER) { + s32 base{}; + while (const auto* destination_data = GetDestinationData(in_params.splitter_id, base++)) { + if (!destination_data->IsConfigured()) { + continue; + } + + const auto& dest_mix = mix_context.GetInfo(destination_data->GetMixId()); + const auto& dest_in_params = dest_mix.GetInParams(); + const auto mix_index = (base - 1) % in_params.buffer_count + in_params.buffer_offset; + for (std::size_t i = 0; i < static_cast<std::size_t>(dest_in_params.buffer_count); + i++) { + const auto mixed_volume = in_params.volume * destination_data->GetMixVolume(i); + if (mixed_volume != 0.0f) { + GenerateMixCommand(dest_in_params.buffer_offset + i, mix_index, mixed_volume, + in_params.node_id); + } + } + } + } +} + +void CommandGenerator::GenerateMixCommand(std::size_t output_offset, std::size_t input_offset, + float volume, s32 node_id) { + + if (dumping_frame) { + LOG_DEBUG(Audio, + "(DSP_TRACE) GenerateMixCommand node_id={}, input={}, output={}, volume={}", + node_id, input_offset, output_offset, volume); + } + + auto* output = GetMixBuffer(output_offset); + const auto* input = GetMixBuffer(input_offset); + + const s32 gain = static_cast<s32>(volume * 32768.0f); + // Mix with loop unrolling + if (worker_params.sample_count % 4 == 0) { + ApplyMix<4>(output, input, gain, worker_params.sample_count); + } else if (worker_params.sample_count % 2 == 0) { + ApplyMix<2>(output, input, gain, worker_params.sample_count); + } else { + ApplyMix<1>(output, input, gain, worker_params.sample_count); + } +} + +void CommandGenerator::GenerateFinalMixCommand() { + if (dumping_frame) { + LOG_DEBUG(Audio, "(DSP_TRACE) GenerateFinalMixCommand"); + } + auto& mix_info = mix_context.GetFinalMixInfo(); + const auto& in_params = mix_info.GetInParams(); + + GenerateDepopForMixBuffersCommand(in_params.buffer_count, in_params.buffer_offset, + in_params.sample_rate); + + GenerateEffectCommand(mix_info); + + for (s32 i = 0; i < in_params.buffer_count; i++) { + const s32 gain = static_cast<s32>(in_params.volume * 32768.0f); + if (dumping_frame) { + LOG_DEBUG( + Audio, + "(DSP_TRACE) ApplyGainWithoutDelta node_id={}, input={}, output={}, volume={}", + in_params.node_id, in_params.buffer_offset + i, in_params.buffer_offset + i, + in_params.volume); + } + ApplyGainWithoutDelta(GetMixBuffer(in_params.buffer_offset + i), + GetMixBuffer(in_params.buffer_offset + i), gain, + worker_params.sample_count); + } +} + +s32 CommandGenerator::DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state, + s32 sample_count, s32 channel, std::size_t mix_offset) { + const auto& in_params = voice_info.GetInParams(); + const auto& wave_buffer = in_params.wave_buffer[dsp_state.wave_buffer_index]; + if (wave_buffer.buffer_address == 0) { + return 0; + } + if (wave_buffer.buffer_size == 0) { + return 0; + } + if (wave_buffer.end_sample_offset < wave_buffer.start_sample_offset) { + return 0; + } + const auto samples_remaining = + (wave_buffer.end_sample_offset - wave_buffer.start_sample_offset) - dsp_state.offset; + const auto start_offset = + ((wave_buffer.start_sample_offset + dsp_state.offset) * in_params.channel_count) * + sizeof(s16); + const auto buffer_pos = wave_buffer.buffer_address + start_offset; + const auto samples_processed = std::min(sample_count, samples_remaining); + + if (in_params.channel_count == 1) { + std::vector<s16> buffer(samples_processed); + memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(s16)); + for (std::size_t i = 0; i < buffer.size(); i++) { + sample_buffer[mix_offset + i] = buffer[i]; + } + } else { + const auto channel_count = in_params.channel_count; + std::vector<s16> buffer(samples_processed * channel_count); + memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(s16)); + + for (std::size_t i = 0; i < static_cast<std::size_t>(samples_processed); i++) { + sample_buffer[mix_offset + i] = buffer[i * channel_count + channel]; + } + } + + return samples_processed; +} + +s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state, + s32 sample_count, s32 channel, std::size_t mix_offset) { + const auto& in_params = voice_info.GetInParams(); + const auto& wave_buffer = in_params.wave_buffer[dsp_state.wave_buffer_index]; + if (wave_buffer.buffer_address == 0) { + return 0; + } + if (wave_buffer.buffer_size == 0) { + return 0; + } + if (wave_buffer.end_sample_offset < wave_buffer.start_sample_offset) { + return 0; + } + + static constexpr std::array<int, 16> SIGNED_NIBBLES{ + 0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1, + }; + + constexpr std::size_t FRAME_LEN = 8; + constexpr std::size_t NIBBLES_PER_SAMPLE = 16; + constexpr std::size_t SAMPLES_PER_FRAME = 14; + + auto frame_header = dsp_state.context.header; + s32 idx = (frame_header >> 4) & 0xf; + s32 scale = frame_header & 0xf; + s16 yn1 = dsp_state.context.yn1; + s16 yn2 = dsp_state.context.yn2; + + Codec::ADPCM_Coeff coeffs; + memory.ReadBlock(in_params.additional_params_address, coeffs.data(), + sizeof(Codec::ADPCM_Coeff)); + + s32 coef1 = coeffs[idx * 2]; + s32 coef2 = coeffs[idx * 2 + 1]; + + const auto samples_remaining = + (wave_buffer.end_sample_offset - wave_buffer.start_sample_offset) - dsp_state.offset; + const auto samples_processed = std::min(sample_count, samples_remaining); + const auto sample_pos = wave_buffer.start_sample_offset + dsp_state.offset; + + const auto samples_remaining_in_frame = sample_pos % SAMPLES_PER_FRAME; + auto position_in_frame = ((sample_pos / SAMPLES_PER_FRAME) * NIBBLES_PER_SAMPLE) + + samples_remaining_in_frame + (samples_remaining_in_frame != 0 ? 2 : 0); + + const auto decode_sample = [&](const int nibble) -> s16 { + const int xn = nibble * (1 << scale); + // We first transform everything into 11 bit fixed point, perform the second order + // digital filter, then transform back. + // 0x400 == 0.5 in 11 bit fixed point. + // Filter: y[n] = x[n] + 0.5 + c1 * y[n-1] + c2 * y[n-2] + int val = ((xn << 11) + 0x400 + coef1 * yn1 + coef2 * yn2) >> 11; + // Clamp to output range. + val = std::clamp<s32>(val, -32768, 32767); + // Advance output feedback. + yn2 = yn1; + yn1 = static_cast<s16>(val); + return yn1; + }; + + std::size_t buffer_offset{}; + std::vector<u8> buffer( + std::max((samples_processed / FRAME_LEN) * SAMPLES_PER_FRAME, FRAME_LEN)); + memory.ReadBlock(wave_buffer.buffer_address + (position_in_frame / 2), buffer.data(), + buffer.size()); + std::size_t cur_mix_offset = mix_offset; + + auto remaining_samples = samples_processed; + while (remaining_samples > 0) { + if (position_in_frame % NIBBLES_PER_SAMPLE == 0) { + // Read header + frame_header = buffer[buffer_offset++]; + idx = (frame_header >> 4) & 0xf; + scale = frame_header & 0xf; + coef1 = coeffs[idx * 2]; + coef2 = coeffs[idx * 2 + 1]; + position_in_frame += 2; + + // Decode entire frame + if (remaining_samples >= static_cast<int>(SAMPLES_PER_FRAME)) { + for (std::size_t i = 0; i < SAMPLES_PER_FRAME / 2; i++) { + // Sample 1 + const s32 s0 = SIGNED_NIBBLES[buffer[buffer_offset] >> 4]; + const s32 s1 = SIGNED_NIBBLES[buffer[buffer_offset++] & 0xf]; + const s16 sample_1 = decode_sample(s0); + const s16 sample_2 = decode_sample(s1); + sample_buffer[cur_mix_offset++] = sample_1; + sample_buffer[cur_mix_offset++] = sample_2; + } + remaining_samples -= static_cast<int>(SAMPLES_PER_FRAME); + position_in_frame += SAMPLES_PER_FRAME; + continue; + } + } + // Decode mid frame + s32 current_nibble = buffer[buffer_offset]; + if (position_in_frame++ & 0x1) { + current_nibble &= 0xf; + buffer_offset++; + } else { + current_nibble >>= 4; + } + const s16 sample = decode_sample(SIGNED_NIBBLES[current_nibble]); + sample_buffer[cur_mix_offset++] = sample; + remaining_samples--; + } + + dsp_state.context.header = frame_header; + dsp_state.context.yn1 = yn1; + dsp_state.context.yn2 = yn2; + + return samples_processed; +} + +s32* CommandGenerator::GetMixBuffer(std::size_t index) { + return mix_buffer.data() + (index * worker_params.sample_count); +} + +const s32* CommandGenerator::GetMixBuffer(std::size_t index) const { + return mix_buffer.data() + (index * worker_params.sample_count); +} + +std::size_t CommandGenerator::GetMixChannelBufferOffset(s32 channel) const { + return worker_params.mix_buffer_count + channel; +} + +std::size_t CommandGenerator::GetTotalMixBufferCount() const { + return worker_params.mix_buffer_count + AudioCommon::MAX_CHANNEL_COUNT; +} + +s32* CommandGenerator::GetChannelMixBuffer(s32 channel) { + return GetMixBuffer(worker_params.mix_buffer_count + channel); +} + +const s32* CommandGenerator::GetChannelMixBuffer(s32 channel) const { + return GetMixBuffer(worker_params.mix_buffer_count + channel); +} + +void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* output, + VoiceState& dsp_state, s32 channel, + s32 target_sample_rate, s32 sample_count, + s32 node_id) { + const auto& in_params = voice_info.GetInParams(); + if (dumping_frame) { + LOG_DEBUG(Audio, + "(DSP_TRACE) DecodeFromWaveBuffers, node_id={}, channel={}, " + "format={}, sample_count={}, sample_rate={}, mix_id={}, splitter_id={}", + node_id, channel, in_params.sample_format, sample_count, in_params.sample_rate, + in_params.mix_id, in_params.splitter_info_id); + } + ASSERT_OR_EXECUTE(output != nullptr, { return; }); + + const auto resample_rate = static_cast<s32>( + static_cast<float>(in_params.sample_rate) / static_cast<float>(target_sample_rate) * + static_cast<float>(static_cast<s32>(in_params.pitch * 32768.0f))); + if (dsp_state.fraction + sample_count * resample_rate > + static_cast<s32>(SCALED_MIX_BUFFER_SIZE - 4ULL)) { + return; + } + + auto min_required_samples = + std::min(static_cast<s32>(SCALED_MIX_BUFFER_SIZE) - dsp_state.fraction, resample_rate); + if (min_required_samples >= sample_count) { + min_required_samples = sample_count; + } + + std::size_t temp_mix_offset{}; + bool is_buffer_completed{false}; + auto samples_remaining = sample_count; + while (samples_remaining > 0 && !is_buffer_completed) { + const auto samples_to_output = std::min(samples_remaining, min_required_samples); + const auto samples_to_read = (samples_to_output * resample_rate + dsp_state.fraction) >> 15; + + if (!in_params.behavior_flags.is_pitch_and_src_skipped) { + // Append sample histtory for resampler + for (std::size_t i = 0; i < AudioCommon::MAX_SAMPLE_HISTORY; i++) { + sample_buffer[temp_mix_offset + i] = dsp_state.sample_history[i]; + } + temp_mix_offset += 4; + } + + s32 samples_read{}; + while (samples_read < samples_to_read) { + const auto& wave_buffer = in_params.wave_buffer[dsp_state.wave_buffer_index]; + // No more data can be read + if (!dsp_state.is_wave_buffer_valid[dsp_state.wave_buffer_index]) { + is_buffer_completed = true; + break; + } + + if (in_params.sample_format == SampleFormat::Adpcm && dsp_state.offset == 0 && + wave_buffer.context_address != 0 && wave_buffer.context_size != 0) { + // TODO(ogniK): ADPCM loop context + } + + s32 samples_decoded{0}; + switch (in_params.sample_format) { + case SampleFormat::Pcm16: + samples_decoded = DecodePcm16(voice_info, dsp_state, samples_to_read - samples_read, + channel, temp_mix_offset); + break; + case SampleFormat::Adpcm: + samples_decoded = DecodeAdpcm(voice_info, dsp_state, samples_to_read - samples_read, + channel, temp_mix_offset); + break; + default: + UNREACHABLE_MSG("Unimplemented sample format={}", in_params.sample_format); + } + + temp_mix_offset += samples_decoded; + samples_read += samples_decoded; + dsp_state.offset += samples_decoded; + dsp_state.played_sample_count += samples_decoded; + + if (dsp_state.offset >= + (wave_buffer.end_sample_offset - wave_buffer.start_sample_offset) || + samples_decoded == 0) { + // Reset our sample offset + dsp_state.offset = 0; + if (wave_buffer.is_looping) { + if (samples_decoded == 0) { + // End of our buffer + is_buffer_completed = true; + break; + } + + if (in_params.behavior_flags.is_played_samples_reset_at_loop_point.Value()) { + dsp_state.played_sample_count = 0; + } + } else { + + // Update our wave buffer states + dsp_state.is_wave_buffer_valid[dsp_state.wave_buffer_index] = false; + dsp_state.wave_buffer_consumed++; + dsp_state.wave_buffer_index = + (dsp_state.wave_buffer_index + 1) % AudioCommon::MAX_WAVE_BUFFERS; + if (wave_buffer.end_of_stream) { + dsp_state.played_sample_count = 0; + } + } + } + } + + if (in_params.behavior_flags.is_pitch_and_src_skipped.Value()) { + // No need to resample + std::memcpy(output, sample_buffer.data(), samples_read * sizeof(s32)); + } else { + std::fill(sample_buffer.begin() + temp_mix_offset, + sample_buffer.begin() + temp_mix_offset + (samples_to_read - samples_read), + 0); + AudioCore::Resample(output, sample_buffer.data(), resample_rate, dsp_state.fraction, + samples_to_output); + // Resample + for (std::size_t i = 0; i < AudioCommon::MAX_SAMPLE_HISTORY; i++) { + dsp_state.sample_history[i] = sample_buffer[samples_to_read + i]; + } + } + output += samples_to_output; + samples_remaining -= samples_to_output; + } +} + +} // namespace AudioCore diff --git a/src/audio_core/command_generator.h b/src/audio_core/command_generator.h new file mode 100644 index 000000000..53e57748b --- /dev/null +++ b/src/audio_core/command_generator.h @@ -0,0 +1,102 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include "audio_core/common.h" +#include "audio_core/voice_context.h" +#include "common/common_types.h" + +namespace Core::Memory { +class Memory; +} + +namespace AudioCore { +class MixContext; +class SplitterContext; +class ServerSplitterDestinationData; +class ServerMixInfo; +class EffectContext; +class EffectBase; +struct AuxInfoDSP; +using MixVolumeBuffer = std::array<float, AudioCommon::MAX_MIX_BUFFERS>; + +class CommandGenerator { +public: + explicit CommandGenerator(AudioCommon::AudioRendererParameter& worker_params, + VoiceContext& voice_context, MixContext& mix_context, + SplitterContext& splitter_context, EffectContext& effect_context, + Core::Memory::Memory& memory); + ~CommandGenerator(); + + void ClearMixBuffers(); + void GenerateVoiceCommands(); + void GenerateVoiceCommand(ServerVoiceInfo& voice_info); + void GenerateSubMixCommands(); + void GenerateFinalMixCommands(); + void PreCommand(); + void PostCommand(); + + s32* GetChannelMixBuffer(s32 channel); + const s32* GetChannelMixBuffer(s32 channel) const; + s32* GetMixBuffer(std::size_t index); + const s32* GetMixBuffer(std::size_t index) const; + std::size_t GetMixChannelBufferOffset(s32 channel) const; + + std::size_t GetTotalMixBufferCount() const; + +private: + void GenerateDataSourceCommand(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 channel); + void GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voice_info, VoiceState& dsp_state, + s32 mix_buffer_count, s32 channel); + void GenerateVolumeRampCommand(float last_volume, float current_volume, s32 channel, + s32 node_id); + void GenerateVoiceMixCommand(const MixVolumeBuffer& mix_volumes, + const MixVolumeBuffer& last_mix_volumes, VoiceState& dsp_state, + s32 mix_buffer_offset, s32 mix_buffer_count, s32 voice_index, + s32 node_id); + void GenerateSubMixCommand(ServerMixInfo& mix_info); + void GenerateMixCommands(ServerMixInfo& mix_info); + void GenerateMixCommand(std::size_t output_offset, std::size_t input_offset, float volume, + s32 node_id); + void GenerateFinalMixCommand(); + void GenerateBiquadFilterCommand(s32 mix_buffer, const BiquadFilterParameter& params, + std::array<s64, 2>& state, std::size_t input_offset, + std::size_t output_offset, s32 sample_count, s32 node_id); + void GenerateDepopPrepareCommand(VoiceState& dsp_state, std::size_t mix_buffer_count, + std::size_t mix_buffer_offset); + void GenerateDepopForMixBuffersCommand(std::size_t mix_buffer_count, + std::size_t mix_buffer_offset, s32 sample_rate); + void GenerateEffectCommand(ServerMixInfo& mix_info); + void GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled); + void GenerateBiquadFilterEffectCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled); + void GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled); + ServerSplitterDestinationData* GetDestinationData(s32 splitter_id, s32 index); + + s32 WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples, const s32* data, + u32 sample_count, u32 write_offset, u32 write_count); + s32 ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples, s32* out_data, + u32 sample_count, u32 read_offset, u32 read_count); + + // DSP Code + s32 DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_count, + s32 channel, std::size_t mix_offset); + s32 DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_count, + s32 channel, std::size_t mix_offset); + void DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* output, VoiceState& dsp_state, + s32 channel, s32 target_sample_rate, s32 sample_count, s32 node_id); + + AudioCommon::AudioRendererParameter& worker_params; + VoiceContext& voice_context; + MixContext& mix_context; + SplitterContext& splitter_context; + EffectContext& effect_context; + Core::Memory::Memory& memory; + std::vector<s32> mix_buffer{}; + std::vector<s32> sample_buffer{}; + std::vector<s32> depop_buffer{}; + bool dumping_frame{false}; +}; +} // namespace AudioCore diff --git a/src/audio_core/common.h b/src/audio_core/common.h new file mode 100644 index 000000000..7b4a1e9e8 --- /dev/null +++ b/src/audio_core/common.h @@ -0,0 +1,108 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" +#include "core/hle/result.h" + +namespace AudioCommon { +namespace Audren { +constexpr ResultCode ERR_INVALID_PARAMETERS{ErrorModule::Audio, 41}; +constexpr ResultCode ERR_SPLITTER_SORT_FAILED{ErrorModule::Audio, 43}; +} // namespace Audren + +constexpr u32_le CURRENT_PROCESS_REVISION = Common::MakeMagic('R', 'E', 'V', '8'); +constexpr std::size_t MAX_MIX_BUFFERS = 24; +constexpr std::size_t MAX_BIQUAD_FILTERS = 2; +constexpr std::size_t MAX_CHANNEL_COUNT = 6; +constexpr std::size_t MAX_WAVE_BUFFERS = 4; +constexpr std::size_t MAX_SAMPLE_HISTORY = 4; +constexpr u32 STREAM_SAMPLE_RATE = 48000; +constexpr u32 STREAM_NUM_CHANNELS = 6; +constexpr s32 NO_SPLITTER = -1; +constexpr s32 NO_MIX = 0x7fffffff; +constexpr s32 NO_FINAL_MIX = std::numeric_limits<s32>::min(); +constexpr s32 FINAL_MIX = 0; +constexpr s32 NO_EFFECT_ORDER = -1; +constexpr std::size_t TEMP_MIX_BASE_SIZE = 0x3f00; // TODO(ogniK): Work out this constant +// Any size checks seem to take the sample history into account +// and our const ends up being 0x3f04, the 4 bytes are most +// likely the sample history +constexpr std::size_t TOTAL_TEMP_MIX_SIZE = TEMP_MIX_BASE_SIZE + AudioCommon::MAX_SAMPLE_HISTORY; + +static constexpr u32 VersionFromRevision(u32_le rev) { + // "REV7" -> 7 + return ((rev >> 24) & 0xff) - 0x30; +} + +static constexpr bool IsRevisionSupported(u32 required, u32_le user_revision) { + const auto base = VersionFromRevision(user_revision); + return required <= base; +} + +static constexpr bool IsValidRevision(u32_le revision) { + const auto base = VersionFromRevision(revision); + constexpr auto max_rev = VersionFromRevision(CURRENT_PROCESS_REVISION); + return base <= max_rev; +} + +static constexpr bool CanConsumeBuffer(std::size_t size, std::size_t offset, std::size_t required) { + if (offset > size) { + return false; + } + if (size < required) { + return false; + } + if ((size - offset) < required) { + return false; + } + return true; +} + +struct UpdateDataSizes { + u32_le behavior{}; + u32_le memory_pool{}; + u32_le voice{}; + u32_le voice_channel_resource{}; + u32_le effect{}; + u32_le mixer{}; + u32_le sink{}; + u32_le performance{}; + u32_le splitter{}; + u32_le render_info{}; + INSERT_PADDING_WORDS(4); +}; +static_assert(sizeof(UpdateDataSizes) == 0x38, "UpdateDataSizes is an invalid size"); + +struct UpdateDataHeader { + u32_le revision{}; + UpdateDataSizes size{}; + u32_le total_size{}; +}; +static_assert(sizeof(UpdateDataHeader) == 0x40, "UpdateDataHeader is an invalid size"); + +struct AudioRendererParameter { + u32_le sample_rate; + u32_le sample_count; + u32_le mix_buffer_count; + u32_le submix_count; + u32_le voice_count; + u32_le sink_count; + u32_le effect_count; + u32_le performance_frame_count; + u8 is_voice_drop_enabled; + u8 unknown_21; + u8 unknown_22; + u8 execution_mode; + u32_le splitter_count; + u32_le num_splitter_send_channels; + u32_le unknown_30; + u32_le revision; +}; +static_assert(sizeof(AudioRendererParameter) == 52, "AudioRendererParameter is an invalid size"); + +} // namespace AudioCommon diff --git a/src/audio_core/cubeb_sink.cpp b/src/audio_core/cubeb_sink.cpp index c4e0e30fe..6eaa60815 100644 --- a/src/audio_core/cubeb_sink.cpp +++ b/src/audio_core/cubeb_sink.cpp @@ -23,14 +23,24 @@ class CubebSinkStream final : public SinkStream { public: CubebSinkStream(cubeb* ctx, u32 sample_rate, u32 num_channels_, cubeb_devid output_device, const std::string& name) - : ctx{ctx}, num_channels{std::min(num_channels_, 2u)}, time_stretch{sample_rate, + : ctx{ctx}, num_channels{std::min(num_channels_, 6u)}, time_stretch{sample_rate, num_channels} { cubeb_stream_params params{}; params.rate = sample_rate; params.channels = num_channels; params.format = CUBEB_SAMPLE_S16NE; - params.layout = num_channels == 1 ? CUBEB_LAYOUT_MONO : CUBEB_LAYOUT_STEREO; + switch (num_channels) { + case 1: + params.layout = CUBEB_LAYOUT_MONO; + break; + case 2: + params.layout = CUBEB_LAYOUT_STEREO; + break; + case 6: + params.layout = CUBEB_LAYOUT_3F2_LFE; + break; + } u32 minimum_latency{}; if (cubeb_get_min_latency(ctx, ¶ms, &minimum_latency) != CUBEB_OK) { @@ -78,13 +88,15 @@ public: const s16 surround_left{samples[i + 4]}; const s16 surround_right{samples[i + 5]}; // Not used in the ATSC reference implementation - [[maybe_unused]] const s16 low_frequency_effects { samples[i + 3] }; + [[maybe_unused]] const s16 low_frequency_effects{samples[i + 3]}; constexpr s32 clev{707}; // center mixing level coefficient constexpr s32 slev{707}; // surround mixing level coefficient - buf.push_back(left + (clev * center / 1000) + (slev * surround_left / 1000)); - buf.push_back(right + (clev * center / 1000) + (slev * surround_right / 1000)); + buf.push_back(static_cast<s16>(left + (clev * center / 1000) + + (slev * surround_left / 1000))); + buf.push_back(static_cast<s16>(right + (clev * center / 1000) + + (slev * surround_right / 1000))); } queue.Push(buf); return; @@ -182,8 +194,8 @@ SinkStream& CubebSink::AcquireSinkStream(u32 sample_rate, u32 num_channels, long CubebSinkStream::DataCallback(cubeb_stream* stream, void* user_data, const void* input_buffer, void* output_buffer, long num_frames) { - CubebSinkStream* impl = static_cast<CubebSinkStream*>(user_data); - u8* buffer = reinterpret_cast<u8*>(output_buffer); + auto* impl = static_cast<CubebSinkStream*>(user_data); + auto* buffer = static_cast<u8*>(output_buffer); if (!impl) { return {}; @@ -193,7 +205,8 @@ long CubebSinkStream::DataCallback(cubeb_stream* stream, void* user_data, const const std::size_t samples_to_write = num_channels * num_frames; std::size_t samples_written; - if (Settings::values.enable_audio_stretching) { + /* + if (Settings::values.enable_audio_stretching.GetValue()) { const std::vector<s16> in{impl->queue.Pop()}; const std::size_t num_in{in.size() / num_channels}; s16* const out{reinterpret_cast<s16*>(buffer)}; @@ -207,7 +220,8 @@ long CubebSinkStream::DataCallback(cubeb_stream* stream, void* user_data, const } } else { samples_written = impl->queue.Pop(buffer, samples_to_write); - } + }*/ + samples_written = impl->queue.Pop(buffer, samples_to_write); if (samples_written >= num_channels) { std::memcpy(&impl->last_frame[0], buffer + (samples_written - num_channels) * sizeof(s16), diff --git a/src/audio_core/effect_context.cpp b/src/audio_core/effect_context.cpp new file mode 100644 index 000000000..4d9cdf524 --- /dev/null +++ b/src/audio_core/effect_context.cpp @@ -0,0 +1,299 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include "audio_core/effect_context.h" + +namespace AudioCore { +namespace { +bool ValidChannelCountForEffect(s32 channel_count) { + return channel_count == 1 || channel_count == 2 || channel_count == 4 || channel_count == 6; +} +} // namespace + +EffectContext::EffectContext(std::size_t effect_count) : effect_count(effect_count) { + effects.reserve(effect_count); + std::generate_n(std::back_inserter(effects), effect_count, + [] { return std::make_unique<EffectStubbed>(); }); +} +EffectContext::~EffectContext() = default; + +std::size_t EffectContext::GetCount() const { + return effect_count; +} + +EffectBase* EffectContext::GetInfo(std::size_t i) { + return effects.at(i).get(); +} + +EffectBase* EffectContext::RetargetEffect(std::size_t i, EffectType effect) { + switch (effect) { + case EffectType::Invalid: + effects[i] = std::make_unique<EffectStubbed>(); + break; + case EffectType::BufferMixer: + effects[i] = std::make_unique<EffectBufferMixer>(); + break; + case EffectType::Aux: + effects[i] = std::make_unique<EffectAuxInfo>(); + break; + case EffectType::Delay: + effects[i] = std::make_unique<EffectDelay>(); + break; + case EffectType::Reverb: + effects[i] = std::make_unique<EffectReverb>(); + break; + case EffectType::I3dl2Reverb: + effects[i] = std::make_unique<EffectI3dl2Reverb>(); + break; + case EffectType::BiquadFilter: + effects[i] = std::make_unique<EffectBiquadFilter>(); + break; + default: + UNREACHABLE_MSG("Unimplemented effect {}", effect); + effects[i] = std::make_unique<EffectStubbed>(); + } + return GetInfo(i); +} + +const EffectBase* EffectContext::GetInfo(std::size_t i) const { + return effects.at(i).get(); +} + +EffectStubbed::EffectStubbed() : EffectBase::EffectBase(EffectType::Invalid) {} +EffectStubbed::~EffectStubbed() = default; + +void EffectStubbed::Update(EffectInfo::InParams& in_params) {} +void EffectStubbed::UpdateForCommandGeneration() {} + +EffectBase::EffectBase(EffectType effect_type) : effect_type(effect_type) {} +EffectBase::~EffectBase() = default; + +UsageState EffectBase::GetUsage() const { + return usage; +} + +EffectType EffectBase::GetType() const { + return effect_type; +} + +bool EffectBase::IsEnabled() const { + return enabled; +} + +s32 EffectBase::GetMixID() const { + return mix_id; +} + +s32 EffectBase::GetProcessingOrder() const { + return processing_order; +} + +EffectI3dl2Reverb::EffectI3dl2Reverb() : EffectGeneric::EffectGeneric(EffectType::I3dl2Reverb) {} +EffectI3dl2Reverb::~EffectI3dl2Reverb() = default; + +void EffectI3dl2Reverb::Update(EffectInfo::InParams& in_params) { + auto& internal_params = GetParams(); + const auto* reverb_params = reinterpret_cast<I3dl2ReverbParams*>(in_params.raw.data()); + if (!ValidChannelCountForEffect(reverb_params->max_channels)) { + UNREACHABLE_MSG("Invalid reverb max channel count {}", reverb_params->max_channels); + return; + } + + const auto last_status = internal_params.status; + mix_id = in_params.mix_id; + processing_order = in_params.processing_order; + internal_params = *reverb_params; + if (!ValidChannelCountForEffect(reverb_params->channel_count)) { + internal_params.channel_count = internal_params.max_channels; + } + enabled = in_params.is_enabled; + if (last_status != ParameterStatus::Updated) { + internal_params.status = last_status; + } + + if (in_params.is_new || skipped) { + usage = UsageState::Initialized; + internal_params.status = ParameterStatus::Initialized; + skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0; + } +} + +void EffectI3dl2Reverb::UpdateForCommandGeneration() { + if (enabled) { + usage = UsageState::Running; + } else { + usage = UsageState::Stopped; + } + GetParams().status = ParameterStatus::Updated; +} + +EffectBiquadFilter::EffectBiquadFilter() : EffectGeneric::EffectGeneric(EffectType::BiquadFilter) {} +EffectBiquadFilter::~EffectBiquadFilter() = default; + +void EffectBiquadFilter::Update(EffectInfo::InParams& in_params) { + auto& internal_params = GetParams(); + const auto* biquad_params = reinterpret_cast<BiquadFilterParams*>(in_params.raw.data()); + mix_id = in_params.mix_id; + processing_order = in_params.processing_order; + internal_params = *biquad_params; + enabled = in_params.is_enabled; +} + +void EffectBiquadFilter::UpdateForCommandGeneration() { + if (enabled) { + usage = UsageState::Running; + } else { + usage = UsageState::Stopped; + } + GetParams().status = ParameterStatus::Updated; +} + +EffectAuxInfo::EffectAuxInfo() : EffectGeneric::EffectGeneric(EffectType::Aux) {} +EffectAuxInfo::~EffectAuxInfo() = default; + +void EffectAuxInfo::Update(EffectInfo::InParams& in_params) { + const auto* aux_params = reinterpret_cast<AuxInfo*>(in_params.raw.data()); + mix_id = in_params.mix_id; + processing_order = in_params.processing_order; + GetParams() = *aux_params; + enabled = in_params.is_enabled; + + if (in_params.is_new || skipped) { + skipped = aux_params->send_buffer_info == 0 || aux_params->return_buffer_info == 0; + if (skipped) { + return; + } + + // There's two AuxInfos which are an identical size, the first one is managed by the cpu, + // the second is managed by the dsp. All we care about is managing the DSP one + send_info = aux_params->send_buffer_info + sizeof(AuxInfoDSP); + send_buffer = aux_params->send_buffer_info + (sizeof(AuxInfoDSP) * 2); + + recv_info = aux_params->return_buffer_info + sizeof(AuxInfoDSP); + recv_buffer = aux_params->return_buffer_info + (sizeof(AuxInfoDSP) * 2); + } +} + +void EffectAuxInfo::UpdateForCommandGeneration() { + if (enabled) { + usage = UsageState::Running; + } else { + usage = UsageState::Stopped; + } +} + +VAddr EffectAuxInfo::GetSendInfo() const { + return send_info; +} + +VAddr EffectAuxInfo::GetSendBuffer() const { + return send_buffer; +} + +VAddr EffectAuxInfo::GetRecvInfo() const { + return recv_info; +} + +VAddr EffectAuxInfo::GetRecvBuffer() const { + return recv_buffer; +} + +EffectDelay::EffectDelay() : EffectGeneric::EffectGeneric(EffectType::Delay) {} +EffectDelay::~EffectDelay() = default; + +void EffectDelay::Update(EffectInfo::InParams& in_params) { + const auto* delay_params = reinterpret_cast<DelayParams*>(in_params.raw.data()); + auto& internal_params = GetParams(); + if (!ValidChannelCountForEffect(delay_params->max_channels)) { + return; + } + + const auto last_status = internal_params.status; + mix_id = in_params.mix_id; + processing_order = in_params.processing_order; + internal_params = *delay_params; + if (!ValidChannelCountForEffect(delay_params->channels)) { + internal_params.channels = internal_params.max_channels; + } + enabled = in_params.is_enabled; + + if (last_status != ParameterStatus::Updated) { + internal_params.status = last_status; + } + + if (in_params.is_new || skipped) { + usage = UsageState::Initialized; + internal_params.status = ParameterStatus::Initialized; + skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0; + } +} + +void EffectDelay::UpdateForCommandGeneration() { + if (enabled) { + usage = UsageState::Running; + } else { + usage = UsageState::Stopped; + } + GetParams().status = ParameterStatus::Updated; +} + +EffectBufferMixer::EffectBufferMixer() : EffectGeneric::EffectGeneric(EffectType::BufferMixer) {} +EffectBufferMixer::~EffectBufferMixer() = default; + +void EffectBufferMixer::Update(EffectInfo::InParams& in_params) { + mix_id = in_params.mix_id; + processing_order = in_params.processing_order; + GetParams() = *reinterpret_cast<BufferMixerParams*>(in_params.raw.data()); + enabled = in_params.is_enabled; +} + +void EffectBufferMixer::UpdateForCommandGeneration() { + if (enabled) { + usage = UsageState::Running; + } else { + usage = UsageState::Stopped; + } +} + +EffectReverb::EffectReverb() : EffectGeneric::EffectGeneric(EffectType::Reverb) {} +EffectReverb::~EffectReverb() = default; + +void EffectReverb::Update(EffectInfo::InParams& in_params) { + const auto* reverb_params = reinterpret_cast<ReverbParams*>(in_params.raw.data()); + auto& internal_params = GetParams(); + if (!ValidChannelCountForEffect(reverb_params->max_channels)) { + return; + } + + const auto last_status = internal_params.status; + mix_id = in_params.mix_id; + processing_order = in_params.processing_order; + internal_params = *reverb_params; + if (!ValidChannelCountForEffect(reverb_params->channels)) { + internal_params.channels = internal_params.max_channels; + } + enabled = in_params.is_enabled; + + if (last_status != ParameterStatus::Updated) { + internal_params.status = last_status; + } + + if (in_params.is_new || skipped) { + usage = UsageState::Initialized; + internal_params.status = ParameterStatus::Initialized; + skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0; + } +} + +void EffectReverb::UpdateForCommandGeneration() { + if (enabled) { + usage = UsageState::Running; + } else { + usage = UsageState::Stopped; + } + GetParams().status = ParameterStatus::Updated; +} + +} // namespace AudioCore diff --git a/src/audio_core/effect_context.h b/src/audio_core/effect_context.h new file mode 100644 index 000000000..2c4ce53ef --- /dev/null +++ b/src/audio_core/effect_context.h @@ -0,0 +1,321 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <memory> +#include <vector> +#include "audio_core/common.h" +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" + +namespace AudioCore { +enum class EffectType : u8 { + Invalid = 0, + BufferMixer = 1, + Aux = 2, + Delay = 3, + Reverb = 4, + I3dl2Reverb = 5, + BiquadFilter = 6, +}; + +enum class UsageStatus : u8 { + Invalid = 0, + New = 1, + Initialized = 2, + Used = 3, + Removed = 4, +}; + +enum class UsageState { + Invalid = 0, + Initialized = 1, + Running = 2, + Stopped = 3, +}; + +enum class ParameterStatus : u8 { + Initialized = 0, + Updating = 1, + Updated = 2, +}; + +struct BufferMixerParams { + std::array<s8, AudioCommon::MAX_MIX_BUFFERS> input{}; + std::array<s8, AudioCommon::MAX_MIX_BUFFERS> output{}; + std::array<float_le, AudioCommon::MAX_MIX_BUFFERS> volume{}; + s32_le count{}; +}; +static_assert(sizeof(BufferMixerParams) == 0x94, "BufferMixerParams is an invalid size"); + +struct AuxInfoDSP { + u32_le read_offset{}; + u32_le write_offset{}; + u32_le remaining{}; + INSERT_PADDING_WORDS(13); +}; +static_assert(sizeof(AuxInfoDSP) == 0x40, "AuxInfoDSP is an invalid size"); + +struct AuxInfo { + std::array<s8, AudioCommon::MAX_MIX_BUFFERS> input_mix_buffers{}; + std::array<s8, AudioCommon::MAX_MIX_BUFFERS> output_mix_buffers{}; + u32_le count{}; + s32_le sample_rate{}; + s32_le sample_count{}; + s32_le mix_buffer_count{}; + u64_le send_buffer_info{}; + u64_le send_buffer_base{}; + + u64_le return_buffer_info{}; + u64_le return_buffer_base{}; +}; +static_assert(sizeof(AuxInfo) == 0x60, "AuxInfo is an invalid size"); + +struct I3dl2ReverbParams { + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> input{}; + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> output{}; + u16_le max_channels{}; + u16_le channel_count{}; + INSERT_PADDING_BYTES(1); + u32_le sample_rate{}; + f32 room_hf{}; + f32 hf_reference{}; + f32 decay_time{}; + f32 hf_decay_ratio{}; + f32 room{}; + f32 reflection{}; + f32 reverb{}; + f32 diffusion{}; + f32 reflection_delay{}; + f32 reverb_delay{}; + f32 density{}; + f32 dry_gain{}; + ParameterStatus status{}; + INSERT_PADDING_BYTES(3); +}; +static_assert(sizeof(I3dl2ReverbParams) == 0x4c, "I3dl2ReverbParams is an invalid size"); + +struct BiquadFilterParams { + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> input{}; + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> output{}; + std::array<s16_le, 3> numerator; + std::array<s16_le, 2> denominator; + s8 channel_count{}; + ParameterStatus status{}; +}; +static_assert(sizeof(BiquadFilterParams) == 0x18, "BiquadFilterParams is an invalid size"); + +struct DelayParams { + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> input{}; + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> output{}; + u16_le max_channels{}; + u16_le channels{}; + s32_le max_delay{}; + s32_le delay{}; + s32_le sample_rate{}; + s32_le gain{}; + s32_le feedback_gain{}; + s32_le out_gain{}; + s32_le dry_gain{}; + s32_le channel_spread{}; + s32_le low_pass{}; + ParameterStatus status{}; + INSERT_PADDING_BYTES(3); +}; +static_assert(sizeof(DelayParams) == 0x38, "DelayParams is an invalid size"); + +struct ReverbParams { + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> input{}; + std::array<s8, AudioCommon::MAX_CHANNEL_COUNT> output{}; + u16_le max_channels{}; + u16_le channels{}; + s32_le sample_rate{}; + s32_le mode0{}; + s32_le mode0_gain{}; + s32_le pre_delay{}; + s32_le mode1{}; + s32_le mode1_gain{}; + s32_le decay{}; + s32_le hf_decay_ratio{}; + s32_le coloration{}; + s32_le reverb_gain{}; + s32_le out_gain{}; + s32_le dry_gain{}; + ParameterStatus status{}; + INSERT_PADDING_BYTES(3); +}; +static_assert(sizeof(ReverbParams) == 0x44, "ReverbParams is an invalid size"); + +class EffectInfo { +public: + struct InParams { + EffectType type{}; + u8 is_new{}; + u8 is_enabled{}; + INSERT_PADDING_BYTES(1); + s32_le mix_id{}; + u64_le buffer_address{}; + u64_le buffer_size{}; + s32_le processing_order{}; + INSERT_PADDING_BYTES(4); + union { + std::array<u8, 0xa0> raw; + }; + }; + static_assert(sizeof(InParams) == 0xc0, "InParams is an invalid size"); + + struct OutParams { + UsageStatus status{}; + INSERT_PADDING_BYTES(15); + }; + static_assert(sizeof(OutParams) == 0x10, "OutParams is an invalid size"); +}; + +struct AuxAddress { + VAddr send_dsp_info{}; + VAddr send_buffer_base{}; + VAddr return_dsp_info{}; + VAddr return_buffer_base{}; +}; + +class EffectBase { +public: + explicit EffectBase(EffectType effect_type); + virtual ~EffectBase(); + + virtual void Update(EffectInfo::InParams& in_params) = 0; + virtual void UpdateForCommandGeneration() = 0; + UsageState GetUsage() const; + EffectType GetType() const; + bool IsEnabled() const; + s32 GetMixID() const; + s32 GetProcessingOrder() const; + +protected: + UsageState usage{UsageState::Invalid}; + EffectType effect_type{}; + s32 mix_id{}; + s32 processing_order{}; + bool enabled = false; +}; + +template <typename T> +class EffectGeneric : public EffectBase { +public: + explicit EffectGeneric(EffectType effect_type) : EffectBase(effect_type) {} + + T& GetParams() { + return internal_params; + } + + const I3dl2ReverbParams& GetParams() const { + return internal_params; + } + +private: + T internal_params{}; +}; + +class EffectStubbed : public EffectBase { +public: + explicit EffectStubbed(); + ~EffectStubbed() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; +}; + +class EffectI3dl2Reverb : public EffectGeneric<I3dl2ReverbParams> { +public: + explicit EffectI3dl2Reverb(); + ~EffectI3dl2Reverb() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; + +private: + bool skipped = false; +}; + +class EffectBiquadFilter : public EffectGeneric<BiquadFilterParams> { +public: + explicit EffectBiquadFilter(); + ~EffectBiquadFilter() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; +}; + +class EffectAuxInfo : public EffectGeneric<AuxInfo> { +public: + explicit EffectAuxInfo(); + ~EffectAuxInfo() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; + VAddr GetSendInfo() const; + VAddr GetSendBuffer() const; + VAddr GetRecvInfo() const; + VAddr GetRecvBuffer() const; + +private: + VAddr send_info{}; + VAddr send_buffer{}; + VAddr recv_info{}; + VAddr recv_buffer{}; + bool skipped = false; + AuxAddress addresses{}; +}; + +class EffectDelay : public EffectGeneric<DelayParams> { +public: + explicit EffectDelay(); + ~EffectDelay() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; + +private: + bool skipped = false; +}; + +class EffectBufferMixer : public EffectGeneric<BufferMixerParams> { +public: + explicit EffectBufferMixer(); + ~EffectBufferMixer() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; +}; + +class EffectReverb : public EffectGeneric<ReverbParams> { +public: + explicit EffectReverb(); + ~EffectReverb() override; + + void Update(EffectInfo::InParams& in_params) override; + void UpdateForCommandGeneration() override; + +private: + bool skipped = false; +}; + +class EffectContext { +public: + explicit EffectContext(std::size_t effect_count); + ~EffectContext(); + + std::size_t GetCount() const; + EffectBase* GetInfo(std::size_t i); + EffectBase* RetargetEffect(std::size_t i, EffectType effect); + const EffectBase* GetInfo(std::size_t i) const; + +private: + std::size_t effect_count{}; + std::vector<std::unique_ptr<EffectBase>> effects; +}; +} // namespace AudioCore diff --git a/src/audio_core/info_updater.cpp b/src/audio_core/info_updater.cpp new file mode 100644 index 000000000..2940e53a9 --- /dev/null +++ b/src/audio_core/info_updater.cpp @@ -0,0 +1,516 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/behavior_info.h" +#include "audio_core/effect_context.h" +#include "audio_core/info_updater.h" +#include "audio_core/memory_pool.h" +#include "audio_core/mix_context.h" +#include "audio_core/sink_context.h" +#include "audio_core/splitter_context.h" +#include "audio_core/voice_context.h" +#include "common/logging/log.h" + +namespace AudioCore { + +InfoUpdater::InfoUpdater(const std::vector<u8>& in_params, std::vector<u8>& out_params, + BehaviorInfo& behavior_info) + : in_params(in_params), out_params(out_params), behavior_info(behavior_info) { + ASSERT( + AudioCommon::CanConsumeBuffer(in_params.size(), 0, sizeof(AudioCommon::UpdateDataHeader))); + std::memcpy(&input_header, in_params.data(), sizeof(AudioCommon::UpdateDataHeader)); + output_header.total_size = sizeof(AudioCommon::UpdateDataHeader); +} + +InfoUpdater::~InfoUpdater() = default; + +bool InfoUpdater::UpdateBehaviorInfo(BehaviorInfo& in_behavior_info) { + if (input_header.size.behavior != sizeof(BehaviorInfo::InParams)) { + LOG_ERROR(Audio, "Behavior info is an invalid size, expecting 0x{:X} but got 0x{:X}", + sizeof(BehaviorInfo::InParams), input_header.size.behavior); + return false; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, + sizeof(BehaviorInfo::InParams))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + BehaviorInfo::InParams behavior_in{}; + std::memcpy(&behavior_in, in_params.data() + input_offset, sizeof(BehaviorInfo::InParams)); + input_offset += sizeof(BehaviorInfo::InParams); + + // Make sure it's an audio revision we can actually support + if (!AudioCommon::IsValidRevision(behavior_in.revision)) { + LOG_ERROR(Audio, "Invalid input revision, revision=0x{:08X}", behavior_in.revision); + return false; + } + + // Make sure that our behavior info revision matches the input + if (in_behavior_info.GetUserRevision() != behavior_in.revision) { + LOG_ERROR(Audio, + "User revision differs from input revision, expecting 0x{:08X} but got 0x{:08X}", + in_behavior_info.GetUserRevision(), behavior_in.revision); + return false; + } + + // Update behavior info flags + in_behavior_info.ClearError(); + in_behavior_info.UpdateFlags(behavior_in.flags); + + return true; +} + +bool InfoUpdater::UpdateMemoryPools(std::vector<ServerMemoryPoolInfo>& memory_pool_info) { + const auto memory_pool_count = memory_pool_info.size(); + const auto total_memory_pool_in = sizeof(ServerMemoryPoolInfo::InParams) * memory_pool_count; + const auto total_memory_pool_out = sizeof(ServerMemoryPoolInfo::OutParams) * memory_pool_count; + + if (input_header.size.memory_pool != total_memory_pool_in) { + LOG_ERROR(Audio, "Memory pools are an invalid size, expecting 0x{:X} but got 0x{:X}", + total_memory_pool_in, input_header.size.memory_pool); + return false; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, total_memory_pool_in)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::vector<ServerMemoryPoolInfo::InParams> mempool_in(memory_pool_count); + std::vector<ServerMemoryPoolInfo::OutParams> mempool_out(memory_pool_count); + + std::memcpy(mempool_in.data(), in_params.data() + input_offset, total_memory_pool_in); + input_offset += total_memory_pool_in; + + // Update our memory pools + for (std::size_t i = 0; i < memory_pool_count; i++) { + if (!memory_pool_info[i].Update(mempool_in[i], mempool_out[i])) { + LOG_ERROR(Audio, "Failed to update memory pool {}!", i); + return false; + } + } + + if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, + sizeof(BehaviorInfo::InParams))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::memcpy(out_params.data() + output_offset, mempool_out.data(), total_memory_pool_out); + output_offset += total_memory_pool_out; + output_header.size.memory_pool = static_cast<u32>(total_memory_pool_out); + return true; +} + +bool InfoUpdater::UpdateVoiceChannelResources(VoiceContext& voice_context) { + const auto voice_count = voice_context.GetVoiceCount(); + const auto voice_size = voice_count * sizeof(VoiceChannelResource::InParams); + std::vector<VoiceChannelResource::InParams> resources_in(voice_count); + + if (input_header.size.voice_channel_resource != voice_size) { + LOG_ERROR(Audio, "VoiceChannelResource is an invalid size, expecting 0x{:X} but got 0x{:X}", + voice_size, input_header.size.voice_channel_resource); + return false; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, voice_size)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::memcpy(resources_in.data(), in_params.data() + input_offset, voice_size); + input_offset += voice_size; + + // Update our channel resources + for (std::size_t i = 0; i < voice_count; i++) { + // Grab our channel resource + auto& resource = voice_context.GetChannelResource(i); + resource.Update(resources_in[i]); + } + + return true; +} + +bool InfoUpdater::UpdateVoices(VoiceContext& voice_context, + std::vector<ServerMemoryPoolInfo>& memory_pool_info, + VAddr audio_codec_dsp_addr) { + const auto voice_count = voice_context.GetVoiceCount(); + std::vector<VoiceInfo::InParams> voice_in(voice_count); + std::vector<VoiceInfo::OutParams> voice_out(voice_count); + + const auto voice_in_size = voice_count * sizeof(VoiceInfo::InParams); + const auto voice_out_size = voice_count * sizeof(VoiceInfo::OutParams); + + if (input_header.size.voice != voice_in_size) { + LOG_ERROR(Audio, "Voices are an invalid size, expecting 0x{:X} but got 0x{:X}", + voice_in_size, input_header.size.voice); + return false; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, voice_in_size)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::memcpy(voice_in.data(), in_params.data() + input_offset, voice_in_size); + input_offset += voice_in_size; + + // Set all voices to not be in use + for (std::size_t i = 0; i < voice_count; i++) { + voice_context.GetInfo(i).GetInParams().in_use = false; + } + + // Update our voices + for (std::size_t i = 0; i < voice_count; i++) { + auto& in_params = voice_in[i]; + const auto channel_count = static_cast<std::size_t>(in_params.channel_count); + // Skip if it's not currently in use + if (!in_params.is_in_use) { + continue; + } + // Voice states for each channel + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT> voice_states{}; + ASSERT(static_cast<std::size_t>(in_params.id) < voice_count); + + // Grab our current voice info + auto& voice_info = voice_context.GetInfo(static_cast<std::size_t>(in_params.id)); + + ASSERT(channel_count <= AudioCommon::MAX_CHANNEL_COUNT); + + // Get all our channel voice states + for (std::size_t channel = 0; channel < channel_count; channel++) { + voice_states[channel] = + &voice_context.GetState(in_params.voice_channel_resource_ids[channel]); + } + + if (in_params.is_new) { + // Default our values for our voice + voice_info.Initialize(); + if (channel_count == 0 || channel_count > AudioCommon::MAX_CHANNEL_COUNT) { + continue; + } + + // Zero out our voice states + for (std::size_t channel = 0; channel < channel_count; channel++) { + std::memset(voice_states[channel], 0, sizeof(VoiceState)); + } + } + + // Update our voice + voice_info.UpdateParameters(in_params, behavior_info); + // TODO(ogniK): Handle mapping errors with behavior info based on in params response + + // Update our wave buffers + voice_info.UpdateWaveBuffers(in_params, voice_states, behavior_info); + voice_info.WriteOutStatus(voice_out[i], in_params, voice_states); + } + + if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, voice_out_size)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + std::memcpy(out_params.data() + output_offset, voice_out.data(), voice_out_size); + output_offset += voice_out_size; + output_header.size.voice = static_cast<u32>(voice_out_size); + return true; +} + +bool InfoUpdater::UpdateEffects(EffectContext& effect_context, bool is_active) { + const auto effect_count = effect_context.GetCount(); + std::vector<EffectInfo::InParams> effect_in(effect_count); + std::vector<EffectInfo::OutParams> effect_out(effect_count); + + const auto total_effect_in = effect_count * sizeof(EffectInfo::InParams); + const auto total_effect_out = effect_count * sizeof(EffectInfo::OutParams); + + if (input_header.size.effect != total_effect_in) { + LOG_ERROR(Audio, "Effects are an invalid size, expecting 0x{:X} but got 0x{:X}", + total_effect_in, input_header.size.effect); + return false; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, total_effect_in)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::memcpy(effect_in.data(), in_params.data() + input_offset, total_effect_in); + input_offset += total_effect_in; + + // Update effects + for (std::size_t i = 0; i < effect_count; i++) { + auto* info = effect_context.GetInfo(i); + if (effect_in[i].type != info->GetType()) { + info = effect_context.RetargetEffect(i, effect_in[i].type); + } + + info->Update(effect_in[i]); + + if ((!is_active && info->GetUsage() != UsageState::Initialized) || + info->GetUsage() == UsageState::Stopped) { + effect_out[i].status = UsageStatus::Removed; + } else { + effect_out[i].status = UsageStatus::Used; + } + } + + if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, total_effect_out)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::memcpy(out_params.data() + output_offset, effect_out.data(), total_effect_out); + output_offset += total_effect_out; + output_header.size.effect = static_cast<u32>(total_effect_out); + + return true; +} + +bool InfoUpdater::UpdateSplitterInfo(SplitterContext& splitter_context) { + std::size_t start_offset = input_offset; + std::size_t bytes_read{}; + // Update splitter context + if (!splitter_context.Update(in_params, input_offset, bytes_read)) { + LOG_ERROR(Audio, "Failed to update splitter context!"); + return false; + } + + const auto consumed = input_offset - start_offset; + + if (input_header.size.splitter != consumed) { + LOG_ERROR(Audio, "Splitters is an invalid size, expecting 0x{:X} but got 0x{:X}", + bytes_read, input_header.size.splitter); + return false; + } + + return true; +} + +ResultCode InfoUpdater::UpdateMixes(MixContext& mix_context, std::size_t mix_buffer_count, + SplitterContext& splitter_context, + EffectContext& effect_context) { + std::vector<MixInfo::InParams> mix_in_params; + + if (!behavior_info.IsMixInParameterDirtyOnlyUpdateSupported()) { + // If we're not dirty, get ALL mix in parameters + const auto context_mix_count = mix_context.GetCount(); + const auto total_mix_in = context_mix_count * sizeof(MixInfo::InParams); + if (input_header.size.mixer != total_mix_in) { + LOG_ERROR(Audio, "Mixer is an invalid size, expecting 0x{:X} but got 0x{:X}", + total_mix_in, input_header.size.mixer); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, total_mix_in)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + + mix_in_params.resize(context_mix_count); + std::memcpy(mix_in_params.data(), in_params.data() + input_offset, total_mix_in); + + input_offset += total_mix_in; + } else { + // Only update the "dirty" mixes + MixInfo::DirtyHeader dirty_header{}; + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, + sizeof(MixInfo::DirtyHeader))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + + std::memcpy(&dirty_header, in_params.data() + input_offset, sizeof(MixInfo::DirtyHeader)); + input_offset += sizeof(MixInfo::DirtyHeader); + + const auto total_mix_in = + dirty_header.mixer_count * sizeof(MixInfo::InParams) + sizeof(MixInfo::DirtyHeader); + + if (input_header.size.mixer != total_mix_in) { + LOG_ERROR(Audio, "Mixer is an invalid size, expecting 0x{:X} but got 0x{:X}", + total_mix_in, input_header.size.mixer); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + + if (dirty_header.mixer_count != 0) { + mix_in_params.resize(dirty_header.mixer_count); + std::memcpy(mix_in_params.data(), in_params.data() + input_offset, + mix_in_params.size() * sizeof(MixInfo::InParams)); + input_offset += mix_in_params.size() * sizeof(MixInfo::InParams); + } + } + + // Get our total input count + const auto mix_count = mix_in_params.size(); + + if (!behavior_info.IsMixInParameterDirtyOnlyUpdateSupported()) { + // Only verify our buffer count if we're not dirty + std::size_t total_buffer_count{}; + for (std::size_t i = 0; i < mix_count; i++) { + const auto& in = mix_in_params[i]; + total_buffer_count += in.buffer_count; + if (static_cast<std::size_t>(in.dest_mix_id) > mix_count && + in.dest_mix_id != AudioCommon::NO_MIX && in.mix_id != AudioCommon::FINAL_MIX) { + LOG_ERROR( + Audio, + "Invalid mix destination, mix_id={:X}, dest_mix_id={:X}, mix_buffer_count={:X}", + in.mix_id, in.dest_mix_id, mix_buffer_count); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + } + + if (total_buffer_count > mix_buffer_count) { + LOG_ERROR(Audio, + "Too many mix buffers used! mix_buffer_count={:X}, requesting_buffers={:X}", + mix_buffer_count, total_buffer_count); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + } + + if (mix_buffer_count == 0) { + LOG_ERROR(Audio, "No mix buffers!"); + return AudioCommon::Audren::ERR_INVALID_PARAMETERS; + } + + bool should_sort = false; + for (std::size_t i = 0; i < mix_count; i++) { + const auto& mix_in = mix_in_params[i]; + std::size_t target_mix{}; + if (behavior_info.IsMixInParameterDirtyOnlyUpdateSupported()) { + target_mix = mix_in.mix_id; + } else { + // Non dirty supported games just use i instead of the actual mix_id + target_mix = i; + } + auto& mix_info = mix_context.GetInfo(target_mix); + auto& mix_info_params = mix_info.GetInParams(); + if (mix_info_params.in_use != mix_in.in_use) { + mix_info_params.in_use = mix_in.in_use; + mix_info.ResetEffectProcessingOrder(); + should_sort = true; + } + + if (mix_in.in_use) { + should_sort |= mix_info.Update(mix_context.GetEdgeMatrix(), mix_in, behavior_info, + splitter_context, effect_context); + } + } + + if (should_sort && behavior_info.IsSplitterSupported()) { + // Sort our splitter data + if (!mix_context.TsortInfo(splitter_context)) { + return AudioCommon::Audren::ERR_SPLITTER_SORT_FAILED; + } + } + + // TODO(ogniK): Sort when splitter is suppoorted + + return RESULT_SUCCESS; +} + +bool InfoUpdater::UpdateSinks(SinkContext& sink_context) { + const auto sink_count = sink_context.GetCount(); + std::vector<SinkInfo::InParams> sink_in_params(sink_count); + const auto total_sink_in = sink_count * sizeof(SinkInfo::InParams); + + if (input_header.size.sink != total_sink_in) { + LOG_ERROR(Audio, "Sinks are an invalid size, expecting 0x{:X} but got 0x{:X}", + total_sink_in, input_header.size.effect); + return false; + } + + if (!AudioCommon::CanConsumeBuffer(in_params.size(), input_offset, total_sink_in)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + std::memcpy(sink_in_params.data(), in_params.data() + input_offset, total_sink_in); + input_offset += total_sink_in; + + // TODO(ogniK): Properly update sinks + if (!sink_in_params.empty()) { + sink_context.UpdateMainSink(sink_in_params[0]); + } + + output_header.size.sink = static_cast<u32>(0x20 * sink_count); + output_offset += 0x20 * sink_count; + return true; +} + +bool InfoUpdater::UpdatePerformanceBuffer() { + output_header.size.performance = 0x10; + output_offset += 0x10; + return true; +} + +bool InfoUpdater::UpdateErrorInfo(BehaviorInfo& in_behavior_info) { + const auto total_beahvior_info_out = sizeof(BehaviorInfo::OutParams); + + if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, total_beahvior_info_out)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + + BehaviorInfo::OutParams behavior_info_out{}; + behavior_info.CopyErrorInfo(behavior_info_out); + + std::memcpy(out_params.data() + output_offset, &behavior_info_out, total_beahvior_info_out); + output_offset += total_beahvior_info_out; + output_header.size.behavior = total_beahvior_info_out; + + return true; +} + +struct RendererInfo { + u64_le elasped_frame_count{}; + INSERT_PADDING_WORDS(2); +}; +static_assert(sizeof(RendererInfo) == 0x10, "RendererInfo is an invalid size"); + +bool InfoUpdater::UpdateRendererInfo(std::size_t elapsed_frame_count) { + const auto total_renderer_info_out = sizeof(RendererInfo); + if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, total_renderer_info_out)) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + RendererInfo out{}; + out.elasped_frame_count = elapsed_frame_count; + std::memcpy(out_params.data() + output_offset, &out, total_renderer_info_out); + output_offset += total_renderer_info_out; + output_header.size.render_info = total_renderer_info_out; + + return true; +} + +bool InfoUpdater::CheckConsumedSize() const { + if (output_offset != out_params.size()) { + LOG_ERROR(Audio, "Output is not consumed! Consumed {}, but requires {}. {} bytes remaining", + output_offset, out_params.size(), out_params.size() - output_offset); + return false; + } + /*if (input_offset != in_params.size()) { + LOG_ERROR(Audio, "Input is not consumed!"); + return false; + }*/ + return true; +} + +bool InfoUpdater::WriteOutputHeader() { + if (!AudioCommon::CanConsumeBuffer(out_params.size(), 0, + sizeof(AudioCommon::UpdateDataHeader))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + output_header.revision = AudioCommon::CURRENT_PROCESS_REVISION; + const auto& sz = output_header.size; + output_header.total_size += sz.behavior + sz.memory_pool + sz.voice + + sz.voice_channel_resource + sz.effect + sz.mixer + sz.sink + + sz.performance + sz.splitter + sz.render_info; + + std::memcpy(out_params.data(), &output_header, sizeof(AudioCommon::UpdateDataHeader)); + return true; +} + +} // namespace AudioCore diff --git a/src/audio_core/info_updater.h b/src/audio_core/info_updater.h new file mode 100644 index 000000000..06f9d770f --- /dev/null +++ b/src/audio_core/info_updater.h @@ -0,0 +1,58 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "audio_core/common.h" +#include "common/common_types.h" + +namespace AudioCore { + +class BehaviorInfo; +class ServerMemoryPoolInfo; +class VoiceContext; +class EffectContext; +class MixContext; +class SinkContext; +class SplitterContext; + +class InfoUpdater { +public: + // TODO(ogniK): Pass process handle when we support it + InfoUpdater(const std::vector<u8>& in_params, std::vector<u8>& out_params, + BehaviorInfo& behavior_info); + ~InfoUpdater(); + + bool UpdateBehaviorInfo(BehaviorInfo& in_behavior_info); + bool UpdateMemoryPools(std::vector<ServerMemoryPoolInfo>& memory_pool_info); + bool UpdateVoiceChannelResources(VoiceContext& voice_context); + bool UpdateVoices(VoiceContext& voice_context, + std::vector<ServerMemoryPoolInfo>& memory_pool_info, + VAddr audio_codec_dsp_addr); + bool UpdateEffects(EffectContext& effect_context, bool is_active); + bool UpdateSplitterInfo(SplitterContext& splitter_context); + ResultCode UpdateMixes(MixContext& mix_context, std::size_t mix_buffer_count, + SplitterContext& splitter_context, EffectContext& effect_context); + bool UpdateSinks(SinkContext& sink_context); + bool UpdatePerformanceBuffer(); + bool UpdateErrorInfo(BehaviorInfo& in_behavior_info); + bool UpdateRendererInfo(std::size_t elapsed_frame_count); + bool CheckConsumedSize() const; + + bool WriteOutputHeader(); + +private: + const std::vector<u8>& in_params; + std::vector<u8>& out_params; + BehaviorInfo& behavior_info; + + AudioCommon::UpdateDataHeader input_header{}; + AudioCommon::UpdateDataHeader output_header{}; + + std::size_t input_offset{sizeof(AudioCommon::UpdateDataHeader)}; + std::size_t output_offset{sizeof(AudioCommon::UpdateDataHeader)}; +}; + +} // namespace AudioCore diff --git a/src/audio_core/memory_pool.cpp b/src/audio_core/memory_pool.cpp new file mode 100644 index 000000000..5a3453063 --- /dev/null +++ b/src/audio_core/memory_pool.cpp @@ -0,0 +1,62 @@ + +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/memory_pool.h" +#include "common/logging/log.h" + +namespace AudioCore { + +ServerMemoryPoolInfo::ServerMemoryPoolInfo() = default; +ServerMemoryPoolInfo::~ServerMemoryPoolInfo() = default; +bool ServerMemoryPoolInfo::Update(const ServerMemoryPoolInfo::InParams& in_params, + ServerMemoryPoolInfo::OutParams& out_params) { + // Our state does not need to be changed + if (in_params.state != ServerMemoryPoolInfo::State::RequestAttach && + in_params.state != ServerMemoryPoolInfo::State::RequestDetach) { + return true; + } + + // Address or size is null + if (in_params.address == 0 || in_params.size == 0) { + LOG_ERROR(Audio, "Memory pool address or size is zero! address={:X}, size={:X}", + in_params.address, in_params.size); + return false; + } + + // Address or size is not aligned + if ((in_params.address % 0x1000) != 0 || (in_params.size % 0x1000) != 0) { + LOG_ERROR(Audio, "Memory pool address or size is not aligned! address={:X}, size={:X}", + in_params.address, in_params.size); + return false; + } + + if (in_params.state == ServerMemoryPoolInfo::State::RequestAttach) { + cpu_address = in_params.address; + size = in_params.size; + used = true; + out_params.state = ServerMemoryPoolInfo::State::Attached; + } else { + // Unexpected address + if (cpu_address != in_params.address) { + LOG_ERROR(Audio, "Memory pool address differs! Expecting {:X} but address is {:X}", + cpu_address, in_params.address); + return false; + } + + if (size != in_params.size) { + LOG_ERROR(Audio, "Memory pool size differs! Expecting {:X} but size is {:X}", size, + in_params.size); + return false; + } + + cpu_address = 0; + size = 0; + used = false; + out_params.state = ServerMemoryPoolInfo::State::Detached; + } + return true; +} + +} // namespace AudioCore diff --git a/src/audio_core/memory_pool.h b/src/audio_core/memory_pool.h new file mode 100644 index 000000000..8ac503f1c --- /dev/null +++ b/src/audio_core/memory_pool.h @@ -0,0 +1,53 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" + +namespace AudioCore { + +class ServerMemoryPoolInfo { +public: + ServerMemoryPoolInfo(); + ~ServerMemoryPoolInfo(); + + enum class State : u32_le { + Invalid = 0x0, + Aquired = 0x1, + RequestDetach = 0x2, + Detached = 0x3, + RequestAttach = 0x4, + Attached = 0x5, + Released = 0x6, + }; + + struct InParams { + u64_le address{}; + u64_le size{}; + ServerMemoryPoolInfo::State state{}; + INSERT_PADDING_WORDS(3); + }; + static_assert(sizeof(ServerMemoryPoolInfo::InParams) == 0x20, "InParams are an invalid size"); + + struct OutParams { + ServerMemoryPoolInfo::State state{}; + INSERT_PADDING_WORDS(3); + }; + static_assert(sizeof(ServerMemoryPoolInfo::OutParams) == 0x10, "OutParams are an invalid size"); + + bool Update(const ServerMemoryPoolInfo::InParams& in_params, + ServerMemoryPoolInfo::OutParams& out_params); + +private: + // There's another entry here which is the DSP address, however since we're not talking to the + // DSP we can just use the same address provided by the guest without needing to remap + u64_le cpu_address{}; + u64_le size{}; + bool used{}; +}; + +} // namespace AudioCore diff --git a/src/audio_core/mix_context.cpp b/src/audio_core/mix_context.cpp new file mode 100644 index 000000000..4bca72eb0 --- /dev/null +++ b/src/audio_core/mix_context.cpp @@ -0,0 +1,296 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/behavior_info.h" +#include "audio_core/common.h" +#include "audio_core/effect_context.h" +#include "audio_core/mix_context.h" +#include "audio_core/splitter_context.h" + +namespace AudioCore { +MixContext::MixContext() = default; +MixContext::~MixContext() = default; + +void MixContext::Initialize(const BehaviorInfo& behavior_info, std::size_t mix_count, + std::size_t effect_count) { + info_count = mix_count; + infos.resize(info_count); + auto& final_mix = GetInfo(AudioCommon::FINAL_MIX); + final_mix.GetInParams().mix_id = AudioCommon::FINAL_MIX; + sorted_info.reserve(infos.size()); + for (auto& info : infos) { + sorted_info.push_back(&info); + } + + for (auto& info : infos) { + info.SetEffectCount(effect_count); + } + + // Only initialize our edge matrix and node states if splitters are supported + if (behavior_info.IsSplitterSupported()) { + node_states.Initialize(mix_count); + edge_matrix.Initialize(mix_count); + } +} + +void MixContext::UpdateDistancesFromFinalMix() { + // Set all distances to be invalid + for (std::size_t i = 0; i < info_count; i++) { + GetInfo(i).GetInParams().final_mix_distance = AudioCommon::NO_FINAL_MIX; + } + + for (std::size_t i = 0; i < info_count; i++) { + auto& info = GetInfo(i); + auto& in_params = info.GetInParams(); + // Populate our sorted info + sorted_info[i] = &info; + + if (!in_params.in_use) { + continue; + } + + auto mix_id = in_params.mix_id; + // Needs to be referenced out of scope + s32 distance_to_final_mix{AudioCommon::FINAL_MIX}; + for (; distance_to_final_mix < static_cast<s32>(info_count); distance_to_final_mix++) { + if (mix_id == AudioCommon::FINAL_MIX) { + // If we're at the final mix, we're done + break; + } else if (mix_id == AudioCommon::NO_MIX) { + // If we have no more mix ids, we're done + distance_to_final_mix = AudioCommon::NO_FINAL_MIX; + break; + } else { + const auto& dest_mix = GetInfo(mix_id); + const auto dest_mix_distance = dest_mix.GetInParams().final_mix_distance; + + if (dest_mix_distance == AudioCommon::NO_FINAL_MIX) { + // If our current mix isn't pointing to a final mix, follow through + mix_id = dest_mix.GetInParams().dest_mix_id; + } else { + // Our current mix + 1 = final distance + distance_to_final_mix = dest_mix_distance + 1; + break; + } + } + } + + // If we're out of range for our distance, mark it as no final mix + if (distance_to_final_mix >= static_cast<s32>(info_count)) { + distance_to_final_mix = AudioCommon::NO_FINAL_MIX; + } + + in_params.final_mix_distance = distance_to_final_mix; + } +} + +void MixContext::CalcMixBufferOffset() { + s32 offset{}; + for (std::size_t i = 0; i < info_count; i++) { + auto& info = GetSortedInfo(i); + auto& in_params = info.GetInParams(); + if (in_params.in_use) { + // Only update if in use + in_params.buffer_offset = offset; + offset += in_params.buffer_count; + } + } +} + +void MixContext::SortInfo() { + // Get the distance to the final mix + UpdateDistancesFromFinalMix(); + + // Sort based on the distance to the final mix + std::sort(sorted_info.begin(), sorted_info.end(), + [](const ServerMixInfo* lhs, const ServerMixInfo* rhs) { + return lhs->GetInParams().final_mix_distance > + rhs->GetInParams().final_mix_distance; + }); + + // Calculate the mix buffer offset + CalcMixBufferOffset(); +} + +bool MixContext::TsortInfo(SplitterContext& splitter_context) { + // If we're not using mixes, just calculate the mix buffer offset + if (!splitter_context.UsingSplitter()) { + CalcMixBufferOffset(); + return true; + } + // Sort our node states + if (!node_states.Tsort(edge_matrix)) { + return false; + } + + // Get our sorted list + const auto sorted_list = node_states.GetIndexList(); + std::size_t info_id{}; + for (auto itr = sorted_list.rbegin(); itr != sorted_list.rend(); ++itr) { + // Set our sorted info + sorted_info[info_id++] = &GetInfo(*itr); + } + + // Calculate the mix buffer offset + CalcMixBufferOffset(); + return true; +} + +std::size_t MixContext::GetCount() const { + return info_count; +} + +ServerMixInfo& MixContext::GetInfo(std::size_t i) { + ASSERT(i < info_count); + return infos.at(i); +} + +const ServerMixInfo& MixContext::GetInfo(std::size_t i) const { + ASSERT(i < info_count); + return infos.at(i); +} + +ServerMixInfo& MixContext::GetSortedInfo(std::size_t i) { + ASSERT(i < info_count); + return *sorted_info.at(i); +} + +const ServerMixInfo& MixContext::GetSortedInfo(std::size_t i) const { + ASSERT(i < info_count); + return *sorted_info.at(i); +} + +ServerMixInfo& MixContext::GetFinalMixInfo() { + return infos.at(AudioCommon::FINAL_MIX); +} + +const ServerMixInfo& MixContext::GetFinalMixInfo() const { + return infos.at(AudioCommon::FINAL_MIX); +} + +EdgeMatrix& MixContext::GetEdgeMatrix() { + return edge_matrix; +} + +const EdgeMatrix& MixContext::GetEdgeMatrix() const { + return edge_matrix; +} + +ServerMixInfo::ServerMixInfo() { + Cleanup(); +} +ServerMixInfo::~ServerMixInfo() = default; + +const ServerMixInfo::InParams& ServerMixInfo::GetInParams() const { + return in_params; +} + +ServerMixInfo::InParams& ServerMixInfo::GetInParams() { + return in_params; +} + +bool ServerMixInfo::Update(EdgeMatrix& edge_matrix, const MixInfo::InParams& mix_in, + BehaviorInfo& behavior_info, SplitterContext& splitter_context, + EffectContext& effect_context) { + in_params.volume = mix_in.volume; + in_params.sample_rate = mix_in.sample_rate; + in_params.buffer_count = mix_in.buffer_count; + in_params.in_use = mix_in.in_use; + in_params.mix_id = mix_in.mix_id; + in_params.node_id = mix_in.node_id; + for (std::size_t i = 0; i < mix_in.mix_volume.size(); i++) { + std::copy(mix_in.mix_volume[i].begin(), mix_in.mix_volume[i].end(), + in_params.mix_volume[i].begin()); + } + + bool require_sort = false; + + if (behavior_info.IsSplitterSupported()) { + require_sort = UpdateConnection(edge_matrix, mix_in, splitter_context); + } else { + in_params.dest_mix_id = mix_in.dest_mix_id; + in_params.splitter_id = AudioCommon::NO_SPLITTER; + } + + ResetEffectProcessingOrder(); + const auto effect_count = effect_context.GetCount(); + for (std::size_t i = 0; i < effect_count; i++) { + auto* effect_info = effect_context.GetInfo(i); + if (effect_info->GetMixID() == in_params.mix_id) { + effect_processing_order[effect_info->GetProcessingOrder()] = static_cast<s32>(i); + } + } + + // TODO(ogniK): Update effect processing order + return require_sort; +} + +bool ServerMixInfo::HasAnyConnection() const { + return in_params.splitter_id != AudioCommon::NO_SPLITTER || + in_params.mix_id != AudioCommon::NO_MIX; +} + +void ServerMixInfo::Cleanup() { + in_params.volume = 0.0f; + in_params.sample_rate = 0; + in_params.buffer_count = 0; + in_params.in_use = false; + in_params.mix_id = AudioCommon::NO_MIX; + in_params.node_id = 0; + in_params.buffer_offset = 0; + in_params.dest_mix_id = AudioCommon::NO_MIX; + in_params.splitter_id = AudioCommon::NO_SPLITTER; + std::memset(in_params.mix_volume.data(), 0, sizeof(float) * in_params.mix_volume.size()); +} + +void ServerMixInfo::SetEffectCount(std::size_t count) { + effect_processing_order.resize(count); + ResetEffectProcessingOrder(); +} + +void ServerMixInfo::ResetEffectProcessingOrder() { + for (auto& order : effect_processing_order) { + order = AudioCommon::NO_EFFECT_ORDER; + } +} + +s32 ServerMixInfo::GetEffectOrder(std::size_t i) const { + return effect_processing_order.at(i); +} + +bool ServerMixInfo::UpdateConnection(EdgeMatrix& edge_matrix, const MixInfo::InParams& mix_in, + SplitterContext& splitter_context) { + // Mixes are identical + if (in_params.dest_mix_id == mix_in.dest_mix_id && + in_params.splitter_id == mix_in.splitter_id && + ((in_params.splitter_id == AudioCommon::NO_SPLITTER) || + !splitter_context.GetInfo(in_params.splitter_id).HasNewConnection())) { + return false; + } + // Remove current edges for mix id + edge_matrix.RemoveEdges(in_params.mix_id); + if (mix_in.dest_mix_id != AudioCommon::NO_MIX) { + // If we have a valid destination mix id, set our edge matrix + edge_matrix.Connect(in_params.mix_id, mix_in.dest_mix_id); + } else if (mix_in.splitter_id != AudioCommon::NO_SPLITTER) { + // Recurse our splitter linked and set our edges + auto& splitter_info = splitter_context.GetInfo(mix_in.splitter_id); + const auto length = splitter_info.GetLength(); + for (s32 i = 0; i < length; i++) { + const auto* splitter_destination = + splitter_context.GetDestinationData(mix_in.splitter_id, i); + if (splitter_destination == nullptr) { + continue; + } + if (splitter_destination->ValidMixId()) { + edge_matrix.Connect(in_params.mix_id, splitter_destination->GetMixId()); + } + } + } + in_params.dest_mix_id = mix_in.dest_mix_id; + in_params.splitter_id = mix_in.splitter_id; + return true; +} + +} // namespace AudioCore diff --git a/src/audio_core/mix_context.h b/src/audio_core/mix_context.h new file mode 100644 index 000000000..6a588eeb4 --- /dev/null +++ b/src/audio_core/mix_context.h @@ -0,0 +1,114 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <vector> +#include "audio_core/common.h" +#include "audio_core/splitter_context.h" +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace AudioCore { +class BehaviorInfo; +class EffectContext; + +class MixInfo { +public: + struct DirtyHeader { + u32_le magic{}; + u32_le mixer_count{}; + INSERT_PADDING_BYTES(0x18); + }; + static_assert(sizeof(DirtyHeader) == 0x20, "MixInfo::DirtyHeader is an invalid size"); + + struct InParams { + float_le volume{}; + s32_le sample_rate{}; + s32_le buffer_count{}; + bool in_use{}; + INSERT_PADDING_BYTES(3); + s32_le mix_id{}; + s32_le effect_count{}; + u32_le node_id{}; + INSERT_PADDING_WORDS(2); + std::array<std::array<float_le, AudioCommon::MAX_MIX_BUFFERS>, AudioCommon::MAX_MIX_BUFFERS> + mix_volume{}; + s32_le dest_mix_id{}; + s32_le splitter_id{}; + INSERT_PADDING_WORDS(1); + }; + static_assert(sizeof(MixInfo::InParams) == 0x930, "MixInfo::InParams is an invalid size"); +}; + +class ServerMixInfo { +public: + struct InParams { + float volume{}; + s32 sample_rate{}; + s32 buffer_count{}; + bool in_use{}; + s32 mix_id{}; + u32 node_id{}; + std::array<std::array<float_le, AudioCommon::MAX_MIX_BUFFERS>, AudioCommon::MAX_MIX_BUFFERS> + mix_volume{}; + s32 dest_mix_id{}; + s32 splitter_id{}; + s32 buffer_offset{}; + s32 final_mix_distance{}; + }; + ServerMixInfo(); + ~ServerMixInfo(); + + const ServerMixInfo::InParams& GetInParams() const; + ServerMixInfo::InParams& GetInParams(); + + bool Update(EdgeMatrix& edge_matrix, const MixInfo::InParams& mix_in, + BehaviorInfo& behavior_info, SplitterContext& splitter_context, + EffectContext& effect_context); + bool HasAnyConnection() const; + void Cleanup(); + void SetEffectCount(std::size_t count); + void ResetEffectProcessingOrder(); + s32 GetEffectOrder(std::size_t i) const; + +private: + std::vector<s32> effect_processing_order; + InParams in_params{}; + bool UpdateConnection(EdgeMatrix& edge_matrix, const MixInfo::InParams& mix_in, + SplitterContext& splitter_context); +}; + +class MixContext { +public: + MixContext(); + ~MixContext(); + + void Initialize(const BehaviorInfo& behavior_info, std::size_t mix_count, + std::size_t effect_count); + void SortInfo(); + bool TsortInfo(SplitterContext& splitter_context); + + std::size_t GetCount() const; + ServerMixInfo& GetInfo(std::size_t i); + const ServerMixInfo& GetInfo(std::size_t i) const; + ServerMixInfo& GetSortedInfo(std::size_t i); + const ServerMixInfo& GetSortedInfo(std::size_t i) const; + ServerMixInfo& GetFinalMixInfo(); + const ServerMixInfo& GetFinalMixInfo() const; + EdgeMatrix& GetEdgeMatrix(); + const EdgeMatrix& GetEdgeMatrix() const; + +private: + void CalcMixBufferOffset(); + void UpdateDistancesFromFinalMix(); + + NodeStates node_states{}; + EdgeMatrix edge_matrix{}; + std::size_t info_count{}; + std::vector<ServerMixInfo> infos{}; + std::vector<ServerMixInfo*> sorted_info{}; +}; +} // namespace AudioCore diff --git a/src/audio_core/sink_context.cpp b/src/audio_core/sink_context.cpp new file mode 100644 index 000000000..0882b411a --- /dev/null +++ b/src/audio_core/sink_context.cpp @@ -0,0 +1,31 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/sink_context.h" + +namespace AudioCore { +SinkContext::SinkContext(std::size_t sink_count) : sink_count(sink_count) {} +SinkContext::~SinkContext() = default; + +std::size_t SinkContext::GetCount() const { + return sink_count; +} + +void SinkContext::UpdateMainSink(SinkInfo::InParams& in) { + in_use = in.in_use; + use_count = in.device.input_count; + std::memcpy(buffers.data(), in.device.input.data(), AudioCommon::MAX_CHANNEL_COUNT); +} + +bool SinkContext::InUse() const { + return in_use; +} + +std::vector<u8> SinkContext::OutputBuffers() const { + std::vector<u8> buffer_ret(use_count); + std::memcpy(buffer_ret.data(), buffers.data(), use_count); + return buffer_ret; +} + +} // namespace AudioCore diff --git a/src/audio_core/sink_context.h b/src/audio_core/sink_context.h new file mode 100644 index 000000000..d7aa72ba7 --- /dev/null +++ b/src/audio_core/sink_context.h @@ -0,0 +1,89 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "audio_core/common.h" +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" + +namespace AudioCore { + +enum class SinkTypes : u8 { + Invalid = 0, + Device = 1, + Circular = 2, +}; + +enum class SinkSampleFormat : u32_le { + None = 0, + Pcm8 = 1, + Pcm16 = 2, + Pcm24 = 3, + Pcm32 = 4, + PcmFloat = 5, + Adpcm = 6, +}; + +class SinkInfo { +public: + struct CircularBufferIn { + u64_le address; + u32_le size; + u32_le input_count; + u32_le sample_count; + u32_le previous_position; + SinkSampleFormat sample_format; + std::array<u8, AudioCommon::MAX_CHANNEL_COUNT> input; + bool in_use; + INSERT_UNION_PADDING_BYTES(5); + }; + static_assert(sizeof(SinkInfo::CircularBufferIn) == 0x28, + "SinkInfo::CircularBufferIn is in invalid size"); + + struct DeviceIn { + std::array<u8, 255> device_name; + INSERT_UNION_PADDING_BYTES(1); + s32_le input_count; + std::array<u8, AudioCommon::MAX_CHANNEL_COUNT> input; + INSERT_UNION_PADDING_BYTES(1); + bool down_matrix_enabled; + std::array<float_le, 4> down_matrix_coef; + }; + static_assert(sizeof(SinkInfo::DeviceIn) == 0x11c, "SinkInfo::DeviceIn is an invalid size"); + + struct InParams { + SinkTypes type{}; + bool in_use{}; + INSERT_PADDING_BYTES(2); + u32_le node_id{}; + INSERT_PADDING_WORDS(6); + union { + // std::array<u8, 0x120> raw{}; + SinkInfo::DeviceIn device; + SinkInfo::CircularBufferIn circular_buffer; + }; + }; + static_assert(sizeof(SinkInfo::InParams) == 0x140, "SinkInfo::InParams are an invalid size!"); +}; + +class SinkContext { +public: + explicit SinkContext(std::size_t sink_count); + ~SinkContext(); + + std::size_t GetCount() const; + + void UpdateMainSink(SinkInfo::InParams& in); + bool InUse() const; + std::vector<u8> OutputBuffers() const; + +private: + bool in_use{false}; + s32 use_count{}; + std::array<u8, AudioCommon::MAX_CHANNEL_COUNT> buffers{}; + std::size_t sink_count{}; +}; +} // namespace AudioCore diff --git a/src/audio_core/splitter_context.cpp b/src/audio_core/splitter_context.cpp new file mode 100644 index 000000000..f21b53147 --- /dev/null +++ b/src/audio_core/splitter_context.cpp @@ -0,0 +1,617 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/behavior_info.h" +#include "audio_core/splitter_context.h" +#include "common/alignment.h" +#include "common/assert.h" +#include "common/logging/log.h" + +namespace AudioCore { + +ServerSplitterDestinationData::ServerSplitterDestinationData(s32 id) : id(id) {} +ServerSplitterDestinationData::~ServerSplitterDestinationData() = default; + +void ServerSplitterDestinationData::Update(SplitterInfo::InDestinationParams& header) { + // Log error as these are not actually failure states + if (header.magic != SplitterMagic::DataHeader) { + LOG_ERROR(Audio, "Splitter destination header is invalid!"); + return; + } + + // Incorrect splitter id + if (header.splitter_id != id) { + LOG_ERROR(Audio, "Splitter destination ids do not match!"); + return; + } + + mix_id = header.mix_id; + // Copy our mix volumes + std::copy(header.mix_volumes.begin(), header.mix_volumes.end(), current_mix_volumes.begin()); + if (!in_use && header.in_use) { + // Update mix volumes + std::copy(current_mix_volumes.begin(), current_mix_volumes.end(), last_mix_volumes.begin()); + needs_update = false; + } + in_use = header.in_use; +} + +ServerSplitterDestinationData* ServerSplitterDestinationData::GetNextDestination() { + return next; +} + +const ServerSplitterDestinationData* ServerSplitterDestinationData::GetNextDestination() const { + return next; +} + +void ServerSplitterDestinationData::SetNextDestination(ServerSplitterDestinationData* dest) { + next = dest; +} + +bool ServerSplitterDestinationData::ValidMixId() const { + return GetMixId() != AudioCommon::NO_MIX; +} + +s32 ServerSplitterDestinationData::GetMixId() const { + return mix_id; +} + +bool ServerSplitterDestinationData::IsConfigured() const { + return in_use && ValidMixId(); +} + +float ServerSplitterDestinationData::GetMixVolume(std::size_t i) const { + ASSERT(i < AudioCommon::MAX_MIX_BUFFERS); + return current_mix_volumes.at(i); +} + +const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& +ServerSplitterDestinationData::CurrentMixVolumes() const { + return current_mix_volumes; +} + +const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& +ServerSplitterDestinationData::LastMixVolumes() const { + return last_mix_volumes; +} + +void ServerSplitterDestinationData::MarkDirty() { + needs_update = true; +} + +void ServerSplitterDestinationData::UpdateInternalState() { + if (in_use && needs_update) { + std::copy(current_mix_volumes.begin(), current_mix_volumes.end(), last_mix_volumes.begin()); + } + needs_update = false; +} + +ServerSplitterInfo::ServerSplitterInfo(s32 id) : id(id) {} +ServerSplitterInfo::~ServerSplitterInfo() = default; + +void ServerSplitterInfo::InitializeInfos() { + send_length = 0; + head = nullptr; + new_connection = true; +} + +void ServerSplitterInfo::ClearNewConnectionFlag() { + new_connection = false; +} + +std::size_t ServerSplitterInfo::Update(SplitterInfo::InInfoPrams& header) { + if (header.send_id != id) { + return 0; + } + + sample_rate = header.sample_rate; + new_connection = true; + // We need to update the size here due to the splitter bug being present and providing an + // incorrect size. We're suppose to also update the header here but we just ignore and continue + return (sizeof(s32_le) * (header.length - 1)) + (sizeof(s32_le) * 3); +} + +ServerSplitterDestinationData* ServerSplitterInfo::GetHead() { + return head; +} + +const ServerSplitterDestinationData* ServerSplitterInfo::GetHead() const { + return head; +} + +ServerSplitterDestinationData* ServerSplitterInfo::GetData(std::size_t depth) { + auto current_head = head; + for (std::size_t i = 0; i < depth; i++) { + if (current_head == nullptr) { + return nullptr; + } + current_head = current_head->GetNextDestination(); + } + return current_head; +} + +const ServerSplitterDestinationData* ServerSplitterInfo::GetData(std::size_t depth) const { + auto current_head = head; + for (std::size_t i = 0; i < depth; i++) { + if (current_head == nullptr) { + return nullptr; + } + current_head = current_head->GetNextDestination(); + } + return current_head; +} + +bool ServerSplitterInfo::HasNewConnection() const { + return new_connection; +} + +s32 ServerSplitterInfo::GetLength() const { + return send_length; +} + +void ServerSplitterInfo::SetHead(ServerSplitterDestinationData* new_head) { + head = new_head; +} + +void ServerSplitterInfo::SetHeadDepth(s32 length) { + send_length = length; +} + +SplitterContext::SplitterContext() = default; +SplitterContext::~SplitterContext() = default; + +void SplitterContext::Initialize(BehaviorInfo& behavior_info, std::size_t _info_count, + std::size_t _data_count) { + if (!behavior_info.IsSplitterSupported() || _data_count == 0 || _info_count == 0) { + Setup(0, 0, false); + return; + } + // Only initialize if we're using splitters + Setup(_info_count, _data_count, behavior_info.IsSplitterBugFixed()); +} + +bool SplitterContext::Update(const std::vector<u8>& input, std::size_t& input_offset, + std::size_t& bytes_read) { + const auto UpdateOffsets = [&](std::size_t read) { + input_offset += read; + bytes_read += read; + }; + + if (info_count == 0 || data_count == 0) { + bytes_read = 0; + return true; + } + + if (!AudioCommon::CanConsumeBuffer(input.size(), input_offset, + sizeof(SplitterInfo::InHeader))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + SplitterInfo::InHeader header{}; + std::memcpy(&header, input.data() + input_offset, sizeof(SplitterInfo::InHeader)); + UpdateOffsets(sizeof(SplitterInfo::InHeader)); + + if (header.magic != SplitterMagic::SplitterHeader) { + LOG_ERROR(Audio, "Invalid header magic! Expecting {:X} but got {:X}", + SplitterMagic::SplitterHeader, header.magic); + return false; + } + + // Clear all connections + for (auto& info : infos) { + info.ClearNewConnectionFlag(); + } + + UpdateInfo(input, input_offset, bytes_read, header.info_count); + UpdateData(input, input_offset, bytes_read, header.data_count); + const auto aligned_bytes_read = Common::AlignUp(bytes_read, 16); + input_offset += aligned_bytes_read - bytes_read; + bytes_read = aligned_bytes_read; + return true; +} + +bool SplitterContext::UsingSplitter() const { + return info_count > 0 && data_count > 0; +} + +ServerSplitterInfo& SplitterContext::GetInfo(std::size_t i) { + ASSERT(i < info_count); + return infos.at(i); +} + +const ServerSplitterInfo& SplitterContext::GetInfo(std::size_t i) const { + ASSERT(i < info_count); + return infos.at(i); +} + +ServerSplitterDestinationData& SplitterContext::GetData(std::size_t i) { + ASSERT(i < data_count); + return datas.at(i); +} + +const ServerSplitterDestinationData& SplitterContext::GetData(std::size_t i) const { + ASSERT(i < data_count); + return datas.at(i); +} + +ServerSplitterDestinationData* SplitterContext::GetDestinationData(std::size_t info, + std::size_t data) { + ASSERT(info < info_count); + auto& cur_info = GetInfo(info); + return cur_info.GetData(data); +} + +const ServerSplitterDestinationData* SplitterContext::GetDestinationData(std::size_t info, + std::size_t data) const { + ASSERT(info < info_count); + auto& cur_info = GetInfo(info); + return cur_info.GetData(data); +} + +void SplitterContext::UpdateInternalState() { + if (data_count == 0) { + return; + } + + for (auto& data : datas) { + data.UpdateInternalState(); + } +} + +std::size_t SplitterContext::GetInfoCount() const { + return info_count; +} + +std::size_t SplitterContext::GetDataCount() const { + return data_count; +} + +void SplitterContext::Setup(std::size_t _info_count, std::size_t _data_count, + bool is_splitter_bug_fixed) { + + info_count = _info_count; + data_count = _data_count; + + for (std::size_t i = 0; i < info_count; i++) { + auto& splitter = infos.emplace_back(static_cast<s32>(i)); + splitter.InitializeInfos(); + } + for (std::size_t i = 0; i < data_count; i++) { + datas.emplace_back(static_cast<s32>(i)); + } + + bug_fixed = is_splitter_bug_fixed; +} + +bool SplitterContext::UpdateInfo(const std::vector<u8>& input, std::size_t& input_offset, + std::size_t& bytes_read, s32 in_splitter_count) { + const auto UpdateOffsets = [&](std::size_t read) { + input_offset += read; + bytes_read += read; + }; + + for (s32 i = 0; i < in_splitter_count; i++) { + if (!AudioCommon::CanConsumeBuffer(input.size(), input_offset, + sizeof(SplitterInfo::InInfoPrams))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + SplitterInfo::InInfoPrams header{}; + std::memcpy(&header, input.data() + input_offset, sizeof(SplitterInfo::InInfoPrams)); + + // Logged as warning as these don't actually cause a bailout for some reason + if (header.magic != SplitterMagic::InfoHeader) { + LOG_ERROR(Audio, "Bad splitter data header"); + break; + } + + if (header.send_id < 0 || static_cast<std::size_t>(header.send_id) > info_count) { + LOG_ERROR(Audio, "Bad splitter data id"); + break; + } + + UpdateOffsets(sizeof(SplitterInfo::InInfoPrams)); + auto& info = GetInfo(header.send_id); + if (!RecomposeDestination(info, header, input, input_offset)) { + LOG_ERROR(Audio, "Failed to recompose destination for splitter!"); + return false; + } + const std::size_t read = info.Update(header); + bytes_read += read; + input_offset += read; + } + return true; +} + +bool SplitterContext::UpdateData(const std::vector<u8>& input, std::size_t& input_offset, + std::size_t& bytes_read, s32 in_data_count) { + const auto UpdateOffsets = [&](std::size_t read) { + input_offset += read; + bytes_read += read; + }; + + for (s32 i = 0; i < in_data_count; i++) { + if (!AudioCommon::CanConsumeBuffer(input.size(), input_offset, + sizeof(SplitterInfo::InDestinationParams))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + SplitterInfo::InDestinationParams header{}; + std::memcpy(&header, input.data() + input_offset, + sizeof(SplitterInfo::InDestinationParams)); + UpdateOffsets(sizeof(SplitterInfo::InDestinationParams)); + + // Logged as warning as these don't actually cause a bailout for some reason + if (header.magic != SplitterMagic::DataHeader) { + LOG_ERROR(Audio, "Bad splitter data header"); + break; + } + + if (header.splitter_id < 0 || static_cast<std::size_t>(header.splitter_id) > data_count) { + LOG_ERROR(Audio, "Bad splitter data id"); + break; + } + GetData(header.splitter_id).Update(header); + } + return true; +} + +bool SplitterContext::RecomposeDestination(ServerSplitterInfo& info, + SplitterInfo::InInfoPrams& header, + const std::vector<u8>& input, + const std::size_t& input_offset) { + // Clear our current destinations + auto* current_head = info.GetHead(); + while (current_head != nullptr) { + auto next_head = current_head->GetNextDestination(); + current_head->SetNextDestination(nullptr); + current_head = next_head; + } + info.SetHead(nullptr); + + s32 size = header.length; + // If the splitter bug is present, calculate fixed size + if (!bug_fixed) { + if (info_count > 0) { + const auto factor = data_count / info_count; + size = std::min(header.length, static_cast<s32>(factor)); + } else { + size = 0; + } + } + + if (size < 1) { + LOG_ERROR(Audio, "Invalid splitter info size! size={:X}", size); + return true; + } + + auto* start_head = &GetData(header.resource_id_base); + current_head = start_head; + std::vector<s32_le> resource_ids(size - 1); + if (!AudioCommon::CanConsumeBuffer(input.size(), input_offset, + resource_ids.size() * sizeof(s32_le))) { + LOG_ERROR(Audio, "Buffer is an invalid size!"); + return false; + } + std::memcpy(resource_ids.data(), input.data() + input_offset, + resource_ids.size() * sizeof(s32_le)); + + for (auto resource_id : resource_ids) { + auto* head = &GetData(resource_id); + current_head->SetNextDestination(head); + current_head = head; + } + + info.SetHead(start_head); + info.SetHeadDepth(size); + + return true; +} + +NodeStates::NodeStates() = default; +NodeStates::~NodeStates() = default; + +void NodeStates::Initialize(std::size_t node_count_) { + // Setup our work parameters + node_count = node_count_; + was_node_found.resize(node_count); + was_node_completed.resize(node_count); + index_list.resize(node_count); + index_stack.Reset(node_count * node_count); +} + +bool NodeStates::Tsort(EdgeMatrix& edge_matrix) { + return DepthFirstSearch(edge_matrix); +} + +std::size_t NodeStates::GetIndexPos() const { + return index_pos; +} + +const std::vector<s32>& NodeStates::GetIndexList() const { + return index_list; +} + +void NodeStates::PushTsortResult(s32 index) { + ASSERT(index < static_cast<s32>(node_count)); + index_list[index_pos++] = index; +} + +bool NodeStates::DepthFirstSearch(EdgeMatrix& edge_matrix) { + ResetState(); + for (std::size_t i = 0; i < node_count; i++) { + const auto node_id = static_cast<s32>(i); + + // If we don't have a state, send to our index stack for work + if (GetState(i) == NodeStates::State::NoState) { + index_stack.push(node_id); + } + + // While we have work to do in our stack + while (index_stack.Count() > 0) { + // Get the current node + const auto current_stack_index = index_stack.top(); + // Check if we've seen the node yet + const auto index_state = GetState(current_stack_index); + if (index_state == NodeStates::State::NoState) { + // Mark the node as seen + UpdateState(NodeStates::State::InFound, current_stack_index); + } else if (index_state == NodeStates::State::InFound) { + // We've seen this node before, mark it as completed + UpdateState(NodeStates::State::InCompleted, current_stack_index); + // Update our index list + PushTsortResult(current_stack_index); + // Pop the stack + index_stack.pop(); + continue; + } else if (index_state == NodeStates::State::InCompleted) { + // If our node is already sorted, clear it + index_stack.pop(); + continue; + } + + const auto node_count = edge_matrix.GetNodeCount(); + for (s32 j = 0; j < static_cast<s32>(node_count); j++) { + // Check if our node is connected to our edge matrix + if (!edge_matrix.Connected(current_stack_index, j)) { + continue; + } + + // Check if our node exists + const auto node_state = GetState(j); + if (node_state == NodeStates::State::NoState) { + // Add more work + index_stack.push(j); + } else if (node_state == NodeStates::State::InFound) { + UNREACHABLE_MSG("Node start marked as found"); + ResetState(); + return false; + } + } + } + } + return true; +} + +void NodeStates::ResetState() { + // Reset to the start of our index stack + index_pos = 0; + for (std::size_t i = 0; i < node_count; i++) { + // Mark all nodes as not found + was_node_found[i] = false; + // Mark all nodes as uncompleted + was_node_completed[i] = false; + // Mark all indexes as invalid + index_list[i] = -1; + } +} + +void NodeStates::UpdateState(NodeStates::State state, std::size_t i) { + switch (state) { + case NodeStates::State::NoState: + was_node_found[i] = false; + was_node_completed[i] = false; + break; + case NodeStates::State::InFound: + was_node_found[i] = true; + was_node_completed[i] = false; + break; + case NodeStates::State::InCompleted: + was_node_found[i] = false; + was_node_completed[i] = true; + break; + } +} + +NodeStates::State NodeStates::GetState(std::size_t i) { + ASSERT(i < node_count); + if (was_node_found[i]) { + // If our node exists in our found list + return NodeStates::State::InFound; + } else if (was_node_completed[i]) { + // If node is in the completed list + return NodeStates::State::InCompleted; + } else { + // If in neither + return NodeStates::State::NoState; + } +} + +NodeStates::Stack::Stack() = default; +NodeStates::Stack::~Stack() = default; + +void NodeStates::Stack::Reset(std::size_t size) { + // Mark our stack as empty + stack.resize(size); + stack_size = size; + stack_pos = 0; + std::fill(stack.begin(), stack.end(), 0); +} + +void NodeStates::Stack::push(s32 val) { + ASSERT(stack_pos < stack_size); + stack[stack_pos++] = val; +} + +std::size_t NodeStates::Stack::Count() const { + return stack_pos; +} + +s32 NodeStates::Stack::top() const { + ASSERT(stack_pos > 0); + return stack[stack_pos - 1]; +} + +s32 NodeStates::Stack::pop() { + ASSERT(stack_pos > 0); + stack_pos--; + return stack[stack_pos]; +} + +EdgeMatrix::EdgeMatrix() = default; +EdgeMatrix::~EdgeMatrix() = default; + +void EdgeMatrix::Initialize(std::size_t _node_count) { + node_count = _node_count; + edge_matrix.resize(node_count * node_count); +} + +bool EdgeMatrix::Connected(s32 a, s32 b) { + return GetState(a, b); +} + +void EdgeMatrix::Connect(s32 a, s32 b) { + SetState(a, b, true); +} + +void EdgeMatrix::Disconnect(s32 a, s32 b) { + SetState(a, b, false); +} + +void EdgeMatrix::RemoveEdges(s32 edge) { + for (std::size_t i = 0; i < node_count; i++) { + SetState(edge, static_cast<s32>(i), false); + } +} + +std::size_t EdgeMatrix::GetNodeCount() const { + return node_count; +} + +void EdgeMatrix::SetState(s32 a, s32 b, bool state) { + ASSERT(InRange(a, b)); + edge_matrix.at(a * node_count + b) = state; +} + +bool EdgeMatrix::GetState(s32 a, s32 b) { + ASSERT(InRange(a, b)); + return edge_matrix.at(a * node_count + b); +} + +bool EdgeMatrix::InRange(s32 a, s32 b) const { + const std::size_t pos = a * node_count + b; + return pos < (node_count * node_count); +} + +} // namespace AudioCore diff --git a/src/audio_core/splitter_context.h b/src/audio_core/splitter_context.h new file mode 100644 index 000000000..ea6239fdb --- /dev/null +++ b/src/audio_core/splitter_context.h @@ -0,0 +1,221 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <stack> +#include <vector> +#include "audio_core/common.h" +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" + +namespace AudioCore { +class BehaviorInfo; + +class EdgeMatrix { +public: + EdgeMatrix(); + ~EdgeMatrix(); + + void Initialize(std::size_t _node_count); + bool Connected(s32 a, s32 b); + void Connect(s32 a, s32 b); + void Disconnect(s32 a, s32 b); + void RemoveEdges(s32 edge); + std::size_t GetNodeCount() const; + +private: + void SetState(s32 a, s32 b, bool state); + bool GetState(s32 a, s32 b); + + bool InRange(s32 a, s32 b) const; + std::vector<bool> edge_matrix{}; + std::size_t node_count{}; +}; + +class NodeStates { +public: + enum class State { + NoState = 0, + InFound = 1, + InCompleted = 2, + }; + + // Looks to be a fixed size stack. Placed within the NodeStates class based on symbols + class Stack { + public: + Stack(); + ~Stack(); + + void Reset(std::size_t size); + void push(s32 val); + std::size_t Count() const; + s32 top() const; + s32 pop(); + + private: + std::vector<s32> stack{}; + std::size_t stack_size{}; + std::size_t stack_pos{}; + }; + NodeStates(); + ~NodeStates(); + + void Initialize(std::size_t _node_count); + bool Tsort(EdgeMatrix& edge_matrix); + std::size_t GetIndexPos() const; + const std::vector<s32>& GetIndexList() const; + +private: + void PushTsortResult(s32 index); + bool DepthFirstSearch(EdgeMatrix& edge_matrix); + void ResetState(); + void UpdateState(NodeStates::State state, std::size_t i); + NodeStates::State GetState(std::size_t i); + + std::size_t node_count{}; + std::vector<bool> was_node_found{}; + std::vector<bool> was_node_completed{}; + std::size_t index_pos{}; + std::vector<s32> index_list{}; + NodeStates::Stack index_stack{}; +}; + +enum class SplitterMagic : u32_le { + SplitterHeader = Common::MakeMagic('S', 'N', 'D', 'H'), + DataHeader = Common::MakeMagic('S', 'N', 'D', 'D'), + InfoHeader = Common::MakeMagic('S', 'N', 'D', 'I'), +}; + +class SplitterInfo { +public: + struct InHeader { + SplitterMagic magic{}; + s32_le info_count{}; + s32_le data_count{}; + INSERT_PADDING_WORDS(5); + }; + static_assert(sizeof(SplitterInfo::InHeader) == 0x20, + "SplitterInfo::InHeader is an invalid size"); + + struct InInfoPrams { + SplitterMagic magic{}; + s32_le send_id{}; + s32_le sample_rate{}; + s32_le length{}; + s32_le resource_id_base{}; + }; + static_assert(sizeof(SplitterInfo::InInfoPrams) == 0x14, + "SplitterInfo::InInfoPrams is an invalid size"); + + struct InDestinationParams { + SplitterMagic magic{}; + s32_le splitter_id{}; + std::array<float_le, AudioCommon::MAX_MIX_BUFFERS> mix_volumes{}; + s32_le mix_id{}; + bool in_use{}; + INSERT_PADDING_BYTES(3); + }; + static_assert(sizeof(SplitterInfo::InDestinationParams) == 0x70, + "SplitterInfo::InDestinationParams is an invalid size"); +}; + +class ServerSplitterDestinationData { +public: + explicit ServerSplitterDestinationData(s32 id); + ~ServerSplitterDestinationData(); + + void Update(SplitterInfo::InDestinationParams& header); + + ServerSplitterDestinationData* GetNextDestination(); + const ServerSplitterDestinationData* GetNextDestination() const; + void SetNextDestination(ServerSplitterDestinationData* dest); + bool ValidMixId() const; + s32 GetMixId() const; + bool IsConfigured() const; + float GetMixVolume(std::size_t i) const; + const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& CurrentMixVolumes() const; + const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& LastMixVolumes() const; + void MarkDirty(); + void UpdateInternalState(); + +private: + bool needs_update{}; + bool in_use{}; + s32 id{}; + s32 mix_id{}; + std::array<float, AudioCommon::MAX_MIX_BUFFERS> current_mix_volumes{}; + std::array<float, AudioCommon::MAX_MIX_BUFFERS> last_mix_volumes{}; + ServerSplitterDestinationData* next = nullptr; +}; + +class ServerSplitterInfo { +public: + explicit ServerSplitterInfo(s32 id); + ~ServerSplitterInfo(); + + void InitializeInfos(); + void ClearNewConnectionFlag(); + std::size_t Update(SplitterInfo::InInfoPrams& header); + + ServerSplitterDestinationData* GetHead(); + const ServerSplitterDestinationData* GetHead() const; + ServerSplitterDestinationData* GetData(std::size_t depth); + const ServerSplitterDestinationData* GetData(std::size_t depth) const; + + bool HasNewConnection() const; + s32 GetLength() const; + + void SetHead(ServerSplitterDestinationData* new_head); + void SetHeadDepth(s32 length); + +private: + s32 sample_rate{}; + s32 id{}; + s32 send_length{}; + ServerSplitterDestinationData* head = nullptr; + bool new_connection{}; +}; + +class SplitterContext { +public: + SplitterContext(); + ~SplitterContext(); + + void Initialize(BehaviorInfo& behavior_info, std::size_t splitter_count, + std::size_t data_count); + + bool Update(const std::vector<u8>& input, std::size_t& input_offset, std::size_t& bytes_read); + bool UsingSplitter() const; + + ServerSplitterInfo& GetInfo(std::size_t i); + const ServerSplitterInfo& GetInfo(std::size_t i) const; + ServerSplitterDestinationData& GetData(std::size_t i); + const ServerSplitterDestinationData& GetData(std::size_t i) const; + ServerSplitterDestinationData* GetDestinationData(std::size_t info, std::size_t data); + const ServerSplitterDestinationData* GetDestinationData(std::size_t info, + std::size_t data) const; + void UpdateInternalState(); + + std::size_t GetInfoCount() const; + std::size_t GetDataCount() const; + +private: + void Setup(std::size_t info_count, std::size_t data_count, bool is_splitter_bug_fixed); + bool UpdateInfo(const std::vector<u8>& input, std::size_t& input_offset, + std::size_t& bytes_read, s32 in_splitter_count); + bool UpdateData(const std::vector<u8>& input, std::size_t& input_offset, + std::size_t& bytes_read, s32 in_data_count); + bool RecomposeDestination(ServerSplitterInfo& info, SplitterInfo::InInfoPrams& header, + const std::vector<u8>& input, const std::size_t& input_offset); + + std::vector<ServerSplitterInfo> infos{}; + std::vector<ServerSplitterDestinationData> datas{}; + + std::size_t info_count{}; + std::size_t data_count{}; + bool bug_fixed{}; +}; +} // namespace AudioCore diff --git a/src/audio_core/stream.cpp b/src/audio_core/stream.cpp index 4ca98f8ea..4bbb1e0c4 100644 --- a/src/audio_core/stream.cpp +++ b/src/audio_core/stream.cpp @@ -12,7 +12,6 @@ #include "common/assert.h" #include "common/logging/log.h" #include "core/core_timing.h" -#include "core/core_timing_util.h" #include "core/settings.h" namespace AudioCore { @@ -36,9 +35,10 @@ Stream::Stream(Core::Timing::CoreTiming& core_timing, u32 sample_rate, Format fo ReleaseCallback&& release_callback, SinkStream& sink_stream, std::string&& name_) : sample_rate{sample_rate}, format{format}, release_callback{std::move(release_callback)}, sink_stream{sink_stream}, core_timing{core_timing}, name{std::move(name_)} { - - release_event = Core::Timing::CreateEvent( - name, [this](u64 userdata, s64 cycles_late) { ReleaseActiveBuffer(); }); + release_event = + Core::Timing::CreateEvent(name, [this](std::uintptr_t, std::chrono::nanoseconds ns_late) { + ReleaseActiveBuffer(ns_late); + }); } void Stream::Play() { @@ -59,15 +59,13 @@ Stream::State Stream::GetState() const { return state; } -s64 Stream::GetBufferReleaseCycles(const Buffer& buffer) const { +std::chrono::nanoseconds Stream::GetBufferReleaseNS(const Buffer& buffer) const { const std::size_t num_samples{buffer.GetSamples().size() / GetNumChannels()}; - const auto us = - std::chrono::microseconds((static_cast<u64>(num_samples) * 1000000) / sample_rate); - return Core::Timing::usToCycles(us); + return std::chrono::nanoseconds((static_cast<u64>(num_samples) * 1000000000ULL) / sample_rate); } static void VolumeAdjustSamples(std::vector<s16>& samples, float game_volume) { - const float volume{std::clamp(Settings::values.volume - (1.0f - game_volume), 0.0f, 1.0f)}; + const float volume{std::clamp(Settings::Volume() - (1.0f - game_volume), 0.0f, 1.0f)}; if (volume == 1.0f) { return; @@ -80,7 +78,7 @@ static void VolumeAdjustSamples(std::vector<s16>& samples, float game_volume) { } } -void Stream::PlayNextBuffer() { +void Stream::PlayNextBuffer(std::chrono::nanoseconds ns_late) { if (!IsPlaying()) { // Ensure we are in playing state before playing the next buffer sink_stream.Flush(); @@ -105,14 +103,14 @@ void Stream::PlayNextBuffer() { sink_stream.EnqueueSamples(GetNumChannels(), active_buffer->GetSamples()); - core_timing.ScheduleEvent(GetBufferReleaseCycles(*active_buffer), release_event, {}); + core_timing.ScheduleEvent(GetBufferReleaseNS(*active_buffer) - ns_late, release_event, {}); } -void Stream::ReleaseActiveBuffer() { +void Stream::ReleaseActiveBuffer(std::chrono::nanoseconds ns_late) { ASSERT(active_buffer); released_buffers.push(std::move(active_buffer)); release_callback(); - PlayNextBuffer(); + PlayNextBuffer(ns_late); } bool Stream::QueueBuffer(BufferPtr&& buffer) { diff --git a/src/audio_core/stream.h b/src/audio_core/stream.h index 1708a4d98..6437b8591 100644 --- a/src/audio_core/stream.h +++ b/src/audio_core/stream.h @@ -4,6 +4,7 @@ #pragma once +#include <chrono> #include <functional> #include <memory> #include <string> @@ -90,13 +91,13 @@ public: private: /// Plays the next queued buffer in the audio stream, starting playback if necessary - void PlayNextBuffer(); + void PlayNextBuffer(std::chrono::nanoseconds ns_late = {}); /// Releases the actively playing buffer, signalling that it has been completed - void ReleaseActiveBuffer(); + void ReleaseActiveBuffer(std::chrono::nanoseconds ns_late = {}); /// Gets the number of core cycles when the specified buffer will be released - s64 GetBufferReleaseCycles(const Buffer& buffer) const; + std::chrono::nanoseconds GetBufferReleaseNS(const Buffer& buffer) const; u32 sample_rate; ///< Sample rate of the stream Format format; ///< Format of the stream diff --git a/src/audio_core/voice_context.cpp b/src/audio_core/voice_context.cpp new file mode 100644 index 000000000..c46ee55f1 --- /dev/null +++ b/src/audio_core/voice_context.cpp @@ -0,0 +1,529 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/behavior_info.h" +#include "audio_core/voice_context.h" +#include "core/memory.h" + +namespace AudioCore { + +ServerVoiceChannelResource::ServerVoiceChannelResource(s32 id) : id(id) {} +ServerVoiceChannelResource::~ServerVoiceChannelResource() = default; + +bool ServerVoiceChannelResource::InUse() const { + return in_use; +} + +float ServerVoiceChannelResource::GetCurrentMixVolumeAt(std::size_t i) const { + ASSERT(i < AudioCommon::MAX_MIX_BUFFERS); + return mix_volume.at(i); +} + +float ServerVoiceChannelResource::GetLastMixVolumeAt(std::size_t i) const { + ASSERT(i < AudioCommon::MAX_MIX_BUFFERS); + return last_mix_volume.at(i); +} + +void ServerVoiceChannelResource::Update(VoiceChannelResource::InParams& in_params) { + in_use = in_params.in_use; + // Update our mix volumes only if it's in use + if (in_params.in_use) { + mix_volume = in_params.mix_volume; + } +} + +void ServerVoiceChannelResource::UpdateLastMixVolumes() { + last_mix_volume = mix_volume; +} + +const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& +ServerVoiceChannelResource::GetCurrentMixVolume() const { + return mix_volume; +} + +const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& +ServerVoiceChannelResource::GetLastMixVolume() const { + return last_mix_volume; +} + +ServerVoiceInfo::ServerVoiceInfo() { + Initialize(); +} +ServerVoiceInfo::~ServerVoiceInfo() = default; + +void ServerVoiceInfo::Initialize() { + in_params.in_use = false; + in_params.node_id = 0; + in_params.id = 0; + in_params.current_playstate = ServerPlayState::Stop; + in_params.priority = 255; + in_params.sample_rate = 0; + in_params.sample_format = SampleFormat::Invalid; + in_params.channel_count = 0; + in_params.pitch = 0.0f; + in_params.volume = 0.0f; + in_params.last_volume = 0.0f; + in_params.biquad_filter.fill({}); + in_params.wave_buffer_count = 0; + in_params.wave_bufffer_head = 0; + in_params.mix_id = AudioCommon::NO_MIX; + in_params.splitter_info_id = AudioCommon::NO_SPLITTER; + in_params.additional_params_address = 0; + in_params.additional_params_size = 0; + in_params.is_new = false; + out_params.played_sample_count = 0; + out_params.wave_buffer_consumed = 0; + in_params.voice_drop_flag = false; + in_params.buffer_mapped = false; + in_params.wave_buffer_flush_request_count = 0; + in_params.was_biquad_filter_enabled.fill(false); + + for (auto& wave_buffer : in_params.wave_buffer) { + wave_buffer.start_sample_offset = 0; + wave_buffer.end_sample_offset = 0; + wave_buffer.is_looping = false; + wave_buffer.end_of_stream = false; + wave_buffer.buffer_address = 0; + wave_buffer.buffer_size = 0; + wave_buffer.context_address = 0; + wave_buffer.context_size = 0; + wave_buffer.sent_to_dsp = true; + } + + stored_samples.clear(); +} + +void ServerVoiceInfo::UpdateParameters(const VoiceInfo::InParams& voice_in, + BehaviorInfo& behavior_info) { + in_params.in_use = voice_in.is_in_use; + in_params.id = voice_in.id; + in_params.node_id = voice_in.node_id; + in_params.last_playstate = in_params.current_playstate; + switch (voice_in.play_state) { + case PlayState::Paused: + in_params.current_playstate = ServerPlayState::Paused; + break; + case PlayState::Stopped: + if (in_params.current_playstate != ServerPlayState::Stop) { + in_params.current_playstate = ServerPlayState::RequestStop; + } + break; + case PlayState::Started: + in_params.current_playstate = ServerPlayState::Play; + break; + default: + UNREACHABLE_MSG("Unknown playstate {}", voice_in.play_state); + break; + } + + in_params.priority = voice_in.priority; + in_params.sorting_order = voice_in.sorting_order; + in_params.sample_rate = voice_in.sample_rate; + in_params.sample_format = voice_in.sample_format; + in_params.channel_count = voice_in.channel_count; + in_params.pitch = voice_in.pitch; + in_params.volume = voice_in.volume; + in_params.biquad_filter = voice_in.biquad_filter; + in_params.wave_buffer_count = voice_in.wave_buffer_count; + in_params.wave_bufffer_head = voice_in.wave_buffer_head; + if (behavior_info.IsFlushVoiceWaveBuffersSupported()) { + const auto in_request_count = in_params.wave_buffer_flush_request_count; + const auto voice_request_count = voice_in.wave_buffer_flush_request_count; + in_params.wave_buffer_flush_request_count = + static_cast<u8>(in_request_count + voice_request_count); + } + in_params.mix_id = voice_in.mix_id; + if (behavior_info.IsSplitterSupported()) { + in_params.splitter_info_id = voice_in.splitter_info_id; + } else { + in_params.splitter_info_id = AudioCommon::NO_SPLITTER; + } + + std::memcpy(in_params.voice_channel_resource_id.data(), + voice_in.voice_channel_resource_ids.data(), + sizeof(s32) * in_params.voice_channel_resource_id.size()); + + if (behavior_info.IsVoicePlayedSampleCountResetAtLoopPointSupported()) { + in_params.behavior_flags.is_played_samples_reset_at_loop_point = + voice_in.behavior_flags.is_played_samples_reset_at_loop_point; + } else { + in_params.behavior_flags.is_played_samples_reset_at_loop_point.Assign(0); + } + if (behavior_info.IsVoicePitchAndSrcSkippedSupported()) { + in_params.behavior_flags.is_pitch_and_src_skipped = + voice_in.behavior_flags.is_pitch_and_src_skipped; + } else { + in_params.behavior_flags.is_pitch_and_src_skipped.Assign(0); + } + + if (voice_in.is_voice_drop_flag_clear_requested) { + in_params.voice_drop_flag = false; + } + + if (in_params.additional_params_address != voice_in.additional_params_address || + in_params.additional_params_size != voice_in.additional_params_size) { + in_params.additional_params_address = voice_in.additional_params_address; + in_params.additional_params_size = voice_in.additional_params_size; + // TODO(ogniK): Reattach buffer, do we actually need to? Maybe just signal to the DSP that + // our context is new + } +} + +void ServerVoiceInfo::UpdateWaveBuffers( + const VoiceInfo::InParams& voice_in, + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& voice_states, + BehaviorInfo& behavior_info) { + if (voice_in.is_new) { + // Initialize our wave buffers + for (auto& wave_buffer : in_params.wave_buffer) { + wave_buffer.start_sample_offset = 0; + wave_buffer.end_sample_offset = 0; + wave_buffer.is_looping = false; + wave_buffer.end_of_stream = false; + wave_buffer.buffer_address = 0; + wave_buffer.buffer_size = 0; + wave_buffer.context_address = 0; + wave_buffer.context_size = 0; + wave_buffer.sent_to_dsp = true; + } + + // Mark all our wave buffers as invalid + for (std::size_t channel = 0; channel < static_cast<std::size_t>(in_params.channel_count); + channel++) { + for (auto& is_valid : voice_states[channel]->is_wave_buffer_valid) { + is_valid = false; + } + } + } + + // Update our wave buffers + for (std::size_t i = 0; i < AudioCommon::MAX_WAVE_BUFFERS; i++) { + // Assume that we have at least 1 channel voice state + const auto have_valid_wave_buffer = voice_states[0]->is_wave_buffer_valid[i]; + + UpdateWaveBuffer(in_params.wave_buffer[i], voice_in.wave_buffer[i], in_params.sample_format, + have_valid_wave_buffer, behavior_info); + } +} + +void ServerVoiceInfo::UpdateWaveBuffer(ServerWaveBuffer& out_wavebuffer, + const WaveBuffer& in_wave_buffer, SampleFormat sample_format, + bool is_buffer_valid, BehaviorInfo& behavior_info) { + if (!is_buffer_valid && out_wavebuffer.sent_to_dsp) { + out_wavebuffer.buffer_address = 0; + out_wavebuffer.buffer_size = 0; + } + + if (!in_wave_buffer.sent_to_server || !in_params.buffer_mapped) { + // Validate sample offset sizings + if (sample_format == SampleFormat::Pcm16) { + const auto buffer_size = in_wave_buffer.buffer_size; + if (in_wave_buffer.start_sample_offset < 0 || in_wave_buffer.end_sample_offset < 0 || + (buffer_size < (sizeof(s16) * in_wave_buffer.start_sample_offset)) || + (buffer_size < (sizeof(s16) * in_wave_buffer.end_sample_offset))) { + // TODO(ogniK): Write error info + return; + } + } + // TODO(ogniK): ADPCM Size error + + out_wavebuffer.sent_to_dsp = false; + out_wavebuffer.start_sample_offset = in_wave_buffer.start_sample_offset; + out_wavebuffer.end_sample_offset = in_wave_buffer.end_sample_offset; + out_wavebuffer.is_looping = in_wave_buffer.is_looping; + out_wavebuffer.end_of_stream = in_wave_buffer.end_of_stream; + + out_wavebuffer.buffer_address = in_wave_buffer.buffer_address; + out_wavebuffer.buffer_size = in_wave_buffer.buffer_size; + out_wavebuffer.context_address = in_wave_buffer.context_address; + out_wavebuffer.context_size = in_wave_buffer.context_size; + in_params.buffer_mapped = + in_wave_buffer.buffer_address != 0 && in_wave_buffer.buffer_size != 0; + // TODO(ogniK): Pool mapper attachment + // TODO(ogniK): IsAdpcmLoopContextBugFixed + } +} + +void ServerVoiceInfo::WriteOutStatus( + VoiceInfo::OutParams& voice_out, VoiceInfo::InParams& voice_in, + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& voice_states) { + if (voice_in.is_new) { + in_params.is_new = true; + voice_out.wave_buffer_consumed = 0; + voice_out.played_sample_count = 0; + voice_out.voice_dropped = false; + } else if (!in_params.is_new) { + voice_out.wave_buffer_consumed = voice_states[0]->wave_buffer_consumed; + voice_out.played_sample_count = voice_states[0]->played_sample_count; + voice_out.voice_dropped = in_params.voice_drop_flag; + } else { + voice_out.wave_buffer_consumed = 0; + voice_out.played_sample_count = 0; + voice_out.voice_dropped = false; + } +} + +const ServerVoiceInfo::InParams& ServerVoiceInfo::GetInParams() const { + return in_params; +} + +ServerVoiceInfo::InParams& ServerVoiceInfo::GetInParams() { + return in_params; +} + +const ServerVoiceInfo::OutParams& ServerVoiceInfo::GetOutParams() const { + return out_params; +} + +ServerVoiceInfo::OutParams& ServerVoiceInfo::GetOutParams() { + return out_params; +} + +bool ServerVoiceInfo::ShouldSkip() const { + // TODO(ogniK): Handle unmapped wave buffers or parameters + return !in_params.in_use || (in_params.wave_buffer_count == 0) || in_params.voice_drop_flag; +} + +bool ServerVoiceInfo::UpdateForCommandGeneration(VoiceContext& voice_context) { + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT> dsp_voice_states{}; + if (in_params.is_new) { + ResetResources(voice_context); + in_params.last_volume = in_params.volume; + in_params.is_new = false; + } + + const s32 channel_count = in_params.channel_count; + for (s32 i = 0; i < channel_count; i++) { + const auto channel_resource = in_params.voice_channel_resource_id[i]; + dsp_voice_states[i] = + &voice_context.GetDspSharedState(static_cast<std::size_t>(channel_resource)); + } + return UpdateParametersForCommandGeneration(dsp_voice_states); +} + +void ServerVoiceInfo::ResetResources(VoiceContext& voice_context) { + const s32 channel_count = in_params.channel_count; + for (s32 i = 0; i < channel_count; i++) { + const auto channel_resource = in_params.voice_channel_resource_id[i]; + auto& dsp_state = + voice_context.GetDspSharedState(static_cast<std::size_t>(channel_resource)); + dsp_state = {}; + voice_context.GetChannelResource(static_cast<std::size_t>(channel_resource)) + .UpdateLastMixVolumes(); + } +} + +bool ServerVoiceInfo::UpdateParametersForCommandGeneration( + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& dsp_voice_states) { + const s32 channel_count = in_params.channel_count; + if (in_params.wave_buffer_flush_request_count > 0) { + FlushWaveBuffers(in_params.wave_buffer_flush_request_count, dsp_voice_states, + channel_count); + in_params.wave_buffer_flush_request_count = 0; + } + + switch (in_params.current_playstate) { + case ServerPlayState::Play: { + for (std::size_t i = 0; i < AudioCommon::MAX_WAVE_BUFFERS; i++) { + if (!in_params.wave_buffer[i].sent_to_dsp) { + for (s32 channel = 0; channel < channel_count; channel++) { + dsp_voice_states[channel]->is_wave_buffer_valid[i] = true; + } + in_params.wave_buffer[i].sent_to_dsp = true; + } + } + in_params.should_depop = false; + return HasValidWaveBuffer(dsp_voice_states[0]); + } + case ServerPlayState::Paused: + case ServerPlayState::Stop: { + in_params.should_depop = in_params.last_playstate == ServerPlayState::Play; + return in_params.should_depop; + } + case ServerPlayState::RequestStop: { + for (std::size_t i = 0; i < AudioCommon::MAX_WAVE_BUFFERS; i++) { + in_params.wave_buffer[i].sent_to_dsp = true; + for (s32 channel = 0; channel < channel_count; channel++) { + auto* dsp_state = dsp_voice_states[channel]; + + if (dsp_state->is_wave_buffer_valid[i]) { + dsp_state->wave_buffer_index = + (dsp_state->wave_buffer_index + 1) % AudioCommon::MAX_WAVE_BUFFERS; + dsp_state->wave_buffer_consumed++; + } + + dsp_state->is_wave_buffer_valid[i] = false; + } + } + + for (s32 channel = 0; channel < channel_count; channel++) { + auto* dsp_state = dsp_voice_states[channel]; + dsp_state->offset = 0; + dsp_state->played_sample_count = 0; + dsp_state->fraction = 0; + dsp_state->sample_history.fill(0); + dsp_state->context = {}; + } + + in_params.current_playstate = ServerPlayState::Stop; + in_params.should_depop = in_params.last_playstate == ServerPlayState::Play; + return in_params.should_depop; + } + default: + UNREACHABLE_MSG("Invalid playstate {}", in_params.current_playstate); + } + + return false; +} + +void ServerVoiceInfo::FlushWaveBuffers( + u8 flush_count, std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& dsp_voice_states, + s32 channel_count) { + auto wave_head = in_params.wave_bufffer_head; + + for (u8 i = 0; i < flush_count; i++) { + in_params.wave_buffer[wave_head].sent_to_dsp = true; + for (s32 channel = 0; channel < channel_count; channel++) { + auto* dsp_state = dsp_voice_states[channel]; + dsp_state->wave_buffer_consumed++; + dsp_state->is_wave_buffer_valid[wave_head] = false; + dsp_state->wave_buffer_index = + (dsp_state->wave_buffer_index + 1) % AudioCommon::MAX_WAVE_BUFFERS; + } + wave_head = (wave_head + 1) % AudioCommon::MAX_WAVE_BUFFERS; + } +} + +bool ServerVoiceInfo::HasValidWaveBuffer(const VoiceState* state) const { + const auto& valid_wb = state->is_wave_buffer_valid; + return std::find(valid_wb.begin(), valid_wb.end(), true) != valid_wb.end(); +} + +VoiceContext::VoiceContext(std::size_t voice_count) : voice_count(voice_count) { + for (std::size_t i = 0; i < voice_count; i++) { + voice_channel_resources.emplace_back(static_cast<s32>(i)); + sorted_voice_info.push_back(&voice_info.emplace_back()); + voice_states.emplace_back(); + dsp_voice_states.emplace_back(); + } +} + +VoiceContext::~VoiceContext() { + sorted_voice_info.clear(); +} + +std::size_t VoiceContext::GetVoiceCount() const { + return voice_count; +} + +ServerVoiceChannelResource& VoiceContext::GetChannelResource(std::size_t i) { + ASSERT(i < voice_count); + return voice_channel_resources.at(i); +} + +const ServerVoiceChannelResource& VoiceContext::GetChannelResource(std::size_t i) const { + ASSERT(i < voice_count); + return voice_channel_resources.at(i); +} + +VoiceState& VoiceContext::GetState(std::size_t i) { + ASSERT(i < voice_count); + return voice_states.at(i); +} + +const VoiceState& VoiceContext::GetState(std::size_t i) const { + ASSERT(i < voice_count); + return voice_states.at(i); +} + +VoiceState& VoiceContext::GetDspSharedState(std::size_t i) { + ASSERT(i < voice_count); + return dsp_voice_states.at(i); +} + +const VoiceState& VoiceContext::GetDspSharedState(std::size_t i) const { + ASSERT(i < voice_count); + return dsp_voice_states.at(i); +} + +ServerVoiceInfo& VoiceContext::GetInfo(std::size_t i) { + ASSERT(i < voice_count); + return voice_info.at(i); +} + +const ServerVoiceInfo& VoiceContext::GetInfo(std::size_t i) const { + ASSERT(i < voice_count); + return voice_info.at(i); +} + +ServerVoiceInfo& VoiceContext::GetSortedInfo(std::size_t i) { + ASSERT(i < voice_count); + return *sorted_voice_info.at(i); +} + +const ServerVoiceInfo& VoiceContext::GetSortedInfo(std::size_t i) const { + ASSERT(i < voice_count); + return *sorted_voice_info.at(i); +} + +s32 VoiceContext::DecodePcm16(s32* output_buffer, ServerWaveBuffer* wave_buffer, s32 channel, + s32 channel_count, s32 buffer_offset, s32 sample_count, + Core::Memory::Memory& memory) { + if (wave_buffer->buffer_address == 0) { + return 0; + } + if (wave_buffer->buffer_size == 0) { + return 0; + } + if (wave_buffer->end_sample_offset < wave_buffer->start_sample_offset) { + return 0; + } + + const auto samples_remaining = + (wave_buffer->end_sample_offset - wave_buffer->start_sample_offset) - buffer_offset; + const auto start_offset = (wave_buffer->start_sample_offset + buffer_offset) * channel_count; + const auto buffer_pos = wave_buffer->buffer_address + start_offset; + + s16* buffer_data = reinterpret_cast<s16*>(memory.GetPointer(buffer_pos)); + + const auto samples_processed = std::min(sample_count, samples_remaining); + + // Fast path + if (channel_count == 1) { + for (std::ptrdiff_t i = 0; i < samples_processed; i++) { + output_buffer[i] = buffer_data[i]; + } + } else { + for (std::ptrdiff_t i = 0; i < samples_processed; i++) { + output_buffer[i] = buffer_data[i * channel_count + channel]; + } + } + + return samples_processed; +} + +void VoiceContext::SortInfo() { + for (std::size_t i = 0; i < voice_count; i++) { + sorted_voice_info[i] = &voice_info[i]; + } + + std::sort(sorted_voice_info.begin(), sorted_voice_info.end(), + [](const ServerVoiceInfo* lhs, const ServerVoiceInfo* rhs) { + const auto& lhs_in = lhs->GetInParams(); + const auto& rhs_in = rhs->GetInParams(); + // Sort by priority + if (lhs_in.priority != rhs_in.priority) { + return lhs_in.priority > rhs_in.priority; + } else { + // If the priorities match, sort by sorting order + return lhs_in.sorting_order > rhs_in.sorting_order; + } + }); +} + +void VoiceContext::UpdateStateByDspShared() { + voice_states = dsp_voice_states; +} + +} // namespace AudioCore diff --git a/src/audio_core/voice_context.h b/src/audio_core/voice_context.h new file mode 100644 index 000000000..59d3d7dfb --- /dev/null +++ b/src/audio_core/voice_context.h @@ -0,0 +1,296 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include "audio_core/algorithm/interpolate.h" +#include "audio_core/codec.h" +#include "audio_core/common.h" +#include "common/bit_field.h" +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Core::Memory { +class Memory; +} + +namespace AudioCore { + +class BehaviorInfo; +class VoiceContext; + +enum class SampleFormat : u8 { + Invalid = 0, + Pcm8 = 1, + Pcm16 = 2, + Pcm24 = 3, + Pcm32 = 4, + PcmFloat = 5, + Adpcm = 6, +}; + +enum class PlayState : u8 { + Started = 0, + Stopped = 1, + Paused = 2, +}; + +enum class ServerPlayState { + Play = 0, + Stop = 1, + RequestStop = 2, + Paused = 3, +}; + +struct BiquadFilterParameter { + bool enabled{}; + INSERT_PADDING_BYTES(1); + std::array<s16, 3> numerator{}; + std::array<s16, 2> denominator{}; +}; +static_assert(sizeof(BiquadFilterParameter) == 0xc, "BiquadFilterParameter is an invalid size"); + +struct WaveBuffer { + u64_le buffer_address{}; + u64_le buffer_size{}; + s32_le start_sample_offset{}; + s32_le end_sample_offset{}; + u8 is_looping{}; + u8 end_of_stream{}; + u8 sent_to_server{}; + INSERT_PADDING_BYTES(5); + u64 context_address{}; + u64 context_size{}; + INSERT_PADDING_BYTES(8); +}; +static_assert(sizeof(WaveBuffer) == 0x38, "WaveBuffer is an invalid size"); + +struct ServerWaveBuffer { + VAddr buffer_address{}; + std::size_t buffer_size{}; + s32 start_sample_offset{}; + s32 end_sample_offset{}; + bool is_looping{}; + bool end_of_stream{}; + VAddr context_address{}; + std::size_t context_size{}; + bool sent_to_dsp{true}; +}; + +struct BehaviorFlags { + BitField<0, 1, u16> is_played_samples_reset_at_loop_point; + BitField<1, 1, u16> is_pitch_and_src_skipped; +}; +static_assert(sizeof(BehaviorFlags) == 0x4, "BehaviorFlags is an invalid size"); + +struct ADPCMContext { + u16 header{}; + s16 yn1{}; + s16 yn2{}; +}; +static_assert(sizeof(ADPCMContext) == 0x6, "ADPCMContext is an invalid size"); + +struct VoiceState { + s64 played_sample_count{}; + s32 offset{}; + s32 wave_buffer_index{}; + std::array<bool, AudioCommon::MAX_WAVE_BUFFERS> is_wave_buffer_valid{}; + s32 wave_buffer_consumed{}; + std::array<s32, AudioCommon::MAX_SAMPLE_HISTORY> sample_history{}; + s32 fraction{}; + VAddr context_address{}; + Codec::ADPCM_Coeff coeff{}; + ADPCMContext context{}; + std::array<s64, 2> biquad_filter_state{}; + std::array<s32, AudioCommon::MAX_MIX_BUFFERS> previous_samples{}; + u32 external_context_size{}; + bool is_external_context_used{}; + bool voice_dropped{}; +}; + +class VoiceChannelResource { +public: + struct InParams { + s32_le id{}; + std::array<float_le, AudioCommon::MAX_MIX_BUFFERS> mix_volume{}; + bool in_use{}; + INSERT_PADDING_BYTES(11); + }; + static_assert(sizeof(VoiceChannelResource::InParams) == 0x70, "InParams is an invalid size"); +}; + +class ServerVoiceChannelResource { +public: + explicit ServerVoiceChannelResource(s32 id); + ~ServerVoiceChannelResource(); + + bool InUse() const; + float GetCurrentMixVolumeAt(std::size_t i) const; + float GetLastMixVolumeAt(std::size_t i) const; + void Update(VoiceChannelResource::InParams& in_params); + void UpdateLastMixVolumes(); + + const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& GetCurrentMixVolume() const; + const std::array<float, AudioCommon::MAX_MIX_BUFFERS>& GetLastMixVolume() const; + +private: + s32 id{}; + std::array<float, AudioCommon::MAX_MIX_BUFFERS> mix_volume{}; + std::array<float, AudioCommon::MAX_MIX_BUFFERS> last_mix_volume{}; + bool in_use{}; +}; + +class VoiceInfo { +public: + struct InParams { + s32_le id{}; + u32_le node_id{}; + u8 is_new{}; + u8 is_in_use{}; + PlayState play_state{}; + SampleFormat sample_format{}; + s32_le sample_rate{}; + s32_le priority{}; + s32_le sorting_order{}; + s32_le channel_count{}; + float_le pitch{}; + float_le volume{}; + std::array<BiquadFilterParameter, 2> biquad_filter{}; + s32_le wave_buffer_count{}; + s16_le wave_buffer_head{}; + INSERT_PADDING_BYTES(6); + u64_le additional_params_address{}; + u64_le additional_params_size{}; + s32_le mix_id{}; + s32_le splitter_info_id{}; + std::array<WaveBuffer, 4> wave_buffer{}; + std::array<u32_le, 6> voice_channel_resource_ids{}; + // TODO(ogniK): Remaining flags + u8 is_voice_drop_flag_clear_requested{}; + u8 wave_buffer_flush_request_count{}; + INSERT_PADDING_BYTES(2); + BehaviorFlags behavior_flags{}; + INSERT_PADDING_BYTES(16); + }; + static_assert(sizeof(VoiceInfo::InParams) == 0x170, "InParams is an invalid size"); + + struct OutParams { + u64_le played_sample_count{}; + u32_le wave_buffer_consumed{}; + u8 voice_dropped{}; + INSERT_PADDING_BYTES(3); + }; + static_assert(sizeof(VoiceInfo::OutParams) == 0x10, "OutParams is an invalid size"); +}; + +class ServerVoiceInfo { +public: + struct InParams { + bool in_use{}; + bool is_new{}; + bool should_depop{}; + SampleFormat sample_format{}; + s32 sample_rate{}; + s32 channel_count{}; + s32 id{}; + s32 node_id{}; + s32 mix_id{}; + ServerPlayState current_playstate{}; + ServerPlayState last_playstate{}; + s32 priority{}; + s32 sorting_order{}; + float pitch{}; + float volume{}; + float last_volume{}; + std::array<BiquadFilterParameter, AudioCommon::MAX_BIQUAD_FILTERS> biquad_filter{}; + s32 wave_buffer_count{}; + s16 wave_bufffer_head{}; + INSERT_PADDING_BYTES(2); + BehaviorFlags behavior_flags{}; + VAddr additional_params_address{}; + std::size_t additional_params_size{}; + std::array<ServerWaveBuffer, AudioCommon::MAX_WAVE_BUFFERS> wave_buffer{}; + std::array<s32, AudioCommon::MAX_CHANNEL_COUNT> voice_channel_resource_id{}; + s32 splitter_info_id{}; + u8 wave_buffer_flush_request_count{}; + bool voice_drop_flag{}; + bool buffer_mapped{}; + std::array<bool, AudioCommon::MAX_BIQUAD_FILTERS> was_biquad_filter_enabled{}; + }; + + struct OutParams { + s64 played_sample_count{}; + s32 wave_buffer_consumed{}; + }; + + ServerVoiceInfo(); + ~ServerVoiceInfo(); + void Initialize(); + void UpdateParameters(const VoiceInfo::InParams& voice_in, BehaviorInfo& behavior_info); + void UpdateWaveBuffers(const VoiceInfo::InParams& voice_in, + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& voice_states, + BehaviorInfo& behavior_info); + void UpdateWaveBuffer(ServerWaveBuffer& out_wavebuffer, const WaveBuffer& in_wave_buffer, + SampleFormat sample_format, bool is_buffer_valid, + BehaviorInfo& behavior_info); + void WriteOutStatus(VoiceInfo::OutParams& voice_out, VoiceInfo::InParams& voice_in, + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& voice_states); + + const InParams& GetInParams() const; + InParams& GetInParams(); + + const OutParams& GetOutParams() const; + OutParams& GetOutParams(); + + bool ShouldSkip() const; + bool UpdateForCommandGeneration(VoiceContext& voice_context); + void ResetResources(VoiceContext& voice_context); + bool UpdateParametersForCommandGeneration( + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& dsp_voice_states); + void FlushWaveBuffers(u8 flush_count, + std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>& dsp_voice_states, + s32 channel_count); + +private: + std::vector<s16> stored_samples; + InParams in_params{}; + OutParams out_params{}; + + bool HasValidWaveBuffer(const VoiceState* state) const; +}; + +class VoiceContext { +public: + VoiceContext(std::size_t voice_count); + ~VoiceContext(); + + std::size_t GetVoiceCount() const; + ServerVoiceChannelResource& GetChannelResource(std::size_t i); + const ServerVoiceChannelResource& GetChannelResource(std::size_t i) const; + VoiceState& GetState(std::size_t i); + const VoiceState& GetState(std::size_t i) const; + VoiceState& GetDspSharedState(std::size_t i); + const VoiceState& GetDspSharedState(std::size_t i) const; + ServerVoiceInfo& GetInfo(std::size_t i); + const ServerVoiceInfo& GetInfo(std::size_t i) const; + ServerVoiceInfo& GetSortedInfo(std::size_t i); + const ServerVoiceInfo& GetSortedInfo(std::size_t i) const; + + s32 DecodePcm16(s32* output_buffer, ServerWaveBuffer* wave_buffer, s32 channel, + s32 channel_count, s32 buffer_offset, s32 sample_count, + Core::Memory::Memory& memory); + void SortInfo(); + void UpdateStateByDspShared(); + +private: + std::size_t voice_count{}; + std::vector<ServerVoiceChannelResource> voice_channel_resources{}; + std::vector<VoiceState> voice_states{}; + std::vector<VoiceState> dsp_voice_states{}; + std::vector<ServerVoiceInfo> voice_info{}; + std::vector<ServerVoiceInfo*> sorted_voice_info{}; +}; + +} // namespace AudioCore diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt index eeceaa655..d20e6c3b5 100644 --- a/src/common/CMakeLists.txt +++ b/src/common/CMakeLists.txt @@ -32,6 +32,8 @@ add_custom_command(OUTPUT scm_rev.cpp DEPENDS # WARNING! It was too much work to try and make a common location for this list, # so if you need to change it, please update CMakeModules/GenerateSCMRev.cmake as well + "${VIDEO_CORE}/renderer_opengl/gl_arb_decompiler.cpp" + "${VIDEO_CORE}/renderer_opengl/gl_arb_decompiler.h" "${VIDEO_CORE}/renderer_opengl/gl_shader_cache.cpp" "${VIDEO_CORE}/renderer_opengl/gl_shader_cache.h" "${VIDEO_CORE}/renderer_opengl/gl_shader_decompiler.cpp" @@ -96,8 +98,11 @@ add_library(common STATIC algorithm.h alignment.h assert.h + atomic_ops.cpp + atomic_ops.h detached_tasks.cpp detached_tasks.h + bit_cast.h bit_field.h bit_util.h cityhash.cpp @@ -106,8 +111,11 @@ add_library(common STATIC common_funcs.h common_paths.h common_types.h + concepts.h dynamic_library.cpp dynamic_library.h + fiber.cpp + fiber.h file_util.cpp file_util.h hash.h @@ -123,6 +131,8 @@ add_library(common STATIC lz4_compression.cpp lz4_compression.h math_util.h + memory_detect.cpp + memory_detect.h memory_hook.cpp memory_hook.h microprofile.cpp @@ -139,6 +149,10 @@ add_library(common STATIC scm_rev.cpp scm_rev.h scope_exit.h + spin_lock.cpp + spin_lock.h + stream.cpp + stream.h string_util.cpp string_util.h swap.h @@ -148,6 +162,8 @@ add_library(common STATIC thread.h thread_queue_list.h threadsafe_queue.h + time_zone.cpp + time_zone.h timer.cpp timer.h uint128.cpp @@ -155,7 +171,10 @@ add_library(common STATIC uuid.cpp uuid.h vector_math.h - web_result.h + virtual_buffer.cpp + virtual_buffer.h + wall_clock.cpp + wall_clock.h zstd_compression.cpp zstd_compression.h ) @@ -165,10 +184,36 @@ if(ARCHITECTURE_x86_64) PRIVATE x64/cpu_detect.cpp x64/cpu_detect.h + x64/native_clock.cpp + x64/native_clock.h + x64/xbyak_abi.h + x64/xbyak_util.h ) endif() +if (MSVC) + target_compile_definitions(common PRIVATE + # The standard library doesn't provide any replacement for codecvt yet + # so we can disable this deprecation warning for the time being. + _SILENCE_CXX17_CODECVT_HEADER_DEPRECATION_WARNING + ) + target_compile_options(common PRIVATE + /W4 + /WX + ) +else() + target_compile_options(common PRIVATE + -Werror + ) +endif() + create_target_directory_groups(common) +find_package(Boost 1.71 COMPONENTS context headers REQUIRED) -target_link_libraries(common PUBLIC Boost::boost fmt microprofile) -target_link_libraries(common PRIVATE lz4_static libzstd_static) +target_link_libraries(common PUBLIC ${Boost_LIBRARIES} fmt::fmt microprofile) +target_link_libraries(common PRIVATE lz4::lz4 xbyak) +if (MSVC) + target_link_libraries(common PRIVATE zstd::zstd) +else() + target_link_libraries(common PRIVATE zstd) +endif() diff --git a/src/common/algorithm.h b/src/common/algorithm.h index e21b1373c..4804a3421 100644 --- a/src/common/algorithm.h +++ b/src/common/algorithm.h @@ -15,7 +15,8 @@ namespace Common { template <class ForwardIt, class T, class Compare = std::less<>> -ForwardIt BinaryFind(ForwardIt first, ForwardIt last, const T& value, Compare comp = {}) { +[[nodiscard]] ForwardIt BinaryFind(ForwardIt first, ForwardIt last, const T& value, + Compare comp = {}) { // Note: BOTH type T and the type after ForwardIt is dereferenced // must be implicitly convertible to BOTH Type1 and Type2, used in Compare. // This is stricter than lower_bound requirement (see above) diff --git a/src/common/alignment.h b/src/common/alignment.h index cdd4833f8..5040043de 100644 --- a/src/common/alignment.h +++ b/src/common/alignment.h @@ -3,41 +3,50 @@ #pragma once #include <cstddef> -#include <memory> +#include <new> #include <type_traits> namespace Common { template <typename T> -constexpr T AlignUp(T value, std::size_t size) { +[[nodiscard]] constexpr T AlignUp(T value, std::size_t size) { static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); - return static_cast<T>(value + (size - value % size) % size); + auto mod{static_cast<T>(value % size)}; + value -= mod; + return static_cast<T>(mod == T{0} ? value : value + size); } template <typename T> -constexpr T AlignDown(T value, std::size_t size) { +[[nodiscard]] constexpr T AlignDown(T value, std::size_t size) { static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); return static_cast<T>(value - value % size); } template <typename T> -constexpr T AlignBits(T value, std::size_t align) { +[[nodiscard]] constexpr T AlignBits(T value, std::size_t align) { static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); return static_cast<T>((value + ((1ULL << align) - 1)) >> align << align); } template <typename T> -constexpr bool Is4KBAligned(T value) { +[[nodiscard]] constexpr bool Is4KBAligned(T value) { static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); return (value & 0xFFF) == 0; } template <typename T> -constexpr bool IsWordAligned(T value) { +[[nodiscard]] constexpr bool IsWordAligned(T value) { static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); return (value & 0b11) == 0; } +template <typename T> +[[nodiscard]] constexpr bool IsAligned(T value, std::size_t alignment) { + using U = typename std::make_unsigned<T>::type; + const U mask = static_cast<U>(alignment - 1); + return (value & mask) == 0; +} + template <typename T, std::size_t Align = 16> class AlignmentAllocator { public: @@ -45,66 +54,28 @@ public: using size_type = std::size_t; using difference_type = std::ptrdiff_t; - using pointer = T*; - using const_pointer = const T*; - - using reference = T&; - using const_reference = const T&; - using propagate_on_container_copy_assignment = std::true_type; using propagate_on_container_move_assignment = std::true_type; using propagate_on_container_swap = std::true_type; using is_always_equal = std::true_type; -public: constexpr AlignmentAllocator() noexcept = default; template <typename T2> constexpr AlignmentAllocator(const AlignmentAllocator<T2, Align>&) noexcept {} - pointer address(reference r) noexcept { - return std::addressof(r); - } - - const_pointer address(const_reference r) const noexcept { - return std::addressof(r); - } - - pointer allocate(size_type n) { - return static_cast<pointer>(::operator new (n, std::align_val_t{Align})); + [[nodiscard]] T* allocate(size_type n) { + return static_cast<T*>(::operator new (n * sizeof(T), std::align_val_t{Align})); } - void deallocate(pointer p, size_type) { - ::operator delete (p, std::align_val_t{Align}); - } - - void construct(pointer p, const value_type& wert) { - new (p) value_type(wert); - } - - void destroy(pointer p) { - p->~value_type(); - } - - size_type max_size() const noexcept { - return size_type(-1) / sizeof(value_type); + void deallocate(T* p, size_type n) { + ::operator delete (p, n * sizeof(T), std::align_val_t{Align}); } template <typename T2> struct rebind { using other = AlignmentAllocator<T2, Align>; }; - - bool operator!=(const AlignmentAllocator<T, Align>& other) const noexcept { - return !(*this == other); - } - - // Returns true if and only if storage allocated from *this - // can be deallocated from other, and vice versa. - // Always returns true for stateless allocators. - bool operator==(const AlignmentAllocator<T, Align>& other) const noexcept { - return true; - } }; } // namespace Common diff --git a/src/common/assert.h b/src/common/assert.h index 5b67c5c52..06d7b5612 100644 --- a/src/common/assert.h +++ b/src/common/assert.h @@ -17,11 +17,12 @@ // enough for our purposes. template <typename Fn> #if defined(_MSC_VER) -__declspec(noinline, noreturn) +[[msvc::noinline, noreturn]] #elif defined(__GNUC__) - __attribute__((noinline, noreturn, cold)) +[[gnu::cold, gnu::noinline, noreturn]] #endif - static void assert_noinline_call(const Fn& fn) { +static void +assert_noinline_call(const Fn& fn) { fn(); Crash(); exit(1); // Keeps GCC's mouth shut about this actually returning diff --git a/src/common/atomic_ops.cpp b/src/common/atomic_ops.cpp new file mode 100644 index 000000000..1612d0e67 --- /dev/null +++ b/src/common/atomic_ops.cpp @@ -0,0 +1,75 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> + +#include "common/atomic_ops.h" + +#if _MSC_VER +#include <intrin.h> +#endif + +namespace Common { + +#if _MSC_VER + +bool AtomicCompareAndSwap(volatile u8* pointer, u8 value, u8 expected) { + const u8 result = + _InterlockedCompareExchange8(reinterpret_cast<volatile char*>(pointer), value, expected); + return result == expected; +} + +bool AtomicCompareAndSwap(volatile u16* pointer, u16 value, u16 expected) { + const u16 result = + _InterlockedCompareExchange16(reinterpret_cast<volatile short*>(pointer), value, expected); + return result == expected; +} + +bool AtomicCompareAndSwap(volatile u32* pointer, u32 value, u32 expected) { + const u32 result = + _InterlockedCompareExchange(reinterpret_cast<volatile long*>(pointer), value, expected); + return result == expected; +} + +bool AtomicCompareAndSwap(volatile u64* pointer, u64 value, u64 expected) { + const u64 result = _InterlockedCompareExchange64(reinterpret_cast<volatile __int64*>(pointer), + value, expected); + return result == expected; +} + +bool AtomicCompareAndSwap(volatile u64* pointer, u128 value, u128 expected) { + return _InterlockedCompareExchange128(reinterpret_cast<volatile __int64*>(pointer), value[1], + value[0], + reinterpret_cast<__int64*>(expected.data())) != 0; +} + +#else + +bool AtomicCompareAndSwap(volatile u8* pointer, u8 value, u8 expected) { + return __sync_bool_compare_and_swap(pointer, expected, value); +} + +bool AtomicCompareAndSwap(volatile u16* pointer, u16 value, u16 expected) { + return __sync_bool_compare_and_swap(pointer, expected, value); +} + +bool AtomicCompareAndSwap(volatile u32* pointer, u32 value, u32 expected) { + return __sync_bool_compare_and_swap(pointer, expected, value); +} + +bool AtomicCompareAndSwap(volatile u64* pointer, u64 value, u64 expected) { + return __sync_bool_compare_and_swap(pointer, expected, value); +} + +bool AtomicCompareAndSwap(volatile u64* pointer, u128 value, u128 expected) { + unsigned __int128 value_a; + unsigned __int128 expected_a; + std::memcpy(&value_a, value.data(), sizeof(u128)); + std::memcpy(&expected_a, expected.data(), sizeof(u128)); + return __sync_bool_compare_and_swap((unsigned __int128*)pointer, expected_a, value_a); +} + +#endif + +} // namespace Common diff --git a/src/common/atomic_ops.h b/src/common/atomic_ops.h new file mode 100644 index 000000000..b46888589 --- /dev/null +++ b/src/common/atomic_ops.h @@ -0,0 +1,17 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace Common { + +[[nodiscard]] bool AtomicCompareAndSwap(volatile u8* pointer, u8 value, u8 expected); +[[nodiscard]] bool AtomicCompareAndSwap(volatile u16* pointer, u16 value, u16 expected); +[[nodiscard]] bool AtomicCompareAndSwap(volatile u32* pointer, u32 value, u32 expected); +[[nodiscard]] bool AtomicCompareAndSwap(volatile u64* pointer, u64 value, u64 expected); +[[nodiscard]] bool AtomicCompareAndSwap(volatile u64* pointer, u128 value, u128 expected); + +} // namespace Common diff --git a/src/common/bit_cast.h b/src/common/bit_cast.h new file mode 100644 index 000000000..a32a063d1 --- /dev/null +++ b/src/common/bit_cast.h @@ -0,0 +1,22 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstring> +#include <type_traits> + +namespace Common { + +template <typename To, typename From> +[[nodiscard]] std::enable_if_t<sizeof(To) == sizeof(From) && std::is_trivially_copyable_v<From> && + std::is_trivially_copyable_v<To>, + To> +BitCast(const From& src) noexcept { + To dst; + std::memcpy(&dst, &src, sizeof(To)); + return dst; +} + +} // namespace Common diff --git a/src/common/bit_field.h b/src/common/bit_field.h index fd2bbbd99..0f0661172 100644 --- a/src/common/bit_field.h +++ b/src/common/bit_field.h @@ -36,13 +36,6 @@ #include "common/common_funcs.h" #include "common/swap.h" -// Inlining -#ifdef _WIN32 -#define FORCE_INLINE __forceinline -#else -#define FORCE_INLINE inline __attribute__((always_inline)) -#endif - /* * Abstract bitfield class * @@ -142,8 +135,8 @@ public: * containing several bitfields can be assembled by formatting each of their values and ORing * the results together. */ - static constexpr FORCE_INLINE StorageType FormatValue(const T& value) { - return ((StorageType)value << position) & mask; + [[nodiscard]] static constexpr StorageType FormatValue(const T& value) { + return (static_cast<StorageType>(value) << position) & mask; } /** @@ -151,7 +144,7 @@ public: * (such as Value() or operator T), but this can be used to extract a value from a bitfield * union in a constexpr context. */ - static constexpr FORCE_INLINE T ExtractValue(const StorageType& storage) { + [[nodiscard]] static constexpr T ExtractValue(const StorageType& storage) { if constexpr (std::numeric_limits<UnderlyingType>::is_signed) { std::size_t shift = 8 * sizeof(T) - bits; return static_cast<T>(static_cast<UnderlyingType>(storage << (shift - position)) >> @@ -175,19 +168,19 @@ public: constexpr BitField(BitField&&) noexcept = default; constexpr BitField& operator=(BitField&&) noexcept = default; - constexpr operator T() const { + [[nodiscard]] constexpr operator T() const { return Value(); } constexpr void Assign(const T& value) { - storage = (static_cast<StorageType>(storage) & ~mask) | FormatValue(value); + storage = static_cast<StorageType>((storage & ~mask) | FormatValue(value)); } - constexpr T Value() const { + [[nodiscard]] constexpr T Value() const { return ExtractValue(storage); } - constexpr explicit operator bool() const { + [[nodiscard]] constexpr explicit operator bool() const { return Value() != 0; } diff --git a/src/common/bit_util.h b/src/common/bit_util.h index 6f7d5a947..29f59a9a3 100644 --- a/src/common/bit_util.h +++ b/src/common/bit_util.h @@ -17,12 +17,12 @@ namespace Common { /// Gets the size of a specified type T in bits. template <typename T> -constexpr std::size_t BitSize() { +[[nodiscard]] constexpr std::size_t BitSize() { return sizeof(T) * CHAR_BIT; } #ifdef _MSC_VER -inline u32 CountLeadingZeroes32(u32 value) { +[[nodiscard]] inline u32 CountLeadingZeroes32(u32 value) { unsigned long leading_zero = 0; if (_BitScanReverse(&leading_zero, value) != 0) { @@ -32,7 +32,7 @@ inline u32 CountLeadingZeroes32(u32 value) { return 32; } -inline u32 CountLeadingZeroes64(u64 value) { +[[nodiscard]] inline u32 CountLeadingZeroes64(u64 value) { unsigned long leading_zero = 0; if (_BitScanReverse64(&leading_zero, value) != 0) { @@ -42,7 +42,7 @@ inline u32 CountLeadingZeroes64(u64 value) { return 64; } #else -inline u32 CountLeadingZeroes32(u32 value) { +[[nodiscard]] inline u32 CountLeadingZeroes32(u32 value) { if (value == 0) { return 32; } @@ -50,7 +50,7 @@ inline u32 CountLeadingZeroes32(u32 value) { return static_cast<u32>(__builtin_clz(value)); } -inline u32 CountLeadingZeroes64(u64 value) { +[[nodiscard]] inline u32 CountLeadingZeroes64(u64 value) { if (value == 0) { return 64; } @@ -60,7 +60,7 @@ inline u32 CountLeadingZeroes64(u64 value) { #endif #ifdef _MSC_VER -inline u32 CountTrailingZeroes32(u32 value) { +[[nodiscard]] inline u32 CountTrailingZeroes32(u32 value) { unsigned long trailing_zero = 0; if (_BitScanForward(&trailing_zero, value) != 0) { @@ -70,7 +70,7 @@ inline u32 CountTrailingZeroes32(u32 value) { return 32; } -inline u32 CountTrailingZeroes64(u64 value) { +[[nodiscard]] inline u32 CountTrailingZeroes64(u64 value) { unsigned long trailing_zero = 0; if (_BitScanForward64(&trailing_zero, value) != 0) { @@ -80,7 +80,7 @@ inline u32 CountTrailingZeroes64(u64 value) { return 64; } #else -inline u32 CountTrailingZeroes32(u32 value) { +[[nodiscard]] inline u32 CountTrailingZeroes32(u32 value) { if (value == 0) { return 32; } @@ -88,7 +88,7 @@ inline u32 CountTrailingZeroes32(u32 value) { return static_cast<u32>(__builtin_ctz(value)); } -inline u32 CountTrailingZeroes64(u64 value) { +[[nodiscard]] inline u32 CountTrailingZeroes64(u64 value) { if (value == 0) { return 64; } @@ -99,13 +99,13 @@ inline u32 CountTrailingZeroes64(u64 value) { #ifdef _MSC_VER -inline u32 MostSignificantBit32(const u32 value) { +[[nodiscard]] inline u32 MostSignificantBit32(const u32 value) { unsigned long result; _BitScanReverse(&result, value); return static_cast<u32>(result); } -inline u32 MostSignificantBit64(const u64 value) { +[[nodiscard]] inline u32 MostSignificantBit64(const u64 value) { unsigned long result; _BitScanReverse64(&result, value); return static_cast<u32>(result); @@ -113,30 +113,30 @@ inline u32 MostSignificantBit64(const u64 value) { #else -inline u32 MostSignificantBit32(const u32 value) { +[[nodiscard]] inline u32 MostSignificantBit32(const u32 value) { return 31U - static_cast<u32>(__builtin_clz(value)); } -inline u32 MostSignificantBit64(const u64 value) { +[[nodiscard]] inline u32 MostSignificantBit64(const u64 value) { return 63U - static_cast<u32>(__builtin_clzll(value)); } #endif -inline u32 Log2Floor32(const u32 value) { +[[nodiscard]] inline u32 Log2Floor32(const u32 value) { return MostSignificantBit32(value); } -inline u32 Log2Ceil32(const u32 value) { +[[nodiscard]] inline u32 Log2Ceil32(const u32 value) { const u32 log2_f = Log2Floor32(value); return log2_f + ((value ^ (1U << log2_f)) != 0U); } -inline u32 Log2Floor64(const u64 value) { +[[nodiscard]] inline u32 Log2Floor64(const u64 value) { return MostSignificantBit64(value); } -inline u32 Log2Ceil64(const u64 value) { +[[nodiscard]] inline u32 Log2Ceil64(const u64 value) { const u64 log2_f = static_cast<u64>(Log2Floor64(value)); return static_cast<u32>(log2_f + ((value ^ (1ULL << log2_f)) != 0ULL)); } diff --git a/src/common/cityhash.h b/src/common/cityhash.h index 4b94f8e18..a00804e01 100644 --- a/src/common/cityhash.h +++ b/src/common/cityhash.h @@ -61,42 +61,43 @@ #pragma once +#include <cstddef> +#include <cstdint> #include <utility> -#include <stdint.h> -#include <stdlib.h> // for std::size_t. namespace Common { -typedef std::pair<uint64_t, uint64_t> uint128; +using uint128 = std::pair<uint64_t, uint64_t>; -inline uint64_t Uint128Low64(const uint128& x) { +[[nodiscard]] inline uint64_t Uint128Low64(const uint128& x) { return x.first; } -inline uint64_t Uint128High64(const uint128& x) { +[[nodiscard]] inline uint64_t Uint128High64(const uint128& x) { return x.second; } // Hash function for a byte array. -uint64_t CityHash64(const char* buf, std::size_t len); +[[nodiscard]] uint64_t CityHash64(const char* buf, std::size_t len); // Hash function for a byte array. For convenience, a 64-bit seed is also // hashed into the result. -uint64_t CityHash64WithSeed(const char* buf, std::size_t len, uint64_t seed); +[[nodiscard]] uint64_t CityHash64WithSeed(const char* buf, std::size_t len, uint64_t seed); // Hash function for a byte array. For convenience, two seeds are also // hashed into the result. -uint64_t CityHash64WithSeeds(const char* buf, std::size_t len, uint64_t seed0, uint64_t seed1); +[[nodiscard]] uint64_t CityHash64WithSeeds(const char* buf, std::size_t len, uint64_t seed0, + uint64_t seed1); // Hash function for a byte array. -uint128 CityHash128(const char* s, std::size_t len); +[[nodiscard]] uint128 CityHash128(const char* s, std::size_t len); // Hash function for a byte array. For convenience, a 128-bit seed is also // hashed into the result. -uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed); +[[nodiscard]] uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed); // Hash 128 input bits down to 64 bits of output. // This is intended to be a reasonably good hash function. -inline uint64_t Hash128to64(const uint128& x) { +[[nodiscard]] inline uint64_t Hash128to64(const uint128& x) { // Murmur-inspired hashing. const uint64_t kMul = 0x9ddfea08eb382d69ULL; uint64_t a = (Uint128Low64(x) ^ Uint128High64(x)) * kMul; diff --git a/src/common/color.h b/src/common/color.h index 3a2222077..bbcac858e 100644 --- a/src/common/color.h +++ b/src/common/color.h @@ -10,45 +10,45 @@ #include "common/swap.h" #include "common/vector_math.h" -namespace Color { +namespace Common::Color { /// Convert a 1-bit color component to 8 bit -constexpr u8 Convert1To8(u8 value) { +[[nodiscard]] constexpr u8 Convert1To8(u8 value) { return value * 255; } /// Convert a 4-bit color component to 8 bit -constexpr u8 Convert4To8(u8 value) { +[[nodiscard]] constexpr u8 Convert4To8(u8 value) { return (value << 4) | value; } /// Convert a 5-bit color component to 8 bit -constexpr u8 Convert5To8(u8 value) { +[[nodiscard]] constexpr u8 Convert5To8(u8 value) { return (value << 3) | (value >> 2); } /// Convert a 6-bit color component to 8 bit -constexpr u8 Convert6To8(u8 value) { +[[nodiscard]] constexpr u8 Convert6To8(u8 value) { return (value << 2) | (value >> 4); } /// Convert a 8-bit color component to 1 bit -constexpr u8 Convert8To1(u8 value) { +[[nodiscard]] constexpr u8 Convert8To1(u8 value) { return value >> 7; } /// Convert a 8-bit color component to 4 bit -constexpr u8 Convert8To4(u8 value) { +[[nodiscard]] constexpr u8 Convert8To4(u8 value) { return value >> 4; } /// Convert a 8-bit color component to 5 bit -constexpr u8 Convert8To5(u8 value) { +[[nodiscard]] constexpr u8 Convert8To5(u8 value) { return value >> 3; } /// Convert a 8-bit color component to 6 bit -constexpr u8 Convert8To6(u8 value) { +[[nodiscard]] constexpr u8 Convert8To6(u8 value) { return value >> 2; } @@ -57,7 +57,7 @@ constexpr u8 Convert8To6(u8 value) { * @param bytes Pointer to encoded source color * @return Result color decoded as Common::Vec4<u8> */ -inline Common::Vec4<u8> DecodeRGBA8(const u8* bytes) { +[[nodiscard]] inline Common::Vec4<u8> DecodeRGBA8(const u8* bytes) { return {bytes[3], bytes[2], bytes[1], bytes[0]}; } @@ -66,7 +66,7 @@ inline Common::Vec4<u8> DecodeRGBA8(const u8* bytes) { * @param bytes Pointer to encoded source color * @return Result color decoded as Common::Vec4<u8> */ -inline Common::Vec4<u8> DecodeRGB8(const u8* bytes) { +[[nodiscard]] inline Common::Vec4<u8> DecodeRGB8(const u8* bytes) { return {bytes[2], bytes[1], bytes[0], 255}; } @@ -75,7 +75,7 @@ inline Common::Vec4<u8> DecodeRGB8(const u8* bytes) { * @param bytes Pointer to encoded source color * @return Result color decoded as Common::Vec4<u8> */ -inline Common::Vec4<u8> DecodeRG8(const u8* bytes) { +[[nodiscard]] inline Common::Vec4<u8> DecodeRG8(const u8* bytes) { return {bytes[1], bytes[0], 0, 255}; } @@ -84,7 +84,7 @@ inline Common::Vec4<u8> DecodeRG8(const u8* bytes) { * @param bytes Pointer to encoded source color * @return Result color decoded as Common::Vec4<u8> */ -inline Common::Vec4<u8> DecodeRGB565(const u8* bytes) { +[[nodiscard]] inline Common::Vec4<u8> DecodeRGB565(const u8* bytes) { u16_le pixel; std::memcpy(&pixel, bytes, sizeof(pixel)); return {Convert5To8((pixel >> 11) & 0x1F), Convert6To8((pixel >> 5) & 0x3F), @@ -96,7 +96,7 @@ inline Common::Vec4<u8> DecodeRGB565(const u8* bytes) { * @param bytes Pointer to encoded source color * @return Result color decoded as Common::Vec4<u8> */ -inline Common::Vec4<u8> DecodeRGB5A1(const u8* bytes) { +[[nodiscard]] inline Common::Vec4<u8> DecodeRGB5A1(const u8* bytes) { u16_le pixel; std::memcpy(&pixel, bytes, sizeof(pixel)); return {Convert5To8((pixel >> 11) & 0x1F), Convert5To8((pixel >> 6) & 0x1F), @@ -108,7 +108,7 @@ inline Common::Vec4<u8> DecodeRGB5A1(const u8* bytes) { * @param bytes Pointer to encoded source color * @return Result color decoded as Common::Vec4<u8> */ -inline Common::Vec4<u8> DecodeRGBA4(const u8* bytes) { +[[nodiscard]] inline Common::Vec4<u8> DecodeRGBA4(const u8* bytes) { u16_le pixel; std::memcpy(&pixel, bytes, sizeof(pixel)); return {Convert4To8((pixel >> 12) & 0xF), Convert4To8((pixel >> 8) & 0xF), @@ -120,7 +120,7 @@ inline Common::Vec4<u8> DecodeRGBA4(const u8* bytes) { * @param bytes Pointer to encoded source value * @return Depth value as an u32 */ -inline u32 DecodeD16(const u8* bytes) { +[[nodiscard]] inline u32 DecodeD16(const u8* bytes) { u16_le data; std::memcpy(&data, bytes, sizeof(data)); return data; @@ -131,7 +131,7 @@ inline u32 DecodeD16(const u8* bytes) { * @param bytes Pointer to encoded source value * @return Depth value as an u32 */ -inline u32 DecodeD24(const u8* bytes) { +[[nodiscard]] inline u32 DecodeD24(const u8* bytes) { return (bytes[2] << 16) | (bytes[1] << 8) | bytes[0]; } @@ -140,7 +140,7 @@ inline u32 DecodeD24(const u8* bytes) { * @param bytes Pointer to encoded source values * @return Resulting values stored as a Common::Vec2 */ -inline Common::Vec2<u32> DecodeD24S8(const u8* bytes) { +[[nodiscard]] inline Common::Vec2<u32> DecodeD24S8(const u8* bytes) { return {static_cast<u32>((bytes[2] << 16) | (bytes[1] << 8) | bytes[0]), bytes[3]}; } @@ -268,4 +268,4 @@ inline void EncodeX24S8(u8 stencil, u8* bytes) { bytes[3] = stencil; } -} // namespace Color +} // namespace Common::Color diff --git a/src/common/common_funcs.h b/src/common/common_funcs.h index 052254678..367b6bf6e 100644 --- a/src/common/common_funcs.h +++ b/src/common/common_funcs.h @@ -53,11 +53,49 @@ __declspec(dllimport) void __stdcall DebugBreak(void); // Call directly after the command or use the error num. // This function might change the error code. // Defined in Misc.cpp. -std::string GetLastErrorMsg(); +[[nodiscard]] std::string GetLastErrorMsg(); + +#define DECLARE_ENUM_FLAG_OPERATORS(type) \ + [[nodiscard]] constexpr type operator|(type a, type b) noexcept { \ + using T = std::underlying_type_t<type>; \ + return static_cast<type>(static_cast<T>(a) | static_cast<T>(b)); \ + } \ + [[nodiscard]] constexpr type operator&(type a, type b) noexcept { \ + using T = std::underlying_type_t<type>; \ + return static_cast<type>(static_cast<T>(a) & static_cast<T>(b)); \ + } \ + [[nodiscard]] constexpr type operator^(type a, type b) noexcept { \ + using T = std::underlying_type_t<type>; \ + return static_cast<type>(static_cast<T>(a) ^ static_cast<T>(b)); \ + } \ + constexpr type& operator|=(type& a, type b) noexcept { \ + a = a | b; \ + return a; \ + } \ + constexpr type& operator&=(type& a, type b) noexcept { \ + a = a & b; \ + return a; \ + } \ + constexpr type& operator^=(type& a, type b) noexcept { \ + a = a ^ b; \ + return a; \ + } \ + [[nodiscard]] constexpr type operator~(type key) noexcept { \ + using T = std::underlying_type_t<type>; \ + return static_cast<type>(~static_cast<T>(key)); \ + } \ + [[nodiscard]] constexpr bool True(type key) noexcept { \ + using T = std::underlying_type_t<type>; \ + return static_cast<T>(key) != 0; \ + } \ + [[nodiscard]] constexpr bool False(type key) noexcept { \ + using T = std::underlying_type_t<type>; \ + return static_cast<T>(key) == 0; \ + } namespace Common { -constexpr u32 MakeMagic(char a, char b, char c, char d) { +[[nodiscard]] constexpr u32 MakeMagic(char a, char b, char c, char d) { return u32(a) | u32(b) << 8 | u32(c) << 16 | u32(d) << 24; } diff --git a/src/common/common_paths.h b/src/common/common_paths.h index 076752d3b..3c593d5f6 100644 --- a/src/common/common_paths.h +++ b/src/common/common_paths.h @@ -35,6 +35,7 @@ #define KEYS_DIR "keys" #define LOAD_DIR "load" #define DUMP_DIR "dump" +#define SCREENSHOTS_DIR "screenshots" #define SHADER_DIR "shader" #define LOG_DIR "log" diff --git a/src/common/concepts.h b/src/common/concepts.h new file mode 100644 index 000000000..5bef3ad67 --- /dev/null +++ b/src/common/concepts.h @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <type_traits> + +namespace Common { + +// Check if type is like an STL container +template <typename T> +concept IsSTLContainer = requires(T t) { + typename T::value_type; + typename T::iterator; + typename T::const_iterator; + // TODO(ogniK): Replace below is std::same_as<void> when MSVC supports it. + t.begin(); + t.end(); + t.cbegin(); + t.cend(); + t.data(); + t.size(); +}; + +// TODO: Replace with std::derived_from when the <concepts> header +// is available on all supported platforms. +template <typename Derived, typename Base> +concept DerivedFrom = requires { + std::is_base_of_v<Base, Derived>; + std::is_convertible_v<const volatile Derived*, const volatile Base*>; +}; + +} // namespace Common diff --git a/src/common/detached_tasks.cpp b/src/common/detached_tasks.cpp index f268d6021..f2b4939df 100644 --- a/src/common/detached_tasks.cpp +++ b/src/common/detached_tasks.cpp @@ -34,8 +34,7 @@ void DetachedTasks::AddTask(std::function<void()> task) { std::unique_lock lock{instance->mutex}; --instance->count; std::notify_all_at_thread_exit(instance->cv, std::move(lock)); - }) - .detach(); + }).detach(); } } // namespace Common diff --git a/src/common/dynamic_library.cpp b/src/common/dynamic_library.cpp index 7ab54e9e4..7f0a10521 100644 --- a/src/common/dynamic_library.cpp +++ b/src/common/dynamic_library.cpp @@ -21,7 +21,7 @@ namespace Common { DynamicLibrary::DynamicLibrary() = default; DynamicLibrary::DynamicLibrary(const char* filename) { - Open(filename); + void(Open(filename)); } DynamicLibrary::DynamicLibrary(DynamicLibrary&& rhs) noexcept diff --git a/src/common/dynamic_library.h b/src/common/dynamic_library.h index 2a06372fd..3512da940 100644 --- a/src/common/dynamic_library.h +++ b/src/common/dynamic_library.h @@ -33,7 +33,7 @@ public: ~DynamicLibrary(); /// Returns the specified library name with the platform-specific suffix added. - static std::string GetUnprefixedFilename(const char* filename); + [[nodiscard]] static std::string GetUnprefixedFilename(const char* filename); /// Returns the specified library name in platform-specific format. /// Major/minor versions will not be included if set to -1. @@ -41,28 +41,29 @@ public: /// Windows: LIBNAME-MAJOR-MINOR.dll /// Linux: libLIBNAME.so.MAJOR.MINOR /// Mac: libLIBNAME.MAJOR.MINOR.dylib - static std::string GetVersionedFilename(const char* libname, int major = -1, int minor = -1); + [[nodiscard]] static std::string GetVersionedFilename(const char* libname, int major = -1, + int minor = -1); /// Returns true if a module is loaded, otherwise false. - bool IsOpen() const { + [[nodiscard]] bool IsOpen() const { return handle != nullptr; } /// Loads (or replaces) the handle with the specified library file name. /// Returns true if the library was loaded and can be used. - bool Open(const char* filename); + [[nodiscard]] bool Open(const char* filename); /// Unloads the library, any function pointers from this library are no longer valid. void Close(); /// Returns the address of the specified symbol (function or variable) as an untyped pointer. /// If the specified symbol does not exist in this library, nullptr is returned. - void* GetSymbolAddress(const char* name) const; + [[nodiscard]] void* GetSymbolAddress(const char* name) const; /// Obtains the address of the specified symbol, automatically casting to the correct type. /// Returns true if the symbol was found and assigned, otherwise false. template <typename T> - bool GetSymbol(const char* name, T* ptr) const { + [[nodiscard]] bool GetSymbol(const char* name, T* ptr) const { *ptr = reinterpret_cast<T>(GetSymbolAddress(name)); return *ptr != nullptr; } diff --git a/src/common/fiber.cpp b/src/common/fiber.cpp new file mode 100644 index 000000000..3e3029cd1 --- /dev/null +++ b/src/common/fiber.cpp @@ -0,0 +1,233 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "common/fiber.h" +#include "common/spin_lock.h" + +#if defined(_WIN32) || defined(WIN32) +#include <windows.h> +#else +#include <boost/context/detail/fcontext.hpp> +#endif + +namespace Common { + +constexpr std::size_t default_stack_size = 256 * 1024; // 256kb + +struct Fiber::FiberImpl { + SpinLock guard{}; + std::function<void(void*)> entry_point; + std::function<void(void*)> rewind_point; + void* rewind_parameter{}; + void* start_parameter{}; + std::shared_ptr<Fiber> previous_fiber; + bool is_thread_fiber{}; + bool released{}; + +#if defined(_WIN32) || defined(WIN32) + LPVOID handle = nullptr; + LPVOID rewind_handle = nullptr; +#else + alignas(64) std::array<u8, default_stack_size> stack; + alignas(64) std::array<u8, default_stack_size> rewind_stack; + u8* stack_limit; + u8* rewind_stack_limit; + boost::context::detail::fcontext_t context; + boost::context::detail::fcontext_t rewind_context; +#endif +}; + +void Fiber::SetStartParameter(void* new_parameter) { + impl->start_parameter = new_parameter; +} + +void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param) { + impl->rewind_point = std::move(rewind_func); + impl->rewind_parameter = rewind_param; +} + +#if defined(_WIN32) || defined(WIN32) + +void Fiber::Start() { + ASSERT(impl->previous_fiber != nullptr); + impl->previous_fiber->impl->guard.unlock(); + impl->previous_fiber.reset(); + impl->entry_point(impl->start_parameter); + UNREACHABLE(); +} + +void Fiber::OnRewind() { + ASSERT(impl->handle != nullptr); + DeleteFiber(impl->handle); + impl->handle = impl->rewind_handle; + impl->rewind_handle = nullptr; + impl->rewind_point(impl->rewind_parameter); + UNREACHABLE(); +} + +void Fiber::FiberStartFunc(void* fiber_parameter) { + auto* fiber = static_cast<Fiber*>(fiber_parameter); + fiber->Start(); +} + +void Fiber::RewindStartFunc(void* fiber_parameter) { + auto* fiber = static_cast<Fiber*>(fiber_parameter); + fiber->OnRewind(); +} + +Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter) + : impl{std::make_unique<FiberImpl>()} { + impl->entry_point = std::move(entry_point_func); + impl->start_parameter = start_parameter; + impl->handle = CreateFiber(default_stack_size, &FiberStartFunc, this); +} + +Fiber::Fiber() : impl{std::make_unique<FiberImpl>()} {} + +Fiber::~Fiber() { + if (impl->released) { + return; + } + // Make sure the Fiber is not being used + const bool locked = impl->guard.try_lock(); + ASSERT_MSG(locked, "Destroying a fiber that's still running"); + if (locked) { + impl->guard.unlock(); + } + DeleteFiber(impl->handle); +} + +void Fiber::Exit() { + ASSERT_MSG(impl->is_thread_fiber, "Exitting non main thread fiber"); + if (!impl->is_thread_fiber) { + return; + } + ConvertFiberToThread(); + impl->guard.unlock(); + impl->released = true; +} + +void Fiber::Rewind() { + ASSERT(impl->rewind_point); + ASSERT(impl->rewind_handle == nullptr); + impl->rewind_handle = CreateFiber(default_stack_size, &RewindStartFunc, this); + SwitchToFiber(impl->rewind_handle); +} + +void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) { + ASSERT_MSG(from != nullptr, "Yielding fiber is null!"); + ASSERT_MSG(to != nullptr, "Next fiber is null!"); + to->impl->guard.lock(); + to->impl->previous_fiber = from; + SwitchToFiber(to->impl->handle); + ASSERT(from->impl->previous_fiber != nullptr); + from->impl->previous_fiber->impl->guard.unlock(); + from->impl->previous_fiber.reset(); +} + +std::shared_ptr<Fiber> Fiber::ThreadToFiber() { + std::shared_ptr<Fiber> fiber = std::shared_ptr<Fiber>{new Fiber()}; + fiber->impl->guard.lock(); + fiber->impl->handle = ConvertThreadToFiber(nullptr); + fiber->impl->is_thread_fiber = true; + return fiber; +} + +#else + +void Fiber::Start(boost::context::detail::transfer_t& transfer) { + ASSERT(impl->previous_fiber != nullptr); + impl->previous_fiber->impl->context = transfer.fctx; + impl->previous_fiber->impl->guard.unlock(); + impl->previous_fiber.reset(); + impl->entry_point(impl->start_parameter); + UNREACHABLE(); +} + +void Fiber::OnRewind([[maybe_unused]] boost::context::detail::transfer_t& transfer) { + ASSERT(impl->context != nullptr); + impl->context = impl->rewind_context; + impl->rewind_context = nullptr; + u8* tmp = impl->stack_limit; + impl->stack_limit = impl->rewind_stack_limit; + impl->rewind_stack_limit = tmp; + impl->rewind_point(impl->rewind_parameter); + UNREACHABLE(); +} + +void Fiber::FiberStartFunc(boost::context::detail::transfer_t transfer) { + auto* fiber = static_cast<Fiber*>(transfer.data); + fiber->Start(transfer); +} + +void Fiber::RewindStartFunc(boost::context::detail::transfer_t transfer) { + auto* fiber = static_cast<Fiber*>(transfer.data); + fiber->OnRewind(transfer); +} + +Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter) + : impl{std::make_unique<FiberImpl>()} { + impl->entry_point = std::move(entry_point_func); + impl->start_parameter = start_parameter; + impl->stack_limit = impl->stack.data(); + impl->rewind_stack_limit = impl->rewind_stack.data(); + u8* stack_base = impl->stack_limit + default_stack_size; + impl->context = + boost::context::detail::make_fcontext(stack_base, impl->stack.size(), FiberStartFunc); +} + +Fiber::Fiber() : impl{std::make_unique<FiberImpl>()} {} + +Fiber::~Fiber() { + if (impl->released) { + return; + } + // Make sure the Fiber is not being used + const bool locked = impl->guard.try_lock(); + ASSERT_MSG(locked, "Destroying a fiber that's still running"); + if (locked) { + impl->guard.unlock(); + } +} + +void Fiber::Exit() { + ASSERT_MSG(impl->is_thread_fiber, "Exitting non main thread fiber"); + if (!impl->is_thread_fiber) { + return; + } + impl->guard.unlock(); + impl->released = true; +} + +void Fiber::Rewind() { + ASSERT(impl->rewind_point); + ASSERT(impl->rewind_context == nullptr); + u8* stack_base = impl->rewind_stack_limit + default_stack_size; + impl->rewind_context = + boost::context::detail::make_fcontext(stack_base, impl->stack.size(), RewindStartFunc); + boost::context::detail::jump_fcontext(impl->rewind_context, this); +} + +void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) { + ASSERT_MSG(from != nullptr, "Yielding fiber is null!"); + ASSERT_MSG(to != nullptr, "Next fiber is null!"); + to->impl->guard.lock(); + to->impl->previous_fiber = from; + auto transfer = boost::context::detail::jump_fcontext(to->impl->context, to.get()); + ASSERT(from->impl->previous_fiber != nullptr); + from->impl->previous_fiber->impl->context = transfer.fctx; + from->impl->previous_fiber->impl->guard.unlock(); + from->impl->previous_fiber.reset(); +} + +std::shared_ptr<Fiber> Fiber::ThreadToFiber() { + std::shared_ptr<Fiber> fiber = std::shared_ptr<Fiber>{new Fiber()}; + fiber->impl->guard.lock(); + fiber->impl->is_thread_fiber = true; + return fiber; +} + +#endif +} // namespace Common diff --git a/src/common/fiber.h b/src/common/fiber.h new file mode 100644 index 000000000..5323e8579 --- /dev/null +++ b/src/common/fiber.h @@ -0,0 +1,78 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <functional> +#include <memory> + +#if !defined(_WIN32) && !defined(WIN32) +namespace boost::context::detail { +struct transfer_t; +} +#endif + +namespace Common { + +/** + * Fiber class + * a fiber is a userspace thread with it's own context. They can be used to + * implement coroutines, emulated threading systems and certain asynchronous + * patterns. + * + * This class implements fibers at a low level, thus allowing greater freedom + * to implement such patterns. This fiber class is 'threadsafe' only one fiber + * can be running at a time and threads will be locked while trying to yield to + * a running fiber until it yields. WARNING exchanging two running fibers between + * threads will cause a deadlock. In order to prevent a deadlock, each thread should + * have an intermediary fiber, you switch to the intermediary fiber of the current + * thread and then from it switch to the expected fiber. This way you can exchange + * 2 fibers within 2 different threads. + */ +class Fiber { +public: + Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter); + ~Fiber(); + + Fiber(const Fiber&) = delete; + Fiber& operator=(const Fiber&) = delete; + + Fiber(Fiber&&) = default; + Fiber& operator=(Fiber&&) = default; + + /// Yields control from Fiber 'from' to Fiber 'to' + /// Fiber 'from' must be the currently running fiber. + static void YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to); + [[nodiscard]] static std::shared_ptr<Fiber> ThreadToFiber(); + + void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param); + + void Rewind(); + + /// Only call from main thread's fiber + void Exit(); + + /// Changes the start parameter of the fiber. Has no effect if the fiber already started + void SetStartParameter(void* new_parameter); + +private: + Fiber(); + +#if defined(_WIN32) || defined(WIN32) + void OnRewind(); + void Start(); + static void FiberStartFunc(void* fiber_parameter); + static void RewindStartFunc(void* fiber_parameter); +#else + void OnRewind(boost::context::detail::transfer_t& transfer); + void Start(boost::context::detail::transfer_t& transfer); + static void FiberStartFunc(boost::context::detail::transfer_t transfer); + static void RewindStartFunc(boost::context::detail::transfer_t transfer); +#endif + + struct FiberImpl; + std::unique_ptr<FiberImpl> impl; +}; + +} // namespace Common diff --git a/src/common/file_util.cpp b/src/common/file_util.cpp index 35eee0096..18fbfa25b 100644 --- a/src/common/file_util.cpp +++ b/src/common/file_util.cpp @@ -74,7 +74,7 @@ // This namespace has various generic functions related to files and paths. // The code still needs a ton of cleanup. // REMEMBER: strdup considered harmful! -namespace FileUtil { +namespace Common::FS { // Remove any ending forward slashes from directory paths // Modifies argument. @@ -196,7 +196,7 @@ bool CreateFullPath(const std::string& fullPath) { int panicCounter = 100; LOG_TRACE(Common_Filesystem, "path {}", fullPath); - if (FileUtil::Exists(fullPath)) { + if (Exists(fullPath)) { LOG_DEBUG(Common_Filesystem, "path exists {}", fullPath); return true; } @@ -212,7 +212,7 @@ bool CreateFullPath(const std::string& fullPath) { // Include the '/' so the first call is CreateDir("/") rather than CreateDir("") std::string const subPath(fullPath.substr(0, position + 1)); - if (!FileUtil::IsDirectory(subPath) && !FileUtil::CreateDir(subPath)) { + if (!IsDirectory(subPath) && !CreateDir(subPath)) { LOG_ERROR(Common, "CreateFullPath: directory creation failed"); return false; } @@ -231,7 +231,7 @@ bool DeleteDir(const std::string& filename) { LOG_TRACE(Common_Filesystem, "directory {}", filename); // check if a directory - if (!FileUtil::IsDirectory(filename)) { + if (!IsDirectory(filename)) { LOG_ERROR(Common_Filesystem, "Not a directory {}", filename); return false; } @@ -371,7 +371,7 @@ u64 GetSize(FILE* f) { bool CreateEmptyFile(const std::string& filename) { LOG_TRACE(Common_Filesystem, "{}", filename); - if (!FileUtil::IOFile(filename, "wb").IsOpen()) { + if (!IOFile(filename, "wb").IsOpen()) { LOG_ERROR(Common_Filesystem, "failed {}: {}", filename, GetLastErrorMsg()); return false; } @@ -472,13 +472,14 @@ u64 ScanDirectoryTree(const std::string& directory, FSTEntry& parent_entry, } bool DeleteDirRecursively(const std::string& directory, unsigned int recursion) { - const auto callback = [recursion](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { - std::string new_path = directory + DIR_SEP_CHR + virtual_name; + const auto callback = [recursion](u64*, const std::string& directory, + const std::string& virtual_name) { + const std::string new_path = directory + DIR_SEP_CHR + virtual_name; if (IsDirectory(new_path)) { - if (recursion == 0) + if (recursion == 0) { return false; + } return DeleteDirRecursively(new_path, recursion - 1); } return Delete(new_path); @@ -488,29 +489,35 @@ bool DeleteDirRecursively(const std::string& directory, unsigned int recursion) return false; // Delete the outermost directory - FileUtil::DeleteDir(directory); + DeleteDir(directory); return true; } -void CopyDir(const std::string& source_path, const std::string& dest_path) { +void CopyDir([[maybe_unused]] const std::string& source_path, + [[maybe_unused]] const std::string& dest_path) { #ifndef _WIN32 - if (source_path == dest_path) + if (source_path == dest_path) { return; - if (!FileUtil::Exists(source_path)) + } + if (!Exists(source_path)) { return; - if (!FileUtil::Exists(dest_path)) - FileUtil::CreateFullPath(dest_path); + } + if (!Exists(dest_path)) { + CreateFullPath(dest_path); + } DIR* dirp = opendir(source_path.c_str()); - if (!dirp) + if (!dirp) { return; + } while (struct dirent* result = readdir(dirp)) { const std::string virtualName(result->d_name); // check for "." and ".." if (((virtualName[0] == '.') && (virtualName[1] == '\0')) || - ((virtualName[0] == '.') && (virtualName[1] == '.') && (virtualName[2] == '\0'))) + ((virtualName[0] == '.') && (virtualName[1] == '.') && (virtualName[2] == '\0'))) { continue; + } std::string source, dest; source = source_path + virtualName; @@ -518,11 +525,13 @@ void CopyDir(const std::string& source_path, const std::string& dest_path) { if (IsDirectory(source)) { source += '/'; dest += '/'; - if (!FileUtil::Exists(dest)) - FileUtil::CreateFullPath(dest); + if (!Exists(dest)) { + CreateFullPath(dest); + } CopyDir(source, dest); - } else if (!FileUtil::Exists(dest)) - FileUtil::Copy(source, dest); + } else if (!Exists(dest)) { + Copy(source, dest); + } } closedir(dirp); #endif @@ -538,7 +547,7 @@ std::optional<std::string> GetCurrentDir() { if (!dir) { #endif LOG_ERROR(Common_Filesystem, "GetCurrentDirectory failed: {}", GetLastErrorMsg()); - return {}; + return std::nullopt; } #ifdef _WIN32 std::string strDir = Common::UTF16ToUTF8(dir); @@ -668,7 +677,7 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) { if (user_path.empty()) { #ifdef _WIN32 user_path = GetExeDirectory() + DIR_SEP USERDATA_DIR DIR_SEP; - if (!FileUtil::IsDirectory(user_path)) { + if (!IsDirectory(user_path)) { user_path = AppDataRoamingDirectory() + DIR_SEP EMU_DATA_DIR DIR_SEP; } else { LOG_INFO(Common_Filesystem, "Using the local user directory"); @@ -677,7 +686,7 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) { paths.emplace(UserPath::ConfigDir, user_path + CONFIG_DIR DIR_SEP); paths.emplace(UserPath::CacheDir, user_path + CACHE_DIR DIR_SEP); #else - if (FileUtil::Exists(ROOT_DIR DIR_SEP USERDATA_DIR)) { + if (Exists(ROOT_DIR DIR_SEP USERDATA_DIR)) { user_path = ROOT_DIR DIR_SEP USERDATA_DIR DIR_SEP; paths.emplace(UserPath::ConfigDir, user_path + CONFIG_DIR DIR_SEP); paths.emplace(UserPath::CacheDir, user_path + CACHE_DIR DIR_SEP); @@ -695,6 +704,7 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) { paths.emplace(UserPath::NANDDir, user_path + NAND_DIR DIR_SEP); paths.emplace(UserPath::LoadDir, user_path + LOAD_DIR DIR_SEP); paths.emplace(UserPath::DumpDir, user_path + DUMP_DIR DIR_SEP); + paths.emplace(UserPath::ScreenshotsDir, user_path + SCREENSHOTS_DIR DIR_SEP); paths.emplace(UserPath::ShaderDir, user_path + SHADER_DIR DIR_SEP); paths.emplace(UserPath::SysDataDir, user_path + SYSDATA_DIR DIR_SEP); paths.emplace(UserPath::KeysDir, user_path + KEYS_DIR DIR_SEP); @@ -703,7 +713,7 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path) { } if (!new_path.empty()) { - if (!FileUtil::IsDirectory(new_path)) { + if (!IsDirectory(new_path)) { LOG_ERROR(Common_Filesystem, "Invalid path specified {}", new_path); return paths[path]; } else { @@ -764,21 +774,23 @@ std::size_t ReadFileToString(bool text_file, const std::string& filename, std::s void SplitFilename83(const std::string& filename, std::array<char, 9>& short_name, std::array<char, 4>& extension) { - const std::string forbidden_characters = ".\"/\\[]:;=, "; + static constexpr std::string_view forbidden_characters = ".\"/\\[]:;=, "; // On a FAT32 partition, 8.3 names are stored as a 11 bytes array, filled with spaces. short_name = {{' ', ' ', ' ', ' ', ' ', ' ', ' ', ' ', '\0'}}; extension = {{' ', ' ', ' ', '\0'}}; - std::string::size_type point = filename.rfind('.'); - if (point == filename.size() - 1) + auto point = filename.rfind('.'); + if (point == filename.size() - 1) { point = filename.rfind('.', point); + } // Get short name. int j = 0; for (char letter : filename.substr(0, point)) { - if (forbidden_characters.find(letter, 0) != std::string::npos) + if (forbidden_characters.find(letter, 0) != std::string::npos) { continue; + } if (j == 8) { // TODO(Link Mauve): also do that for filenames containing a space. // TODO(Link Mauve): handle multiple files having the same short name. @@ -786,14 +798,15 @@ void SplitFilename83(const std::string& filename, std::array<char, 9>& short_nam short_name[7] = '1'; break; } - short_name[j++] = toupper(letter); + short_name[j++] = static_cast<char>(std::toupper(letter)); } // Get extension. if (point != std::string::npos) { j = 0; - for (char letter : filename.substr(point + 1, 3)) - extension[j++] = toupper(letter); + for (char letter : filename.substr(point + 1, 3)) { + extension[j++] = static_cast<char>(std::toupper(letter)); + } } } @@ -888,16 +901,23 @@ std::string SanitizePath(std::string_view path_, DirectorySeparator directory_se } std::replace(path.begin(), path.end(), type1, type2); - path.erase(std::unique(path.begin(), path.end(), + + auto start = path.begin(); +#ifdef _WIN32 + // allow network paths which start with a double backslash (e.g. \\server\share) + if (start != path.end()) + ++start; +#endif + path.erase(std::unique(start, path.end(), [type2](char c1, char c2) { return c1 == type2 && c2 == type2; }), path.end()); return std::string(RemoveTrailingSlash(path)); } -IOFile::IOFile() {} +IOFile::IOFile() = default; IOFile::IOFile(const std::string& filename, const char openmode[], int flags) { - Open(filename, openmode, flags); + void(Open(filename, openmode, flags)); } IOFile::~IOFile() { @@ -938,17 +958,18 @@ bool IOFile::Open(const std::string& filename, const char openmode[], int flags) } bool IOFile::Close() { - if (!IsOpen() || 0 != std::fclose(m_file)) + if (!IsOpen() || 0 != std::fclose(m_file)) { return false; + } m_file = nullptr; return true; } u64 IOFile::GetSize() const { - if (IsOpen()) - return FileUtil::GetSize(m_file); - + if (IsOpen()) { + return FS::GetSize(m_file); + } return 0; } @@ -957,9 +978,9 @@ bool IOFile::Seek(s64 off, int origin) const { } u64 IOFile::Tell() const { - if (IsOpen()) + if (IsOpen()) { return ftello(m_file); - + } return std::numeric_limits<u64>::max(); } @@ -967,6 +988,34 @@ bool IOFile::Flush() { return IsOpen() && 0 == std::fflush(m_file); } +std::size_t IOFile::ReadImpl(void* data, std::size_t length, std::size_t data_size) const { + if (!IsOpen()) { + return std::numeric_limits<std::size_t>::max(); + } + + if (length == 0) { + return 0; + } + + DEBUG_ASSERT(data != nullptr); + + return std::fread(data, data_size, length, m_file); +} + +std::size_t IOFile::WriteImpl(const void* data, std::size_t length, std::size_t data_size) { + if (!IsOpen()) { + return std::numeric_limits<std::size_t>::max(); + } + + if (length == 0) { + return 0; + } + + DEBUG_ASSERT(data != nullptr); + + return std::fwrite(data, data_size, length, m_file); +} + bool IOFile::Resize(u64 size) { return IsOpen() && 0 == #ifdef _WIN32 @@ -980,4 +1029,4 @@ bool IOFile::Resize(u64 size) { ; } -} // namespace FileUtil +} // namespace Common::FS diff --git a/src/common/file_util.h b/src/common/file_util.h index cde7ddf2d..840cde2a6 100644 --- a/src/common/file_util.h +++ b/src/common/file_util.h @@ -19,7 +19,7 @@ #include "common/string_util.h" #endif -namespace FileUtil { +namespace Common::FS { // User paths for GetUserPath enum class UserPath { @@ -32,6 +32,7 @@ enum class UserPath { SDMCDir, LoadDir, DumpDir, + ScreenshotsDir, ShaderDir, SysDataDir, UserDir, @@ -47,19 +48,19 @@ struct FSTEntry { }; // Returns true if file filename exists -bool Exists(const std::string& filename); +[[nodiscard]] bool Exists(const std::string& filename); // Returns true if filename is a directory -bool IsDirectory(const std::string& filename); +[[nodiscard]] bool IsDirectory(const std::string& filename); // Returns the size of filename (64bit) -u64 GetSize(const std::string& filename); +[[nodiscard]] u64 GetSize(const std::string& filename); // Overloaded GetSize, accepts file descriptor -u64 GetSize(const int fd); +[[nodiscard]] u64 GetSize(int fd); // Overloaded GetSize, accepts FILE* -u64 GetSize(FILE* f); +[[nodiscard]] u64 GetSize(FILE* f); // Returns true if successful, or path already exists. bool CreateDir(const std::string& filename); @@ -119,7 +120,7 @@ u64 ScanDirectoryTree(const std::string& directory, FSTEntry& parent_entry, bool DeleteDirRecursively(const std::string& directory, unsigned int recursion = 256); // Returns the current directory -std::optional<std::string> GetCurrentDir(); +[[nodiscard]] std::optional<std::string> GetCurrentDir(); // Create directory and copy contents (does not overwrite existing files) void CopyDir(const std::string& source_path, const std::string& dest_path); @@ -131,20 +132,20 @@ bool SetCurrentDir(const std::string& directory); // directory. To be used in "multi-user" mode (that is, installed). const std::string& GetUserPath(UserPath path, const std::string& new_path = ""); -std::string GetHactoolConfigurationPath(); +[[nodiscard]] std::string GetHactoolConfigurationPath(); -std::string GetNANDRegistrationDir(bool system = false); +[[nodiscard]] std::string GetNANDRegistrationDir(bool system = false); // Returns the path to where the sys file are -std::string GetSysDirectory(); +[[nodiscard]] std::string GetSysDirectory(); #ifdef __APPLE__ -std::string GetBundleDirectory(); +[[nodiscard]] std::string GetBundleDirectory(); #endif #ifdef _WIN32 -const std::string& GetExeDirectory(); -std::string AppDataRoamingDirectory(); +[[nodiscard]] const std::string& GetExeDirectory(); +[[nodiscard]] std::string AppDataRoamingDirectory(); #endif std::size_t WriteStringToFile(bool text_file, const std::string& filename, std::string_view str); @@ -163,38 +164,55 @@ void SplitFilename83(const std::string& filename, std::array<char, 9>& short_nam // Splits the path on '/' or '\' and put the components into a vector // i.e. "C:\Users\Yuzu\Documents\save.bin" becomes {"C:", "Users", "Yuzu", "Documents", "save.bin" } -std::vector<std::string> SplitPathComponents(std::string_view filename); +[[nodiscard]] std::vector<std::string> SplitPathComponents(std::string_view filename); // Gets all of the text up to the last '/' or '\' in the path. -std::string_view GetParentPath(std::string_view path); +[[nodiscard]] std::string_view GetParentPath(std::string_view path); // Gets all of the text after the first '/' or '\' in the path. -std::string_view GetPathWithoutTop(std::string_view path); +[[nodiscard]] std::string_view GetPathWithoutTop(std::string_view path); // Gets the filename of the path -std::string_view GetFilename(std::string_view path); +[[nodiscard]] std::string_view GetFilename(std::string_view path); // Gets the extension of the filename -std::string_view GetExtensionFromFilename(std::string_view name); +[[nodiscard]] std::string_view GetExtensionFromFilename(std::string_view name); // Removes the final '/' or '\' if one exists -std::string_view RemoveTrailingSlash(std::string_view path); +[[nodiscard]] std::string_view RemoveTrailingSlash(std::string_view path); // Creates a new vector containing indices [first, last) from the original. template <typename T> -std::vector<T> SliceVector(const std::vector<T>& vector, std::size_t first, std::size_t last) { - if (first >= last) +[[nodiscard]] std::vector<T> SliceVector(const std::vector<T>& vector, std::size_t first, + std::size_t last) { + if (first >= last) { return {}; + } last = std::min<std::size_t>(last, vector.size()); return std::vector<T>(vector.begin() + first, vector.begin() + first + last); } -enum class DirectorySeparator { ForwardSlash, BackwardSlash, PlatformDefault }; +enum class DirectorySeparator { + ForwardSlash, + BackwardSlash, + PlatformDefault, +}; // Removes trailing slash, makes all '\\' into '/', and removes duplicate '/'. Makes '/' into '\\' // depending if directory_separator is BackwardSlash or PlatformDefault and running on windows -std::string SanitizePath(std::string_view path, - DirectorySeparator directory_separator = DirectorySeparator::ForwardSlash); +[[nodiscard]] std::string SanitizePath( + std::string_view path, + DirectorySeparator directory_separator = DirectorySeparator::ForwardSlash); + +// To deal with Windows being dumb at Unicode +template <typename T> +void OpenFStream(T& fstream, const std::string& filename, std::ios_base::openmode openmode) { +#ifdef _MSC_VER + fstream.open(Common::UTF8ToUTF16W(filename), openmode); +#else + fstream.open(filename, openmode); +#endif +} // simple wrapper for cstdlib file functions to // hopefully will make error checking easier @@ -222,22 +240,15 @@ public: static_assert(std::is_trivially_copyable_v<T>, "Given array does not consist of trivially copyable objects"); - if (!IsOpen()) { - return std::numeric_limits<std::size_t>::max(); - } - - return std::fread(data, sizeof(T), length, m_file); + return ReadImpl(data, length, sizeof(T)); } template <typename T> std::size_t WriteArray(const T* data, std::size_t length) { static_assert(std::is_trivially_copyable_v<T>, "Given array does not consist of trivially copyable objects"); - if (!IsOpen()) { - return std::numeric_limits<std::size_t>::max(); - } - return std::fwrite(data, sizeof(T), length, m_file); + return WriteImpl(data, length, sizeof(T)); } template <typename T> @@ -262,13 +273,13 @@ public: return WriteArray(str.data(), str.length()); } - bool IsOpen() const { + [[nodiscard]] bool IsOpen() const { return nullptr != m_file; } bool Seek(s64 off, int origin) const; - u64 Tell() const; - u64 GetSize() const; + [[nodiscard]] u64 Tell() const; + [[nodiscard]] u64 GetSize() const; bool Resize(u64 size); bool Flush(); @@ -278,17 +289,10 @@ public: } private: + std::size_t ReadImpl(void* data, std::size_t length, std::size_t data_size) const; + std::size_t WriteImpl(const void* data, std::size_t length, std::size_t data_size); + std::FILE* m_file = nullptr; }; -} // namespace FileUtil - -// To deal with Windows being dumb at unicode: -template <typename T> -void OpenFStream(T& fstream, const std::string& filename, std::ios_base::openmode openmode) { -#ifdef _MSC_VER - fstream.open(Common::UTF8ToUTF16W(filename), openmode); -#else - fstream.open(filename, openmode); -#endif -} +} // namespace Common::FS diff --git a/src/common/hash.h b/src/common/hash.h index b2538f3ea..298930702 100644 --- a/src/common/hash.h +++ b/src/common/hash.h @@ -5,36 +5,11 @@ #pragma once #include <cstddef> -#include <cstring> #include <utility> #include <boost/functional/hash.hpp> -#include "common/cityhash.h" -#include "common/common_types.h" namespace Common { -/** - * Computes a 64-bit hash over the specified block of data - * @param data Block of data to compute hash over - * @param len Length of data (in bytes) to compute hash over - * @returns 64-bit hash value that was computed over the data block - */ -static inline u64 ComputeHash64(const void* data, std::size_t len) { - return CityHash64(static_cast<const char*>(data), len); -} - -/** - * Computes a 64-bit hash of a struct. In addition to being trivially copyable, it is also critical - * that either the struct includes no padding, or that any padding is initialized to a known value - * by memsetting the struct to 0 before filling it in. - */ -template <typename T> -static inline u64 ComputeStructHash64(const T& data) { - static_assert(std::is_trivially_copyable_v<T>, - "Type passed to ComputeStructHash64 must be trivially copyable"); - return ComputeHash64(&data, sizeof(data)); -} - struct PairHash { template <class T1, class T2> std::size_t operator()(const std::pair<T1, T2>& pair) const noexcept { diff --git a/src/common/hex_util.cpp b/src/common/hex_util.cpp index c2f6cf0f6..74f52dd11 100644 --- a/src/common/hex_util.cpp +++ b/src/common/hex_util.cpp @@ -3,21 +3,9 @@ // Refer to the license.txt file included. #include "common/hex_util.h" -#include "common/logging/log.h" namespace Common { -u8 ToHexNibble(char c1) { - if (c1 >= 65 && c1 <= 70) - return c1 - 55; - if (c1 >= 97 && c1 <= 102) - return c1 - 87; - if (c1 >= 48 && c1 <= 57) - return c1 - 48; - LOG_ERROR(Common, "Invalid hex digit: 0x{:02X}", c1); - return 0; -} - std::vector<u8> HexStringToVector(std::string_view str, bool little_endian) { std::vector<u8> out(str.size() / 2); if (little_endian) { @@ -30,26 +18,4 @@ std::vector<u8> HexStringToVector(std::string_view str, bool little_endian) { return out; } -std::array<u8, 16> operator""_array16(const char* str, std::size_t len) { - if (len != 32) { - LOG_ERROR(Common, - "Attempting to parse string to array that is not of correct size (expected=32, " - "actual={}).", - len); - return {}; - } - return HexStringToArray<16>(str); -} - -std::array<u8, 32> operator""_array32(const char* str, std::size_t len) { - if (len != 64) { - LOG_ERROR(Common, - "Attempting to parse string to array that is not of correct size (expected=64, " - "actual={}).", - len); - return {}; - } - return HexStringToArray<32>(str); -} - } // namespace Common diff --git a/src/common/hex_util.h b/src/common/hex_util.h index bb4736f96..a8d414fb8 100644 --- a/src/common/hex_util.h +++ b/src/common/hex_util.h @@ -14,25 +14,37 @@ namespace Common { -u8 ToHexNibble(char c1); +[[nodiscard]] constexpr u8 ToHexNibble(char c) { + if (c >= 65 && c <= 70) { + return static_cast<u8>(c - 55); + } + + if (c >= 97 && c <= 102) { + return static_cast<u8>(c - 87); + } + + return static_cast<u8>(c - 48); +} -std::vector<u8> HexStringToVector(std::string_view str, bool little_endian); +[[nodiscard]] std::vector<u8> HexStringToVector(std::string_view str, bool little_endian); template <std::size_t Size, bool le = false> -std::array<u8, Size> HexStringToArray(std::string_view str) { +[[nodiscard]] constexpr std::array<u8, Size> HexStringToArray(std::string_view str) { std::array<u8, Size> out{}; if constexpr (le) { - for (std::size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2) - out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]); + for (std::size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2) { + out[i / 2] = static_cast<u8>((ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1])); + } } else { - for (std::size_t i = 0; i < 2 * Size; i += 2) - out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]); + for (std::size_t i = 0; i < 2 * Size; i += 2) { + out[i / 2] = static_cast<u8>((ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1])); + } } return out; } template <typename ContiguousContainer> -std::string HexToString(const ContiguousContainer& data, bool upper = true) { +[[nodiscard]] std::string HexToString(const ContiguousContainer& data, bool upper = true) { static_assert(std::is_same_v<typename ContiguousContainer::value_type, u8>, "Underlying type within the contiguous container must be u8."); @@ -48,7 +60,12 @@ std::string HexToString(const ContiguousContainer& data, bool upper = true) { return out; } -std::array<u8, 0x10> operator"" _array16(const char* str, std::size_t len); -std::array<u8, 0x20> operator"" _array32(const char* str, std::size_t len); +[[nodiscard]] constexpr std::array<u8, 16> AsArray(const char (&data)[17]) { + return HexStringToArray<16>(data); +} + +[[nodiscard]] constexpr std::array<u8, 32> AsArray(const char (&data)[65]) { + return HexStringToArray<32>(data); +} } // namespace Common diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp index 04bc3128f..631f64d05 100644 --- a/src/common/logging/backend.cpp +++ b/src/common/logging/backend.cpp @@ -23,6 +23,7 @@ #include "common/logging/text_formatter.h" #include "common/string_util.h" #include "common/threadsafe_queue.h" +#include "core/settings.h" namespace Log { @@ -113,19 +114,19 @@ private: Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr, const char* function, std::string message) const { using std::chrono::duration_cast; + using std::chrono::microseconds; using std::chrono::steady_clock; - Entry entry; - entry.timestamp = - duration_cast<std::chrono::microseconds>(steady_clock::now() - time_origin); - entry.log_class = log_class; - entry.log_level = log_level; - entry.filename = filename; - entry.line_num = line_nr; - entry.function = function; - entry.message = std::move(message); - - return entry; + return { + .timestamp = duration_cast<microseconds>(steady_clock::now() - time_origin), + .log_class = log_class, + .log_level = log_level, + .filename = filename, + .line_num = line_nr, + .function = function, + .message = std::move(message), + .final_entry = false, + }; } std::mutex writing_mutex; @@ -152,10 +153,19 @@ FileBackend::FileBackend(const std::string& filename) void FileBackend::Write(const Entry& entry) { // prevent logs from going over the maximum size (in case its spamming and the user doesn't // know) - constexpr std::size_t MAX_BYTES_WRITTEN = 50 * 1024L * 1024L; - if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) { + constexpr std::size_t MAX_BYTES_WRITTEN = 100 * 1024 * 1024; + constexpr std::size_t MAX_BYTES_WRITTEN_EXTENDED = 1024 * 1024 * 1024; + + if (!file.IsOpen()) { return; } + + if (Settings::values.extended_logging && bytes_written > MAX_BYTES_WRITTEN_EXTENDED) { + return; + } else if (!Settings::values.extended_logging && bytes_written > MAX_BYTES_WRITTEN) { + return; + } + bytes_written += file.WriteString(FormatLogMessage(entry).append(1, '\n')); if (entry.log_level >= Level::Error) { file.Flush(); @@ -222,6 +232,7 @@ void DebuggerBackend::Write(const Entry& entry) { SUB(Service, NPNS) \ SUB(Service, NS) \ SUB(Service, NVDRV) \ + SUB(Service, OLSC) \ SUB(Service, PCIE) \ SUB(Service, PCTL) \ SUB(Service, PCV) \ @@ -274,7 +285,6 @@ const char* GetLogClassName(Class log_class) { case Class::Count: break; } - UNREACHABLE(); return "Invalid"; } @@ -293,7 +303,6 @@ const char* GetLevelName(Level log_level) { break; } #undef LVL - UNREACHABLE(); return "Invalid"; } diff --git a/src/common/logging/backend.h b/src/common/logging/backend.h index fc338c70d..da1c2f185 100644 --- a/src/common/logging/backend.h +++ b/src/common/logging/backend.h @@ -21,19 +21,13 @@ class Filter; */ struct Entry { std::chrono::microseconds timestamp; - Class log_class; - Level log_level; - const char* filename; - unsigned int line_num; + Class log_class{}; + Level log_level{}; + const char* filename = nullptr; + unsigned int line_num = 0; std::string function; std::string message; bool final_entry = false; - - Entry() = default; - Entry(Entry&& o) = default; - - Entry& operator=(Entry&& o) = default; - Entry& operator=(const Entry& o) = default; }; /** @@ -100,7 +94,7 @@ public: void Write(const Entry& entry) override; private: - FileUtil::IOFile file; + Common::FS::IOFile file; std::size_t bytes_written; }; diff --git a/src/common/logging/log.h b/src/common/logging/log.h index 13a4f1e30..835894918 100644 --- a/src/common/logging/log.h +++ b/src/common/logging/log.h @@ -95,6 +95,7 @@ enum class Class : ClassType { Service_NPNS, ///< The NPNS service Service_NS, ///< The NS services Service_NVDRV, ///< The NVDRV (Nvidia driver) service + Service_OLSC, ///< The OLSC service Service_PCIE, ///< The PCIe service Service_PCTL, ///< The PCTL (Parental control) service Service_PCV, ///< The PCV service diff --git a/src/common/lz4_compression.cpp b/src/common/lz4_compression.cpp index ade6759bb..25700015a 100644 --- a/src/common/lz4_compression.cpp +++ b/src/common/lz4_compression.cpp @@ -14,19 +14,19 @@ std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size) { ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size"); const auto source_size_int = static_cast<int>(source_size); - const int max_compressed_size = LZ4_compressBound(source_size_int); + const auto max_compressed_size = static_cast<std::size_t>(LZ4_compressBound(source_size_int)); std::vector<u8> compressed(max_compressed_size); - const int compressed_size = LZ4_compress_default(reinterpret_cast<const char*>(source), - reinterpret_cast<char*>(compressed.data()), - source_size_int, max_compressed_size); + const int compressed_size = LZ4_compress_default( + reinterpret_cast<const char*>(source), reinterpret_cast<char*>(compressed.data()), + source_size_int, static_cast<int>(max_compressed_size)); if (compressed_size <= 0) { // Compression failed return {}; } - compressed.resize(compressed_size); + compressed.resize(static_cast<std::size_t>(compressed_size)); return compressed; } @@ -38,19 +38,19 @@ std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, compression_level = std::clamp(compression_level, LZ4HC_CLEVEL_MIN, LZ4HC_CLEVEL_MAX); const auto source_size_int = static_cast<int>(source_size); - const int max_compressed_size = LZ4_compressBound(source_size_int); + const auto max_compressed_size = static_cast<std::size_t>(LZ4_compressBound(source_size_int)); std::vector<u8> compressed(max_compressed_size); const int compressed_size = LZ4_compress_HC( reinterpret_cast<const char*>(source), reinterpret_cast<char*>(compressed.data()), - source_size_int, max_compressed_size, compression_level); + source_size_int, static_cast<int>(max_compressed_size), compression_level); if (compressed_size <= 0) { // Compression failed return {}; } - compressed.resize(compressed_size); + compressed.resize(static_cast<std::size_t>(compressed_size)); return compressed; } diff --git a/src/common/lz4_compression.h b/src/common/lz4_compression.h index 4c16f6e03..87a4be1b0 100644 --- a/src/common/lz4_compression.h +++ b/src/common/lz4_compression.h @@ -13,12 +13,12 @@ namespace Common::Compression { /** * Compresses a source memory region with LZ4 and returns the compressed data in a vector. * - * @param source the uncompressed source memory region. - * @param source_size the size in bytes of the uncompressed source memory region. + * @param source The uncompressed source memory region. + * @param source_size The size of the uncompressed source memory region. * * @return the compressed data. */ -std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size); +[[nodiscard]] std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size); /** * Utilizes the LZ4 subalgorithm LZ4HC with the specified compression level. Higher compression @@ -26,23 +26,24 @@ std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size); * compression level has almost no impact on decompression speed. Data compressed with LZ4HC can * also be decompressed with the default LZ4 decompression. * - * @param source the uncompressed source memory region. - * @param source_size the size in bytes of the uncompressed source memory region. - * @param compression_level the used compression level. Should be between 3 and 12. + * @param source The uncompressed source memory region. + * @param source_size The size of the uncompressed source memory region. + * @param compression_level The used compression level. Should be between 3 and 12. * * @return the compressed data. */ -std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, s32 compression_level); +[[nodiscard]] std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, + s32 compression_level); /** * Utilizes the LZ4 subalgorithm LZ4HC with the highest possible compression level. * - * @param source the uncompressed source memory region. - * @param source_size the size in bytes of the uncompressed source memory region. + * @param source The uncompressed source memory region. + * @param source_size The size of the uncompressed source memory region * * @return the compressed data. */ -std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size); +[[nodiscard]] std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size); /** * Decompresses a source memory region with LZ4 and returns the uncompressed data in a vector. @@ -52,6 +53,7 @@ std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size); * * @return the decompressed data. */ -std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed, std::size_t uncompressed_size); +[[nodiscard]] std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed, + std::size_t uncompressed_size); } // namespace Common::Compression
\ No newline at end of file diff --git a/src/common/math_util.h b/src/common/math_util.h index 83ef0201f..4c38d8040 100644 --- a/src/common/math_util.h +++ b/src/common/math_util.h @@ -9,7 +9,7 @@ namespace Common { -constexpr float PI = 3.14159265f; +constexpr float PI = 3.1415926535f; template <class T> struct Rectangle { @@ -20,40 +20,40 @@ struct Rectangle { constexpr Rectangle() = default; - constexpr Rectangle(T left, T top, T right, T bottom) - : left(left), top(top), right(right), bottom(bottom) {} + constexpr Rectangle(T left_, T top_, T right_, T bottom_) + : left(left_), top(top_), right(right_), bottom(bottom_) {} - T GetWidth() const { + [[nodiscard]] T GetWidth() const { if constexpr (std::is_floating_point_v<T>) { return std::abs(right - left); } else { - return std::abs(static_cast<std::make_signed_t<T>>(right - left)); + return static_cast<T>(std::abs(static_cast<std::make_signed_t<T>>(right - left))); } } - T GetHeight() const { + [[nodiscard]] T GetHeight() const { if constexpr (std::is_floating_point_v<T>) { return std::abs(bottom - top); } else { - return std::abs(static_cast<std::make_signed_t<T>>(bottom - top)); + return static_cast<T>(std::abs(static_cast<std::make_signed_t<T>>(bottom - top))); } } - Rectangle<T> TranslateX(const T x) const { + [[nodiscard]] Rectangle<T> TranslateX(const T x) const { return Rectangle{left + x, top, right + x, bottom}; } - Rectangle<T> TranslateY(const T y) const { + [[nodiscard]] Rectangle<T> TranslateY(const T y) const { return Rectangle{left, top + y, right, bottom + y}; } - Rectangle<T> Scale(const float s) const { + [[nodiscard]] Rectangle<T> Scale(const float s) const { return Rectangle{left, top, static_cast<T>(left + GetWidth() * s), static_cast<T>(top + GetHeight() * s)}; } }; template <typename T> -Rectangle(T, T, T, T)->Rectangle<T>; +Rectangle(T, T, T, T) -> Rectangle<T>; } // namespace Common diff --git a/src/common/memory_detect.cpp b/src/common/memory_detect.cpp new file mode 100644 index 000000000..8cff6ec37 --- /dev/null +++ b/src/common/memory_detect.cpp @@ -0,0 +1,73 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#ifdef _WIN32 +// clang-format off +#include <windows.h> +#include <sysinfoapi.h> +// clang-format on +#else +#include <sys/types.h> +#if defined(__APPLE__) || defined(__FreeBSD__) +#include <sys/sysctl.h> +#elif defined(__linux__) +#include <sys/sysinfo.h> +#else +#include <unistd.h> +#endif +#endif + +#include "common/memory_detect.h" + +namespace Common { + +// Detects the RAM and Swapfile sizes +static MemoryInfo Detect() { + MemoryInfo mem_info{}; + +#ifdef _WIN32 + MEMORYSTATUSEX memorystatus; + memorystatus.dwLength = sizeof(memorystatus); + GlobalMemoryStatusEx(&memorystatus); + mem_info.TotalPhysicalMemory = memorystatus.ullTotalPhys; + mem_info.TotalSwapMemory = memorystatus.ullTotalPageFile - mem_info.TotalPhysicalMemory; +#elif defined(__APPLE__) + u64 ramsize; + struct xsw_usage vmusage; + std::size_t sizeof_ramsize = sizeof(ramsize); + std::size_t sizeof_vmusage = sizeof(vmusage); + // hw and vm are defined in sysctl.h + // https://github.com/apple/darwin-xnu/blob/master/bsd/sys/sysctl.h#L471 + // sysctlbyname(const char *, void *, size_t *, void *, size_t); + sysctlbyname("hw.memsize", &ramsize, &sizeof_ramsize, nullptr, 0); + sysctlbyname("vm.swapusage", &vmusage, &sizeof_vmusage, nullptr, 0); + mem_info.TotalPhysicalMemory = ramsize; + mem_info.TotalSwapMemory = vmusage.xsu_total; +#elif defined(__FreeBSD__) + u_long physmem, swap_total; + std::size_t sizeof_u_long = sizeof(u_long); + // sysctlbyname(const char *, void *, size_t *, const void *, size_t); + sysctlbyname("hw.physmem", &physmem, &sizeof_u_long, nullptr, 0); + sysctlbyname("vm.swap_total", &swap_total, &sizeof_u_long, nullptr, 0); + mem_info.TotalPhysicalMemory = physmem; + mem_info.TotalSwapMemory = swap_total; +#elif defined(__linux__) + struct sysinfo meminfo; + sysinfo(&meminfo); + mem_info.TotalPhysicalMemory = meminfo.totalram; + mem_info.TotalSwapMemory = meminfo.totalswap; +#else + mem_info.TotalPhysicalMemory = sysconf(_SC_PHYS_PAGES) * sysconf(_SC_PAGE_SIZE); + mem_info.TotalSwapMemory = 0; +#endif + + return mem_info; +} + +const MemoryInfo& GetMemInfo() { + static MemoryInfo mem_info = Detect(); + return mem_info; +} + +} // namespace Common
\ No newline at end of file diff --git a/src/common/memory_detect.h b/src/common/memory_detect.h new file mode 100644 index 000000000..0f73751c8 --- /dev/null +++ b/src/common/memory_detect.h @@ -0,0 +1,22 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace Common { + +struct MemoryInfo { + u64 TotalPhysicalMemory{}; + u64 TotalSwapMemory{}; +}; + +/** + * Gets the memory info of the host system + * @return Reference to a MemoryInfo struct with the physical and swap memory sizes in bytes + */ +[[nodiscard]] const MemoryInfo& GetMemInfo(); + +} // namespace Common
\ No newline at end of file diff --git a/src/common/misc.cpp b/src/common/misc.cpp index 68cb86cd1..1d5393597 100644 --- a/src/common/misc.cpp +++ b/src/common/misc.cpp @@ -16,16 +16,23 @@ // Call directly after the command or use the error num. // This function might change the error code. std::string GetLastErrorMsg() { - static const std::size_t buff_size = 255; + static constexpr std::size_t buff_size = 255; char err_str[buff_size]; #ifdef _WIN32 FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM, nullptr, GetLastError(), MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), err_str, buff_size, nullptr); + return std::string(err_str, buff_size); +#elif defined(__GLIBC__) && (_GNU_SOURCE || (_POSIX_C_SOURCE < 200112L && _XOPEN_SOURCE < 600)) + // Thread safe (GNU-specific) + const char* str = strerror_r(errno, err_str, buff_size); + return std::string(str); #else // Thread safe (XSI-compliant) - strerror_r(errno, err_str, buff_size); + const int success = strerror_r(errno, err_str, buff_size); + if (success != 0) { + return {}; + } + return std::string(err_str); #endif - - return std::string(err_str, buff_size); } diff --git a/src/common/multi_level_queue.h b/src/common/multi_level_queue.h index 50acfdbf2..4b305bf40 100644 --- a/src/common/multi_level_queue.h +++ b/src/common/multi_level_queue.h @@ -223,15 +223,15 @@ public: ListShiftForward(levels[priority], n); } - std::size_t depth() const { + [[nodiscard]] std::size_t depth() const { return Depth; } - std::size_t size(u32 priority) const { + [[nodiscard]] std::size_t size(u32 priority) const { return levels[priority].size(); } - std::size_t size() const { + [[nodiscard]] std::size_t size() const { u64 priorities = used_priorities; std::size_t size = 0; while (priorities != 0) { @@ -242,64 +242,64 @@ public: return size; } - bool empty() const { + [[nodiscard]] bool empty() const { return used_priorities == 0; } - bool empty(u32 priority) const { + [[nodiscard]] bool empty(u32 priority) const { return (used_priorities & (1ULL << priority)) == 0; } - u32 highest_priority_set(u32 max_priority = 0) const { + [[nodiscard]] u32 highest_priority_set(u32 max_priority = 0) const { const u64 priorities = max_priority == 0 ? used_priorities : (used_priorities & ~((1ULL << max_priority) - 1)); return priorities == 0 ? Depth : static_cast<u32>(CountTrailingZeroes64(priorities)); } - u32 lowest_priority_set(u32 min_priority = Depth - 1) const { + [[nodiscard]] u32 lowest_priority_set(u32 min_priority = Depth - 1) const { const u64 priorities = min_priority >= Depth - 1 ? used_priorities : (used_priorities & ((1ULL << (min_priority + 1)) - 1)); return priorities == 0 ? Depth : 63 - CountLeadingZeroes64(priorities); } - const_iterator cbegin(u32 max_prio = 0) const { + [[nodiscard]] const_iterator cbegin(u32 max_prio = 0) const { const u32 priority = highest_priority_set(max_prio); return priority == Depth ? cend() : const_iterator{*this, levels[priority].cbegin(), priority}; } - const_iterator begin(u32 max_prio = 0) const { + [[nodiscard]] const_iterator begin(u32 max_prio = 0) const { return cbegin(max_prio); } - iterator begin(u32 max_prio = 0) { + [[nodiscard]] iterator begin(u32 max_prio = 0) { const u32 priority = highest_priority_set(max_prio); return priority == Depth ? end() : iterator{*this, levels[priority].begin(), priority}; } - const_iterator cend(u32 min_prio = Depth - 1) const { + [[nodiscard]] const_iterator cend(u32 min_prio = Depth - 1) const { return min_prio == Depth - 1 ? const_iterator{*this, Depth} : cbegin(min_prio + 1); } - const_iterator end(u32 min_prio = Depth - 1) const { + [[nodiscard]] const_iterator end(u32 min_prio = Depth - 1) const { return cend(min_prio); } - iterator end(u32 min_prio = Depth - 1) { + [[nodiscard]] iterator end(u32 min_prio = Depth - 1) { return min_prio == Depth - 1 ? iterator{*this, Depth} : begin(min_prio + 1); } - T& front(u32 max_priority = 0) { + [[nodiscard]] T& front(u32 max_priority = 0) { const u32 priority = highest_priority_set(max_priority); return levels[priority == Depth ? 0 : priority].front(); } - const T& front(u32 max_priority = 0) const { + [[nodiscard]] const T& front(u32 max_priority = 0) const { const u32 priority = highest_priority_set(max_priority); return levels[priority == Depth ? 0 : priority].front(); } - T back(u32 min_priority = Depth - 1) { + [[nodiscard]] T& back(u32 min_priority = Depth - 1) { const u32 priority = lowest_priority_set(min_priority); // intended return levels[priority == Depth ? 63 : priority].back(); } - const T& back(u32 min_priority = Depth - 1) const { + [[nodiscard]] const T& back(u32 min_priority = Depth - 1) const { const u32 priority = lowest_priority_set(min_priority); // intended return levels[priority == Depth ? 63 : priority].back(); } @@ -329,7 +329,8 @@ private: in_list.splice(position, out_list, element); } - static const_list_iterator ListIterateTo(const std::list<T>& list, const T& element) { + [[nodiscard]] static const_list_iterator ListIterateTo(const std::list<T>& list, + const T& element) { auto it = list.cbegin(); while (it != list.cend() && *it != element) { ++it; diff --git a/src/common/page_table.cpp b/src/common/page_table.cpp index 566b57b62..bccea0894 100644 --- a/src/common/page_table.cpp +++ b/src/common/page_table.cpp @@ -6,36 +6,20 @@ namespace Common { -PageTable::PageTable(std::size_t page_size_in_bits) : page_size_in_bits{page_size_in_bits} {} +PageTable::PageTable() = default; -PageTable::~PageTable() = default; - -void PageTable::Resize(std::size_t address_space_width_in_bits) { - const std::size_t num_page_table_entries = 1ULL - << (address_space_width_in_bits - page_size_in_bits); +PageTable::~PageTable() noexcept = default; +void PageTable::Resize(std::size_t address_space_width_in_bits, std::size_t page_size_in_bits, + bool has_attribute) { + const std::size_t num_page_table_entries{1ULL + << (address_space_width_in_bits - page_size_in_bits)}; pointers.resize(num_page_table_entries); - attributes.resize(num_page_table_entries); - - // The default is a 39-bit address space, which causes an initial 1GB allocation size. If the - // vector size is subsequently decreased (via resize), the vector might not automatically - // actually reallocate/resize its underlying allocation, which wastes up to ~800 MB for - // 36-bit titles. Call shrink_to_fit to reduce capacity to what's actually in use. - - pointers.shrink_to_fit(); - attributes.shrink_to_fit(); -} - -BackingPageTable::BackingPageTable(std::size_t page_size_in_bits) : PageTable{page_size_in_bits} {} - -BackingPageTable::~BackingPageTable() = default; - -void BackingPageTable::Resize(std::size_t address_space_width_in_bits) { - PageTable::Resize(address_space_width_in_bits); - const std::size_t num_page_table_entries = 1ULL - << (address_space_width_in_bits - page_size_in_bits); backing_addr.resize(num_page_table_entries); - backing_addr.shrink_to_fit(); + + if (has_attribute) { + attributes.resize(num_page_table_entries); + } } } // namespace Common diff --git a/src/common/page_table.h b/src/common/page_table.h index dbc272ab7..9754fabf9 100644 --- a/src/common/page_table.h +++ b/src/common/page_table.h @@ -4,10 +4,11 @@ #pragma once -#include <vector> -#include <boost/icl/interval_map.hpp> +#include <tuple> + #include "common/common_types.h" #include "common/memory_hook.h" +#include "common/virtual_buffer.h" namespace Common { @@ -33,11 +34,11 @@ struct SpecialRegion { MemoryHookPointer handler; - bool operator<(const SpecialRegion& other) const { + [[nodiscard]] bool operator<(const SpecialRegion& other) const { return std::tie(type, handler) < std::tie(other.type, other.handler); } - bool operator==(const SpecialRegion& other) const { + [[nodiscard]] bool operator==(const SpecialRegion& other) const { return std::tie(type, handler) == std::tie(other.type, other.handler); } }; @@ -47,49 +48,35 @@ struct SpecialRegion { * mimics the way a real CPU page table works. */ struct PageTable { - explicit PageTable(std::size_t page_size_in_bits); - ~PageTable(); + PageTable(); + ~PageTable() noexcept; + + PageTable(const PageTable&) = delete; + PageTable& operator=(const PageTable&) = delete; + + PageTable(PageTable&&) noexcept = default; + PageTable& operator=(PageTable&&) noexcept = default; /** * Resizes the page table to be able to accomodate enough pages within * a given address space. * * @param address_space_width_in_bits The address size width in bits. + * @param page_size_in_bits The page size in bits. + * @param has_attribute Whether or not this page has any backing attributes. */ - void Resize(std::size_t address_space_width_in_bits); + void Resize(std::size_t address_space_width_in_bits, std::size_t page_size_in_bits, + bool has_attribute); /** * Vector of memory pointers backing each page. An entry can only be non-null if the * corresponding entry in the `attributes` vector is of type `Memory`. */ - std::vector<u8*> pointers; - - /** - * Contains MMIO handlers that back memory regions whose entries in the `attribute` vector is - * of type `Special`. - */ - boost::icl::interval_map<u64, std::set<SpecialRegion>> special_regions; - - /** - * Vector of fine grained page attributes. If it is set to any value other than `Memory`, then - * the corresponding entry in `pointers` MUST be set to null. - */ - std::vector<PageType> attributes; - - const std::size_t page_size_in_bits{}; -}; - -/** - * A more advanced Page Table with the ability to save a backing address when using it - * depends on another MMU. - */ -struct BackingPageTable : PageTable { - explicit BackingPageTable(std::size_t page_size_in_bits); - ~BackingPageTable(); + VirtualBuffer<u8*> pointers; - void Resize(std::size_t address_space_width_in_bits); + VirtualBuffer<u64> backing_addr; - std::vector<u64> backing_addr; + VirtualBuffer<PageType> attributes; }; } // namespace Common diff --git a/src/common/param_package.h b/src/common/param_package.h index 6a0a9b656..c13e45479 100644 --- a/src/common/param_package.h +++ b/src/common/param_package.h @@ -19,19 +19,19 @@ public: explicit ParamPackage(const std::string& serialized); ParamPackage(std::initializer_list<DataType::value_type> list); ParamPackage(const ParamPackage& other) = default; - ParamPackage(ParamPackage&& other) = default; + ParamPackage(ParamPackage&& other) noexcept = default; ParamPackage& operator=(const ParamPackage& other) = default; ParamPackage& operator=(ParamPackage&& other) = default; - std::string Serialize() const; - std::string Get(const std::string& key, const std::string& default_value) const; - int Get(const std::string& key, int default_value) const; - float Get(const std::string& key, float default_value) const; + [[nodiscard]] std::string Serialize() const; + [[nodiscard]] std::string Get(const std::string& key, const std::string& default_value) const; + [[nodiscard]] int Get(const std::string& key, int default_value) const; + [[nodiscard]] float Get(const std::string& key, float default_value) const; void Set(const std::string& key, std::string value); void Set(const std::string& key, int value); void Set(const std::string& key, float value); - bool Has(const std::string& key) const; + [[nodiscard]] bool Has(const std::string& key) const; void Erase(const std::string& key); void Clear(); diff --git a/src/common/quaternion.h b/src/common/quaternion.h index 370198ae0..4d0871eb4 100644 --- a/src/common/quaternion.h +++ b/src/common/quaternion.h @@ -14,35 +14,66 @@ public: Vec3<T> xyz; T w{}; - Quaternion<decltype(-T{})> Inverse() const { + [[nodiscard]] Quaternion<decltype(-T{})> Inverse() const { return {-xyz, w}; } - Quaternion<decltype(T{} + T{})> operator+(const Quaternion& other) const { + [[nodiscard]] Quaternion<decltype(T{} + T{})> operator+(const Quaternion& other) const { return {xyz + other.xyz, w + other.w}; } - Quaternion<decltype(T{} - T{})> operator-(const Quaternion& other) const { + [[nodiscard]] Quaternion<decltype(T{} - T{})> operator-(const Quaternion& other) const { return {xyz - other.xyz, w - other.w}; } - Quaternion<decltype(T{} * T{} - T{} * T{})> operator*(const Quaternion& other) const { + [[nodiscard]] Quaternion<decltype(T{} * T{} - T{} * T{})> operator*( + const Quaternion& other) const { return {xyz * other.w + other.xyz * w + Cross(xyz, other.xyz), w * other.w - Dot(xyz, other.xyz)}; } - Quaternion<T> Normalized() const { + [[nodiscard]] Quaternion<T> Normalized() const { T length = std::sqrt(xyz.Length2() + w * w); return {xyz / length, w / length}; } + + [[nodiscard]] std::array<decltype(-T{}), 16> ToMatrix() const { + const T x2 = xyz[0] * xyz[0]; + const T y2 = xyz[1] * xyz[1]; + const T z2 = xyz[2] * xyz[2]; + + const T xy = xyz[0] * xyz[1]; + const T wz = w * xyz[2]; + const T xz = xyz[0] * xyz[2]; + const T wy = w * xyz[1]; + const T yz = xyz[1] * xyz[2]; + const T wx = w * xyz[0]; + + return {1.0f - 2.0f * (y2 + z2), + 2.0f * (xy + wz), + 2.0f * (xz - wy), + 0.0f, + 2.0f * (xy - wz), + 1.0f - 2.0f * (x2 + z2), + 2.0f * (yz + wx), + 0.0f, + 2.0f * (xz + wy), + 2.0f * (yz - wx), + 1.0f - 2.0f * (x2 + y2), + 0.0f, + 0.0f, + 0.0f, + 0.0f, + 1.0f}; + } }; template <typename T> -auto QuaternionRotate(const Quaternion<T>& q, const Vec3<T>& v) { +[[nodiscard]] auto QuaternionRotate(const Quaternion<T>& q, const Vec3<T>& v) { return v + 2 * Cross(q.xyz, Cross(q.xyz, v) + v * q.w); } -inline Quaternion<float> MakeQuaternion(const Vec3<float>& axis, float angle) { +[[nodiscard]] inline Quaternion<float> MakeQuaternion(const Vec3<float>& axis, float angle) { return {axis * std::sin(angle / 2), std::cos(angle / 2)}; } diff --git a/src/common/ring_buffer.h b/src/common/ring_buffer.h index abe3b4dc2..138fa0131 100644 --- a/src/common/ring_buffer.h +++ b/src/common/ring_buffer.h @@ -91,12 +91,12 @@ public: } /// @returns Number of slots used - std::size_t Size() const { + [[nodiscard]] std::size_t Size() const { return m_write_index.load() - m_read_index.load(); } /// @returns Maximum size of ring buffer - constexpr std::size_t Capacity() const { + [[nodiscard]] constexpr std::size_t Capacity() const { return capacity; } diff --git a/src/common/scope_exit.h b/src/common/scope_exit.h index 1176a72b1..68ef5f197 100644 --- a/src/common/scope_exit.h +++ b/src/common/scope_exit.h @@ -12,10 +12,17 @@ template <typename Func> struct ScopeExitHelper { explicit ScopeExitHelper(Func&& func) : func(std::move(func)) {} ~ScopeExitHelper() { - func(); + if (active) { + func(); + } + } + + void Cancel() { + active = false; } Func func; + bool active{true}; }; template <typename Func> diff --git a/src/common/spin_lock.cpp b/src/common/spin_lock.cpp new file mode 100644 index 000000000..c1524220f --- /dev/null +++ b/src/common/spin_lock.cpp @@ -0,0 +1,54 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/spin_lock.h" + +#if _MSC_VER +#include <intrin.h> +#if _M_AMD64 +#define __x86_64__ 1 +#endif +#if _M_ARM64 +#define __aarch64__ 1 +#endif +#else +#if __x86_64__ +#include <xmmintrin.h> +#endif +#endif + +namespace { + +void ThreadPause() { +#if __x86_64__ + _mm_pause(); +#elif __aarch64__ && _MSC_VER + __yield(); +#elif __aarch64__ + asm("yield"); +#endif +} + +} // Anonymous namespace + +namespace Common { + +void SpinLock::lock() { + while (lck.test_and_set(std::memory_order_acquire)) { + ThreadPause(); + } +} + +void SpinLock::unlock() { + lck.clear(std::memory_order_release); +} + +bool SpinLock::try_lock() { + if (lck.test_and_set(std::memory_order_acquire)) { + return false; + } + return true; +} + +} // namespace Common diff --git a/src/common/spin_lock.h b/src/common/spin_lock.h new file mode 100644 index 000000000..06ac2f5bb --- /dev/null +++ b/src/common/spin_lock.h @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <atomic> + +namespace Common { + +/** + * SpinLock class + * a lock similar to mutex that forces a thread to spin wait instead calling the + * supervisor. Should be used on short sequences of code. + */ +class SpinLock { +public: + SpinLock() = default; + + SpinLock(const SpinLock&) = delete; + SpinLock& operator=(const SpinLock&) = delete; + + SpinLock(SpinLock&&) = delete; + SpinLock& operator=(SpinLock&&) = delete; + + void lock(); + void unlock(); + [[nodiscard]] bool try_lock(); + +private: + std::atomic_flag lck = ATOMIC_FLAG_INIT; +}; + +} // namespace Common diff --git a/src/common/stream.cpp b/src/common/stream.cpp new file mode 100644 index 000000000..bf0496c26 --- /dev/null +++ b/src/common/stream.cpp @@ -0,0 +1,47 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <stdexcept> +#include "common/common_types.h" +#include "common/stream.h" + +namespace Common { + +Stream::Stream() = default; +Stream::~Stream() = default; + +void Stream::Seek(s32 offset, SeekOrigin origin) { + if (origin == SeekOrigin::SetOrigin) { + if (offset < 0) { + position = 0; + } else if (position >= buffer.size()) { + position = buffer.size(); + } else { + position = offset; + } + } else if (origin == SeekOrigin::FromCurrentPos) { + Seek(static_cast<s32>(position) + offset, SeekOrigin::SetOrigin); + } else if (origin == SeekOrigin::FromEnd) { + Seek(static_cast<s32>(buffer.size()) - offset, SeekOrigin::SetOrigin); + } +} + +u8 Stream::ReadByte() { + if (position < buffer.size()) { + return buffer[position++]; + } else { + throw std::out_of_range("Attempting to read a byte not within the buffer range"); + } +} + +void Stream::WriteByte(u8 byte) { + if (position == buffer.size()) { + buffer.push_back(byte); + position++; + } else { + buffer.insert(buffer.begin() + position, byte); + } +} + +} // namespace Common diff --git a/src/common/stream.h b/src/common/stream.h new file mode 100644 index 000000000..0e40692de --- /dev/null +++ b/src/common/stream.h @@ -0,0 +1,56 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "common/common_types.h" + +namespace Common { + +enum class SeekOrigin { + SetOrigin, + FromCurrentPos, + FromEnd, +}; + +class Stream { +public: + /// Stream creates a bitstream and provides common functionality on the stream. + explicit Stream(); + ~Stream(); + + Stream(const Stream&) = delete; + Stream& operator=(const Stream&) = delete; + + Stream(Stream&&) = default; + Stream& operator=(Stream&&) = default; + + /// Reposition bitstream "cursor" to the specified offset from origin + void Seek(s32 offset, SeekOrigin origin); + + /// Reads next byte in the stream buffer and increments position + u8 ReadByte(); + + /// Writes byte at current position + void WriteByte(u8 byte); + + [[nodiscard]] std::size_t GetPosition() const { + return position; + } + + [[nodiscard]] std::vector<u8>& GetBuffer() { + return buffer; + } + + [[nodiscard]] const std::vector<u8>& GetBuffer() const { + return buffer; + } + +private: + std::vector<u8> buffer; + std::size_t position{0}; +}; + +} // namespace Common diff --git a/src/common/string_util.cpp b/src/common/string_util.cpp index 84883a1d3..4cba2aaa4 100644 --- a/src/common/string_util.cpp +++ b/src/common/string_util.cpp @@ -8,6 +8,7 @@ #include <cstdlib> #include <locale> #include <sstream> + #include "common/common_paths.h" #include "common/logging/log.h" #include "common/string_util.h" @@ -21,14 +22,14 @@ namespace Common { /// Make a string lowercase std::string ToLower(std::string str) { std::transform(str.begin(), str.end(), str.begin(), - [](unsigned char c) { return std::tolower(c); }); + [](unsigned char c) { return static_cast<char>(std::tolower(c)); }); return str; } /// Make a string uppercase std::string ToUpper(std::string str) { std::transform(str.begin(), str.end(), str.begin(), - [](unsigned char c) { return std::toupper(c); }); + [](unsigned char c) { return static_cast<char>(std::toupper(c)); }); return str; } diff --git a/src/common/string_util.h b/src/common/string_util.h index 583fd05e6..a32c07c06 100644 --- a/src/common/string_util.h +++ b/src/common/string_util.h @@ -12,19 +12,19 @@ namespace Common { /// Make a string lowercase -std::string ToLower(std::string str); +[[nodiscard]] std::string ToLower(std::string str); /// Make a string uppercase -std::string ToUpper(std::string str); +[[nodiscard]] std::string ToUpper(std::string str); -std::string StringFromBuffer(const std::vector<u8>& data); +[[nodiscard]] std::string StringFromBuffer(const std::vector<u8>& data); -std::string StripSpaces(const std::string& s); -std::string StripQuotes(const std::string& s); +[[nodiscard]] std::string StripSpaces(const std::string& s); +[[nodiscard]] std::string StripQuotes(const std::string& s); -std::string StringFromBool(bool value); +[[nodiscard]] std::string StringFromBool(bool value); -std::string TabsToSpaces(int tab_size, std::string in); +[[nodiscard]] std::string TabsToSpaces(int tab_size, std::string in); void SplitString(const std::string& str, char delim, std::vector<std::string>& output); @@ -34,14 +34,15 @@ bool SplitPath(const std::string& full_path, std::string* _pPath, std::string* _ void BuildCompleteFilename(std::string& _CompleteFilename, const std::string& _Path, const std::string& _Filename); -std::string ReplaceAll(std::string result, const std::string& src, const std::string& dest); +[[nodiscard]] std::string ReplaceAll(std::string result, const std::string& src, + const std::string& dest); -std::string UTF16ToUTF8(const std::u16string& input); -std::u16string UTF8ToUTF16(const std::string& input); +[[nodiscard]] std::string UTF16ToUTF8(const std::u16string& input); +[[nodiscard]] std::u16string UTF8ToUTF16(const std::string& input); #ifdef _WIN32 -std::string UTF16ToUTF8(const std::wstring& input); -std::wstring UTF8ToUTF16W(const std::string& str); +[[nodiscard]] std::string UTF16ToUTF8(const std::wstring& input); +[[nodiscard]] std::wstring UTF8ToUTF16W(const std::string& str); #endif @@ -50,7 +51,7 @@ std::wstring UTF8ToUTF16W(const std::string& str); * `other` for equality. */ template <typename InIt> -bool ComparePartialString(InIt begin, InIt end, const char* other) { +[[nodiscard]] bool ComparePartialString(InIt begin, InIt end, const char* other) { for (; begin != end && *other != '\0'; ++begin, ++other) { if (*begin != *other) { return false; @@ -64,26 +65,15 @@ bool ComparePartialString(InIt begin, InIt end, const char* other) { * Creates a std::string from a fixed-size NUL-terminated char buffer. If the buffer isn't * NUL-terminated then the string ends at max_len characters. */ -std::string StringFromFixedZeroTerminatedBuffer(const char* buffer, std::size_t max_len); +[[nodiscard]] std::string StringFromFixedZeroTerminatedBuffer(const char* buffer, + std::size_t max_len); /** * Creates a UTF-16 std::u16string from a fixed-size NUL-terminated char buffer. If the buffer isn't * null-terminated, then the string ends at the greatest multiple of two less then or equal to * max_len_bytes. */ -std::u16string UTF16StringFromFixedZeroTerminatedBuffer(std::u16string_view buffer, - std::size_t max_len); - -/** - * Attempts to trim an arbitrary prefix from `path`, leaving only the part starting at `root`. It's - * intended to be used to strip a system-specific build directory from the `__FILE__` macro, - * leaving only the path relative to the sources root. - * - * @param path The input file path as a null-terminated string - * @param root The name of the root source directory as a null-terminated string. Path up to and - * including the last occurrence of this name will be stripped - * @return A pointer to the same string passed as `path`, but starting at the trimmed portion - */ -const char* TrimSourcePath(const char* path, const char* root = "src"); +[[nodiscard]] std::u16string UTF16StringFromFixedZeroTerminatedBuffer(std::u16string_view buffer, + std::size_t max_len); } // namespace Common diff --git a/src/common/swap.h b/src/common/swap.h index 71932c2bb..7665942a2 100644 --- a/src/common/swap.h +++ b/src/common/swap.h @@ -17,43 +17,14 @@ #pragma once -#include <type_traits> - #if defined(_MSC_VER) #include <cstdlib> #endif +#include <bit> #include <cstring> +#include <type_traits> #include "common/common_types.h" -// GCC -#ifdef __GNUC__ - -#if __BYTE_ORDER__ && (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) && !defined(COMMON_LITTLE_ENDIAN) -#define COMMON_LITTLE_ENDIAN 1 -#elif __BYTE_ORDER__ && (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) && !defined(COMMON_BIG_ENDIAN) -#define COMMON_BIG_ENDIAN 1 -#endif - -// LLVM/clang -#elif defined(__clang__) - -#if __LITTLE_ENDIAN__ && !defined(COMMON_LITTLE_ENDIAN) -#define COMMON_LITTLE_ENDIAN 1 -#elif __BIG_ENDIAN__ && !defined(COMMON_BIG_ENDIAN) -#define COMMON_BIG_ENDIAN 1 -#endif - -// MSVC -#elif defined(_MSC_VER) && !defined(COMMON_BIG_ENDIAN) && !defined(COMMON_LITTLE_ENDIAN) - -#define COMMON_LITTLE_ENDIAN 1 -#endif - -// Worst case, default to little endian. -#if !COMMON_BIG_ENDIAN && !COMMON_LITTLE_ENDIAN -#define COMMON_LITTLE_ENDIAN 1 -#endif - namespace Common { #ifdef _MSC_VER @@ -675,17 +646,8 @@ struct AddEndian<T, SwapTag> { }; // Alias LETag/BETag as KeepTag/SwapTag depending on the system -#if COMMON_LITTLE_ENDIAN - -using LETag = KeepTag; -using BETag = SwapTag; - -#else - -using BETag = KeepTag; -using LETag = SwapTag; - -#endif +using LETag = std::conditional_t<std::endian::native == std::endian::little, KeepTag, SwapTag>; +using BETag = std::conditional_t<std::endian::native == std::endian::big, KeepTag, SwapTag>; // Aliases for LE types using u16_le = AddEndian<u16, LETag>::type; diff --git a/src/common/telemetry.cpp b/src/common/telemetry.cpp index 200c6489a..6241d08b3 100644 --- a/src/common/telemetry.cpp +++ b/src/common/telemetry.cpp @@ -12,7 +12,7 @@ #include "common/x64/cpu_detect.h" #endif -namespace Telemetry { +namespace Common::Telemetry { void FieldCollection::Accept(VisitorInterface& visitor) const { for (const auto& field : fields) { @@ -60,6 +60,7 @@ void AppendCPUInfo(FieldCollection& fc) { fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AES", Common::GetCPUCaps().aes); fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AVX", Common::GetCPUCaps().avx); fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AVX2", Common::GetCPUCaps().avx2); + fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_AVX512", Common::GetCPUCaps().avx512); fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_BMI1", Common::GetCPUCaps().bmi1); fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_BMI2", Common::GetCPUCaps().bmi2); fc.AddField(FieldType::UserSystem, "CPU_Extension_x64_FMA", Common::GetCPUCaps().fma); @@ -87,4 +88,4 @@ void AppendOSInfo(FieldCollection& fc) { #endif } -} // namespace Telemetry +} // namespace Common::Telemetry diff --git a/src/common/telemetry.h b/src/common/telemetry.h index 854a73fae..a50c5d1de 100644 --- a/src/common/telemetry.h +++ b/src/common/telemetry.h @@ -10,7 +10,7 @@ #include <string> #include "common/common_types.h" -namespace Telemetry { +namespace Common::Telemetry { /// Field type, used for grouping fields together in the final submitted telemetry log enum class FieldType : u8 { @@ -63,30 +63,30 @@ public: void Accept(VisitorInterface& visitor) const override; - const std::string& GetName() const override { + [[nodiscard]] const std::string& GetName() const override { return name; } /** * Returns the type of the field. */ - FieldType GetType() const { + [[nodiscard]] FieldType GetType() const { return type; } /** * Returns the value of the field. */ - const T& GetValue() const { + [[nodiscard]] const T& GetValue() const { return value; } - bool operator==(const Field& other) const { + [[nodiscard]] bool operator==(const Field& other) const { return (type == other.type) && (name == other.name) && (value == other.value); } - bool operator!=(const Field& other) const { - return !(*this == other); + [[nodiscard]] bool operator!=(const Field& other) const { + return !operator==(other); } private: @@ -196,4 +196,4 @@ void AppendCPUInfo(FieldCollection& fc); /// such as platform name, etc. void AppendOSInfo(FieldCollection& fc); -} // namespace Telemetry +} // namespace Common::Telemetry diff --git a/src/common/thread.cpp b/src/common/thread.cpp index 0cd2d10bf..d2c1ac60d 100644 --- a/src/common/thread.cpp +++ b/src/common/thread.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/common_funcs.h" +#include "common/logging/log.h" #include "common/thread.h" #ifdef __APPLE__ #include <mach/mach.h> @@ -19,12 +21,60 @@ #include <unistd.h> #endif +#include <string> + #ifdef __FreeBSD__ #define cpu_set_t cpuset_t #endif namespace Common { +#ifdef _WIN32 + +void SetCurrentThreadPriority(ThreadPriority new_priority) { + auto handle = GetCurrentThread(); + int windows_priority = 0; + switch (new_priority) { + case ThreadPriority::Low: + windows_priority = THREAD_PRIORITY_BELOW_NORMAL; + break; + case ThreadPriority::Normal: + windows_priority = THREAD_PRIORITY_NORMAL; + break; + case ThreadPriority::High: + windows_priority = THREAD_PRIORITY_ABOVE_NORMAL; + break; + case ThreadPriority::VeryHigh: + windows_priority = THREAD_PRIORITY_HIGHEST; + break; + default: + windows_priority = THREAD_PRIORITY_NORMAL; + break; + } + SetThreadPriority(handle, windows_priority); +} + +#else + +void SetCurrentThreadPriority(ThreadPriority new_priority) { + pthread_t this_thread = pthread_self(); + + s32 max_prio = sched_get_priority_max(SCHED_OTHER); + s32 min_prio = sched_get_priority_min(SCHED_OTHER); + u32 level = static_cast<u32>(new_priority) + 1; + + struct sched_param params; + if (max_prio > min_prio) { + params.sched_priority = min_prio + ((max_prio - min_prio) * level) / 4; + } else { + params.sched_priority = min_prio - ((min_prio - max_prio) * level) / 4; + } + + pthread_setschedparam(this_thread, SCHED_OTHER, ¶ms); +} + +#endif + #ifdef _MSC_VER // Sets the debugger-visible name of the current thread. @@ -64,12 +114,26 @@ void SetCurrentThreadName(const char* name) { pthread_set_name_np(pthread_self(), name); #elif defined(__NetBSD__) pthread_setname_np(pthread_self(), "%s", (void*)name); +#elif defined(__linux__) + // Linux limits thread names to 15 characters and will outright reject any + // attempt to set a longer name with ERANGE. + std::string truncated(name, std::min(strlen(name), static_cast<size_t>(15))); + if (int e = pthread_setname_np(pthread_self(), truncated.c_str())) { + errno = e; + LOG_ERROR(Common, "Failed to set thread name to '{}': {}", truncated, GetLastErrorMsg()); + } #else pthread_setname_np(pthread_self(), name); #endif } #endif +#if defined(_WIN32) +void SetCurrentThreadName(const char* name) { + // Do Nothing on MingW +} +#endif + #endif } // namespace Common diff --git a/src/common/thread.h b/src/common/thread.h index 2fc071685..a8c17c71a 100644 --- a/src/common/thread.h +++ b/src/common/thread.h @@ -4,11 +4,13 @@ #pragma once +#include <atomic> #include <chrono> #include <condition_variable> #include <cstddef> #include <mutex> #include <thread> +#include "common/common_types.h" namespace Common { @@ -24,14 +26,13 @@ public: void Wait() { std::unique_lock lk{mutex}; - condvar.wait(lk, [&] { return is_set; }); + condvar.wait(lk, [&] { return is_set.load(); }); is_set = false; } - template <class Duration> - bool WaitFor(const std::chrono::duration<Duration>& time) { + bool WaitFor(const std::chrono::nanoseconds& time) { std::unique_lock lk{mutex}; - if (!condvar.wait_for(lk, time, [this] { return is_set; })) + if (!condvar.wait_for(lk, time, [this] { return is_set.load(); })) return false; is_set = false; return true; @@ -40,7 +41,7 @@ public: template <class Clock, class Duration> bool WaitUntil(const std::chrono::time_point<Clock, Duration>& time) { std::unique_lock lk{mutex}; - if (!condvar.wait_until(lk, time, [this] { return is_set; })) + if (!condvar.wait_until(lk, time, [this] { return is_set.load(); })) return false; is_set = false; return true; @@ -54,9 +55,9 @@ public: } private: - bool is_set = false; std::condition_variable condvar; std::mutex mutex; + std::atomic_bool is_set{false}; }; class Barrier { @@ -86,6 +87,15 @@ private: std::size_t generation = 0; // Incremented once each time the barrier is used }; +enum class ThreadPriority : u32 { + Low = 0, + Normal = 1, + High = 2, + VeryHigh = 3, +}; + +void SetCurrentThreadPriority(ThreadPriority new_priority); + void SetCurrentThreadName(const char* name); } // namespace Common diff --git a/src/common/thread_queue_list.h b/src/common/thread_queue_list.h index 791f99a8c..def9e5d8d 100644 --- a/src/common/thread_queue_list.h +++ b/src/common/thread_queue_list.h @@ -18,14 +18,14 @@ struct ThreadQueueList { using Priority = unsigned int; // Number of priority levels. (Valid levels are [0..NUM_QUEUES).) - static const Priority NUM_QUEUES = N; + static constexpr Priority NUM_QUEUES = N; ThreadQueueList() { first = nullptr; } // Only for debugging, returns priority level. - Priority contains(const T& uid) const { + [[nodiscard]] Priority contains(const T& uid) const { for (Priority i = 0; i < NUM_QUEUES; ++i) { const Queue& cur = queues[i]; if (std::find(cur.data.cbegin(), cur.data.cend(), uid) != cur.data.cend()) { @@ -36,7 +36,7 @@ struct ThreadQueueList { return -1; } - T get_first() const { + [[nodiscard]] T get_first() const { const Queue* cur = first; while (cur != nullptr) { if (!cur->data.empty()) { @@ -49,7 +49,7 @@ struct ThreadQueueList { } template <typename UnaryPredicate> - T get_first_filter(UnaryPredicate filter) const { + [[nodiscard]] T get_first_filter(UnaryPredicate filter) const { const Queue* cur = first; while (cur != nullptr) { if (!cur->data.empty()) { @@ -129,7 +129,7 @@ struct ThreadQueueList { first = nullptr; } - bool empty(Priority priority) const { + [[nodiscard]] bool empty(Priority priority) const { const Queue* cur = &queues[priority]; return cur->data.empty(); } diff --git a/src/common/threadsafe_queue.h b/src/common/threadsafe_queue.h index 8268bbd5c..a4647314a 100644 --- a/src/common/threadsafe_queue.h +++ b/src/common/threadsafe_queue.h @@ -25,15 +25,15 @@ public: delete read_ptr; } - std::size_t Size() const { + [[nodiscard]] std::size_t Size() const { return size.load(); } - bool Empty() const { + [[nodiscard]] bool Empty() const { return Size() == 0; } - T& Front() const { + [[nodiscard]] T& Front() const { return read_ptr->current; } @@ -130,15 +130,15 @@ private: template <typename T> class MPSCQueue { public: - std::size_t Size() const { + [[nodiscard]] std::size_t Size() const { return spsc_queue.Size(); } - bool Empty() const { + [[nodiscard]] bool Empty() const { return spsc_queue.Empty(); } - T& Front() const { + [[nodiscard]] T& Front() const { return spsc_queue.Front(); } diff --git a/src/common/time_zone.cpp b/src/common/time_zone.cpp new file mode 100644 index 000000000..ce239eb63 --- /dev/null +++ b/src/common/time_zone.cpp @@ -0,0 +1,49 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <chrono> +#include <iomanip> +#include <sstream> + +#include "common/logging/log.h" +#include "common/time_zone.h" + +namespace Common::TimeZone { + +std::string GetDefaultTimeZone() { + return "GMT"; +} + +static std::string GetOsTimeZoneOffset() { + const std::time_t t{std::time(nullptr)}; + const std::tm tm{*std::localtime(&t)}; + + std::stringstream ss; + ss << std::put_time(&tm, "%z"); // Get the current timezone offset, e.g. "-400", as a string + + return ss.str(); +} + +static int ConvertOsTimeZoneOffsetToInt(const std::string& timezone) { + try { + return std::stoi(timezone); + } catch (const std::invalid_argument&) { + LOG_CRITICAL(Common, "invalid_argument with {}!", timezone); + return 0; + } catch (const std::out_of_range&) { + LOG_CRITICAL(Common, "out_of_range with {}!", timezone); + return 0; + } +} + +std::chrono::seconds GetCurrentOffsetSeconds() { + const int offset{ConvertOsTimeZoneOffsetToInt(GetOsTimeZoneOffset())}; + + int seconds{(offset / 100) * 60 * 60}; // Convert hour component to seconds + seconds += (offset % 100) * 60; // Convert minute component to seconds + + return std::chrono::seconds{seconds}; +} + +} // namespace Common::TimeZone diff --git a/src/common/time_zone.h b/src/common/time_zone.h new file mode 100644 index 000000000..9f5939ca5 --- /dev/null +++ b/src/common/time_zone.h @@ -0,0 +1,18 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <chrono> +#include <string> + +namespace Common::TimeZone { + +/// Gets the default timezone, i.e. "GMT" +[[nodiscard]] std::string GetDefaultTimeZone(); + +/// Gets the offset of the current timezone (from the default), in seconds +[[nodiscard]] std::chrono::seconds GetCurrentOffsetSeconds(); + +} // namespace Common::TimeZone diff --git a/src/common/timer.cpp b/src/common/timer.cpp index 2dc15e434..d17dc2a50 100644 --- a/src/common/timer.cpp +++ b/src/common/timer.cpp @@ -142,20 +142,18 @@ std::string Timer::GetTimeFormatted() { // ---------------- double Timer::GetDoubleTime() { // Get continuous timestamp - u64 TmpSeconds = static_cast<u64>(Common::Timer::GetTimeSinceJan1970().count()); - double ms = static_cast<u64>(GetTimeMs().count()) % 1000; + auto tmp_seconds = static_cast<u64>(GetTimeSinceJan1970().count()); + const auto ms = static_cast<double>(static_cast<u64>(GetTimeMs().count()) % 1000); // Remove a few years. We only really want enough seconds to make // sure that we are detecting actual actions, perhaps 60 seconds is // enough really, but I leave a year of seconds anyway, in case the // user's clock is incorrect or something like that. - TmpSeconds = TmpSeconds - (38 * 365 * 24 * 60 * 60); + tmp_seconds = tmp_seconds - (38 * 365 * 24 * 60 * 60); // Make a smaller integer that fits in the double - u32 Seconds = static_cast<u32>(TmpSeconds); - double TmpTime = Seconds + ms; - - return TmpTime; + const auto seconds = static_cast<u32>(tmp_seconds); + return seconds + ms; } } // Namespace Common diff --git a/src/common/timer.h b/src/common/timer.h index 27b521baa..8894a143d 100644 --- a/src/common/timer.h +++ b/src/common/timer.h @@ -19,18 +19,18 @@ public: // The time difference is always returned in milliseconds, regardless of alternative internal // representation - std::chrono::milliseconds GetTimeDifference(); + [[nodiscard]] std::chrono::milliseconds GetTimeDifference(); void AddTimeDifference(); - static std::chrono::seconds GetTimeSinceJan1970(); - static std::chrono::seconds GetLocalTimeSinceJan1970(); - static double GetDoubleTime(); + [[nodiscard]] static std::chrono::seconds GetTimeSinceJan1970(); + [[nodiscard]] static std::chrono::seconds GetLocalTimeSinceJan1970(); + [[nodiscard]] static double GetDoubleTime(); - static std::string GetTimeFormatted(); - std::string GetTimeElapsedFormatted() const; - std::chrono::milliseconds GetTimeElapsed(); + [[nodiscard]] static std::string GetTimeFormatted(); + [[nodiscard]] std::string GetTimeElapsedFormatted() const; + [[nodiscard]] std::chrono::milliseconds GetTimeElapsed(); - static std::chrono::milliseconds GetTimeMs(); + [[nodiscard]] static std::chrono::milliseconds GetTimeMs(); private: std::chrono::milliseconds m_LastTime; diff --git a/src/common/uint128.cpp b/src/common/uint128.cpp index 32bf56730..16bf7c828 100644 --- a/src/common/uint128.cpp +++ b/src/common/uint128.cpp @@ -6,12 +6,38 @@ #include <intrin.h> #pragma intrinsic(_umul128) +#pragma intrinsic(_udiv128) #endif #include <cstring> #include "common/uint128.h" namespace Common { +#ifdef _MSC_VER + +u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) { + u128 r{}; + r[0] = _umul128(a, b, &r[1]); + u64 remainder; +#if _MSC_VER < 1923 + return udiv128(r[1], r[0], d, &remainder); +#else + return _udiv128(r[1], r[0], d, &remainder); +#endif +} + +#else + +u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) { + const u64 diva = a / d; + const u64 moda = a % d; + const u64 divb = b / d; + const u64 modb = b % d; + return diva * b + moda * divb + moda * modb / d; +} + +#endif + u128 Multiply64Into128(u64 a, u64 b) { u128 result; #ifdef _MSC_VER diff --git a/src/common/uint128.h b/src/common/uint128.h index a3be2a2cb..969259ab6 100644 --- a/src/common/uint128.h +++ b/src/common/uint128.h @@ -9,11 +9,14 @@ namespace Common { +// This function multiplies 2 u64 values and divides it by a u64 value. +[[nodiscard]] u64 MultiplyAndDivide64(u64 a, u64 b, u64 d); + // This function multiplies 2 u64 values and produces a u128 value; -u128 Multiply64Into128(u64 a, u64 b); +[[nodiscard]] u128 Multiply64Into128(u64 a, u64 b); // This function divides a u128 by a u32 value and produces two u64 values: // the result of division and the remainder -std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor); +[[nodiscard]] std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor); } // namespace Common diff --git a/src/common/uuid.h b/src/common/uuid.h index f6ad064fb..4ab9a25f0 100644 --- a/src/common/uuid.h +++ b/src/common/uuid.h @@ -19,29 +19,34 @@ struct UUID { constexpr explicit UUID(const u128& id) : uuid{id} {} constexpr explicit UUID(const u64 lo, const u64 hi) : uuid{{lo, hi}} {} - constexpr explicit operator bool() const { + [[nodiscard]] constexpr explicit operator bool() const { return uuid[0] != INVALID_UUID[0] && uuid[1] != INVALID_UUID[1]; } - constexpr bool operator==(const UUID& rhs) const { + [[nodiscard]] constexpr bool operator==(const UUID& rhs) const { // TODO(DarkLordZach): Replace with uuid == rhs.uuid with C++20 return uuid[0] == rhs.uuid[0] && uuid[1] == rhs.uuid[1]; } - constexpr bool operator!=(const UUID& rhs) const { + [[nodiscard]] constexpr bool operator!=(const UUID& rhs) const { return !operator==(rhs); } // TODO(ogniK): Properly generate uuids based on RFC-4122 - static UUID Generate(); + [[nodiscard]] static UUID Generate(); // Set the UUID to {0,0} to be considered an invalid user constexpr void Invalidate() { uuid = INVALID_UUID; } - std::string Format() const; - std::string FormatSwitch() const; + // TODO(ogniK): Properly generate a Nintendo ID + [[nodiscard]] constexpr u64 GetNintendoID() const { + return uuid[0]; + } + + [[nodiscard]] std::string Format() const; + [[nodiscard]] std::string FormatSwitch() const; }; static_assert(sizeof(UUID) == 16, "UUID is an invalid size!"); diff --git a/src/common/vector_math.h b/src/common/vector_math.h index 429485329..22dba3c2d 100644 --- a/src/common/vector_math.h +++ b/src/common/vector_math.h @@ -52,15 +52,15 @@ public: constexpr Vec2(const T& x_, const T& y_) : x(x_), y(y_) {} template <typename T2> - constexpr Vec2<T2> Cast() const { + [[nodiscard]] constexpr Vec2<T2> Cast() const { return Vec2<T2>(static_cast<T2>(x), static_cast<T2>(y)); } - static constexpr Vec2 AssignToAll(const T& f) { + [[nodiscard]] static constexpr Vec2 AssignToAll(const T& f) { return Vec2{f, f}; } - constexpr Vec2<decltype(T{} + T{})> operator+(const Vec2& other) const { + [[nodiscard]] constexpr Vec2<decltype(T{} + T{})> operator+(const Vec2& other) const { return {x + other.x, y + other.y}; } constexpr Vec2& operator+=(const Vec2& other) { @@ -68,7 +68,7 @@ public: y += other.y; return *this; } - constexpr Vec2<decltype(T{} - T{})> operator-(const Vec2& other) const { + [[nodiscard]] constexpr Vec2<decltype(T{} - T{})> operator-(const Vec2& other) const { return {x - other.x, y - other.y}; } constexpr Vec2& operator-=(const Vec2& other) { @@ -78,16 +78,22 @@ public: } template <typename U = T> - constexpr Vec2<std::enable_if_t<std::is_signed_v<U>, U>> operator-() const { + [[nodiscard]] constexpr Vec2<std::enable_if_t<std::is_signed_v<U>, U>> operator-() const { return {-x, -y}; } - constexpr Vec2<decltype(T{} * T{})> operator*(const Vec2& other) const { + [[nodiscard]] constexpr Vec2<decltype(T{} * T{})> operator*(const Vec2& other) const { return {x * other.x, y * other.y}; } template <typename V> - constexpr Vec2<decltype(T{} * V{})> operator*(const V& f) const { - return {x * f, y * f}; + [[nodiscard]] constexpr Vec2<decltype(T{} * V{})> operator*(const V& f) const { + using TV = decltype(T{} * V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)), + }; } template <typename V> @@ -97,8 +103,14 @@ public: } template <typename V> - constexpr Vec2<decltype(T{} / V{})> operator/(const V& f) const { - return {x / f, y / f}; + [[nodiscard]] constexpr Vec2<decltype(T{} / V{})> operator/(const V& f) const { + using TV = decltype(T{} / V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)), + }; } template <typename V> @@ -107,18 +119,18 @@ public: return *this; } - constexpr T Length2() const { + [[nodiscard]] constexpr T Length2() const { return x * x + y * y; } // Only implemented for T=float - float Length() const; - float Normalize(); // returns the previous length, which is often useful + [[nodiscard]] float Length() const; + [[nodiscard]] float Normalize(); // returns the previous length, which is often useful - constexpr T& operator[](std::size_t i) { + [[nodiscard]] constexpr T& operator[](std::size_t i) { return *((&x) + i); } - constexpr const T& operator[](std::size_t i) const { + [[nodiscard]] constexpr const T& operator[](std::size_t i) const { return *((&x) + i); } @@ -128,47 +140,50 @@ public: } // Common aliases: UV (texel coordinates), ST (texture coordinates) - constexpr T& u() { + [[nodiscard]] constexpr T& u() { return x; } - constexpr T& v() { + [[nodiscard]] constexpr T& v() { return y; } - constexpr T& s() { + [[nodiscard]] constexpr T& s() { return x; } - constexpr T& t() { + [[nodiscard]] constexpr T& t() { return y; } - constexpr const T& u() const { + [[nodiscard]] constexpr const T& u() const { return x; } - constexpr const T& v() const { + [[nodiscard]] constexpr const T& v() const { return y; } - constexpr const T& s() const { + [[nodiscard]] constexpr const T& s() const { return x; } - constexpr const T& t() const { + [[nodiscard]] constexpr const T& t() const { return y; } // swizzlers - create a subvector of specific components - constexpr Vec2 yx() const { + [[nodiscard]] constexpr Vec2 yx() const { return Vec2(y, x); } - constexpr Vec2 vu() const { + [[nodiscard]] constexpr Vec2 vu() const { return Vec2(y, x); } - constexpr Vec2 ts() const { + [[nodiscard]] constexpr Vec2 ts() const { return Vec2(y, x); } }; template <typename T, typename V> -constexpr Vec2<T> operator*(const V& f, const Vec2<T>& vec) { - return Vec2<T>(f * vec.x, f * vec.y); +[[nodiscard]] constexpr Vec2<T> operator*(const V& f, const Vec2<T>& vec) { + using C = std::common_type_t<T, V>; + + return Vec2<T>(static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.x)), + static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.y))); } using Vec2f = Vec2<float>; @@ -196,15 +211,15 @@ public: constexpr Vec3(const T& x_, const T& y_, const T& z_) : x(x_), y(y_), z(z_) {} template <typename T2> - constexpr Vec3<T2> Cast() const { + [[nodiscard]] constexpr Vec3<T2> Cast() const { return Vec3<T2>(static_cast<T2>(x), static_cast<T2>(y), static_cast<T2>(z)); } - static constexpr Vec3 AssignToAll(const T& f) { + [[nodiscard]] static constexpr Vec3 AssignToAll(const T& f) { return Vec3(f, f, f); } - constexpr Vec3<decltype(T{} + T{})> operator+(const Vec3& other) const { + [[nodiscard]] constexpr Vec3<decltype(T{} + T{})> operator+(const Vec3& other) const { return {x + other.x, y + other.y, z + other.z}; } @@ -215,7 +230,7 @@ public: return *this; } - constexpr Vec3<decltype(T{} - T{})> operator-(const Vec3& other) const { + [[nodiscard]] constexpr Vec3<decltype(T{} - T{})> operator-(const Vec3& other) const { return {x - other.x, y - other.y, z - other.z}; } @@ -227,17 +242,24 @@ public: } template <typename U = T> - constexpr Vec3<std::enable_if_t<std::is_signed_v<U>, U>> operator-() const { + [[nodiscard]] constexpr Vec3<std::enable_if_t<std::is_signed_v<U>, U>> operator-() const { return {-x, -y, -z}; } - constexpr Vec3<decltype(T{} * T{})> operator*(const Vec3& other) const { + [[nodiscard]] constexpr Vec3<decltype(T{} * T{})> operator*(const Vec3& other) const { return {x * other.x, y * other.y, z * other.z}; } template <typename V> - constexpr Vec3<decltype(T{} * V{})> operator*(const V& f) const { - return {x * f, y * f, z * f}; + [[nodiscard]] constexpr Vec3<decltype(T{} * V{})> operator*(const V& f) const { + using TV = decltype(T{} * V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) * static_cast<C>(f)), + }; } template <typename V> @@ -246,8 +268,15 @@ public: return *this; } template <typename V> - constexpr Vec3<decltype(T{} / V{})> operator/(const V& f) const { - return {x / f, y / f, z / f}; + [[nodiscard]] constexpr Vec3<decltype(T{} / V{})> operator/(const V& f) const { + using TV = decltype(T{} / V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) / static_cast<C>(f)), + }; } template <typename V> @@ -256,20 +285,20 @@ public: return *this; } - constexpr T Length2() const { + [[nodiscard]] constexpr T Length2() const { return x * x + y * y + z * z; } // Only implemented for T=float - float Length() const; - Vec3 Normalized() const; - float Normalize(); // returns the previous length, which is often useful + [[nodiscard]] float Length() const; + [[nodiscard]] Vec3 Normalized() const; + [[nodiscard]] float Normalize(); // returns the previous length, which is often useful - constexpr T& operator[](std::size_t i) { + [[nodiscard]] constexpr T& operator[](std::size_t i) { return *((&x) + i); } - constexpr const T& operator[](std::size_t i) const { + [[nodiscard]] constexpr const T& operator[](std::size_t i) const { return *((&x) + i); } @@ -280,63 +309,63 @@ public: } // Common aliases: UVW (texel coordinates), RGB (colors), STQ (texture coordinates) - constexpr T& u() { + [[nodiscard]] constexpr T& u() { return x; } - constexpr T& v() { + [[nodiscard]] constexpr T& v() { return y; } - constexpr T& w() { + [[nodiscard]] constexpr T& w() { return z; } - constexpr T& r() { + [[nodiscard]] constexpr T& r() { return x; } - constexpr T& g() { + [[nodiscard]] constexpr T& g() { return y; } - constexpr T& b() { + [[nodiscard]] constexpr T& b() { return z; } - constexpr T& s() { + [[nodiscard]] constexpr T& s() { return x; } - constexpr T& t() { + [[nodiscard]] constexpr T& t() { return y; } - constexpr T& q() { + [[nodiscard]] constexpr T& q() { return z; } - constexpr const T& u() const { + [[nodiscard]] constexpr const T& u() const { return x; } - constexpr const T& v() const { + [[nodiscard]] constexpr const T& v() const { return y; } - constexpr const T& w() const { + [[nodiscard]] constexpr const T& w() const { return z; } - constexpr const T& r() const { + [[nodiscard]] constexpr const T& r() const { return x; } - constexpr const T& g() const { + [[nodiscard]] constexpr const T& g() const { return y; } - constexpr const T& b() const { + [[nodiscard]] constexpr const T& b() const { return z; } - constexpr const T& s() const { + [[nodiscard]] constexpr const T& s() const { return x; } - constexpr const T& t() const { + [[nodiscard]] constexpr const T& t() const { return y; } - constexpr const T& q() const { + [[nodiscard]] constexpr const T& q() const { return z; } @@ -345,7 +374,7 @@ public: // _DEFINE_SWIZZLER2 defines a single such function, DEFINE_SWIZZLER2 defines all of them for all // component names (x<->r) and permutations (xy<->yx) #define _DEFINE_SWIZZLER2(a, b, name) \ - constexpr Vec2<T> name() const { \ + [[nodiscard]] constexpr Vec2<T> name() const { \ return Vec2<T>(a, b); \ } #define DEFINE_SWIZZLER2(a, b, a2, b2, a3, b3, a4, b4) \ @@ -366,8 +395,12 @@ public: }; template <typename T, typename V> -constexpr Vec3<T> operator*(const V& f, const Vec3<T>& vec) { - return Vec3<T>(f * vec.x, f * vec.y, f * vec.z); +[[nodiscard]] constexpr Vec3<T> operator*(const V& f, const Vec3<T>& vec) { + using C = std::common_type_t<T, V>; + + return Vec3<T>(static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.x)), + static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.y)), + static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.z))); } template <> @@ -402,16 +435,16 @@ public: : x(x_), y(y_), z(z_), w(w_) {} template <typename T2> - constexpr Vec4<T2> Cast() const { + [[nodiscard]] constexpr Vec4<T2> Cast() const { return Vec4<T2>(static_cast<T2>(x), static_cast<T2>(y), static_cast<T2>(z), static_cast<T2>(w)); } - static constexpr Vec4 AssignToAll(const T& f) { + [[nodiscard]] static constexpr Vec4 AssignToAll(const T& f) { return Vec4(f, f, f, f); } - constexpr Vec4<decltype(T{} + T{})> operator+(const Vec4& other) const { + [[nodiscard]] constexpr Vec4<decltype(T{} + T{})> operator+(const Vec4& other) const { return {x + other.x, y + other.y, z + other.z, w + other.w}; } @@ -423,7 +456,7 @@ public: return *this; } - constexpr Vec4<decltype(T{} - T{})> operator-(const Vec4& other) const { + [[nodiscard]] constexpr Vec4<decltype(T{} - T{})> operator-(const Vec4& other) const { return {x - other.x, y - other.y, z - other.z, w - other.w}; } @@ -436,17 +469,25 @@ public: } template <typename U = T> - constexpr Vec4<std::enable_if_t<std::is_signed_v<U>, U>> operator-() const { + [[nodiscard]] constexpr Vec4<std::enable_if_t<std::is_signed_v<U>, U>> operator-() const { return {-x, -y, -z, -w}; } - constexpr Vec4<decltype(T{} * T{})> operator*(const Vec4& other) const { + [[nodiscard]] constexpr Vec4<decltype(T{} * T{})> operator*(const Vec4& other) const { return {x * other.x, y * other.y, z * other.z, w * other.w}; } template <typename V> - constexpr Vec4<decltype(T{} * V{})> operator*(const V& f) const { - return {x * f, y * f, z * f, w * f}; + [[nodiscard]] constexpr Vec4<decltype(T{} * V{})> operator*(const V& f) const { + using TV = decltype(T{} * V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(w) * static_cast<C>(f)), + }; } template <typename V> @@ -456,8 +497,16 @@ public: } template <typename V> - constexpr Vec4<decltype(T{} / V{})> operator/(const V& f) const { - return {x / f, y / f, z / f, w / f}; + [[nodiscard]] constexpr Vec4<decltype(T{} / V{})> operator/(const V& f) const { + using TV = decltype(T{} / V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(w) / static_cast<C>(f)), + }; } template <typename V> @@ -466,15 +515,15 @@ public: return *this; } - constexpr T Length2() const { + [[nodiscard]] constexpr T Length2() const { return x * x + y * y + z * z + w * w; } - constexpr T& operator[](std::size_t i) { + [[nodiscard]] constexpr T& operator[](std::size_t i) { return *((&x) + i); } - constexpr const T& operator[](std::size_t i) const { + [[nodiscard]] constexpr const T& operator[](std::size_t i) const { return *((&x) + i); } @@ -486,29 +535,29 @@ public: } // Common alias: RGBA (colors) - constexpr T& r() { + [[nodiscard]] constexpr T& r() { return x; } - constexpr T& g() { + [[nodiscard]] constexpr T& g() { return y; } - constexpr T& b() { + [[nodiscard]] constexpr T& b() { return z; } - constexpr T& a() { + [[nodiscard]] constexpr T& a() { return w; } - constexpr const T& r() const { + [[nodiscard]] constexpr const T& r() const { return x; } - constexpr const T& g() const { + [[nodiscard]] constexpr const T& g() const { return y; } - constexpr const T& b() const { + [[nodiscard]] constexpr const T& b() const { return z; } - constexpr const T& a() const { + [[nodiscard]] constexpr const T& a() const { return w; } @@ -520,7 +569,7 @@ public: // DEFINE_SWIZZLER2_COMP2 defines two component functions for all component names (x<->r) and // permutations (xy<->yx) #define _DEFINE_SWIZZLER2(a, b, name) \ - constexpr Vec2<T> name() const { \ + [[nodiscard]] constexpr Vec2<T> name() const { \ return Vec2<T>(a, b); \ } #define DEFINE_SWIZZLER2_COMP1(a, a2) \ @@ -547,7 +596,7 @@ public: #undef _DEFINE_SWIZZLER2 #define _DEFINE_SWIZZLER3(a, b, c, name) \ - constexpr Vec3<T> name() const { \ + [[nodiscard]] constexpr Vec3<T> name() const { \ return Vec3<T>(a, b, c); \ } #define DEFINE_SWIZZLER3_COMP1(a, a2) \ @@ -581,8 +630,16 @@ public: }; template <typename T, typename V> -constexpr Vec4<decltype(V{} * T{})> operator*(const V& f, const Vec4<T>& vec) { - return {f * vec.x, f * vec.y, f * vec.z, f * vec.w}; +[[nodiscard]] constexpr Vec4<decltype(V{} * T{})> operator*(const V& f, const Vec4<T>& vec) { + using TV = decltype(V{} * T{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.x)), + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.y)), + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.z)), + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.w)), + }; } using Vec4f = Vec4<float>; @@ -593,39 +650,41 @@ constexpr decltype(T{} * T{} + T{} * T{}) Dot(const Vec2<T>& a, const Vec2<T>& b } template <typename T> -constexpr decltype(T{} * T{} + T{} * T{}) Dot(const Vec3<T>& a, const Vec3<T>& b) { +[[nodiscard]] constexpr decltype(T{} * T{} + T{} * T{}) Dot(const Vec3<T>& a, const Vec3<T>& b) { return a.x * b.x + a.y * b.y + a.z * b.z; } template <typename T> -constexpr decltype(T{} * T{} + T{} * T{}) Dot(const Vec4<T>& a, const Vec4<T>& b) { +[[nodiscard]] constexpr decltype(T{} * T{} + T{} * T{}) Dot(const Vec4<T>& a, const Vec4<T>& b) { return a.x * b.x + a.y * b.y + a.z * b.z + a.w * b.w; } template <typename T> -constexpr Vec3<decltype(T{} * T{} - T{} * T{})> Cross(const Vec3<T>& a, const Vec3<T>& b) { +[[nodiscard]] constexpr Vec3<decltype(T{} * T{} - T{} * T{})> Cross(const Vec3<T>& a, + const Vec3<T>& b) { return {a.y * b.z - a.z * b.y, a.z * b.x - a.x * b.z, a.x * b.y - a.y * b.x}; } // linear interpolation via float: 0.0=begin, 1.0=end template <typename X> -constexpr decltype(X{} * float{} + X{} * float{}) Lerp(const X& begin, const X& end, - const float t) { +[[nodiscard]] constexpr decltype(X{} * float{} + X{} * float{}) Lerp(const X& begin, const X& end, + const float t) { return begin * (1.f - t) + end * t; } // linear interpolation via int: 0=begin, base=end template <typename X, int base> -constexpr decltype((X{} * int{} + X{} * int{}) / base) LerpInt(const X& begin, const X& end, - const int t) { +[[nodiscard]] constexpr decltype((X{} * int{} + X{} * int{}) / base) LerpInt(const X& begin, + const X& end, + const int t) { return (begin * (base - t) + end * t) / base; } // bilinear interpolation. s is for interpolating x00-x01 and x10-x11, and t is for the second // interpolation. template <typename X> -constexpr auto BilinearInterp(const X& x00, const X& x01, const X& x10, const X& x11, const float s, - const float t) { +[[nodiscard]] constexpr auto BilinearInterp(const X& x00, const X& x01, const X& x10, const X& x11, + const float s, const float t) { auto y0 = Lerp(x00, x01, s); auto y1 = Lerp(x10, x11, s); return Lerp(y0, y1, t); @@ -633,42 +692,42 @@ constexpr auto BilinearInterp(const X& x00, const X& x01, const X& x10, const X& // Utility vector factories template <typename T> -constexpr Vec2<T> MakeVec(const T& x, const T& y) { +[[nodiscard]] constexpr Vec2<T> MakeVec(const T& x, const T& y) { return Vec2<T>{x, y}; } template <typename T> -constexpr Vec3<T> MakeVec(const T& x, const T& y, const T& z) { +[[nodiscard]] constexpr Vec3<T> MakeVec(const T& x, const T& y, const T& z) { return Vec3<T>{x, y, z}; } template <typename T> -constexpr Vec4<T> MakeVec(const T& x, const T& y, const Vec2<T>& zw) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const T& x, const T& y, const Vec2<T>& zw) { return MakeVec(x, y, zw[0], zw[1]); } template <typename T> -constexpr Vec3<T> MakeVec(const Vec2<T>& xy, const T& z) { +[[nodiscard]] constexpr Vec3<T> MakeVec(const Vec2<T>& xy, const T& z) { return MakeVec(xy[0], xy[1], z); } template <typename T> -constexpr Vec3<T> MakeVec(const T& x, const Vec2<T>& yz) { +[[nodiscard]] constexpr Vec3<T> MakeVec(const T& x, const Vec2<T>& yz) { return MakeVec(x, yz[0], yz[1]); } template <typename T> -constexpr Vec4<T> MakeVec(const T& x, const T& y, const T& z, const T& w) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const T& x, const T& y, const T& z, const T& w) { return Vec4<T>{x, y, z, w}; } template <typename T> -constexpr Vec4<T> MakeVec(const Vec2<T>& xy, const T& z, const T& w) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const Vec2<T>& xy, const T& z, const T& w) { return MakeVec(xy[0], xy[1], z, w); } template <typename T> -constexpr Vec4<T> MakeVec(const T& x, const Vec2<T>& yz, const T& w) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const T& x, const Vec2<T>& yz, const T& w) { return MakeVec(x, yz[0], yz[1], w); } @@ -676,17 +735,17 @@ constexpr Vec4<T> MakeVec(const T& x, const Vec2<T>& yz, const T& w) { // Even if someone wanted to use an odd object like Vec2<Vec2<T>>, the compiler would error // out soon enough due to misuse of the returned structure. template <typename T> -constexpr Vec4<T> MakeVec(const Vec2<T>& xy, const Vec2<T>& zw) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const Vec2<T>& xy, const Vec2<T>& zw) { return MakeVec(xy[0], xy[1], zw[0], zw[1]); } template <typename T> -constexpr Vec4<T> MakeVec(const Vec3<T>& xyz, const T& w) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const Vec3<T>& xyz, const T& w) { return MakeVec(xyz[0], xyz[1], xyz[2], w); } template <typename T> -constexpr Vec4<T> MakeVec(const T& x, const Vec3<T>& yzw) { +[[nodiscard]] constexpr Vec4<T> MakeVec(const T& x, const Vec3<T>& yzw) { return MakeVec(x, yzw[0], yzw[1], yzw[2]); } diff --git a/src/common/virtual_buffer.cpp b/src/common/virtual_buffer.cpp new file mode 100644 index 000000000..e3ca29258 --- /dev/null +++ b/src/common/virtual_buffer.cpp @@ -0,0 +1,43 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#ifdef _WIN32 +#include <windows.h> +#else +#include <sys/mman.h> +#endif + +#include "common/assert.h" +#include "common/virtual_buffer.h" + +namespace Common { + +void* AllocateMemoryPages(std::size_t size) noexcept { +#ifdef _WIN32 + void* base{VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_READWRITE)}; +#else + void* base{mmap(nullptr, size, PROT_READ | PROT_WRITE, MAP_ANON | MAP_PRIVATE, -1, 0)}; + + if (base == MAP_FAILED) { + base = nullptr; + } +#endif + + ASSERT(base); + + return base; +} + +void FreeMemoryPages(void* base, [[maybe_unused]] std::size_t size) noexcept { + if (!base) { + return; + } +#ifdef _WIN32 + ASSERT(VirtualFree(base, 0, MEM_RELEASE)); +#else + ASSERT(munmap(base, size) == 0); +#endif +} + +} // namespace Common diff --git a/src/common/virtual_buffer.h b/src/common/virtual_buffer.h new file mode 100644 index 000000000..91d430036 --- /dev/null +++ b/src/common/virtual_buffer.h @@ -0,0 +1,82 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <type_traits> +#include <utility> + +namespace Common { + +void* AllocateMemoryPages(std::size_t size) noexcept; +void FreeMemoryPages(void* base, std::size_t size) noexcept; + +template <typename T> +class VirtualBuffer final { +public: + static_assert( + std::is_trivially_constructible_v<T>, + "T must be trivially constructible, as non-trivial constructors will not be executed " + "with the current allocator"); + + constexpr VirtualBuffer() = default; + explicit VirtualBuffer(std::size_t count) : alloc_size{count * sizeof(T)} { + base_ptr = reinterpret_cast<T*>(AllocateMemoryPages(alloc_size)); + } + + ~VirtualBuffer() noexcept { + FreeMemoryPages(base_ptr, alloc_size); + } + + VirtualBuffer(const VirtualBuffer&) = delete; + VirtualBuffer& operator=(const VirtualBuffer&) = delete; + + VirtualBuffer(VirtualBuffer&& other) noexcept + : alloc_size{std::exchange(other.alloc_size, 0)}, base_ptr{std::exchange(other.base_ptr), + nullptr} {} + + VirtualBuffer& operator=(VirtualBuffer&& other) noexcept { + alloc_size = std::exchange(other.alloc_size, 0); + base_ptr = std::exchange(other.base_ptr, nullptr); + return *this; + } + + void resize(std::size_t count) { + const auto new_size = count * sizeof(T); + if (new_size == alloc_size) { + return; + } + + FreeMemoryPages(base_ptr, alloc_size); + + alloc_size = new_size; + base_ptr = reinterpret_cast<T*>(AllocateMemoryPages(alloc_size)); + } + + [[nodiscard]] constexpr const T& operator[](std::size_t index) const { + return base_ptr[index]; + } + + [[nodiscard]] constexpr T& operator[](std::size_t index) { + return base_ptr[index]; + } + + [[nodiscard]] constexpr T* data() { + return base_ptr; + } + + [[nodiscard]] constexpr const T* data() const { + return base_ptr; + } + + [[nodiscard]] constexpr std::size_t size() const { + return alloc_size / sizeof(T); + } + +private: + std::size_t alloc_size{}; + T* base_ptr{}; +}; + +} // namespace Common diff --git a/src/common/wall_clock.cpp b/src/common/wall_clock.cpp new file mode 100644 index 000000000..452a2837e --- /dev/null +++ b/src/common/wall_clock.cpp @@ -0,0 +1,91 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/uint128.h" +#include "common/wall_clock.h" + +#ifdef ARCHITECTURE_x86_64 +#include "common/x64/cpu_detect.h" +#include "common/x64/native_clock.h" +#endif + +namespace Common { + +using base_timer = std::chrono::steady_clock; +using base_time_point = std::chrono::time_point<base_timer>; + +class StandardWallClock final : public WallClock { +public: + StandardWallClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency) + : WallClock(emulated_cpu_frequency, emulated_clock_frequency, false) { + start_time = base_timer::now(); + } + + std::chrono::nanoseconds GetTimeNS() override { + base_time_point current = base_timer::now(); + auto elapsed = current - start_time; + return std::chrono::duration_cast<std::chrono::nanoseconds>(elapsed); + } + + std::chrono::microseconds GetTimeUS() override { + base_time_point current = base_timer::now(); + auto elapsed = current - start_time; + return std::chrono::duration_cast<std::chrono::microseconds>(elapsed); + } + + std::chrono::milliseconds GetTimeMS() override { + base_time_point current = base_timer::now(); + auto elapsed = current - start_time; + return std::chrono::duration_cast<std::chrono::milliseconds>(elapsed); + } + + u64 GetClockCycles() override { + std::chrono::nanoseconds time_now = GetTimeNS(); + const u128 temporary = + Common::Multiply64Into128(time_now.count(), emulated_clock_frequency); + return Common::Divide128On32(temporary, 1000000000).first; + } + + u64 GetCPUCycles() override { + std::chrono::nanoseconds time_now = GetTimeNS(); + const u128 temporary = Common::Multiply64Into128(time_now.count(), emulated_cpu_frequency); + return Common::Divide128On32(temporary, 1000000000).first; + } + + void Pause([[maybe_unused]] bool is_paused) override { + // Do nothing in this clock type. + } + +private: + base_time_point start_time; +}; + +#ifdef ARCHITECTURE_x86_64 + +std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency, + u32 emulated_clock_frequency) { + const auto& caps = GetCPUCaps(); + u64 rtsc_frequency = 0; + if (caps.invariant_tsc) { + rtsc_frequency = EstimateRDTSCFrequency(); + } + if (rtsc_frequency == 0) { + return std::make_unique<StandardWallClock>(emulated_cpu_frequency, + emulated_clock_frequency); + } else { + return std::make_unique<X64::NativeClock>(emulated_cpu_frequency, emulated_clock_frequency, + rtsc_frequency); + } +} + +#else + +std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency, + u32 emulated_clock_frequency) { + return std::make_unique<StandardWallClock>(emulated_cpu_frequency, emulated_clock_frequency); +} + +#endif + +} // namespace Common diff --git a/src/common/wall_clock.h b/src/common/wall_clock.h new file mode 100644 index 000000000..bc7adfbf8 --- /dev/null +++ b/src/common/wall_clock.h @@ -0,0 +1,55 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <chrono> +#include <memory> + +#include "common/common_types.h" + +namespace Common { + +class WallClock { +public: + virtual ~WallClock() = default; + + /// Returns current wall time in nanoseconds + [[nodiscard]] virtual std::chrono::nanoseconds GetTimeNS() = 0; + + /// Returns current wall time in microseconds + [[nodiscard]] virtual std::chrono::microseconds GetTimeUS() = 0; + + /// Returns current wall time in milliseconds + [[nodiscard]] virtual std::chrono::milliseconds GetTimeMS() = 0; + + /// Returns current wall time in emulated clock cycles + [[nodiscard]] virtual u64 GetClockCycles() = 0; + + /// Returns current wall time in emulated cpu cycles + [[nodiscard]] virtual u64 GetCPUCycles() = 0; + + virtual void Pause(bool is_paused) = 0; + + /// Tells if the wall clock, uses the host CPU's hardware clock + [[nodiscard]] bool IsNative() const { + return is_native; + } + +protected: + WallClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency, bool is_native) + : emulated_cpu_frequency{emulated_cpu_frequency}, + emulated_clock_frequency{emulated_clock_frequency}, is_native{is_native} {} + + u64 emulated_cpu_frequency; + u64 emulated_clock_frequency; + +private: + bool is_native; +}; + +[[nodiscard]] std::unique_ptr<WallClock> CreateBestMatchingClock(u32 emulated_cpu_frequency, + u32 emulated_clock_frequency); + +} // namespace Common diff --git a/src/common/x64/cpu_detect.cpp b/src/common/x64/cpu_detect.cpp index c9349a6b4..fccd2eee5 100644 --- a/src/common/x64/cpu_detect.cpp +++ b/src/common/x64/cpu_detect.cpp @@ -62,6 +62,17 @@ static CPUCaps Detect() { std::memcpy(&caps.brand_string[0], &cpu_id[1], sizeof(int)); std::memcpy(&caps.brand_string[4], &cpu_id[3], sizeof(int)); std::memcpy(&caps.brand_string[8], &cpu_id[2], sizeof(int)); + if (cpu_id[1] == 0x756e6547 && cpu_id[2] == 0x6c65746e && cpu_id[3] == 0x49656e69) + caps.manufacturer = Manufacturer::Intel; + else if (cpu_id[1] == 0x68747541 && cpu_id[2] == 0x444d4163 && cpu_id[3] == 0x69746e65) + caps.manufacturer = Manufacturer::AMD; + else if (cpu_id[1] == 0x6f677948 && cpu_id[2] == 0x656e6975 && cpu_id[3] == 0x6e65476e) + caps.manufacturer = Manufacturer::Hygon; + else + caps.manufacturer = Manufacturer::Unknown; + + u32 family = {}; + u32 model = {}; __cpuid(cpu_id, 0x80000000); @@ -73,6 +84,14 @@ static CPUCaps Detect() { // Detect family and other miscellaneous features if (max_std_fn >= 1) { __cpuid(cpu_id, 0x00000001); + family = (cpu_id[0] >> 8) & 0xf; + model = (cpu_id[0] >> 4) & 0xf; + if (family == 0xf) { + family += (cpu_id[0] >> 20) & 0xff; + } + if (family >= 6) { + model += ((cpu_id[0] >> 16) & 0xf) << 4; + } if ((cpu_id[3] >> 25) & 1) caps.sse = true; @@ -110,6 +129,11 @@ static CPUCaps Detect() { caps.bmi1 = true; if ((cpu_id[1] >> 8) & 1) caps.bmi2 = true; + // Checks for AVX512F, AVX512CD, AVX512VL, AVX512DQ, AVX512BW (Intel Skylake-X/SP) + if ((cpu_id[1] >> 16) & 1 && (cpu_id[1] >> 28) & 1 && (cpu_id[1] >> 31) & 1 && + (cpu_id[1] >> 17) & 1 && (cpu_id[1] >> 30) & 1) { + caps.avx512 = caps.avx2; + } } } @@ -130,6 +154,20 @@ static CPUCaps Detect() { caps.fma4 = true; } + if (max_ex_fn >= 0x80000007) { + __cpuid(cpu_id, 0x80000007); + if (cpu_id[3] & (1 << 8)) { + caps.invariant_tsc = true; + } + } + + if (max_std_fn >= 0x16) { + __cpuid(cpu_id, 0x16); + caps.base_frequency = cpu_id[0]; + caps.max_frequency = cpu_id[1]; + caps.bus_frequency = cpu_id[2]; + } + return caps; } diff --git a/src/common/x64/cpu_detect.h b/src/common/x64/cpu_detect.h index 20f2ba234..e3b63302e 100644 --- a/src/common/x64/cpu_detect.h +++ b/src/common/x64/cpu_detect.h @@ -6,8 +6,16 @@ namespace Common { +enum class Manufacturer : u32 { + Intel = 0, + AMD = 1, + Hygon = 2, + Unknown = 3, +}; + /// x86/x64 CPU capabilities that may be detected by this module struct CPUCaps { + Manufacturer manufacturer; char cpu_string[0x21]; char brand_string[0x41]; bool sse; @@ -19,11 +27,16 @@ struct CPUCaps { bool lzcnt; bool avx; bool avx2; + bool avx512; bool bmi1; bool bmi2; bool fma; bool fma4; bool aes; + bool invariant_tsc; + u32 base_frequency; + u32 max_frequency; + u32 bus_frequency; }; /** diff --git a/src/common/x64/native_clock.cpp b/src/common/x64/native_clock.cpp new file mode 100644 index 000000000..424b39b1f --- /dev/null +++ b/src/common/x64/native_clock.cpp @@ -0,0 +1,103 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <chrono> +#include <mutex> +#include <thread> + +#ifdef _MSC_VER +#include <intrin.h> +#else +#include <x86intrin.h> +#endif + +#include "common/uint128.h" +#include "common/x64/native_clock.h" + +namespace Common { + +u64 EstimateRDTSCFrequency() { + const auto milli_10 = std::chrono::milliseconds{10}; + // get current time + _mm_mfence(); + const u64 tscStart = __rdtsc(); + const auto startTime = std::chrono::high_resolution_clock::now(); + // wait roughly 3 seconds + while (true) { + auto milli = std::chrono::duration_cast<std::chrono::milliseconds>( + std::chrono::high_resolution_clock::now() - startTime); + if (milli.count() >= 3000) + break; + std::this_thread::sleep_for(milli_10); + } + const auto endTime = std::chrono::high_resolution_clock::now(); + _mm_mfence(); + const u64 tscEnd = __rdtsc(); + // calculate difference + const u64 timer_diff = + std::chrono::duration_cast<std::chrono::nanoseconds>(endTime - startTime).count(); + const u64 tsc_diff = tscEnd - tscStart; + const u64 tsc_freq = MultiplyAndDivide64(tsc_diff, 1000000000ULL, timer_diff); + return tsc_freq; +} + +namespace X64 { +NativeClock::NativeClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency, + u64 rtsc_frequency) + : WallClock(emulated_cpu_frequency, emulated_clock_frequency, true), rtsc_frequency{ + rtsc_frequency} { + _mm_mfence(); + last_measure = __rdtsc(); + accumulated_ticks = 0U; +} + +u64 NativeClock::GetRTSC() { + std::scoped_lock scope{rtsc_serialize}; + _mm_mfence(); + const u64 current_measure = __rdtsc(); + u64 diff = current_measure - last_measure; + diff = diff & ~static_cast<u64>(static_cast<s64>(diff) >> 63); // max(diff, 0) + if (current_measure > last_measure) { + last_measure = current_measure; + } + accumulated_ticks += diff; + /// The clock cannot be more precise than the guest timer, remove the lower bits + return accumulated_ticks & inaccuracy_mask; +} + +void NativeClock::Pause(bool is_paused) { + if (!is_paused) { + _mm_mfence(); + last_measure = __rdtsc(); + } +} + +std::chrono::nanoseconds NativeClock::GetTimeNS() { + const u64 rtsc_value = GetRTSC(); + return std::chrono::nanoseconds{MultiplyAndDivide64(rtsc_value, 1000000000, rtsc_frequency)}; +} + +std::chrono::microseconds NativeClock::GetTimeUS() { + const u64 rtsc_value = GetRTSC(); + return std::chrono::microseconds{MultiplyAndDivide64(rtsc_value, 1000000, rtsc_frequency)}; +} + +std::chrono::milliseconds NativeClock::GetTimeMS() { + const u64 rtsc_value = GetRTSC(); + return std::chrono::milliseconds{MultiplyAndDivide64(rtsc_value, 1000, rtsc_frequency)}; +} + +u64 NativeClock::GetClockCycles() { + const u64 rtsc_value = GetRTSC(); + return MultiplyAndDivide64(rtsc_value, emulated_clock_frequency, rtsc_frequency); +} + +u64 NativeClock::GetCPUCycles() { + const u64 rtsc_value = GetRTSC(); + return MultiplyAndDivide64(rtsc_value, emulated_cpu_frequency, rtsc_frequency); +} + +} // namespace X64 + +} // namespace Common diff --git a/src/common/x64/native_clock.h b/src/common/x64/native_clock.h new file mode 100644 index 000000000..97aab6ac9 --- /dev/null +++ b/src/common/x64/native_clock.h @@ -0,0 +1,48 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <optional> + +#include "common/spin_lock.h" +#include "common/wall_clock.h" + +namespace Common { + +namespace X64 { +class NativeClock final : public WallClock { +public: + NativeClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency, u64 rtsc_frequency); + + std::chrono::nanoseconds GetTimeNS() override; + + std::chrono::microseconds GetTimeUS() override; + + std::chrono::milliseconds GetTimeMS() override; + + u64 GetClockCycles() override; + + u64 GetCPUCycles() override; + + void Pause(bool is_paused) override; + +private: + u64 GetRTSC(); + + /// value used to reduce the native clocks accuracy as some apss rely on + /// undefined behavior where the level of accuracy in the clock shouldn't + /// be higher. + static constexpr u64 inaccuracy_mask = ~(UINT64_C(0x400) - 1); + + SpinLock rtsc_serialize{}; + u64 last_measure{}; + u64 accumulated_ticks{}; + u64 rtsc_frequency; +}; +} // namespace X64 + +u64 EstimateRDTSCFrequency(); + +} // namespace Common diff --git a/src/common/x64/xbyak_abi.h b/src/common/x64/xbyak_abi.h new file mode 100644 index 000000000..26e4bfda5 --- /dev/null +++ b/src/common/x64/xbyak_abi.h @@ -0,0 +1,229 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <bitset> +#include <initializer_list> +#include <xbyak.h> +#include "common/assert.h" + +namespace Common::X64 { + +constexpr std::size_t RegToIndex(const Xbyak::Reg& reg) { + using Kind = Xbyak::Reg::Kind; + ASSERT_MSG((reg.getKind() & (Kind::REG | Kind::XMM)) != 0, + "RegSet only support GPRs and XMM registers."); + ASSERT_MSG(reg.getIdx() < 16, "RegSet only supports XXM0-15."); + return reg.getIdx() + (reg.getKind() == Kind::REG ? 0 : 16); +} + +constexpr Xbyak::Reg64 IndexToReg64(std::size_t reg_index) { + ASSERT(reg_index < 16); + return Xbyak::Reg64(static_cast<int>(reg_index)); +} + +constexpr Xbyak::Xmm IndexToXmm(std::size_t reg_index) { + ASSERT(reg_index >= 16 && reg_index < 32); + return Xbyak::Xmm(static_cast<int>(reg_index - 16)); +} + +constexpr Xbyak::Reg IndexToReg(std::size_t reg_index) { + if (reg_index < 16) { + return IndexToReg64(reg_index); + } else { + return IndexToXmm(reg_index); + } +} + +inline std::bitset<32> BuildRegSet(std::initializer_list<Xbyak::Reg> regs) { + std::bitset<32> bits; + for (const Xbyak::Reg& reg : regs) { + bits[RegToIndex(reg)] = true; + } + return bits; +} + +constexpr inline std::bitset<32> ABI_ALL_GPRS(0x0000FFFF); +constexpr inline std::bitset<32> ABI_ALL_XMMS(0xFFFF0000); + +#ifdef _WIN32 + +// Microsoft x64 ABI +constexpr inline Xbyak::Reg ABI_RETURN = Xbyak::util::rax; +constexpr inline Xbyak::Reg ABI_PARAM1 = Xbyak::util::rcx; +constexpr inline Xbyak::Reg ABI_PARAM2 = Xbyak::util::rdx; +constexpr inline Xbyak::Reg ABI_PARAM3 = Xbyak::util::r8; +constexpr inline Xbyak::Reg ABI_PARAM4 = Xbyak::util::r9; + +const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({ + // GPRs + Xbyak::util::rcx, + Xbyak::util::rdx, + Xbyak::util::r8, + Xbyak::util::r9, + Xbyak::util::r10, + Xbyak::util::r11, + // XMMs + Xbyak::util::xmm0, + Xbyak::util::xmm1, + Xbyak::util::xmm2, + Xbyak::util::xmm3, + Xbyak::util::xmm4, + Xbyak::util::xmm5, +}); + +const std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({ + // GPRs + Xbyak::util::rbx, + Xbyak::util::rsi, + Xbyak::util::rdi, + Xbyak::util::rbp, + Xbyak::util::r12, + Xbyak::util::r13, + Xbyak::util::r14, + Xbyak::util::r15, + // XMMs + Xbyak::util::xmm6, + Xbyak::util::xmm7, + Xbyak::util::xmm8, + Xbyak::util::xmm9, + Xbyak::util::xmm10, + Xbyak::util::xmm11, + Xbyak::util::xmm12, + Xbyak::util::xmm13, + Xbyak::util::xmm14, + Xbyak::util::xmm15, +}); + +constexpr size_t ABI_SHADOW_SPACE = 0x20; + +#else + +// System V x86-64 ABI +constexpr inline Xbyak::Reg ABI_RETURN = Xbyak::util::rax; +constexpr inline Xbyak::Reg ABI_PARAM1 = Xbyak::util::rdi; +constexpr inline Xbyak::Reg ABI_PARAM2 = Xbyak::util::rsi; +constexpr inline Xbyak::Reg ABI_PARAM3 = Xbyak::util::rdx; +constexpr inline Xbyak::Reg ABI_PARAM4 = Xbyak::util::rcx; + +const std::bitset<32> ABI_ALL_CALLER_SAVED = BuildRegSet({ + // GPRs + Xbyak::util::rcx, + Xbyak::util::rdx, + Xbyak::util::rdi, + Xbyak::util::rsi, + Xbyak::util::r8, + Xbyak::util::r9, + Xbyak::util::r10, + Xbyak::util::r11, + // XMMs + Xbyak::util::xmm0, + Xbyak::util::xmm1, + Xbyak::util::xmm2, + Xbyak::util::xmm3, + Xbyak::util::xmm4, + Xbyak::util::xmm5, + Xbyak::util::xmm6, + Xbyak::util::xmm7, + Xbyak::util::xmm8, + Xbyak::util::xmm9, + Xbyak::util::xmm10, + Xbyak::util::xmm11, + Xbyak::util::xmm12, + Xbyak::util::xmm13, + Xbyak::util::xmm14, + Xbyak::util::xmm15, +}); + +const std::bitset<32> ABI_ALL_CALLEE_SAVED = BuildRegSet({ + // GPRs + Xbyak::util::rbx, + Xbyak::util::rbp, + Xbyak::util::r12, + Xbyak::util::r13, + Xbyak::util::r14, + Xbyak::util::r15, +}); + +constexpr size_t ABI_SHADOW_SPACE = 0; + +#endif + +struct ABIFrameInfo { + s32 subtraction; + s32 xmm_offset; +}; + +inline ABIFrameInfo ABI_CalculateFrameSize(std::bitset<32> regs, size_t rsp_alignment, + size_t needed_frame_size) { + const auto count = (regs & ABI_ALL_GPRS).count(); + rsp_alignment -= count * 8; + size_t subtraction = 0; + const auto xmm_count = (regs & ABI_ALL_XMMS).count(); + if (xmm_count) { + // If we have any XMMs to save, we must align the stack here. + subtraction = rsp_alignment & 0xF; + } + subtraction += 0x10 * xmm_count; + size_t xmm_base_subtraction = subtraction; + subtraction += needed_frame_size; + subtraction += ABI_SHADOW_SPACE; + // Final alignment. + rsp_alignment -= subtraction; + subtraction += rsp_alignment & 0xF; + + return ABIFrameInfo{static_cast<s32>(subtraction), + static_cast<s32>(subtraction - xmm_base_subtraction)}; +} + +inline size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, std::bitset<32> regs, + size_t rsp_alignment, size_t needed_frame_size = 0) { + auto frame_info = ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size); + + for (std::size_t i = 0; i < regs.size(); ++i) { + if (regs[i] && ABI_ALL_GPRS[i]) { + code.push(IndexToReg64(i)); + } + } + + if (frame_info.subtraction != 0) { + code.sub(code.rsp, frame_info.subtraction); + } + + for (std::size_t i = 0; i < regs.size(); ++i) { + if (regs[i] && ABI_ALL_XMMS[i]) { + code.movaps(code.xword[code.rsp + frame_info.xmm_offset], IndexToXmm(i)); + frame_info.xmm_offset += 0x10; + } + } + + return ABI_SHADOW_SPACE; +} + +inline void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, std::bitset<32> regs, + size_t rsp_alignment, size_t needed_frame_size = 0) { + auto frame_info = ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size); + + for (std::size_t i = 0; i < regs.size(); ++i) { + if (regs[i] && ABI_ALL_XMMS[i]) { + code.movaps(IndexToXmm(i), code.xword[code.rsp + frame_info.xmm_offset]); + frame_info.xmm_offset += 0x10; + } + } + + if (frame_info.subtraction != 0) { + code.add(code.rsp, frame_info.subtraction); + } + + // GPRs need to be popped in reverse order + for (std::size_t j = 0; j < regs.size(); ++j) { + const std::size_t i = regs.size() - j - 1; + if (regs[i] && ABI_ALL_GPRS[i]) { + code.pop(IndexToReg64(i)); + } + } +} + +} // namespace Common::X64 diff --git a/src/common/x64/xbyak_util.h b/src/common/x64/xbyak_util.h new file mode 100644 index 000000000..df17f8cbe --- /dev/null +++ b/src/common/x64/xbyak_util.h @@ -0,0 +1,47 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <type_traits> +#include <xbyak.h> +#include "common/x64/xbyak_abi.h" + +namespace Common::X64 { + +// Constants for use with cmpps/cmpss +enum { + CMP_EQ = 0, + CMP_LT = 1, + CMP_LE = 2, + CMP_UNORD = 3, + CMP_NEQ = 4, + CMP_NLT = 5, + CMP_NLE = 6, + CMP_ORD = 7, +}; + +constexpr bool IsWithin2G(uintptr_t ref, uintptr_t target) { + const u64 distance = target - (ref + 5); + return !(distance >= 0x8000'0000ULL && distance <= ~0x8000'0000ULL); +} + +inline bool IsWithin2G(const Xbyak::CodeGenerator& code, uintptr_t target) { + return IsWithin2G(reinterpret_cast<uintptr_t>(code.getCurr()), target); +} + +template <typename T> +inline void CallFarFunction(Xbyak::CodeGenerator& code, const T f) { + static_assert(std::is_pointer_v<T>, "Argument must be a (function) pointer."); + size_t addr = reinterpret_cast<size_t>(f); + if (IsWithin2G(code, addr)) { + code.call(f); + } else { + // ABI_RETURN is a safe temp register to use before a call + code.mov(ABI_RETURN, addr); + code.call(ABI_RETURN); + } +} + +} // namespace Common::X64 diff --git a/src/common/zstd_compression.cpp b/src/common/zstd_compression.cpp index 978526492..5f45459da 100644 --- a/src/common/zstd_compression.cpp +++ b/src/common/zstd_compression.cpp @@ -5,7 +5,6 @@ #include <algorithm> #include <zstd.h> -#include "common/assert.h" #include "common/zstd_compression.h" namespace Common::Compression { diff --git a/src/common/zstd_compression.h b/src/common/zstd_compression.h index e9de941c8..c26a30ab9 100644 --- a/src/common/zstd_compression.h +++ b/src/common/zstd_compression.h @@ -13,24 +13,25 @@ namespace Common::Compression { /** * Compresses a source memory region with Zstandard and returns the compressed data in a vector. * - * @param source the uncompressed source memory region. - * @param source_size the size in bytes of the uncompressed source memory region. - * @param compression_level the used compression level. Should be between 1 and 22. + * @param source The uncompressed source memory region. + * @param source_size The size of the uncompressed source memory region. + * @param compression_level The used compression level. Should be between 1 and 22. * * @return the compressed data. */ -std::vector<u8> CompressDataZSTD(const u8* source, std::size_t source_size, s32 compression_level); +[[nodiscard]] std::vector<u8> CompressDataZSTD(const u8* source, std::size_t source_size, + s32 compression_level); /** * Compresses a source memory region with Zstandard with the default compression level and returns * the compressed data in a vector. * - * @param source the uncompressed source memory region. - * @param source_size the size in bytes of the uncompressed source memory region. + * @param source The uncompressed source memory region. + * @param source_size The size of the uncompressed source memory region. * * @return the compressed data. */ -std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_size); +[[nodiscard]] std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_size); /** * Decompresses a source memory region with Zstandard and returns the uncompressed data in a vector. @@ -39,6 +40,6 @@ std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_siz * * @return the decompressed data. */ -std::vector<u8> DecompressDataZSTD(const std::vector<u8>& compressed); +[[nodiscard]] std::vector<u8> DecompressDataZSTD(const std::vector<u8>& compressed); } // namespace Common::Compression
\ No newline at end of file diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index 66497a386..e370fd225 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -1,22 +1,22 @@ -if (YUZU_ENABLE_BOXCAT) - set(BCAT_BOXCAT_ADDITIONAL_SOURCES hle/service/bcat/backend/boxcat.cpp hle/service/bcat/backend/boxcat.h) -else() - set(BCAT_BOXCAT_ADDITIONAL_SOURCES) -endif() - add_library(core STATIC arm/arm_interface.h arm/arm_interface.cpp + arm/cpu_interrupt_handler.cpp + arm/cpu_interrupt_handler.h + arm/dynarmic/arm_dynarmic_32.cpp + arm/dynarmic/arm_dynarmic_32.h + arm/dynarmic/arm_dynarmic_64.cpp + arm/dynarmic/arm_dynarmic_64.h + arm/dynarmic/arm_dynarmic_cp15.cpp + arm/dynarmic/arm_dynarmic_cp15.h + arm/dynarmic/arm_exclusive_monitor.cpp + arm/dynarmic/arm_exclusive_monitor.h arm/exclusive_monitor.cpp arm/exclusive_monitor.h - arm/unicorn/arm_unicorn.cpp - arm/unicorn/arm_unicorn.h constants.cpp constants.h core.cpp core.h - core_manager.cpp - core_manager.h core_timing.cpp core_timing.h core_timing_util.cpp @@ -35,6 +35,8 @@ add_library(core STATIC crypto/ctr_encryption_layer.h crypto/xts_encryption_layer.cpp crypto/xts_encryption_layer.h + device_memory.cpp + device_memory.h file_sys/bis_factory.cpp file_sys/bis_factory.h file_sys/card_image.cpp @@ -116,6 +118,8 @@ add_library(core STATIC file_sys/vfs_vector.h file_sys/xts_archive.cpp file_sys/xts_archive.h + frontend/applets/controller.cpp + frontend/applets/controller.h frontend/applets/error.cpp frontend/applets/error.h frontend/applets/general_frontend.cpp @@ -152,12 +156,30 @@ add_library(core STATIC hle/kernel/hle_ipc.h hle/kernel/kernel.cpp hle/kernel/kernel.h + hle/kernel/memory/address_space_info.cpp + hle/kernel/memory/address_space_info.h + hle/kernel/memory/memory_block.h + hle/kernel/memory/memory_block_manager.cpp + hle/kernel/memory/memory_block_manager.h + hle/kernel/memory/memory_layout.h + hle/kernel/memory/memory_manager.cpp + hle/kernel/memory/memory_manager.h + hle/kernel/memory/memory_types.h + hle/kernel/memory/page_linked_list.h + hle/kernel/memory/page_heap.cpp + hle/kernel/memory/page_heap.h + hle/kernel/memory/page_table.cpp + hle/kernel/memory/page_table.h + hle/kernel/memory/slab_heap.h + hle/kernel/memory/system_control.cpp + hle/kernel/memory/system_control.h hle/kernel/mutex.cpp hle/kernel/mutex.h hle/kernel/object.cpp hle/kernel/object.h hle/kernel/physical_core.cpp hle/kernel/physical_core.h + hle/kernel/physical_memory.h hle/kernel/process.cpp hle/kernel/process.h hle/kernel/process_capability.cpp @@ -178,6 +200,7 @@ add_library(core STATIC hle/kernel/shared_memory.h hle/kernel/svc.cpp hle/kernel/svc.h + hle/kernel/svc_types.h hle/kernel/svc_wrap.h hle/kernel/synchronization_object.cpp hle/kernel/synchronization_object.h @@ -189,8 +212,6 @@ add_library(core STATIC hle/kernel/time_manager.h hle/kernel/transfer_memory.cpp hle/kernel/transfer_memory.h - hle/kernel/vm_manager.cpp - hle/kernel/vm_manager.h hle/kernel/writable_event.cpp hle/kernel/writable_event.h hle/lock.cpp @@ -217,6 +238,8 @@ add_library(core STATIC hle/service/am/applet_oe.h hle/service/am/applets/applets.cpp hle/service/am/applets/applets.h + hle/service/am/applets/controller.cpp + hle/service/am/applets/controller.h hle/service/am/applets/error.cpp hle/service/am/applets/error.h hle/service/am/applets/general_backend.cpp @@ -272,7 +295,6 @@ add_library(core STATIC hle/service/audio/hwopus.h hle/service/bcat/backend/backend.cpp hle/service/bcat/backend/backend.h - ${BCAT_BOXCAT_ADDITIONAL_SOURCES} hle/service/bcat/bcat.cpp hle/service/bcat/bcat.h hle/service/bcat/module.cpp @@ -372,10 +394,13 @@ add_library(core STATIC hle/service/lm/manager.h hle/service/mig/mig.cpp hle/service/mig/mig.h + hle/service/mii/manager.cpp + hle/service/mii/manager.h hle/service/mii/mii.cpp hle/service/mii/mii.h - hle/service/mii/mii_manager.cpp - hle/service/mii/mii_manager.h + hle/service/mii/raw_data.cpp + hle/service/mii/raw_data.h + hle/service/mii/types.h hle/service/mm/mm_u.cpp hle/service/mm/mm_u.h hle/service/ncm/ncm.cpp @@ -412,6 +437,8 @@ add_library(core STATIC hle/service/nvdrv/devices/nvhost_gpu.h hle/service/nvdrv/devices/nvhost_nvdec.cpp hle/service/nvdrv/devices/nvhost_nvdec.h + hle/service/nvdrv/devices/nvhost_nvdec_common.cpp + hle/service/nvdrv/devices/nvhost_nvdec_common.h hle/service/nvdrv/devices/nvhost_nvjpg.cpp hle/service/nvdrv/devices/nvhost_nvjpg.h hle/service/nvdrv/devices/nvhost_vic.cpp @@ -425,10 +452,14 @@ add_library(core STATIC hle/service/nvdrv/nvdrv.h hle/service/nvdrv/nvmemp.cpp hle/service/nvdrv/nvmemp.h + hle/service/nvdrv/syncpoint_manager.cpp + hle/service/nvdrv/syncpoint_manager.h hle/service/nvflinger/buffer_queue.cpp hle/service/nvflinger/buffer_queue.h hle/service/nvflinger/nvflinger.cpp hle/service/nvflinger/nvflinger.h + hle/service/olsc/olsc.cpp + hle/service/olsc/olsc.h hle/service/pcie/pcie.cpp hle/service/pcie/pcie.h hle/service/pctl/module.cpp @@ -461,6 +492,7 @@ add_library(core STATIC hle/service/sm/controller.h hle/service/sm/sm.cpp hle/service/sm/sm.h + hle/service/sockets/blocking_worker.h hle/service/sockets/bsd.cpp hle/service/sockets/bsd.h hle/service/sockets/ethc.cpp @@ -471,6 +503,8 @@ add_library(core STATIC hle/service/sockets/sfdnsres.h hle/service/sockets/sockets.cpp hle/service/sockets/sockets.h + hle/service/sockets/sockets_translate.cpp + hle/service/sockets/sockets_translate.h hle/service/spl/csrng.cpp hle/service/spl/csrng.h hle/service/spl/module.cpp @@ -556,6 +590,9 @@ add_library(core STATIC memory/dmnt_cheat_vm.h memory.cpp memory.h + network/network.cpp + network/network.h + network/sockets.h perf_stats.cpp perf_stats.h reporter.cpp @@ -568,6 +605,13 @@ add_library(core STATIC tools/freezer.h ) +if (YUZU_ENABLE_BOXCAT) + target_sources(core PRIVATE + hle/service/bcat/backend/boxcat.cpp + hle/service/bcat/backend/boxcat.h + ) +endif() + if (MSVC) target_compile_options(core PRIVATE # 'expression' : signed/unsigned mismatch @@ -583,19 +627,30 @@ if (MSVC) # 'context' : truncation from 'type1' to 'type2' /we4305 ) +else() + target_compile_options(core PRIVATE + -Werror=conversion + -Werror=ignored-qualifiers + -Werror=implicit-fallthrough + -Werror=reorder + -Werror=sign-compare + -Werror=unused-variable + + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + + -Wno-sign-conversion + ) endif() create_target_directory_groups(core) target_link_libraries(core PUBLIC common PRIVATE audio_core video_core) -target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt json-headers mbedtls opus unicorn) +target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls opus zip) if (YUZU_ENABLE_BOXCAT) - get_directory_property(OPENSSL_LIBS - DIRECTORY ${PROJECT_SOURCE_DIR}/externals/libressl - DEFINITION OPENSSL_LIBS) - target_compile_definitions(core PRIVATE -DCPPHTTPLIB_OPENSSL_SUPPORT -DYUZU_ENABLE_BOXCAT) - target_link_libraries(core PRIVATE httplib json-headers ${OPENSSL_LIBS} zip) + target_compile_definitions(core PRIVATE -DYUZU_ENABLE_BOXCAT) + target_link_libraries(core PRIVATE httplib nlohmann_json::nlohmann_json) endif() if (ENABLE_WEB_SERVICE) diff --git a/src/core/arm/arm_interface.cpp b/src/core/arm/arm_interface.cpp index 7e846ddd5..0951e1976 100644 --- a/src/core/arm/arm_interface.cpp +++ b/src/core/arm/arm_interface.cpp @@ -60,7 +60,7 @@ static_assert(sizeof(ELFSymbol) == 0x18, "ELFSymbol has incorrect size."); using Symbols = std::vector<std::pair<ELFSymbol, std::string>>; -Symbols GetSymbols(VAddr text_offset, Memory::Memory& memory) { +Symbols GetSymbols(VAddr text_offset, Core::Memory::Memory& memory) { const auto mod_offset = text_offset + memory.Read32(text_offset + 4); if (mod_offset < text_offset || (mod_offset & 0b11) != 0 || @@ -123,7 +123,7 @@ Symbols GetSymbols(VAddr text_offset, Memory::Memory& memory) { std::optional<std::string> GetSymbolName(const Symbols& symbols, VAddr func_address) { const auto iter = std::find_if(symbols.begin(), symbols.end(), [func_address](const auto& pair) { - const auto& [symbol, name] = pair; + const auto& symbol = pair.first; const auto end_address = symbol.value + symbol.size; return func_address >= symbol.value && func_address < end_address; }); @@ -139,6 +139,71 @@ std::optional<std::string> GetSymbolName(const Symbols& symbols, VAddr func_addr constexpr u64 SEGMENT_BASE = 0x7100000000ull; +std::vector<ARM_Interface::BacktraceEntry> ARM_Interface::GetBacktraceFromContext( + System& system, const ThreadContext64& ctx) { + std::vector<BacktraceEntry> out; + auto& memory = system.Memory(); + + auto fp = ctx.cpu_registers[29]; + auto lr = ctx.cpu_registers[30]; + while (true) { + out.push_back({ + .module = "", + .address = 0, + .original_address = lr, + .offset = 0, + .name = {}, + }); + + if (fp == 0) { + break; + } + + lr = memory.Read64(fp + 8) - 4; + fp = memory.Read64(fp); + } + + std::map<VAddr, std::string> modules; + auto& loader{system.GetAppLoader()}; + if (loader.ReadNSOModules(modules) != Loader::ResultStatus::Success) { + return {}; + } + + std::map<std::string, Symbols> symbols; + for (const auto& module : modules) { + symbols.insert_or_assign(module.second, GetSymbols(module.first, memory)); + } + + for (auto& entry : out) { + VAddr base = 0; + for (auto iter = modules.rbegin(); iter != modules.rend(); ++iter) { + const auto& module{*iter}; + if (entry.original_address >= module.first) { + entry.module = module.second; + base = module.first; + break; + } + } + + entry.offset = entry.original_address - base; + entry.address = SEGMENT_BASE + entry.offset; + + if (entry.module.empty()) + entry.module = "unknown"; + + const auto symbol_set = symbols.find(entry.module); + if (symbol_set != symbols.end()) { + const auto symbol = GetSymbolName(symbol_set->second, entry.offset); + if (symbol.has_value()) { + // TODO(DarkLordZach): Add demangling of symbol names. + entry.name = *symbol; + } + } + } + + return out; +} + std::vector<ARM_Interface::BacktraceEntry> ARM_Interface::GetBacktrace() const { std::vector<BacktraceEntry> out; auto& memory = system.Memory(); @@ -146,7 +211,7 @@ std::vector<ARM_Interface::BacktraceEntry> ARM_Interface::GetBacktrace() const { auto fp = GetReg(29); auto lr = GetReg(30); while (true) { - out.push_back({"", 0, lr, 0}); + out.push_back({"", 0, lr, 0, ""}); if (!fp) { break; } diff --git a/src/core/arm/arm_interface.h b/src/core/arm/arm_interface.h index 57eae839e..1f24051e4 100644 --- a/src/core/arm/arm_interface.h +++ b/src/core/arm/arm_interface.h @@ -7,6 +7,7 @@ #include <array> #include <vector> #include "common/common_types.h" +#include "core/hardware_properties.h" namespace Common { struct PageTable; @@ -18,36 +19,40 @@ enum class VMAPermission : u8; namespace Core { class System; +class CPUInterruptHandler; + +using CPUInterrupts = std::array<CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>; /// Generic ARMv8 CPU interface class ARM_Interface : NonCopyable { public: - explicit ARM_Interface(System& system_) : system{system_} {} + explicit ARM_Interface(System& system_, CPUInterrupts& interrupt_handlers, bool uses_wall_clock) + : system{system_}, interrupt_handlers{interrupt_handlers}, uses_wall_clock{ + uses_wall_clock} {} virtual ~ARM_Interface() = default; struct ThreadContext32 { - std::array<u32, 16> cpu_registers; - u32 cpsr; - std::array<u8, 4> padding; - std::array<u64, 32> fprs; - u32 fpscr; - u32 fpexc; - u32 tpidr; + std::array<u32, 16> cpu_registers{}; + std::array<u32, 64> extension_registers{}; + u32 cpsr{}; + u32 fpscr{}; + u32 fpexc{}; + u32 tpidr{}; }; // Internally within the kernel, it expects the AArch32 version of the // thread context to be 344 bytes in size. - static_assert(sizeof(ThreadContext32) == 0x158); + static_assert(sizeof(ThreadContext32) == 0x150); struct ThreadContext64 { - std::array<u64, 31> cpu_registers; - u64 sp; - u64 pc; - u32 pstate; - std::array<u8, 4> padding; - std::array<u128, 32> vector_registers; - u32 fpcr; - u32 fpsr; - u64 tpidr; + std::array<u64, 31> cpu_registers{}; + u64 sp{}; + u64 pc{}; + u32 pstate{}; + std::array<u8, 4> padding{}; + std::array<u128, 32> vector_registers{}; + u32 fpcr{}; + u32 fpsr{}; + u64 tpidr{}; }; // Internally within the kernel, it expects the AArch64 version of the // thread context to be 800 bytes in size. @@ -143,6 +148,8 @@ public: */ virtual void SetTPIDR_EL0(u64 value) = 0; + virtual void ChangeProcessorID(std::size_t new_core_id) = 0; + virtual void SaveContext(ThreadContext32& ctx) = 0; virtual void SaveContext(ThreadContext64& ctx) = 0; virtual void LoadContext(const ThreadContext32& ctx) = 0; @@ -162,6 +169,9 @@ public: std::string name; }; + static std::vector<BacktraceEntry> GetBacktraceFromContext(System& system, + const ThreadContext64& ctx); + std::vector<BacktraceEntry> GetBacktrace() const; /// fp (= r29) points to the last frame record. @@ -175,6 +185,8 @@ public: protected: /// System context that this ARM interface is running under. System& system; + CPUInterrupts& interrupt_handlers; + bool uses_wall_clock; }; } // namespace Core diff --git a/src/core/arm/cpu_interrupt_handler.cpp b/src/core/arm/cpu_interrupt_handler.cpp new file mode 100644 index 000000000..9c8898700 --- /dev/null +++ b/src/core/arm/cpu_interrupt_handler.cpp @@ -0,0 +1,25 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/thread.h" +#include "core/arm/cpu_interrupt_handler.h" + +namespace Core { + +CPUInterruptHandler::CPUInterruptHandler() : interrupt_event{std::make_unique<Common::Event>()} {} + +CPUInterruptHandler::~CPUInterruptHandler() = default; + +void CPUInterruptHandler::SetInterrupt(bool is_interrupted_) { + if (is_interrupted_) { + interrupt_event->Set(); + } + is_interrupted = is_interrupted_; +} + +void CPUInterruptHandler::AwaitInterrupt() { + interrupt_event->Wait(); +} + +} // namespace Core diff --git a/src/core/arm/cpu_interrupt_handler.h b/src/core/arm/cpu_interrupt_handler.h new file mode 100644 index 000000000..c20c280f1 --- /dev/null +++ b/src/core/arm/cpu_interrupt_handler.h @@ -0,0 +1,40 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <atomic> +#include <memory> + +namespace Common { +class Event; +} + +namespace Core { + +class CPUInterruptHandler { +public: + CPUInterruptHandler(); + ~CPUInterruptHandler(); + + CPUInterruptHandler(const CPUInterruptHandler&) = delete; + CPUInterruptHandler& operator=(const CPUInterruptHandler&) = delete; + + CPUInterruptHandler(CPUInterruptHandler&&) = delete; + CPUInterruptHandler& operator=(CPUInterruptHandler&&) = delete; + + bool IsInterrupted() const { + return is_interrupted; + } + + void SetInterrupt(bool is_interrupted); + + void AwaitInterrupt(); + +private: + std::unique_ptr<Common::Event> interrupt_event; + std::atomic_bool is_interrupted{false}; +}; + +} // namespace Core diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.cpp b/src/core/arm/dynarmic/arm_dynarmic_32.cpp index 187a972ac..6dc03f3b1 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_32.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_32.cpp @@ -7,15 +7,18 @@ #include <dynarmic/A32/a32.h> #include <dynarmic/A32/config.h> #include <dynarmic/A32/context.h> -#include "common/microprofile.h" +#include "common/assert.h" +#include "common/logging/log.h" +#include "common/page_table.h" +#include "core/arm/cpu_interrupt_handler.h" #include "core/arm/dynarmic/arm_dynarmic_32.h" -#include "core/arm/dynarmic/arm_dynarmic_64.h" #include "core/arm/dynarmic/arm_dynarmic_cp15.h" +#include "core/arm/dynarmic/arm_exclusive_monitor.h" #include "core/core.h" -#include "core/core_manager.h" #include "core/core_timing.h" #include "core/hle/kernel/svc.h" #include "core/memory.h" +#include "core/settings.h" namespace Core { @@ -49,8 +52,22 @@ public: parent.system.Memory().Write64(vaddr, value); } + bool MemoryWriteExclusive8(u32 vaddr, u8 value, u8 expected) override { + return parent.system.Memory().WriteExclusive8(vaddr, value, expected); + } + bool MemoryWriteExclusive16(u32 vaddr, u16 value, u16 expected) override { + return parent.system.Memory().WriteExclusive16(vaddr, value, expected); + } + bool MemoryWriteExclusive32(u32 vaddr, u32 value, u32 expected) override { + return parent.system.Memory().WriteExclusive32(vaddr, value, expected); + } + bool MemoryWriteExclusive64(u32 vaddr, u64 value, u64 expected) override { + return parent.system.Memory().WriteExclusive64(vaddr, value, expected); + } + void InterpreterFallback(u32 pc, std::size_t num_instructions) override { - UNIMPLEMENTED(); + UNIMPLEMENTED_MSG("This should never happen, pc = {:08X}, code = {:08X}", pc, + MemoryReadCode(pc)); } void ExceptionRaised(u32 pc, Dynarmic::A32::Exception exception) override { @@ -61,36 +78,46 @@ public: case Dynarmic::A32::Exception::Breakpoint: break; } - LOG_CRITICAL(HW_GPU, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})", + LOG_CRITICAL(Core_ARM, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})", static_cast<std::size_t>(exception), pc, MemoryReadCode(pc)); UNIMPLEMENTED(); } void CallSVC(u32 swi) override { - Kernel::CallSVC(parent.system, swi); + Kernel::Svc::Call(parent.system, swi); } void AddTicks(u64 ticks) override { + if (parent.uses_wall_clock) { + return; + } // Divide the number of ticks by the amount of CPU cores. TODO(Subv): This yields only a // rough approximation of the amount of executed ticks in the system, it may be thrown off // if not all cores are doing a similar amount of work. Instead of doing this, we should // device a way so that timing is consistent across all cores without increasing the ticks 4 // times. - u64 amortized_ticks = (ticks - num_interpreted_instructions) / Core::NUM_CPU_CORES; + u64 amortized_ticks = + (ticks - num_interpreted_instructions) / Core::Hardware::NUM_CPU_CORES; // Always execute at least one tick. amortized_ticks = std::max<u64>(amortized_ticks, 1); parent.system.CoreTiming().AddTicks(amortized_ticks); num_interpreted_instructions = 0; } + u64 GetTicksRemaining() override { - return std::max(parent.system.CoreTiming().GetDowncount(), {}); + if (parent.uses_wall_clock) { + if (!parent.interrupt_handlers[parent.core_index].IsInterrupted()) { + return minimum_run_cycles; + } + return 0U; + } + return std::max<s64>(parent.system.CoreTiming().GetDowncount(), 0); } ARM_Dynarmic_32& parent; std::size_t num_interpreted_instructions{}; - u64 tpidrro_el0{}; - u64 tpidr_el0{}; + static constexpr u64 minimum_run_cycles = 1000U; }; std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& page_table, @@ -99,26 +126,79 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& config.callbacks = cb.get(); // TODO(bunnei): Implement page table for 32-bit // config.page_table = &page_table.pointers; - config.coprocessors[15] = std::make_shared<DynarmicCP15>((u32*)&CP15_regs[0]); + config.coprocessors[15] = cp15; config.define_unpredictable_behaviour = true; + static constexpr std::size_t PAGE_BITS = 12; + static constexpr std::size_t NUM_PAGE_TABLE_ENTRIES = 1 << (32 - PAGE_BITS); + config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>( + page_table.pointers.data()); + config.absolute_offset_page_table = true; + config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; + config.only_detect_misalignment_via_page_table_on_page_boundary = true; + + // Multi-process state + config.processor_id = core_index; + config.global_monitor = &exclusive_monitor.monitor; + + // Timing + config.wall_clock_cntpct = uses_wall_clock; + + // Safe optimizations + if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::DebugMode) { + if (!Settings::values.cpuopt_page_tables) { + config.page_table = nullptr; + } + if (!Settings::values.cpuopt_block_linking) { + config.optimizations &= ~Dynarmic::OptimizationFlag::BlockLinking; + } + if (!Settings::values.cpuopt_return_stack_buffer) { + config.optimizations &= ~Dynarmic::OptimizationFlag::ReturnStackBuffer; + } + if (!Settings::values.cpuopt_fast_dispatcher) { + config.optimizations &= ~Dynarmic::OptimizationFlag::FastDispatch; + } + if (!Settings::values.cpuopt_context_elimination) { + config.optimizations &= ~Dynarmic::OptimizationFlag::GetSetElimination; + } + if (!Settings::values.cpuopt_const_prop) { + config.optimizations &= ~Dynarmic::OptimizationFlag::ConstProp; + } + if (!Settings::values.cpuopt_misc_ir) { + config.optimizations &= ~Dynarmic::OptimizationFlag::MiscIROpt; + } + if (!Settings::values.cpuopt_reduce_misalign_checks) { + config.only_detect_misalignment_via_page_table_on_page_boundary = false; + } + } + + // Unsafe optimizations + if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::Unsafe) { + config.unsafe_optimizations = true; + if (Settings::values.cpuopt_unsafe_unfuse_fma) { + config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_UnfuseFMA; + } + if (Settings::values.cpuopt_unsafe_reduce_fp_error) { + config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_ReducedErrorFP; + } + } + return std::make_unique<Dynarmic::A32::Jit>(config); } -MICROPROFILE_DEFINE(ARM_Jit_Dynarmic_32, "ARM JIT", "Dynarmic", MP_RGB(255, 64, 64)); - void ARM_Dynarmic_32::Run() { - MICROPROFILE_SCOPE(ARM_Jit_Dynarmic_32); jit->Run(); } void ARM_Dynarmic_32::Step() { - cb->InterpreterFallback(jit->Regs()[15], 1); + jit->Step(); } -ARM_Dynarmic_32::ARM_Dynarmic_32(System& system, ExclusiveMonitor& exclusive_monitor, +ARM_Dynarmic_32::ARM_Dynarmic_32(System& system, CPUInterrupts& interrupt_handlers, + bool uses_wall_clock, ExclusiveMonitor& exclusive_monitor, std::size_t core_index) - : ARM_Interface{system}, - cb(std::make_unique<DynarmicCallbacks32>(*this)), core_index{core_index}, + : ARM_Interface{system, interrupt_handlers, uses_wall_clock}, + cb(std::make_unique<DynarmicCallbacks32>(*this)), + cp15(std::make_shared<DynarmicCP15>(*this)), core_index{core_index}, exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} ARM_Dynarmic_32::~ARM_Dynarmic_32() = default; @@ -154,32 +234,40 @@ void ARM_Dynarmic_32::SetPSTATE(u32 cpsr) { } u64 ARM_Dynarmic_32::GetTlsAddress() const { - return CP15_regs[static_cast<std::size_t>(CP15Register::CP15_THREAD_URO)]; + return cp15->uro; } void ARM_Dynarmic_32::SetTlsAddress(VAddr address) { - CP15_regs[static_cast<std::size_t>(CP15Register::CP15_THREAD_URO)] = static_cast<u32>(address); + cp15->uro = static_cast<u32>(address); } u64 ARM_Dynarmic_32::GetTPIDR_EL0() const { - return cb->tpidr_el0; + return cp15->uprw; } void ARM_Dynarmic_32::SetTPIDR_EL0(u64 value) { - cb->tpidr_el0 = value; + cp15->uprw = static_cast<u32>(value); +} + +void ARM_Dynarmic_32::ChangeProcessorID(std::size_t new_core_id) { + jit->ChangeProcessorID(new_core_id); } void ARM_Dynarmic_32::SaveContext(ThreadContext32& ctx) { Dynarmic::A32::Context context; jit->SaveContext(context); ctx.cpu_registers = context.Regs(); + ctx.extension_registers = context.ExtRegs(); ctx.cpsr = context.Cpsr(); + ctx.fpscr = context.Fpscr(); } void ARM_Dynarmic_32::LoadContext(const ThreadContext32& ctx) { Dynarmic::A32::Context context; context.Regs() = ctx.cpu_registers; + context.ExtRegs() = ctx.extension_registers; context.SetCpsr(ctx.cpsr); + context.SetFpscr(ctx.fpscr); jit->LoadContext(context); } @@ -188,10 +276,15 @@ void ARM_Dynarmic_32::PrepareReschedule() { } void ARM_Dynarmic_32::ClearInstructionCache() { + if (!jit) { + return; + } jit->ClearCache(); } -void ARM_Dynarmic_32::ClearExclusiveState() {} +void ARM_Dynarmic_32::ClearExclusiveState() { + jit->ClearExclusiveState(); +} void ARM_Dynarmic_32::PageTableChanged(Common::PageTable& page_table, std::size_t new_address_space_size_in_bits) { diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.h b/src/core/arm/dynarmic/arm_dynarmic_32.h index 143e46e4d..2bab31b92 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_32.h +++ b/src/core/arm/dynarmic/arm_dynarmic_32.h @@ -9,25 +9,28 @@ #include <dynarmic/A32/a32.h> #include <dynarmic/A64/a64.h> -#include <dynarmic/A64/exclusive_monitor.h> +#include <dynarmic/exclusive_monitor.h> #include "common/common_types.h" #include "common/hash.h" #include "core/arm/arm_interface.h" #include "core/arm/exclusive_monitor.h" -namespace Memory { +namespace Core::Memory { class Memory; } namespace Core { +class CPUInterruptHandler; class DynarmicCallbacks32; +class DynarmicCP15; class DynarmicExclusiveMonitor; class System; class ARM_Dynarmic_32 final : public ARM_Interface { public: - ARM_Dynarmic_32(System& system, ExclusiveMonitor& exclusive_monitor, std::size_t core_index); + ARM_Dynarmic_32(System& system, CPUInterrupts& interrupt_handlers, bool uses_wall_clock, + ExclusiveMonitor& exclusive_monitor, std::size_t core_index); ~ARM_Dynarmic_32() override; void SetPC(u64 pc) override; @@ -44,6 +47,7 @@ public: void SetTlsAddress(VAddr address) override; void SetTPIDR_EL0(u64 value) override; u64 GetTPIDR_EL0() const override; + void ChangeProcessorID(std::size_t new_core_id) override; void SaveContext(ThreadContext32& ctx) override; void SaveContext(ThreadContext64& ctx) override {} @@ -66,12 +70,14 @@ private: std::unordered_map<JitCacheKey, std::shared_ptr<Dynarmic::A32::Jit>, Common::PairHash>; friend class DynarmicCallbacks32; + friend class DynarmicCP15; + std::unique_ptr<DynarmicCallbacks32> cb; JitCacheType jit_cache; std::shared_ptr<Dynarmic::A32::Jit> jit; + std::shared_ptr<DynarmicCP15> cp15; std::size_t core_index; DynarmicExclusiveMonitor& exclusive_monitor; - std::array<u32, 84> CP15_regs{}; }; } // namespace Core diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.cpp b/src/core/arm/dynarmic/arm_dynarmic_64.cpp index a53a58ba0..9f170a224 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_64.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_64.cpp @@ -6,20 +6,21 @@ #include <memory> #include <dynarmic/A64/a64.h> #include <dynarmic/A64/config.h> +#include "common/assert.h" #include "common/logging/log.h" -#include "common/microprofile.h" +#include "common/page_table.h" +#include "core/arm/cpu_interrupt_handler.h" #include "core/arm/dynarmic/arm_dynarmic_64.h" +#include "core/arm/dynarmic/arm_exclusive_monitor.h" #include "core/core.h" -#include "core/core_manager.h" #include "core/core_timing.h" -#include "core/core_timing_util.h" #include "core/gdbstub/gdbstub.h" #include "core/hardware_properties.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/svc.h" -#include "core/hle/kernel/vm_manager.h" #include "core/memory.h" +#include "core/settings.h" namespace Core { @@ -64,17 +65,26 @@ public: memory.Write64(vaddr + 8, value[1]); } + bool MemoryWriteExclusive8(u64 vaddr, std::uint8_t value, std::uint8_t expected) override { + return parent.system.Memory().WriteExclusive8(vaddr, value, expected); + } + bool MemoryWriteExclusive16(u64 vaddr, std::uint16_t value, std::uint16_t expected) override { + return parent.system.Memory().WriteExclusive16(vaddr, value, expected); + } + bool MemoryWriteExclusive32(u64 vaddr, std::uint32_t value, std::uint32_t expected) override { + return parent.system.Memory().WriteExclusive32(vaddr, value, expected); + } + bool MemoryWriteExclusive64(u64 vaddr, std::uint64_t value, std::uint64_t expected) override { + return parent.system.Memory().WriteExclusive64(vaddr, value, expected); + } + bool MemoryWriteExclusive128(u64 vaddr, Vector value, Vector expected) override { + return parent.system.Memory().WriteExclusive128(vaddr, value, expected); + } + void InterpreterFallback(u64 pc, std::size_t num_instructions) override { - LOG_INFO(Core_ARM, "Unicorn fallback @ 0x{:X} for {} instructions (instr = {:08X})", pc, - num_instructions, MemoryReadCode(pc)); - - ARM_Interface::ThreadContext64 ctx; - parent.SaveContext(ctx); - parent.inner_unicorn.LoadContext(ctx); - parent.inner_unicorn.ExecuteInstructions(num_instructions); - parent.inner_unicorn.SaveContext(ctx); - parent.LoadContext(ctx); - num_interpreted_instructions += num_instructions; + LOG_ERROR(Core_ARM, + "Unimplemented instruction @ 0x{:X} for {} instructions (instr = {:08X})", pc, + num_instructions, MemoryReadCode(pc)); } void ExceptionRaised(u64 pc, Dynarmic::A64::Exception exception) override { @@ -97,39 +107,50 @@ public: } [[fallthrough]]; default: - ASSERT_MSG(false, "ExceptionRaised(exception = {}, pc = {:X})", - static_cast<std::size_t>(exception), pc); + ASSERT_MSG(false, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})", + static_cast<std::size_t>(exception), pc, MemoryReadCode(pc)); } } void CallSVC(u32 swi) override { - Kernel::CallSVC(parent.system, swi); + Kernel::Svc::Call(parent.system, swi); } void AddTicks(u64 ticks) override { + if (parent.uses_wall_clock) { + return; + } + // Divide the number of ticks by the amount of CPU cores. TODO(Subv): This yields only a // rough approximation of the amount of executed ticks in the system, it may be thrown off // if not all cores are doing a similar amount of work. Instead of doing this, we should // device a way so that timing is consistent across all cores without increasing the ticks 4 // times. - u64 amortized_ticks = (ticks - num_interpreted_instructions) / Core::NUM_CPU_CORES; + u64 amortized_ticks = ticks / Core::Hardware::NUM_CPU_CORES; // Always execute at least one tick. amortized_ticks = std::max<u64>(amortized_ticks, 1); parent.system.CoreTiming().AddTicks(amortized_ticks); - num_interpreted_instructions = 0; } + u64 GetTicksRemaining() override { - return std::max(parent.system.CoreTiming().GetDowncount(), s64{0}); + if (parent.uses_wall_clock) { + if (!parent.interrupt_handlers[parent.core_index].IsInterrupted()) { + return minimum_run_cycles; + } + return 0U; + } + return std::max<s64>(parent.system.CoreTiming().GetDowncount(), 0); } + u64 GetCNTPCT() override { - return Timing::CpuCyclesToClockCycles(parent.system.CoreTiming().GetTicks()); + return parent.system.CoreTiming().GetClockTicks(); } ARM_Dynarmic_64& parent; - std::size_t num_interpreted_instructions = 0; u64 tpidrro_el0 = 0; u64 tpidr_el0 = 0; + static constexpr u64 minimum_run_cycles = 1000U; }; std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& page_table, @@ -144,6 +165,8 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& config.page_table_address_space_bits = address_space_bits; config.silently_mirror_page_table = false; config.absolute_offset_page_table = true; + config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; + config.only_detect_misalignment_via_page_table_on_page_boundary = true; // Multi-process state config.processor_id = core_index; @@ -159,14 +182,52 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& // Unpredictable instructions config.define_unpredictable_behaviour = true; + // Timing + config.wall_clock_cntpct = uses_wall_clock; + + // Safe optimizations + if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::DebugMode) { + if (!Settings::values.cpuopt_page_tables) { + config.page_table = nullptr; + } + if (!Settings::values.cpuopt_block_linking) { + config.optimizations &= ~Dynarmic::OptimizationFlag::BlockLinking; + } + if (!Settings::values.cpuopt_return_stack_buffer) { + config.optimizations &= ~Dynarmic::OptimizationFlag::ReturnStackBuffer; + } + if (!Settings::values.cpuopt_fast_dispatcher) { + config.optimizations &= ~Dynarmic::OptimizationFlag::FastDispatch; + } + if (!Settings::values.cpuopt_context_elimination) { + config.optimizations &= ~Dynarmic::OptimizationFlag::GetSetElimination; + } + if (!Settings::values.cpuopt_const_prop) { + config.optimizations &= ~Dynarmic::OptimizationFlag::ConstProp; + } + if (!Settings::values.cpuopt_misc_ir) { + config.optimizations &= ~Dynarmic::OptimizationFlag::MiscIROpt; + } + if (!Settings::values.cpuopt_reduce_misalign_checks) { + config.only_detect_misalignment_via_page_table_on_page_boundary = false; + } + } + + // Unsafe optimizations + if (Settings::values.cpu_accuracy == Settings::CPUAccuracy::Unsafe) { + config.unsafe_optimizations = true; + if (Settings::values.cpuopt_unsafe_unfuse_fma) { + config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_UnfuseFMA; + } + if (Settings::values.cpuopt_unsafe_reduce_fp_error) { + config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_ReducedErrorFP; + } + } + return std::make_shared<Dynarmic::A64::Jit>(config); } -MICROPROFILE_DEFINE(ARM_Jit_Dynarmic_64, "ARM JIT", "Dynarmic", MP_RGB(255, 64, 64)); - void ARM_Dynarmic_64::Run() { - MICROPROFILE_SCOPE(ARM_Jit_Dynarmic_64); - jit->Run(); } @@ -174,12 +235,12 @@ void ARM_Dynarmic_64::Step() { cb->InterpreterFallback(jit->GetPC(), 1); } -ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, ExclusiveMonitor& exclusive_monitor, +ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, CPUInterrupts& interrupt_handlers, + bool uses_wall_clock, ExclusiveMonitor& exclusive_monitor, std::size_t core_index) - : ARM_Interface{system}, - cb(std::make_unique<DynarmicCallbacks64>(*this)), inner_unicorn{system}, - core_index{core_index}, exclusive_monitor{ - dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} + : ARM_Interface{system, interrupt_handlers, uses_wall_clock}, + cb(std::make_unique<DynarmicCallbacks64>(*this)), core_index{core_index}, + exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} ARM_Dynarmic_64::~ARM_Dynarmic_64() = default; @@ -231,6 +292,10 @@ void ARM_Dynarmic_64::SetTPIDR_EL0(u64 value) { cb->tpidr_el0 = value; } +void ARM_Dynarmic_64::ChangeProcessorID(std::size_t new_core_id) { + jit->ChangeProcessorID(new_core_id); +} + void ARM_Dynarmic_64::SaveContext(ThreadContext64& ctx) { ctx.cpu_registers = jit->GetRegisters(); ctx.sp = jit->GetSP(); @@ -258,6 +323,9 @@ void ARM_Dynarmic_64::PrepareReschedule() { } void ARM_Dynarmic_64::ClearInstructionCache() { + if (!jit) { + return; + } jit->ClearCache(); } @@ -277,44 +345,4 @@ void ARM_Dynarmic_64::PageTableChanged(Common::PageTable& page_table, jit_cache.emplace(key, jit); } -DynarmicExclusiveMonitor::DynarmicExclusiveMonitor(Memory::Memory& memory, std::size_t core_count) - : monitor(core_count), memory{memory} {} - -DynarmicExclusiveMonitor::~DynarmicExclusiveMonitor() = default; - -void DynarmicExclusiveMonitor::SetExclusive(std::size_t core_index, VAddr addr) { - // Size doesn't actually matter. - monitor.Mark(core_index, addr, 16); -} - -void DynarmicExclusiveMonitor::ClearExclusive() { - monitor.Clear(); -} - -bool DynarmicExclusiveMonitor::ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) { - return monitor.DoExclusiveOperation(core_index, vaddr, 1, [&] { memory.Write8(vaddr, value); }); -} - -bool DynarmicExclusiveMonitor::ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) { - return monitor.DoExclusiveOperation(core_index, vaddr, 2, - [&] { memory.Write16(vaddr, value); }); -} - -bool DynarmicExclusiveMonitor::ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) { - return monitor.DoExclusiveOperation(core_index, vaddr, 4, - [&] { memory.Write32(vaddr, value); }); -} - -bool DynarmicExclusiveMonitor::ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) { - return monitor.DoExclusiveOperation(core_index, vaddr, 8, - [&] { memory.Write64(vaddr, value); }); -} - -bool DynarmicExclusiveMonitor::ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) { - return monitor.DoExclusiveOperation(core_index, vaddr, 16, [&] { - memory.Write64(vaddr + 0, value[0]); - memory.Write64(vaddr + 8, value[1]); - }); -} - } // namespace Core diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.h b/src/core/arm/dynarmic/arm_dynarmic_64.h index e71240a96..28e11a17d 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_64.h +++ b/src/core/arm/dynarmic/arm_dynarmic_64.h @@ -8,26 +8,26 @@ #include <unordered_map> #include <dynarmic/A64/a64.h> -#include <dynarmic/A64/exclusive_monitor.h> #include "common/common_types.h" #include "common/hash.h" #include "core/arm/arm_interface.h" #include "core/arm/exclusive_monitor.h" -#include "core/arm/unicorn/arm_unicorn.h" -namespace Memory { +namespace Core::Memory { class Memory; } namespace Core { class DynarmicCallbacks64; +class CPUInterruptHandler; class DynarmicExclusiveMonitor; class System; class ARM_Dynarmic_64 final : public ARM_Interface { public: - ARM_Dynarmic_64(System& system, ExclusiveMonitor& exclusive_monitor, std::size_t core_index); + ARM_Dynarmic_64(System& system, CPUInterrupts& interrupt_handlers, bool uses_wall_clock, + ExclusiveMonitor& exclusive_monitor, std::size_t core_index); ~ARM_Dynarmic_64() override; void SetPC(u64 pc) override; @@ -44,6 +44,7 @@ public: void SetTlsAddress(VAddr address) override; void SetTPIDR_EL0(u64 value) override; u64 GetTPIDR_EL0() const override; + void ChangeProcessorID(std::size_t new_core_id) override; void SaveContext(ThreadContext32& ctx) override {} void SaveContext(ThreadContext64& ctx) override; @@ -69,30 +70,9 @@ private: std::unique_ptr<DynarmicCallbacks64> cb; JitCacheType jit_cache; std::shared_ptr<Dynarmic::A64::Jit> jit; - ARM_Unicorn inner_unicorn; std::size_t core_index; DynarmicExclusiveMonitor& exclusive_monitor; }; -class DynarmicExclusiveMonitor final : public ExclusiveMonitor { -public: - explicit DynarmicExclusiveMonitor(Memory::Memory& memory, std::size_t core_count); - ~DynarmicExclusiveMonitor() override; - - void SetExclusive(std::size_t core_index, VAddr addr) override; - void ClearExclusive() override; - - bool ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) override; - bool ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) override; - bool ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) override; - bool ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) override; - bool ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) override; - -private: - friend class ARM_Dynarmic_64; - Dynarmic::A64::ExclusiveMonitor monitor; - Memory::Memory& memory; -}; - } // namespace Core diff --git a/src/core/arm/dynarmic/arm_dynarmic_cp15.cpp b/src/core/arm/dynarmic/arm_dynarmic_cp15.cpp index 3fdcdebde..caefc09f4 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_cp15.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_cp15.cpp @@ -2,79 +2,132 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <fmt/format.h> +#include "common/logging/log.h" +#include "core/arm/dynarmic/arm_dynarmic_32.h" #include "core/arm/dynarmic/arm_dynarmic_cp15.h" +#include "core/core.h" +#include "core/core_timing.h" +#include "core/core_timing_util.h" using Callback = Dynarmic::A32::Coprocessor::Callback; using CallbackOrAccessOneWord = Dynarmic::A32::Coprocessor::CallbackOrAccessOneWord; using CallbackOrAccessTwoWords = Dynarmic::A32::Coprocessor::CallbackOrAccessTwoWords; +template <> +struct fmt::formatter<Dynarmic::A32::CoprocReg> { + constexpr auto parse(format_parse_context& ctx) { + return ctx.begin(); + } + template <typename FormatContext> + auto format(const Dynarmic::A32::CoprocReg& reg, FormatContext& ctx) { + return format_to(ctx.out(), "cp{}", static_cast<size_t>(reg)); + } +}; + +namespace Core { + +static u32 dummy_value; + std::optional<Callback> DynarmicCP15::CompileInternalOperation(bool two, unsigned opc1, CoprocReg CRd, CoprocReg CRn, CoprocReg CRm, unsigned opc2) { - return {}; + LOG_CRITICAL(Core_ARM, "CP15: cdp{} p15, {}, {}, {}, {}, {}", two ? "2" : "", opc1, CRd, CRn, + CRm, opc2); + return std::nullopt; } CallbackOrAccessOneWord DynarmicCP15::CompileSendOneWord(bool two, unsigned opc1, CoprocReg CRn, CoprocReg CRm, unsigned opc2) { - // TODO(merry): Privileged CP15 registers - if (!two && CRn == CoprocReg::C7 && opc1 == 0 && CRm == CoprocReg::C5 && opc2 == 4) { + // CP15_FLUSH_PREFETCH_BUFFER // This is a dummy write, we ignore the value written here. - return &CP15[static_cast<std::size_t>(CP15Register::CP15_FLUSH_PREFETCH_BUFFER)]; + return &dummy_value; } if (!two && CRn == CoprocReg::C7 && opc1 == 0 && CRm == CoprocReg::C10) { switch (opc2) { case 4: + // CP15_DATA_SYNC_BARRIER // This is a dummy write, we ignore the value written here. - return &CP15[static_cast<std::size_t>(CP15Register::CP15_DATA_SYNC_BARRIER)]; + return &dummy_value; case 5: + // CP15_DATA_MEMORY_BARRIER // This is a dummy write, we ignore the value written here. - return &CP15[static_cast<std::size_t>(CP15Register::CP15_DATA_MEMORY_BARRIER)]; - default: - return {}; + return &dummy_value; } } if (!two && CRn == CoprocReg::C13 && opc1 == 0 && CRm == CoprocReg::C0 && opc2 == 2) { - return &CP15[static_cast<std::size_t>(CP15Register::CP15_THREAD_UPRW)]; + // CP15_THREAD_UPRW + return &uprw; } + LOG_CRITICAL(Core_ARM, "CP15: mcr{} p15, {}, <Rt>, {}, {}, {}", two ? "2" : "", opc1, CRn, CRm, + opc2); return {}; } CallbackOrAccessTwoWords DynarmicCP15::CompileSendTwoWords(bool two, unsigned opc, CoprocReg CRm) { + LOG_CRITICAL(Core_ARM, "CP15: mcrr{} p15, {}, <Rt>, <Rt2>, {}", two ? "2" : "", opc, CRm); return {}; } CallbackOrAccessOneWord DynarmicCP15::CompileGetOneWord(bool two, unsigned opc1, CoprocReg CRn, CoprocReg CRm, unsigned opc2) { - // TODO(merry): Privileged CP15 registers - if (!two && CRn == CoprocReg::C13 && opc1 == 0 && CRm == CoprocReg::C0) { switch (opc2) { case 2: - return &CP15[static_cast<std::size_t>(CP15Register::CP15_THREAD_UPRW)]; + // CP15_THREAD_UPRW + return &uprw; case 3: - return &CP15[static_cast<std::size_t>(CP15Register::CP15_THREAD_URO)]; - default: - return {}; + // CP15_THREAD_URO + return &uro; } } + LOG_CRITICAL(Core_ARM, "CP15: mrc{} p15, {}, <Rt>, {}, {}, {}", two ? "2" : "", opc1, CRn, CRm, + opc2); return {}; } CallbackOrAccessTwoWords DynarmicCP15::CompileGetTwoWords(bool two, unsigned opc, CoprocReg CRm) { + if (!two && opc == 0 && CRm == CoprocReg::C14) { + // CNTPCT + const auto callback = static_cast<u64 (*)(Dynarmic::A32::Jit*, void*, u32, u32)>( + [](Dynarmic::A32::Jit*, void* arg, u32, u32) -> u64 { + ARM_Dynarmic_32& parent = *(ARM_Dynarmic_32*)arg; + return parent.system.CoreTiming().GetClockTicks(); + }); + return Dynarmic::A32::Coprocessor::Callback{callback, (void*)&parent}; + } + + LOG_CRITICAL(Core_ARM, "CP15: mrrc{} p15, {}, <Rt>, <Rt2>, {}", two ? "2" : "", opc, CRm); return {}; } std::optional<Callback> DynarmicCP15::CompileLoadWords(bool two, bool long_transfer, CoprocReg CRd, std::optional<u8> option) { - return {}; + if (option) { + LOG_CRITICAL(Core_ARM, "CP15: mrrc{}{} p15, {}, [...], {}", two ? "2" : "", + long_transfer ? "l" : "", CRd, *option); + } else { + LOG_CRITICAL(Core_ARM, "CP15: mrrc{}{} p15, {}, [...]", two ? "2" : "", + long_transfer ? "l" : "", CRd); + } + return std::nullopt; } std::optional<Callback> DynarmicCP15::CompileStoreWords(bool two, bool long_transfer, CoprocReg CRd, std::optional<u8> option) { - return {}; + if (option) { + LOG_CRITICAL(Core_ARM, "CP15: mrrc{}{} p15, {}, [...], {}", two ? "2" : "", + long_transfer ? "l" : "", CRd, *option); + } else { + LOG_CRITICAL(Core_ARM, "CP15: mrrc{}{} p15, {}, [...]", two ? "2" : "", + long_transfer ? "l" : "", CRd); + } + return std::nullopt; } + +} // namespace Core diff --git a/src/core/arm/dynarmic/arm_dynarmic_cp15.h b/src/core/arm/dynarmic/arm_dynarmic_cp15.h index 07bcde5f9..dc6f4af3a 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_cp15.h +++ b/src/core/arm/dynarmic/arm_dynarmic_cp15.h @@ -10,128 +10,15 @@ #include <dynarmic/A32/coprocessor.h> #include "common/common_types.h" -enum class CP15Register { - // c0 - Information registers - CP15_MAIN_ID, - CP15_CACHE_TYPE, - CP15_TCM_STATUS, - CP15_TLB_TYPE, - CP15_CPU_ID, - CP15_PROCESSOR_FEATURE_0, - CP15_PROCESSOR_FEATURE_1, - CP15_DEBUG_FEATURE_0, - CP15_AUXILIARY_FEATURE_0, - CP15_MEMORY_MODEL_FEATURE_0, - CP15_MEMORY_MODEL_FEATURE_1, - CP15_MEMORY_MODEL_FEATURE_2, - CP15_MEMORY_MODEL_FEATURE_3, - CP15_ISA_FEATURE_0, - CP15_ISA_FEATURE_1, - CP15_ISA_FEATURE_2, - CP15_ISA_FEATURE_3, - CP15_ISA_FEATURE_4, +namespace Core { - // c1 - Control registers - CP15_CONTROL, - CP15_AUXILIARY_CONTROL, - CP15_COPROCESSOR_ACCESS_CONTROL, - - // c2 - Translation table registers - CP15_TRANSLATION_BASE_TABLE_0, - CP15_TRANSLATION_BASE_TABLE_1, - CP15_TRANSLATION_BASE_CONTROL, - CP15_DOMAIN_ACCESS_CONTROL, - CP15_RESERVED, - - // c5 - Fault status registers - CP15_FAULT_STATUS, - CP15_INSTR_FAULT_STATUS, - CP15_COMBINED_DATA_FSR = CP15_FAULT_STATUS, - CP15_INST_FSR, - - // c6 - Fault Address registers - CP15_FAULT_ADDRESS, - CP15_COMBINED_DATA_FAR = CP15_FAULT_ADDRESS, - CP15_WFAR, - CP15_IFAR, - - // c7 - Cache operation registers - CP15_WAIT_FOR_INTERRUPT, - CP15_PHYS_ADDRESS, - CP15_INVALIDATE_INSTR_CACHE, - CP15_INVALIDATE_INSTR_CACHE_USING_MVA, - CP15_INVALIDATE_INSTR_CACHE_USING_INDEX, - CP15_FLUSH_PREFETCH_BUFFER, - CP15_FLUSH_BRANCH_TARGET_CACHE, - CP15_FLUSH_BRANCH_TARGET_CACHE_ENTRY, - CP15_INVALIDATE_DATA_CACHE, - CP15_INVALIDATE_DATA_CACHE_LINE_USING_MVA, - CP15_INVALIDATE_DATA_CACHE_LINE_USING_INDEX, - CP15_INVALIDATE_DATA_AND_INSTR_CACHE, - CP15_CLEAN_DATA_CACHE, - CP15_CLEAN_DATA_CACHE_LINE_USING_MVA, - CP15_CLEAN_DATA_CACHE_LINE_USING_INDEX, - CP15_DATA_SYNC_BARRIER, - CP15_DATA_MEMORY_BARRIER, - CP15_CLEAN_AND_INVALIDATE_DATA_CACHE, - CP15_CLEAN_AND_INVALIDATE_DATA_CACHE_LINE_USING_MVA, - CP15_CLEAN_AND_INVALIDATE_DATA_CACHE_LINE_USING_INDEX, - - // c8 - TLB operations - CP15_INVALIDATE_ITLB, - CP15_INVALIDATE_ITLB_SINGLE_ENTRY, - CP15_INVALIDATE_ITLB_ENTRY_ON_ASID_MATCH, - CP15_INVALIDATE_ITLB_ENTRY_ON_MVA, - CP15_INVALIDATE_DTLB, - CP15_INVALIDATE_DTLB_SINGLE_ENTRY, - CP15_INVALIDATE_DTLB_ENTRY_ON_ASID_MATCH, - CP15_INVALIDATE_DTLB_ENTRY_ON_MVA, - CP15_INVALIDATE_UTLB, - CP15_INVALIDATE_UTLB_SINGLE_ENTRY, - CP15_INVALIDATE_UTLB_ENTRY_ON_ASID_MATCH, - CP15_INVALIDATE_UTLB_ENTRY_ON_MVA, - - // c9 - Data cache lockdown register - CP15_DATA_CACHE_LOCKDOWN, - - // c10 - TLB/Memory map registers - CP15_TLB_LOCKDOWN, - CP15_PRIMARY_REGION_REMAP, - CP15_NORMAL_REGION_REMAP, - - // c13 - Thread related registers - CP15_PID, - CP15_CONTEXT_ID, - CP15_THREAD_UPRW, // Thread ID register - User/Privileged Read/Write - CP15_THREAD_URO, // Thread ID register - User Read Only (Privileged R/W) - CP15_THREAD_PRW, // Thread ID register - Privileged R/W only. - - // c15 - Performance and TLB lockdown registers - CP15_PERFORMANCE_MONITOR_CONTROL, - CP15_CYCLE_COUNTER, - CP15_COUNT_0, - CP15_COUNT_1, - CP15_READ_MAIN_TLB_LOCKDOWN_ENTRY, - CP15_WRITE_MAIN_TLB_LOCKDOWN_ENTRY, - CP15_MAIN_TLB_LOCKDOWN_VIRT_ADDRESS, - CP15_MAIN_TLB_LOCKDOWN_PHYS_ADDRESS, - CP15_MAIN_TLB_LOCKDOWN_ATTRIBUTE, - CP15_TLB_DEBUG_CONTROL, - - // Skyeye defined - CP15_TLB_FAULT_ADDR, - CP15_TLB_FAULT_STATUS, - - // Not an actual register. - // All registers should be defined above this. - CP15_REGISTER_COUNT, -}; +class ARM_Dynarmic_32; class DynarmicCP15 final : public Dynarmic::A32::Coprocessor { public: using CoprocReg = Dynarmic::A32::CoprocReg; - explicit DynarmicCP15(u32* cp15) : CP15(cp15){}; + explicit DynarmicCP15(ARM_Dynarmic_32& parent) : parent(parent) {} std::optional<Callback> CompileInternalOperation(bool two, unsigned opc1, CoprocReg CRd, CoprocReg CRn, CoprocReg CRm, @@ -147,6 +34,9 @@ public: std::optional<Callback> CompileStoreWords(bool two, bool long_transfer, CoprocReg CRd, std::optional<u8> option) override; -private: - u32* CP15{}; + ARM_Dynarmic_32& parent; + u32 uprw = 0; + u32 uro = 0; }; + +} // namespace Core diff --git a/src/core/arm/dynarmic/arm_exclusive_monitor.cpp b/src/core/arm/dynarmic/arm_exclusive_monitor.cpp new file mode 100644 index 000000000..4e209f6a5 --- /dev/null +++ b/src/core/arm/dynarmic/arm_exclusive_monitor.cpp @@ -0,0 +1,76 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cinttypes> +#include <memory> +#include "core/arm/dynarmic/arm_exclusive_monitor.h" +#include "core/memory.h" + +namespace Core { + +DynarmicExclusiveMonitor::DynarmicExclusiveMonitor(Memory::Memory& memory, std::size_t core_count) + : monitor(core_count), memory{memory} {} + +DynarmicExclusiveMonitor::~DynarmicExclusiveMonitor() = default; + +u8 DynarmicExclusiveMonitor::ExclusiveRead8(std::size_t core_index, VAddr addr) { + return monitor.ReadAndMark<u8>(core_index, addr, [&]() -> u8 { return memory.Read8(addr); }); +} + +u16 DynarmicExclusiveMonitor::ExclusiveRead16(std::size_t core_index, VAddr addr) { + return monitor.ReadAndMark<u16>(core_index, addr, [&]() -> u16 { return memory.Read16(addr); }); +} + +u32 DynarmicExclusiveMonitor::ExclusiveRead32(std::size_t core_index, VAddr addr) { + return monitor.ReadAndMark<u32>(core_index, addr, [&]() -> u32 { return memory.Read32(addr); }); +} + +u64 DynarmicExclusiveMonitor::ExclusiveRead64(std::size_t core_index, VAddr addr) { + return monitor.ReadAndMark<u64>(core_index, addr, [&]() -> u64 { return memory.Read64(addr); }); +} + +u128 DynarmicExclusiveMonitor::ExclusiveRead128(std::size_t core_index, VAddr addr) { + return monitor.ReadAndMark<u128>(core_index, addr, [&]() -> u128 { + u128 result; + result[0] = memory.Read64(addr); + result[1] = memory.Read64(addr + 8); + return result; + }); +} + +void DynarmicExclusiveMonitor::ClearExclusive() { + monitor.Clear(); +} + +bool DynarmicExclusiveMonitor::ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) { + return monitor.DoExclusiveOperation<u8>(core_index, vaddr, [&](u8 expected) -> bool { + return memory.WriteExclusive8(vaddr, value, expected); + }); +} + +bool DynarmicExclusiveMonitor::ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) { + return monitor.DoExclusiveOperation<u16>(core_index, vaddr, [&](u16 expected) -> bool { + return memory.WriteExclusive16(vaddr, value, expected); + }); +} + +bool DynarmicExclusiveMonitor::ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) { + return monitor.DoExclusiveOperation<u32>(core_index, vaddr, [&](u32 expected) -> bool { + return memory.WriteExclusive32(vaddr, value, expected); + }); +} + +bool DynarmicExclusiveMonitor::ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) { + return monitor.DoExclusiveOperation<u64>(core_index, vaddr, [&](u64 expected) -> bool { + return memory.WriteExclusive64(vaddr, value, expected); + }); +} + +bool DynarmicExclusiveMonitor::ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) { + return monitor.DoExclusiveOperation<u128>(core_index, vaddr, [&](u128 expected) -> bool { + return memory.WriteExclusive128(vaddr, value, expected); + }); +} + +} // namespace Core diff --git a/src/core/arm/dynarmic/arm_exclusive_monitor.h b/src/core/arm/dynarmic/arm_exclusive_monitor.h new file mode 100644 index 000000000..964f4a55d --- /dev/null +++ b/src/core/arm/dynarmic/arm_exclusive_monitor.h @@ -0,0 +1,48 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <unordered_map> + +#include <dynarmic/exclusive_monitor.h> + +#include "common/common_types.h" +#include "core/arm/dynarmic/arm_dynarmic_32.h" +#include "core/arm/dynarmic/arm_dynarmic_64.h" +#include "core/arm/exclusive_monitor.h" + +namespace Core::Memory { +class Memory; +} + +namespace Core { + +class DynarmicExclusiveMonitor final : public ExclusiveMonitor { +public: + explicit DynarmicExclusiveMonitor(Memory::Memory& memory, std::size_t core_count); + ~DynarmicExclusiveMonitor() override; + + u8 ExclusiveRead8(std::size_t core_index, VAddr addr) override; + u16 ExclusiveRead16(std::size_t core_index, VAddr addr) override; + u32 ExclusiveRead32(std::size_t core_index, VAddr addr) override; + u64 ExclusiveRead64(std::size_t core_index, VAddr addr) override; + u128 ExclusiveRead128(std::size_t core_index, VAddr addr) override; + void ClearExclusive() override; + + bool ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) override; + bool ExclusiveWrite16(std::size_t core_index, VAddr vaddr, u16 value) override; + bool ExclusiveWrite32(std::size_t core_index, VAddr vaddr, u32 value) override; + bool ExclusiveWrite64(std::size_t core_index, VAddr vaddr, u64 value) override; + bool ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) override; + +private: + friend class ARM_Dynarmic_32; + friend class ARM_Dynarmic_64; + Dynarmic::ExclusiveMonitor monitor; + Core::Memory::Memory& memory; +}; + +} // namespace Core diff --git a/src/core/arm/exclusive_monitor.cpp b/src/core/arm/exclusive_monitor.cpp index b32401e0b..d8cba369d 100644 --- a/src/core/arm/exclusive_monitor.cpp +++ b/src/core/arm/exclusive_monitor.cpp @@ -3,7 +3,7 @@ // Refer to the license.txt file included. #ifdef ARCHITECTURE_x86_64 -#include "core/arm/dynarmic/arm_dynarmic_64.h" +#include "core/arm/dynarmic/arm_exclusive_monitor.h" #endif #include "core/arm/exclusive_monitor.h" #include "core/memory.h" diff --git a/src/core/arm/exclusive_monitor.h b/src/core/arm/exclusive_monitor.h index 4ef418b90..62f6e6023 100644 --- a/src/core/arm/exclusive_monitor.h +++ b/src/core/arm/exclusive_monitor.h @@ -8,7 +8,7 @@ #include "common/common_types.h" -namespace Memory { +namespace Core::Memory { class Memory; } @@ -18,7 +18,11 @@ class ExclusiveMonitor { public: virtual ~ExclusiveMonitor(); - virtual void SetExclusive(std::size_t core_index, VAddr addr) = 0; + virtual u8 ExclusiveRead8(std::size_t core_index, VAddr addr) = 0; + virtual u16 ExclusiveRead16(std::size_t core_index, VAddr addr) = 0; + virtual u32 ExclusiveRead32(std::size_t core_index, VAddr addr) = 0; + virtual u64 ExclusiveRead64(std::size_t core_index, VAddr addr) = 0; + virtual u128 ExclusiveRead128(std::size_t core_index, VAddr addr) = 0; virtual void ClearExclusive() = 0; virtual bool ExclusiveWrite8(std::size_t core_index, VAddr vaddr, u8 value) = 0; diff --git a/src/core/arm/unicorn/arm_unicorn.cpp b/src/core/arm/unicorn/arm_unicorn.cpp deleted file mode 100644 index 8a9800a96..000000000 --- a/src/core/arm/unicorn/arm_unicorn.cpp +++ /dev/null @@ -1,274 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <unicorn/arm64.h> -#include "common/assert.h" -#include "common/microprofile.h" -#include "core/arm/unicorn/arm_unicorn.h" -#include "core/core.h" -#include "core/core_timing.h" -#include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/svc.h" - -namespace Core { - -// Load Unicorn DLL once on Windows using RAII -#ifdef _MSC_VER -#include <unicorn_dynload.h> -struct LoadDll { -private: - LoadDll() { - ASSERT(uc_dyn_load(NULL, 0)); - } - ~LoadDll() { - ASSERT(uc_dyn_free()); - } - static LoadDll g_load_dll; -}; -LoadDll LoadDll::g_load_dll; -#endif - -#define CHECKED(expr) \ - do { \ - if (auto _cerr = (expr)) { \ - ASSERT_MSG(false, "Call " #expr " failed with error: {} ({})\n", _cerr, \ - uc_strerror(_cerr)); \ - } \ - } while (0) - -static void CodeHook(uc_engine* uc, uint64_t address, uint32_t size, void* user_data) { - GDBStub::BreakpointAddress bkpt = - GDBStub::GetNextBreakpointFromAddress(address, GDBStub::BreakpointType::Execute); - if (GDBStub::IsMemoryBreak() || - (bkpt.type != GDBStub::BreakpointType::None && address == bkpt.address)) { - auto core = static_cast<ARM_Unicorn*>(user_data); - core->RecordBreak(bkpt); - uc_emu_stop(uc); - } -} - -static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int size, u64 value, - void* user_data) { - auto* const system = static_cast<System*>(user_data); - - ARM_Interface::ThreadContext64 ctx{}; - system->CurrentArmInterface().SaveContext(ctx); - ASSERT_MSG(false, "Attempted to read from unmapped memory: 0x{:X}, pc=0x{:X}, lr=0x{:X}", addr, - ctx.pc, ctx.cpu_registers[30]); - - return false; -} - -ARM_Unicorn::ARM_Unicorn(System& system) : ARM_Interface{system} { - CHECKED(uc_open(UC_ARCH_ARM64, UC_MODE_ARM, &uc)); - - auto fpv = 3 << 20; - CHECKED(uc_reg_write(uc, UC_ARM64_REG_CPACR_EL1, &fpv)); - - uc_hook hook{}; - CHECKED(uc_hook_add(uc, &hook, UC_HOOK_INTR, (void*)InterruptHook, this, 0, UINT64_MAX)); - CHECKED(uc_hook_add(uc, &hook, UC_HOOK_MEM_INVALID, (void*)UnmappedMemoryHook, &system, 0, - UINT64_MAX)); - if (GDBStub::IsServerEnabled()) { - CHECKED(uc_hook_add(uc, &hook, UC_HOOK_CODE, (void*)CodeHook, this, 0, UINT64_MAX)); - last_bkpt_hit = false; - } -} - -ARM_Unicorn::~ARM_Unicorn() { - CHECKED(uc_close(uc)); -} - -void ARM_Unicorn::SetPC(u64 pc) { - CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &pc)); -} - -u64 ARM_Unicorn::GetPC() const { - u64 val{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_PC, &val)); - return val; -} - -u64 ARM_Unicorn::GetReg(int regn) const { - u64 val{}; - auto treg = UC_ARM64_REG_SP; - if (regn <= 28) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X0 + regn); - } else if (regn < 31) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X29 + regn - 29); - } - CHECKED(uc_reg_read(uc, treg, &val)); - return val; -} - -void ARM_Unicorn::SetReg(int regn, u64 val) { - auto treg = UC_ARM64_REG_SP; - if (regn <= 28) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X0 + regn); - } else if (regn < 31) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X29 + regn - 29); - } - CHECKED(uc_reg_write(uc, treg, &val)); -} - -u128 ARM_Unicorn::GetVectorReg(int /*index*/) const { - UNIMPLEMENTED(); - static constexpr u128 res{}; - return res; -} - -void ARM_Unicorn::SetVectorReg(int /*index*/, u128 /*value*/) { - UNIMPLEMENTED(); -} - -u32 ARM_Unicorn::GetPSTATE() const { - u64 nzcv{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &nzcv)); - return static_cast<u32>(nzcv); -} - -void ARM_Unicorn::SetPSTATE(u32 pstate) { - u64 nzcv = pstate; - CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &nzcv)); -} - -VAddr ARM_Unicorn::GetTlsAddress() const { - u64 base{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_TPIDRRO_EL0, &base)); - return base; -} - -void ARM_Unicorn::SetTlsAddress(VAddr base) { - CHECKED(uc_reg_write(uc, UC_ARM64_REG_TPIDRRO_EL0, &base)); -} - -u64 ARM_Unicorn::GetTPIDR_EL0() const { - u64 value{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_TPIDR_EL0, &value)); - return value; -} - -void ARM_Unicorn::SetTPIDR_EL0(u64 value) { - CHECKED(uc_reg_write(uc, UC_ARM64_REG_TPIDR_EL0, &value)); -} - -void ARM_Unicorn::Run() { - if (GDBStub::IsServerEnabled()) { - ExecuteInstructions(std::max(4000000U, 0U)); - } else { - ExecuteInstructions( - std::max(std::size_t(system.CoreTiming().GetDowncount()), std::size_t{0})); - } -} - -void ARM_Unicorn::Step() { - ExecuteInstructions(1); -} - -MICROPROFILE_DEFINE(ARM_Jit_Unicorn, "ARM JIT", "Unicorn", MP_RGB(255, 64, 64)); - -void ARM_Unicorn::ExecuteInstructions(std::size_t num_instructions) { - MICROPROFILE_SCOPE(ARM_Jit_Unicorn); - CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions)); - system.CoreTiming().AddTicks(num_instructions); - if (GDBStub::IsServerEnabled()) { - if (last_bkpt_hit && last_bkpt.type == GDBStub::BreakpointType::Execute) { - uc_reg_write(uc, UC_ARM64_REG_PC, &last_bkpt.address); - } - - Kernel::Thread* const thread = system.CurrentScheduler().GetCurrentThread(); - SaveContext(thread->GetContext64()); - if (last_bkpt_hit || GDBStub::IsMemoryBreak() || GDBStub::GetCpuStepFlag()) { - last_bkpt_hit = false; - GDBStub::Break(); - GDBStub::SendTrap(thread, 5); - } - } -} - -void ARM_Unicorn::SaveContext(ThreadContext64& ctx) { - int uregs[32]; - void* tregs[32]; - - CHECKED(uc_reg_read(uc, UC_ARM64_REG_SP, &ctx.sp)); - CHECKED(uc_reg_read(uc, UC_ARM64_REG_PC, &ctx.pc)); - CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &ctx.pstate)); - - for (auto i = 0; i < 29; ++i) { - uregs[i] = UC_ARM64_REG_X0 + i; - tregs[i] = &ctx.cpu_registers[i]; - } - uregs[29] = UC_ARM64_REG_X29; - tregs[29] = (void*)&ctx.cpu_registers[29]; - uregs[30] = UC_ARM64_REG_X30; - tregs[30] = (void*)&ctx.cpu_registers[30]; - - CHECKED(uc_reg_read_batch(uc, uregs, tregs, 31)); - - for (int i = 0; i < 32; ++i) { - uregs[i] = UC_ARM64_REG_Q0 + i; - tregs[i] = &ctx.vector_registers[i]; - } - - CHECKED(uc_reg_read_batch(uc, uregs, tregs, 32)); -} - -void ARM_Unicorn::LoadContext(const ThreadContext64& ctx) { - int uregs[32]; - void* tregs[32]; - - CHECKED(uc_reg_write(uc, UC_ARM64_REG_SP, &ctx.sp)); - CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &ctx.pc)); - CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &ctx.pstate)); - - for (int i = 0; i < 29; ++i) { - uregs[i] = UC_ARM64_REG_X0 + i; - tregs[i] = (void*)&ctx.cpu_registers[i]; - } - uregs[29] = UC_ARM64_REG_X29; - tregs[29] = (void*)&ctx.cpu_registers[29]; - uregs[30] = UC_ARM64_REG_X30; - tregs[30] = (void*)&ctx.cpu_registers[30]; - - CHECKED(uc_reg_write_batch(uc, uregs, tregs, 31)); - - for (auto i = 0; i < 32; ++i) { - uregs[i] = UC_ARM64_REG_Q0 + i; - tregs[i] = (void*)&ctx.vector_registers[i]; - } - - CHECKED(uc_reg_write_batch(uc, uregs, tregs, 32)); -} - -void ARM_Unicorn::PrepareReschedule() { - CHECKED(uc_emu_stop(uc)); -} - -void ARM_Unicorn::ClearExclusiveState() {} - -void ARM_Unicorn::ClearInstructionCache() {} - -void ARM_Unicorn::RecordBreak(GDBStub::BreakpointAddress bkpt) { - last_bkpt = bkpt; - last_bkpt_hit = true; -} - -void ARM_Unicorn::InterruptHook(uc_engine* uc, u32 int_no, void* user_data) { - u32 esr{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_ESR, &esr)); - - const auto ec = esr >> 26; - const auto iss = esr & 0xFFFFFF; - - auto* const arm_instance = static_cast<ARM_Unicorn*>(user_data); - - switch (ec) { - case 0x15: // SVC - Kernel::CallSVC(arm_instance->system, iss); - break; - } -} - -} // namespace Core diff --git a/src/core/arm/unicorn/arm_unicorn.h b/src/core/arm/unicorn/arm_unicorn.h deleted file mode 100644 index f30d13cb6..000000000 --- a/src/core/arm/unicorn/arm_unicorn.h +++ /dev/null @@ -1,55 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <unicorn/unicorn.h> -#include "common/common_types.h" -#include "core/arm/arm_interface.h" -#include "core/gdbstub/gdbstub.h" - -namespace Core { - -class System; - -class ARM_Unicorn final : public ARM_Interface { -public: - explicit ARM_Unicorn(System& system); - ~ARM_Unicorn() override; - - void SetPC(u64 pc) override; - u64 GetPC() const override; - u64 GetReg(int index) const override; - void SetReg(int index, u64 value) override; - u128 GetVectorReg(int index) const override; - void SetVectorReg(int index, u128 value) override; - u32 GetPSTATE() const override; - void SetPSTATE(u32 pstate) override; - VAddr GetTlsAddress() const override; - void SetTlsAddress(VAddr address) override; - void SetTPIDR_EL0(u64 value) override; - u64 GetTPIDR_EL0() const override; - void PrepareReschedule() override; - void ClearExclusiveState() override; - void ExecuteInstructions(std::size_t num_instructions); - void Run() override; - void Step() override; - void ClearInstructionCache() override; - void PageTableChanged(Common::PageTable&, std::size_t) override {} - void RecordBreak(GDBStub::BreakpointAddress bkpt); - - void SaveContext(ThreadContext32& ctx) override {} - void SaveContext(ThreadContext64& ctx) override; - void LoadContext(const ThreadContext32& ctx) override {} - void LoadContext(const ThreadContext64& ctx) override; - -private: - static void InterruptHook(uc_engine* uc, u32 int_no, void* user_data); - - uc_engine* uc{}; - GDBStub::BreakpointAddress last_bkpt{}; - bool last_bkpt_hit = false; -}; - -} // namespace Core diff --git a/src/core/constants.h b/src/core/constants.h index 6d0ec022a..81c5cb279 100644 --- a/src/core/constants.h +++ b/src/core/constants.h @@ -4,6 +4,7 @@ #pragma once +#include <array> #include "common/common_types.h" // This is to consolidate system-wide constants that are used by multiple components of yuzu. diff --git a/src/core/core.cpp b/src/core/core.cpp index 3bd90d79f..5accdc783 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -8,12 +8,13 @@ #include "common/file_util.h" #include "common/logging/log.h" +#include "common/microprofile.h" #include "common/string_util.h" #include "core/arm/exclusive_monitor.h" #include "core/core.h" -#include "core/core_manager.h" #include "core/core_timing.h" #include "core/cpu_manager.h" +#include "core/device_memory.h" #include "core/file_sys/bis_factory.h" #include "core/file_sys/card_image.h" #include "core/file_sys/mode.h" @@ -39,9 +40,11 @@ #include "core/hle/service/lm/manager.h" #include "core/hle/service/service.h" #include "core/hle/service/sm/sm.h" +#include "core/hle/service/time/time_manager.h" #include "core/loader/loader.h" #include "core/memory.h" #include "core/memory/cheat_engine.h" +#include "core/network/network.h" #include "core/perf_stats.h" #include "core/reporter.h" #include "core/settings.h" @@ -50,6 +53,11 @@ #include "video_core/renderer_base.h" #include "video_core/video_core.h" +MICROPROFILE_DEFINE(ARM_Jit_Dynarmic_CPU0, "ARM JIT", "Dynarmic CPU 0", MP_RGB(255, 64, 64)); +MICROPROFILE_DEFINE(ARM_Jit_Dynarmic_CPU1, "ARM JIT", "Dynarmic CPU 1", MP_RGB(255, 64, 64)); +MICROPROFILE_DEFINE(ARM_Jit_Dynarmic_CPU2, "ARM JIT", "Dynarmic CPU 2", MP_RGB(255, 64, 64)); +MICROPROFILE_DEFINE(ARM_Jit_Dynarmic_CPU3, "ARM JIT", "Dynarmic CPU 3", MP_RGB(255, 64, 64)); + namespace Core { namespace { @@ -106,7 +114,7 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs, return FileSys::ConcatenatedVfsFile::MakeConcatenatedFile(concat, dir->GetName()); } - if (FileUtil::IsDirectory(path)) + if (Common::FS::IsDirectory(path)) return vfs->OpenFile(path + "/" + "main", FileSys::Mode::Read); return vfs->OpenFile(path, FileSys::Mode::Read); @@ -114,25 +122,24 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs, struct System::Impl { explicit Impl(System& system) : kernel{system}, fs_controller{system}, memory{system}, - cpu_manager{system}, reporter{system}, applet_manager{system} {} + cpu_manager{system}, reporter{system}, applet_manager{system}, time_manager{system} {} - CoreManager& CurrentCoreManager() { - return cpu_manager.GetCurrentCoreManager(); - } + ResultStatus Run() { + status = ResultStatus::Success; - Kernel::PhysicalCore& CurrentPhysicalCore() { - const auto index = cpu_manager.GetActiveCoreIndex(); - return kernel.PhysicalCore(index); - } + kernel.Suspend(false); + core_timing.SyncPause(false); + cpu_manager.Pause(false); - Kernel::PhysicalCore& GetPhysicalCore(std::size_t index) { - return kernel.PhysicalCore(index); + return status; } - ResultStatus RunLoop(bool tight_loop) { + ResultStatus Pause() { status = ResultStatus::Success; - cpu_manager.RunLoop(tight_loop); + core_timing.SyncPause(true); + kernel.Suspend(true); + cpu_manager.Pause(true); return status; } @@ -140,14 +147,24 @@ struct System::Impl { ResultStatus Init(System& system, Frontend::EmuWindow& emu_window) { LOG_DEBUG(HW_Memory, "initialized OK"); - core_timing.Initialize(); + device_memory = std::make_unique<Core::DeviceMemory>(); + + is_multicore = Settings::values.use_multi_core.GetValue(); + is_async_gpu = is_multicore || Settings::values.use_asynchronous_gpu_emulation.GetValue(); + + kernel.SetMulticore(is_multicore); + cpu_manager.SetMulticore(is_multicore); + cpu_manager.SetAsyncGpu(is_async_gpu); + core_timing.SetMulticore(is_multicore); + + core_timing.Initialize([&system]() { system.RegisterHostThread(); }); kernel.Initialize(); cpu_manager.Initialize(); const auto current_time = std::chrono::duration_cast<std::chrono::seconds>( std::chrono::system_clock::now().time_since_epoch()); Settings::values.custom_rtc_differential = - Settings::values.custom_rtc.value_or(current_time) - current_time; + Settings::values.custom_rtc.GetValue().value_or(current_time) - current_time; // Create a default fs if one doesn't already exist. if (virtual_filesystem == nullptr) @@ -162,21 +179,30 @@ struct System::Impl { arp_manager.ResetAll(); telemetry_session = std::make_unique<Core::TelemetrySession>(); - service_manager = std::make_shared<Service::SM::ServiceManager>(); - Service::Init(service_manager, system); - GDBStub::DeferStart(); - - interrupt_manager = std::make_unique<Core::Hardware::InterruptManager>(system); gpu_core = VideoCore::CreateGPU(emu_window, system); if (!gpu_core) { return ResultStatus::ErrorVideoCore; } - gpu_core->Renderer().Rasterizer().SetupDirtyFlags(); + + service_manager = std::make_shared<Service::SM::ServiceManager>(kernel); + + Service::Init(service_manager, system); + GDBStub::DeferStart(); + + interrupt_manager = std::make_unique<Core::Hardware::InterruptManager>(system); + + // Initialize time manager, which must happen after kernel is created + time_manager.Initialize(); is_powered_on = true; exit_lock = false; + microprofile_dynarmic[0] = MICROPROFILE_TOKEN(ARM_Jit_Dynarmic_CPU0); + microprofile_dynarmic[1] = MICROPROFILE_TOKEN(ARM_Jit_Dynarmic_CPU1); + microprofile_dynarmic[2] = MICROPROFILE_TOKEN(ARM_Jit_Dynarmic_CPU2); + microprofile_dynarmic[3] = MICROPROFILE_TOKEN(ARM_Jit_Dynarmic_CPU3); + LOG_DEBUG(Core, "Initialized OK"); return ResultStatus::Success; @@ -184,7 +210,7 @@ struct System::Impl { ResultStatus Load(System& system, Frontend::EmuWindow& emu_window, const std::string& filepath) { - app_loader = Loader::GetLoader(GetGameFileFromPath(virtual_filesystem, filepath)); + app_loader = Loader::GetLoader(system, GetGameFileFromPath(virtual_filesystem, filepath)); if (!app_loader) { LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath); return ResultStatus::ErrorGetLoader; @@ -198,10 +224,10 @@ struct System::Impl { return init_result; } - telemetry_session->AddInitialInfo(*app_loader); + telemetry_session->AddInitialInfo(*app_loader, fs_controller, *content_provider); auto main_process = Kernel::Process::Create(system, "main", Kernel::Process::ProcessType::Userland); - const auto [load_result, load_parameters] = app_loader->Load(*main_process); + const auto [load_result, load_parameters] = app_loader->Load(*main_process, system); if (load_result != Loader::ResultStatus::Success) { LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", static_cast<int>(load_result)); Shutdown(); @@ -248,14 +274,14 @@ struct System::Impl { // Log last frame performance stats if game was loded if (perf_stats) { const auto perf_results = GetAndResetPerfStats(); - telemetry_session->AddField(Telemetry::FieldType::Performance, - "Shutdown_EmulationSpeed", + constexpr auto performance = Common::Telemetry::FieldType::Performance; + + telemetry_session->AddField(performance, "Shutdown_EmulationSpeed", perf_results.emulation_speed * 100.0); - telemetry_session->AddField(Telemetry::FieldType::Performance, "Shutdown_Framerate", - perf_results.game_fps); - telemetry_session->AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime", + telemetry_session->AddField(performance, "Shutdown_Framerate", perf_results.game_fps); + telemetry_session->AddField(performance, "Shutdown_Frametime", perf_results.frametime * 1000.0); - telemetry_session->AddField(Telemetry::FieldType::Performance, "Mean_Frametime_MS", + telemetry_session->AddField(performance, "Mean_Frametime_MS", perf_stats->GetMeanFrametime()); } @@ -274,8 +300,7 @@ struct System::Impl { service_manager.reset(); cheat_engine.reset(); telemetry_session.reset(); - perf_stats.reset(); - gpu_core.reset(); + device_memory.reset(); // Close all CPU/threading state cpu_manager.Shutdown(); @@ -286,6 +311,8 @@ struct System::Impl { // Close app loader app_loader.reset(); + gpu_core.reset(); + perf_stats.reset(); // Clear all applets applet_manager.ClearAll(); @@ -311,7 +338,7 @@ struct System::Impl { Service::Glue::ApplicationLaunchProperty launch{}; launch.title_id = process.GetTitleID(); - FileSys::PatchManager pm{launch.title_id}; + FileSys::PatchManager pm{launch.title_id, fs_controller, *content_provider}; launch.version = pm.GetGameVersion().value_or(0); // TODO(DarkLordZach): When FSController/Game Card Support is added, if @@ -346,7 +373,8 @@ struct System::Impl { std::unique_ptr<Loader::AppLoader> app_loader; std::unique_ptr<Tegra::GPU> gpu_core; std::unique_ptr<Hardware::InterruptManager> interrupt_manager; - Memory::Memory memory; + std::unique_ptr<Core::DeviceMemory> device_memory; + Core::Memory::Memory memory; CpuManager cpu_manager; bool is_powered_on = false; bool exit_lock = false; @@ -365,6 +393,7 @@ struct System::Impl { /// Service State Service::Glue::ARPManager arp_manager; Service::LM::Manager lm_manager{reporter}; + Service::Time::TimeManager time_manager; /// Service manager std::shared_ptr<Service::SM::ServiceManager> service_manager; @@ -372,36 +401,53 @@ struct System::Impl { /// Telemetry session for this emulation session std::unique_ptr<Core::TelemetrySession> telemetry_session; + /// Network instance + Network::NetworkInstance network_instance; + ResultStatus status = ResultStatus::Success; std::string status_details = ""; std::unique_ptr<Core::PerfStats> perf_stats; Core::FrameLimiter frame_limiter; + + bool is_multicore{}; + bool is_async_gpu{}; + + std::array<u64, Core::Hardware::NUM_CPU_CORES> dynarmic_ticks{}; + std::array<MicroProfileToken, Core::Hardware::NUM_CPU_CORES> microprofile_dynarmic{}; }; System::System() : impl{std::make_unique<Impl>(*this)} {} System::~System() = default; -CoreManager& System::CurrentCoreManager() { - return impl->CurrentCoreManager(); +CpuManager& System::GetCpuManager() { + return impl->cpu_manager; +} + +const CpuManager& System::GetCpuManager() const { + return impl->cpu_manager; } -const CoreManager& System::CurrentCoreManager() const { - return impl->CurrentCoreManager(); +System::ResultStatus System::Run() { + return impl->Run(); } -System::ResultStatus System::RunLoop(bool tight_loop) { - return impl->RunLoop(tight_loop); +System::ResultStatus System::Pause() { + return impl->Pause(); } System::ResultStatus System::SingleStep() { - return RunLoop(false); + return ResultStatus::Success; } void System::InvalidateCpuInstructionCaches() { impl->kernel.InvalidateAllInstructionCaches(); } +void System::Shutdown() { + impl->Shutdown(); +} + System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath) { return impl->Load(*this, emu_window, filepath); } @@ -411,7 +457,7 @@ bool System::IsPoweredOn() const { } void System::PrepareReschedule() { - impl->CurrentPhysicalCore().Stop(); + // Deprecated, does nothing, kept for backward compatibility. } void System::PrepareReschedule(const u32 core_index) { @@ -431,31 +477,41 @@ const TelemetrySession& System::TelemetrySession() const { } ARM_Interface& System::CurrentArmInterface() { - return impl->CurrentPhysicalCore().ArmInterface(); + return impl->kernel.CurrentScheduler().GetCurrentThread()->ArmInterface(); } const ARM_Interface& System::CurrentArmInterface() const { - return impl->CurrentPhysicalCore().ArmInterface(); + return impl->kernel.CurrentScheduler().GetCurrentThread()->ArmInterface(); } std::size_t System::CurrentCoreIndex() const { - return impl->cpu_manager.GetActiveCoreIndex(); + std::size_t core = impl->kernel.GetCurrentHostThreadID(); + ASSERT(core < Core::Hardware::NUM_CPU_CORES); + return core; } Kernel::Scheduler& System::CurrentScheduler() { - return impl->CurrentPhysicalCore().Scheduler(); + return impl->kernel.CurrentScheduler(); } const Kernel::Scheduler& System::CurrentScheduler() const { - return impl->CurrentPhysicalCore().Scheduler(); + return impl->kernel.CurrentScheduler(); +} + +Kernel::PhysicalCore& System::CurrentPhysicalCore() { + return impl->kernel.CurrentPhysicalCore(); +} + +const Kernel::PhysicalCore& System::CurrentPhysicalCore() const { + return impl->kernel.CurrentPhysicalCore(); } Kernel::Scheduler& System::Scheduler(std::size_t core_index) { - return impl->GetPhysicalCore(core_index).Scheduler(); + return impl->kernel.Scheduler(core_index); } const Kernel::Scheduler& System::Scheduler(std::size_t core_index) const { - return impl->GetPhysicalCore(core_index).Scheduler(); + return impl->kernel.Scheduler(core_index); } /// Gets the global scheduler @@ -472,25 +528,28 @@ Kernel::Process* System::CurrentProcess() { return impl->kernel.CurrentProcess(); } -const Kernel::Process* System::CurrentProcess() const { - return impl->kernel.CurrentProcess(); +Core::DeviceMemory& System::DeviceMemory() { + return *impl->device_memory; } -ARM_Interface& System::ArmInterface(std::size_t core_index) { - return impl->GetPhysicalCore(core_index).ArmInterface(); +const Core::DeviceMemory& System::DeviceMemory() const { + return *impl->device_memory; } -const ARM_Interface& System::ArmInterface(std::size_t core_index) const { - return impl->GetPhysicalCore(core_index).ArmInterface(); +const Kernel::Process* System::CurrentProcess() const { + return impl->kernel.CurrentProcess(); } -CoreManager& System::GetCoreManager(std::size_t core_index) { - return impl->cpu_manager.GetCoreManager(core_index); +ARM_Interface& System::ArmInterface(std::size_t core_index) { + auto* thread = impl->kernel.Scheduler(core_index).GetCurrentThread(); + ASSERT(thread && !thread->IsHLEThread()); + return thread->ArmInterface(); } -const CoreManager& System::GetCoreManager(std::size_t core_index) const { - ASSERT(core_index < NUM_CPU_CORES); - return impl->cpu_manager.GetCoreManager(core_index); +const ARM_Interface& System::ArmInterface(std::size_t core_index) const { + auto* thread = impl->kernel.Scheduler(core_index).GetCurrentThread(); + ASSERT(thread && !thread->IsHLEThread()); + return thread->ArmInterface(); } ExclusiveMonitor& System::Monitor() { @@ -505,7 +564,7 @@ Memory::Memory& System::Memory() { return impl->memory; } -const Memory::Memory& System::Memory() const { +const Core::Memory::Memory& System::Memory() const { return impl->memory; } @@ -577,15 +636,19 @@ const std::string& System::GetStatusDetails() const { return impl->status_details; } -Loader::AppLoader& System::GetAppLoader() const { +Loader::AppLoader& System::GetAppLoader() { + return *impl->app_loader; +} + +const Loader::AppLoader& System::GetAppLoader() const { return *impl->app_loader; } -void System::SetFilesystem(std::shared_ptr<FileSys::VfsFilesystem> vfs) { +void System::SetFilesystem(FileSys::VirtualFilesystem vfs) { impl->virtual_filesystem = std::move(vfs); } -std::shared_ptr<FileSys::VfsFilesystem> System::GetFilesystem() const { +FileSys::VirtualFilesystem System::GetFilesystem() const { return impl->virtual_filesystem; } @@ -669,6 +732,14 @@ const Service::LM::Manager& System::GetLogManager() const { return impl->lm_manager; } +Service::Time::TimeManager& System::GetTimeManager() { + return impl->time_manager; +} + +const Service::Time::TimeManager& System::GetTimeManager() const { + return impl->time_manager; +} + void System::SetExitLock(bool locked) { impl->exit_lock = locked; } @@ -685,14 +756,6 @@ const System::CurrentBuildProcessID& System::GetCurrentProcessBuildID() const { return impl->build_id; } -System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) { - return impl->Init(*this, emu_window); -} - -void System::Shutdown() { - impl->Shutdown(); -} - Service::SM::ServiceManager& System::ServiceManager() { return *impl->service_manager; } @@ -709,4 +772,18 @@ void System::RegisterHostThread() { impl->kernel.RegisterHostThread(); } +void System::EnterDynarmicProfile() { + std::size_t core = impl->kernel.GetCurrentHostThreadID(); + impl->dynarmic_ticks[core] = MicroProfileEnter(impl->microprofile_dynarmic[core]); +} + +void System::ExitDynarmicProfile() { + std::size_t core = impl->kernel.GetCurrentHostThreadID(); + MicroProfileLeave(impl->microprofile_dynarmic[core], impl->dynarmic_ticks[core]); +} + +bool System::IsMulticore() const { + return impl->is_multicore; +} + } // namespace Core diff --git a/src/core/core.h b/src/core/core.h index 8d862a8e6..cd155625c 100644 --- a/src/core/core.h +++ b/src/core/core.h @@ -27,6 +27,7 @@ class VfsFilesystem; namespace Kernel { class GlobalScheduler; class KernelCore; +class PhysicalCore; class Process; class Scheduler; } // namespace Kernel @@ -36,9 +37,10 @@ class AppLoader; enum class ResultStatus : u16; } // namespace Loader -namespace Memory { +namespace Core::Memory { struct CheatEntry; -} // namespace Memory +class Memory; +} // namespace Core::Memory namespace Service { @@ -67,6 +69,10 @@ namespace SM { class ServiceManager; } // namespace SM +namespace Time { +class TimeManager; +} // namespace Time + } // namespace Service namespace Tegra { @@ -86,14 +92,11 @@ namespace Core::Hardware { class InterruptManager; } -namespace Memory { -class Memory; -} - namespace Core { class ARM_Interface; -class CoreManager; +class CpuManager; +class DeviceMemory; class ExclusiveMonitor; class FrameLimiter; class PerfStats; @@ -121,7 +124,7 @@ public: * Gets the instance of the System singleton class. * @returns Reference to the instance of the System singleton class. */ - static System& GetInstance() { + [[deprecated("Use of the global system instance is deprecated")]] static System& GetInstance() { return s_instance; } @@ -138,22 +141,22 @@ public: }; /** - * Run the core CPU loop - * This function runs the core for the specified number of CPU instructions before trying to - * update hardware. This is much faster than SingleStep (and should be equivalent), as the CPU - * is not required to do a full dispatch with each instruction. NOTE: the number of instructions - * requested is not guaranteed to run, as this will be interrupted preemptively if a hardware - * update is requested (e.g. on a thread switch). - * @param tight_loop If false, the CPU single-steps. - * @return Result status, indicating whether or not the operation succeeded. + * Run the OS and Application + * This function will start emulation and run the relevant devices */ - ResultStatus RunLoop(bool tight_loop = true); + [[nodiscard]] ResultStatus Run(); + + /** + * Pause the OS and Application + * This function will pause emulation and stop the relevant devices + */ + [[nodiscard]] ResultStatus Pause(); /** * Step the CPU one instruction * @return Result status, indicating whether or not the operation succeeded. */ - ResultStatus SingleStep(); + [[nodiscard]] ResultStatus SingleStep(); /** * Invalidate the CPU instruction caches @@ -172,20 +175,20 @@ public: * @param filepath String path to the executable application to load on the host file system. * @returns ResultStatus code, indicating if the operation succeeded. */ - ResultStatus Load(Frontend::EmuWindow& emu_window, const std::string& filepath); + [[nodiscard]] ResultStatus Load(Frontend::EmuWindow& emu_window, const std::string& filepath); /** * Indicates if the emulated system is powered on (all subsystems initialized and able to run an * application). * @returns True if the emulated system is powered on, otherwise false. */ - bool IsPoweredOn() const; + [[nodiscard]] bool IsPoweredOn() const; /// Gets a reference to the telemetry session for this emulation session. - Core::TelemetrySession& TelemetrySession(); + [[nodiscard]] Core::TelemetrySession& TelemetrySession(); /// Gets a reference to the telemetry session for this emulation session. - const Core::TelemetrySession& TelemetrySession() const; + [[nodiscard]] const Core::TelemetrySession& TelemetrySession() const; /// Prepare the core emulation for a reschedule void PrepareReschedule(); @@ -194,171 +197,178 @@ public: void PrepareReschedule(u32 core_index); /// Gets and resets core performance statistics - PerfStatsResults GetAndResetPerfStats(); + [[nodiscard]] PerfStatsResults GetAndResetPerfStats(); /// Gets an ARM interface to the CPU core that is currently running - ARM_Interface& CurrentArmInterface(); + [[nodiscard]] ARM_Interface& CurrentArmInterface(); /// Gets an ARM interface to the CPU core that is currently running - const ARM_Interface& CurrentArmInterface() const; + [[nodiscard]] const ARM_Interface& CurrentArmInterface() const; /// Gets the index of the currently running CPU core - std::size_t CurrentCoreIndex() const; + [[nodiscard]] std::size_t CurrentCoreIndex() const; /// Gets the scheduler for the CPU core that is currently running - Kernel::Scheduler& CurrentScheduler(); + [[nodiscard]] Kernel::Scheduler& CurrentScheduler(); /// Gets the scheduler for the CPU core that is currently running - const Kernel::Scheduler& CurrentScheduler() const; + [[nodiscard]] const Kernel::Scheduler& CurrentScheduler() const; + + /// Gets the physical core for the CPU core that is currently running + [[nodiscard]] Kernel::PhysicalCore& CurrentPhysicalCore(); + + /// Gets the physical core for the CPU core that is currently running + [[nodiscard]] const Kernel::PhysicalCore& CurrentPhysicalCore() const; /// Gets a reference to an ARM interface for the CPU core with the specified index - ARM_Interface& ArmInterface(std::size_t core_index); + [[nodiscard]] ARM_Interface& ArmInterface(std::size_t core_index); /// Gets a const reference to an ARM interface from the CPU core with the specified index - const ARM_Interface& ArmInterface(std::size_t core_index) const; + [[nodiscard]] const ARM_Interface& ArmInterface(std::size_t core_index) const; - /// Gets a CPU interface to the CPU core with the specified index - CoreManager& GetCoreManager(std::size_t core_index); + /// Gets a reference to the underlying CPU manager. + [[nodiscard]] CpuManager& GetCpuManager(); - /// Gets a CPU interface to the CPU core with the specified index - const CoreManager& GetCoreManager(std::size_t core_index) const; + /// Gets a const reference to the underlying CPU manager + [[nodiscard]] const CpuManager& GetCpuManager() const; /// Gets a reference to the exclusive monitor - ExclusiveMonitor& Monitor(); + [[nodiscard]] ExclusiveMonitor& Monitor(); /// Gets a constant reference to the exclusive monitor - const ExclusiveMonitor& Monitor() const; + [[nodiscard]] const ExclusiveMonitor& Monitor() const; /// Gets a mutable reference to the system memory instance. - Memory::Memory& Memory(); + [[nodiscard]] Core::Memory::Memory& Memory(); /// Gets a constant reference to the system memory instance. - const Memory::Memory& Memory() const; + [[nodiscard]] const Core::Memory::Memory& Memory() const; /// Gets a mutable reference to the GPU interface - Tegra::GPU& GPU(); + [[nodiscard]] Tegra::GPU& GPU(); /// Gets an immutable reference to the GPU interface. - const Tegra::GPU& GPU() const; + [[nodiscard]] const Tegra::GPU& GPU() const; /// Gets a mutable reference to the renderer. - VideoCore::RendererBase& Renderer(); + [[nodiscard]] VideoCore::RendererBase& Renderer(); /// Gets an immutable reference to the renderer. - const VideoCore::RendererBase& Renderer() const; + [[nodiscard]] const VideoCore::RendererBase& Renderer() const; /// Gets the scheduler for the CPU core with the specified index - Kernel::Scheduler& Scheduler(std::size_t core_index); + [[nodiscard]] Kernel::Scheduler& Scheduler(std::size_t core_index); /// Gets the scheduler for the CPU core with the specified index - const Kernel::Scheduler& Scheduler(std::size_t core_index) const; + [[nodiscard]] const Kernel::Scheduler& Scheduler(std::size_t core_index) const; /// Gets the global scheduler - Kernel::GlobalScheduler& GlobalScheduler(); + [[nodiscard]] Kernel::GlobalScheduler& GlobalScheduler(); /// Gets the global scheduler - const Kernel::GlobalScheduler& GlobalScheduler() const; + [[nodiscard]] const Kernel::GlobalScheduler& GlobalScheduler() const; + + /// Gets the manager for the guest device memory + [[nodiscard]] Core::DeviceMemory& DeviceMemory(); + + /// Gets the manager for the guest device memory + [[nodiscard]] const Core::DeviceMemory& DeviceMemory() const; /// Provides a pointer to the current process - Kernel::Process* CurrentProcess(); + [[nodiscard]] Kernel::Process* CurrentProcess(); /// Provides a constant pointer to the current process. - const Kernel::Process* CurrentProcess() const; + [[nodiscard]] const Kernel::Process* CurrentProcess() const; /// Provides a reference to the core timing instance. - Timing::CoreTiming& CoreTiming(); + [[nodiscard]] Timing::CoreTiming& CoreTiming(); /// Provides a constant reference to the core timing instance. - const Timing::CoreTiming& CoreTiming() const; + [[nodiscard]] const Timing::CoreTiming& CoreTiming() const; /// Provides a reference to the interrupt manager instance. - Core::Hardware::InterruptManager& InterruptManager(); + [[nodiscard]] Core::Hardware::InterruptManager& InterruptManager(); /// Provides a constant reference to the interrupt manager instance. - const Core::Hardware::InterruptManager& InterruptManager() const; + [[nodiscard]] const Core::Hardware::InterruptManager& InterruptManager() const; /// Provides a reference to the kernel instance. - Kernel::KernelCore& Kernel(); + [[nodiscard]] Kernel::KernelCore& Kernel(); /// Provides a constant reference to the kernel instance. - const Kernel::KernelCore& Kernel() const; + [[nodiscard]] const Kernel::KernelCore& Kernel() const; /// Provides a reference to the internal PerfStats instance. - Core::PerfStats& GetPerfStats(); + [[nodiscard]] Core::PerfStats& GetPerfStats(); /// Provides a constant reference to the internal PerfStats instance. - const Core::PerfStats& GetPerfStats() const; + [[nodiscard]] const Core::PerfStats& GetPerfStats() const; /// Provides a reference to the frame limiter; - Core::FrameLimiter& FrameLimiter(); + [[nodiscard]] Core::FrameLimiter& FrameLimiter(); /// Provides a constant referent to the frame limiter - const Core::FrameLimiter& FrameLimiter() const; + [[nodiscard]] const Core::FrameLimiter& FrameLimiter() const; /// Gets the name of the current game - Loader::ResultStatus GetGameName(std::string& out) const; + [[nodiscard]] Loader::ResultStatus GetGameName(std::string& out) const; void SetStatus(ResultStatus new_status, const char* details); - const std::string& GetStatusDetails() const; + [[nodiscard]] const std::string& GetStatusDetails() const; - Loader::AppLoader& GetAppLoader() const; + [[nodiscard]] Loader::AppLoader& GetAppLoader(); + [[nodiscard]] const Loader::AppLoader& GetAppLoader() const; - Service::SM::ServiceManager& ServiceManager(); - const Service::SM::ServiceManager& ServiceManager() const; + [[nodiscard]] Service::SM::ServiceManager& ServiceManager(); + [[nodiscard]] const Service::SM::ServiceManager& ServiceManager() const; - void SetFilesystem(std::shared_ptr<FileSys::VfsFilesystem> vfs); + void SetFilesystem(FileSys::VirtualFilesystem vfs); - std::shared_ptr<FileSys::VfsFilesystem> GetFilesystem() const; + [[nodiscard]] FileSys::VirtualFilesystem GetFilesystem() const; void RegisterCheatList(const std::vector<Memory::CheatEntry>& list, const std::array<u8, 0x20>& build_id, VAddr main_region_begin, u64 main_region_size); void SetAppletFrontendSet(Service::AM::Applets::AppletFrontendSet&& set); - void SetDefaultAppletFrontendSet(); - Service::AM::Applets::AppletManager& GetAppletManager(); - - const Service::AM::Applets::AppletManager& GetAppletManager() const; + [[nodiscard]] Service::AM::Applets::AppletManager& GetAppletManager(); + [[nodiscard]] const Service::AM::Applets::AppletManager& GetAppletManager() const; void SetContentProvider(std::unique_ptr<FileSys::ContentProviderUnion> provider); - FileSys::ContentProvider& GetContentProvider(); + [[nodiscard]] FileSys::ContentProvider& GetContentProvider(); + [[nodiscard]] const FileSys::ContentProvider& GetContentProvider() const; - const FileSys::ContentProvider& GetContentProvider() const; - - Service::FileSystem::FileSystemController& GetFileSystemController(); - - const Service::FileSystem::FileSystemController& GetFileSystemController() const; + [[nodiscard]] Service::FileSystem::FileSystemController& GetFileSystemController(); + [[nodiscard]] const Service::FileSystem::FileSystemController& GetFileSystemController() const; void RegisterContentProvider(FileSys::ContentProviderUnionSlot slot, FileSys::ContentProvider* provider); void ClearContentProvider(FileSys::ContentProviderUnionSlot slot); - const Reporter& GetReporter() const; - - Service::Glue::ARPManager& GetARPManager(); - - const Service::Glue::ARPManager& GetARPManager() const; + [[nodiscard]] const Reporter& GetReporter() const; - Service::APM::Controller& GetAPMController(); + [[nodiscard]] Service::Glue::ARPManager& GetARPManager(); + [[nodiscard]] const Service::Glue::ARPManager& GetARPManager() const; - const Service::APM::Controller& GetAPMController() const; + [[nodiscard]] Service::APM::Controller& GetAPMController(); + [[nodiscard]] const Service::APM::Controller& GetAPMController() const; - Service::LM::Manager& GetLogManager(); + [[nodiscard]] Service::LM::Manager& GetLogManager(); + [[nodiscard]] const Service::LM::Manager& GetLogManager() const; - const Service::LM::Manager& GetLogManager() const; + [[nodiscard]] Service::Time::TimeManager& GetTimeManager(); + [[nodiscard]] const Service::Time::TimeManager& GetTimeManager() const; void SetExitLock(bool locked); - - bool GetExitLock() const; + [[nodiscard]] bool GetExitLock() const; void SetCurrentProcessBuildID(const CurrentBuildProcessID& id); - - const CurrentBuildProcessID& GetCurrentProcessBuildID() const; + [[nodiscard]] const CurrentBuildProcessID& GetCurrentProcessBuildID() const; /// Register a host thread as an emulated CPU Core. void RegisterCoreThread(std::size_t id); @@ -366,22 +376,17 @@ public: /// Register a host thread as an auxiliary thread. void RegisterHostThread(); -private: - System(); + /// Enter Dynarmic Microprofile + void EnterDynarmicProfile(); - /// Returns the currently running CPU core - CoreManager& CurrentCoreManager(); + /// Exit Dynarmic Microprofile + void ExitDynarmicProfile(); - /// Returns the currently running CPU core - const CoreManager& CurrentCoreManager() const; + /// Tells if system is running on multicore. + [[nodiscard]] bool IsMulticore() const; - /** - * Initialize the emulated system. - * @param emu_window Reference to the host-system window used for video output and keyboard - * input. - * @return ResultStatus code, indicating if the operation succeeded. - */ - ResultStatus Init(Frontend::EmuWindow& emu_window); +private: + System(); struct Impl; std::unique_ptr<Impl> impl; diff --git a/src/core/core_manager.cpp b/src/core/core_manager.cpp deleted file mode 100644 index b6b797c80..000000000 --- a/src/core/core_manager.cpp +++ /dev/null @@ -1,67 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <condition_variable> -#include <mutex> - -#include "common/logging/log.h" -#include "core/arm/exclusive_monitor.h" -#include "core/arm/unicorn/arm_unicorn.h" -#include "core/core.h" -#include "core/core_manager.h" -#include "core/core_timing.h" -#include "core/hle/kernel/kernel.h" -#include "core/hle/kernel/physical_core.h" -#include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/thread.h" -#include "core/hle/lock.h" -#include "core/settings.h" - -namespace Core { - -CoreManager::CoreManager(System& system, std::size_t core_index) - : global_scheduler{system.GlobalScheduler()}, physical_core{system.Kernel().PhysicalCore( - core_index)}, - core_timing{system.CoreTiming()}, core_index{core_index} {} - -CoreManager::~CoreManager() = default; - -void CoreManager::RunLoop(bool tight_loop) { - Reschedule(); - - // If we don't have a currently active thread then don't execute instructions, - // instead advance to the next event and try to yield to the next thread - if (Kernel::GetCurrentThread() == nullptr) { - LOG_TRACE(Core, "Core-{} idling", core_index); - core_timing.Idle(); - } else { - if (tight_loop) { - physical_core.Run(); - } else { - physical_core.Step(); - } - } - core_timing.Advance(); - - Reschedule(); -} - -void CoreManager::SingleStep() { - return RunLoop(false); -} - -void CoreManager::PrepareReschedule() { - physical_core.Stop(); -} - -void CoreManager::Reschedule() { - // Lock the global kernel mutex when we manipulate the HLE state - std::lock_guard lock(HLE::g_hle_lock); - - global_scheduler.SelectThread(core_index); - - physical_core.Scheduler().TryDoContextSwitch(); -} - -} // namespace Core diff --git a/src/core/core_manager.h b/src/core/core_manager.h deleted file mode 100644 index b14e723d7..000000000 --- a/src/core/core_manager.h +++ /dev/null @@ -1,63 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <atomic> -#include <cstddef> -#include <memory> -#include "common/common_types.h" - -namespace Kernel { -class GlobalScheduler; -class PhysicalCore; -} // namespace Kernel - -namespace Core { -class System; -} - -namespace Core::Timing { -class CoreTiming; -} - -namespace Memory { -class Memory; -} - -namespace Core { - -constexpr unsigned NUM_CPU_CORES{4}; - -class CoreManager { -public: - CoreManager(System& system, std::size_t core_index); - ~CoreManager(); - - void RunLoop(bool tight_loop = true); - - void SingleStep(); - - void PrepareReschedule(); - - bool IsMainCore() const { - return core_index == 0; - } - - std::size_t CoreIndex() const { - return core_index; - } - -private: - void Reschedule(); - - Kernel::GlobalScheduler& global_scheduler; - Kernel::PhysicalCore& physical_core; - Timing::CoreTiming& core_timing; - - std::atomic<bool> reschedule_pending = false; - std::size_t core_index; -}; - -} // namespace Core diff --git a/src/core/core_timing.cpp b/src/core/core_timing.cpp index 46d4178c4..e6c8461a5 100644 --- a/src/core/core_timing.cpp +++ b/src/core/core_timing.cpp @@ -1,31 +1,29 @@ -// Copyright 2008 Dolphin Emulator Project / 2017 Citra Emulator Project -// Licensed under GPLv2+ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "core/core_timing.h" - #include <algorithm> #include <mutex> #include <string> #include <tuple> -#include "common/assert.h" -#include "common/thread.h" +#include "common/microprofile.h" +#include "core/core_timing.h" #include "core/core_timing_util.h" #include "core/hardware_properties.h" namespace Core::Timing { -constexpr int MAX_SLICE_LENGTH = 10000; +constexpr s64 MAX_SLICE_LENGTH = 4000; std::shared_ptr<EventType> CreateEvent(std::string name, TimedCallback&& callback) { return std::make_shared<EventType>(std::move(callback), std::move(name)); } struct CoreTiming::Event { - s64 time; + u64 time; u64 fifo_order; - u64 userdata; + std::uintptr_t user_data; std::weak_ptr<EventType> type; // Sort by time, unless the times are the same, in which case sort by @@ -39,53 +37,92 @@ struct CoreTiming::Event { } }; -CoreTiming::CoreTiming() = default; -CoreTiming::~CoreTiming() = default; +CoreTiming::CoreTiming() + : clock{Common::CreateBestMatchingClock(Hardware::BASE_CLOCK_RATE, Hardware::CNTFREQ)} {} -void CoreTiming::Initialize() { - downcounts.fill(MAX_SLICE_LENGTH); - time_slice.fill(MAX_SLICE_LENGTH); - slice_length = MAX_SLICE_LENGTH; - global_timer = 0; - idled_cycles = 0; - current_context = 0; +CoreTiming::~CoreTiming() = default; - // The time between CoreTiming being initialized and the first call to Advance() is considered - // the slice boundary between slice -1 and slice 0. Dispatcher loops must call Advance() before - // executing the first cycle of each slice to prepare the slice length and downcount for - // that slice. - is_global_timer_sane = true; +void CoreTiming::ThreadEntry(CoreTiming& instance) { + constexpr char name[] = "yuzu:HostTiming"; + MicroProfileOnThreadCreate(name); + Common::SetCurrentThreadName(name); + Common::SetCurrentThreadPriority(Common::ThreadPriority::VeryHigh); + instance.on_thread_init(); + instance.ThreadLoop(); +} +void CoreTiming::Initialize(std::function<void()>&& on_thread_init_) { + on_thread_init = std::move(on_thread_init_); event_fifo_id = 0; - - const auto empty_timed_callback = [](u64, s64) {}; + shutting_down = false; + ticks = 0; + const auto empty_timed_callback = [](std::uintptr_t, std::chrono::nanoseconds) {}; ev_lost = CreateEvent("_lost_event", empty_timed_callback); + if (is_multicore) { + timer_thread = std::make_unique<std::thread>(ThreadEntry, std::ref(*this)); + } } void CoreTiming::Shutdown() { + paused = true; + shutting_down = true; + pause_event.Set(); + event.Set(); + if (timer_thread) { + timer_thread->join(); + } ClearPendingEvents(); + timer_thread.reset(); + has_started = false; } -void CoreTiming::ScheduleEvent(s64 cycles_into_future, const std::shared_ptr<EventType>& event_type, - u64 userdata) { - std::lock_guard guard{inner_mutex}; - const s64 timeout = GetTicks() + cycles_into_future; +void CoreTiming::Pause(bool is_paused) { + paused = is_paused; + pause_event.Set(); +} - // If this event needs to be scheduled before the next advance(), force one early - if (!is_global_timer_sane) { - ForceExceptionCheck(cycles_into_future); +void CoreTiming::SyncPause(bool is_paused) { + if (is_paused == paused && paused_set == paused) { + return; } + Pause(is_paused); + if (timer_thread) { + if (!is_paused) { + pause_event.Set(); + } + event.Set(); + while (paused_set != is_paused) + ; + } +} - event_queue.emplace_back(Event{timeout, event_fifo_id++, userdata, event_type}); +bool CoreTiming::IsRunning() const { + return !paused_set; +} - std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>()); +bool CoreTiming::HasPendingEvents() const { + return !(wait_set && event_queue.empty()); } -void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type, u64 userdata) { - std::lock_guard guard{inner_mutex}; +void CoreTiming::ScheduleEvent(std::chrono::nanoseconds ns_into_future, + const std::shared_ptr<EventType>& event_type, + std::uintptr_t user_data) { + { + std::scoped_lock scope{basic_lock}; + const u64 timeout = static_cast<u64>((GetGlobalTimeNs() + ns_into_future).count()); + + event_queue.emplace_back(Event{timeout, event_fifo_id++, user_data, event_type}); + std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>()); + } + event.Set(); +} + +void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type, + std::uintptr_t user_data) { + std::scoped_lock scope{basic_lock}; const auto itr = std::remove_if(event_queue.begin(), event_queue.end(), [&](const Event& e) { - return e.type.lock().get() == event_type.get() && e.userdata == userdata; + return e.type.lock().get() == event_type.get() && e.user_data == user_data; }); // Removing random items breaks the invariant so we have to re-establish it. @@ -95,21 +132,39 @@ void CoreTiming::UnscheduleEvent(const std::shared_ptr<EventType>& event_type, u } } -u64 CoreTiming::GetTicks() const { - u64 ticks = static_cast<u64>(global_timer); - if (!is_global_timer_sane) { - ticks += accumulated_ticks; +void CoreTiming::AddTicks(u64 ticks) { + this->ticks += ticks; + downcount -= static_cast<s64>(ticks); +} + +void CoreTiming::Idle() { + if (!event_queue.empty()) { + const u64 next_event_time = event_queue.front().time; + const u64 next_ticks = nsToCycles(std::chrono::nanoseconds(next_event_time)) + 10U; + if (next_ticks > ticks) { + ticks = next_ticks; + } + return; } - return ticks; + ticks += 1000U; } -u64 CoreTiming::GetIdleTicks() const { - return static_cast<u64>(idled_cycles); +void CoreTiming::ResetTicks() { + downcount = MAX_SLICE_LENGTH; } -void CoreTiming::AddTicks(u64 ticks) { - accumulated_ticks += ticks; - downcounts[current_context] -= static_cast<s64>(ticks); +u64 CoreTiming::GetCPUTicks() const { + if (is_multicore) { + return clock->GetCPUCycles(); + } + return ticks; +} + +u64 CoreTiming::GetClockTicks() const { + if (is_multicore) { + return clock->GetClockCycles(); + } + return CpuCyclesToClockCycles(ticks); } void CoreTiming::ClearPendingEvents() { @@ -117,7 +172,7 @@ void CoreTiming::ClearPendingEvents() { } void CoreTiming::RemoveEvent(const std::shared_ptr<EventType>& event_type) { - std::lock_guard guard{inner_mutex}; + std::scoped_lock lock{basic_lock}; const auto itr = std::remove_if(event_queue.begin(), event_queue.end(), [&](const Event& e) { return e.type.lock().get() == event_type.get(); @@ -130,97 +185,69 @@ void CoreTiming::RemoveEvent(const std::shared_ptr<EventType>& event_type) { } } -void CoreTiming::ForceExceptionCheck(s64 cycles) { - cycles = std::max<s64>(0, cycles); - if (downcounts[current_context] <= cycles) { - return; - } - - // downcount is always (much) smaller than MAX_INT so we can safely cast cycles to an int - // here. Account for cycles already executed by adjusting the g.slice_length - downcounts[current_context] = static_cast<int>(cycles); -} - -std::optional<u64> CoreTiming::NextAvailableCore(const s64 needed_ticks) const { - const u64 original_context = current_context; - u64 next_context = (original_context + 1) % num_cpu_cores; - while (next_context != original_context) { - if (time_slice[next_context] >= needed_ticks) { - return {next_context}; - } else if (time_slice[next_context] >= 0) { - return std::nullopt; - } - next_context = (next_context + 1) % num_cpu_cores; - } - return std::nullopt; -} - -void CoreTiming::Advance() { - std::unique_lock<std::mutex> guard(inner_mutex); - - const u64 cycles_executed = accumulated_ticks; - time_slice[current_context] = std::max<s64>(0, time_slice[current_context] - accumulated_ticks); - global_timer += cycles_executed; - - is_global_timer_sane = true; +std::optional<s64> CoreTiming::Advance() { + std::scoped_lock lock{advance_lock, basic_lock}; + global_timer = GetGlobalTimeNs().count(); while (!event_queue.empty() && event_queue.front().time <= global_timer) { Event evt = std::move(event_queue.front()); std::pop_heap(event_queue.begin(), event_queue.end(), std::greater<>()); event_queue.pop_back(); - inner_mutex.unlock(); + basic_lock.unlock(); - if (auto event_type{evt.type.lock()}) { - event_type->callback(evt.userdata, global_timer - evt.time); + if (const auto event_type{evt.type.lock()}) { + event_type->callback( + evt.user_data, std::chrono::nanoseconds{static_cast<s64>(global_timer - evt.time)}); } - inner_mutex.lock(); + basic_lock.lock(); + global_timer = GetGlobalTimeNs().count(); } - is_global_timer_sane = false; - - // Still events left (scheduled in the future) if (!event_queue.empty()) { - const s64 needed_ticks = - std::min<s64>(event_queue.front().time - global_timer, MAX_SLICE_LENGTH); - const auto next_core = NextAvailableCore(needed_ticks); - if (next_core) { - downcounts[*next_core] = needed_ticks; - } + const s64 next_time = event_queue.front().time - global_timer; + return next_time; + } else { + return std::nullopt; } - - accumulated_ticks = 0; - - downcounts[current_context] = time_slice[current_context]; } -void CoreTiming::ResetRun() { - downcounts.fill(MAX_SLICE_LENGTH); - time_slice.fill(MAX_SLICE_LENGTH); - current_context = 0; - // Still events left (scheduled in the future) - if (!event_queue.empty()) { - const s64 needed_ticks = - std::min<s64>(event_queue.front().time - global_timer, MAX_SLICE_LENGTH); - downcounts[current_context] = needed_ticks; +void CoreTiming::ThreadLoop() { + has_started = true; + while (!shutting_down) { + while (!paused) { + paused_set = false; + const auto next_time = Advance(); + if (next_time) { + if (*next_time > 0) { + std::chrono::nanoseconds next_time_ns = std::chrono::nanoseconds(*next_time); + event.WaitFor(next_time_ns); + } + } else { + wait_set = true; + event.Wait(); + } + wait_set = false; + } + paused_set = true; + clock->Pause(true); + pause_event.Wait(); + clock->Pause(false); } - - is_global_timer_sane = false; - accumulated_ticks = 0; } -void CoreTiming::Idle() { - accumulated_ticks += downcounts[current_context]; - idled_cycles += downcounts[current_context]; - downcounts[current_context] = 0; +std::chrono::nanoseconds CoreTiming::GetGlobalTimeNs() const { + if (is_multicore) { + return clock->GetTimeNS(); + } + return CyclesToNs(ticks); } std::chrono::microseconds CoreTiming::GetGlobalTimeUs() const { - return std::chrono::microseconds{GetTicks() * 1000000 / Hardware::BASE_CLOCK_RATE}; -} - -s64 CoreTiming::GetDowncount() const { - return downcounts[current_context]; + if (is_multicore) { + return clock->GetTimeUS(); + } + return CyclesToUs(ticks); } } // namespace Core::Timing diff --git a/src/core/core_timing.h b/src/core/core_timing.h index d50f4eb8a..b0b6036e4 100644 --- a/src/core/core_timing.h +++ b/src/core/core_timing.h @@ -1,24 +1,29 @@ -// Copyright 2008 Dolphin Emulator Project / 2017 Citra Emulator Project -// Licensed under GPLv2+ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version // Refer to the license.txt file included. #pragma once +#include <atomic> #include <chrono> #include <functional> #include <memory> #include <mutex> #include <optional> #include <string> +#include <thread> #include <vector> #include "common/common_types.h" -#include "common/threadsafe_queue.h" +#include "common/spin_lock.h" +#include "common/thread.h" +#include "common/wall_clock.h" namespace Core::Timing { /// A callback that may be scheduled for a particular core timing event. -using TimedCallback = std::function<void(u64 userdata, s64 cycles_late)>; +using TimedCallback = + std::function<void(std::uintptr_t user_data, std::chrono::nanoseconds ns_late)>; /// Contains the characteristics of a particular event. struct EventType { @@ -36,12 +41,12 @@ struct EventType { * in main CPU clock cycles. * * To schedule an event, you first have to register its type. This is where you pass in the - * callback. You then schedule events using the type id you get back. + * callback. You then schedule events using the type ID you get back. * - * The int cyclesLate that the callbacks get is how many cycles late it was. + * The s64 ns_late that the callbacks get is how many ns late it was. * So to schedule a new event on a regular basis: * inside callback: - * ScheduleEvent(periodInCycles - cyclesLate, callback, "whatever") + * ScheduleEvent(period_in_ns - ns_late, callback, "whatever") */ class CoreTiming { public: @@ -56,58 +61,71 @@ public: /// CoreTiming begins at the boundary of timing slice -1. An initial call to Advance() is /// required to end slice - 1 and start slice 0 before the first cycle of code is executed. - void Initialize(); + void Initialize(std::function<void()>&& on_thread_init_); /// Tears down all timing related functionality. void Shutdown(); - /// After the first Advance, the slice lengths and the downcount will be reduced whenever an - /// event is scheduled earlier than the current values. - /// - /// Scheduling from a callback will not update the downcount until the Advance() completes. - void ScheduleEvent(s64 cycles_into_future, const std::shared_ptr<EventType>& event_type, - u64 userdata = 0); + /// Sets if emulation is multicore or single core, must be set before Initialize + void SetMulticore(bool is_multicore) { + this->is_multicore = is_multicore; + } - void UnscheduleEvent(const std::shared_ptr<EventType>& event_type, u64 userdata); + /// Check if it's using host timing. + bool IsHostTiming() const { + return is_multicore; + } - /// We only permit one event of each type in the queue at a time. - void RemoveEvent(const std::shared_ptr<EventType>& event_type); + /// Pauses/Unpauses the execution of the timer thread. + void Pause(bool is_paused); - void ForceExceptionCheck(s64 cycles); + /// Pauses/Unpauses the execution of the timer thread and waits until paused. + void SyncPause(bool is_paused); - /// This should only be called from the emu thread, if you are calling it any other thread, - /// you are doing something evil - u64 GetTicks() const; + /// Checks if core timing is running. + bool IsRunning() const; - u64 GetIdleTicks() const; + /// Checks if the timer thread has started. + bool HasStarted() const { + return has_started; + } + + /// Checks if there are any pending time events. + bool HasPendingEvents() const; + + /// Schedules an event in core timing + void ScheduleEvent(std::chrono::nanoseconds ns_into_future, + const std::shared_ptr<EventType>& event_type, std::uintptr_t user_data = 0); + + void UnscheduleEvent(const std::shared_ptr<EventType>& event_type, std::uintptr_t user_data); + + /// We only permit one event of each type in the queue at a time. + void RemoveEvent(const std::shared_ptr<EventType>& event_type); void AddTicks(u64 ticks); - /// Advance must be called at the beginning of dispatcher loops, not the end. Advance() ends - /// the previous timing slice and begins the next one, you must Advance from the previous - /// slice to the current one before executing any cycles. CoreTiming starts in slice -1 so an - /// Advance() is required to initialize the slice length before the first cycle of emulated - /// instructions is executed. - void Advance(); + void ResetTicks(); - /// Pretend that the main CPU has executed enough cycles to reach the next event. void Idle(); - std::chrono::microseconds GetGlobalTimeUs() const; + s64 GetDowncount() const { + return downcount; + } - void ResetRun(); + /// Returns current time in emulated CPU cycles + u64 GetCPUTicks() const; - s64 GetDowncount() const; + /// Returns current time in emulated in Clock cycles + u64 GetClockTicks() const; - void SwitchContext(u64 new_context) { - current_context = new_context; - } + /// Returns current time in microseconds. + std::chrono::microseconds GetGlobalTimeUs() const; - bool CanCurrentContextRun() const { - return time_slice[current_context] > 0; - } + /// Returns current time in nanoseconds. + std::chrono::nanoseconds GetGlobalTimeNs() const; - std::optional<u64> NextAvailableCore(const s64 needed_ticks) const; + /// Checks for events manually and returns time in nanoseconds for next event, threadsafe. + std::optional<s64> Advance(); private: struct Event; @@ -115,21 +133,12 @@ private: /// Clear all pending events. This should ONLY be done on exit. void ClearPendingEvents(); - static constexpr u64 num_cpu_cores = 4; + static void ThreadEntry(CoreTiming& instance); + void ThreadLoop(); - s64 global_timer = 0; - s64 idled_cycles = 0; - s64 slice_length = 0; - u64 accumulated_ticks = 0; - std::array<s64, num_cpu_cores> downcounts{}; - // Slice of time assigned to each core per run. - std::array<s64, num_cpu_cores> time_slice{}; - u64 current_context = 0; + std::unique_ptr<Common::WallClock> clock; - // Are we in a function that has been called from Advance() - // If events are scheduled from a function that gets called from Advance(), - // don't change slice_length and downcount. - bool is_global_timer_sane = false; + u64 global_timer = 0; // The queue is a min-heap using std::make_heap/push_heap/pop_heap. // We don't use std::priority_queue because we need to be able to serialize, unserialize and @@ -139,8 +148,23 @@ private: u64 event_fifo_id = 0; std::shared_ptr<EventType> ev_lost; - - std::mutex inner_mutex; + Common::Event event{}; + Common::Event pause_event{}; + Common::SpinLock basic_lock{}; + Common::SpinLock advance_lock{}; + std::unique_ptr<std::thread> timer_thread; + std::atomic<bool> paused{}; + std::atomic<bool> paused_set{}; + std::atomic<bool> wait_set{}; + std::atomic<bool> shutting_down{}; + std::atomic<bool> has_started{}; + std::function<void()> on_thread_init{}; + + bool is_multicore{}; + + /// Cycle timing + u64 ticks{}; + s64 downcount{}; }; /// Creates a core timing event with the given name and callback. diff --git a/src/core/core_timing_util.cpp b/src/core/core_timing_util.cpp index de50d3b14..8ce8e602e 100644 --- a/src/core/core_timing_util.cpp +++ b/src/core/core_timing_util.cpp @@ -8,6 +8,7 @@ #include <limits> #include "common/logging/log.h" #include "common/uint128.h" +#include "core/hardware_properties.h" namespace Core::Timing { @@ -38,15 +39,23 @@ s64 usToCycles(std::chrono::microseconds us) { } s64 nsToCycles(std::chrono::nanoseconds ns) { - if (static_cast<u64>(ns.count() / 1000000000) > MAX_VALUE_TO_MULTIPLY) { - LOG_ERROR(Core_Timing, "Integer overflow, use max value"); - return std::numeric_limits<s64>::max(); - } - if (static_cast<u64>(ns.count()) > MAX_VALUE_TO_MULTIPLY) { - LOG_DEBUG(Core_Timing, "Time very big, do rounding"); - return Hardware::BASE_CLOCK_RATE * (ns.count() / 1000000000); - } - return (Hardware::BASE_CLOCK_RATE * ns.count()) / 1000000000; + const u128 temporal = Common::Multiply64Into128(ns.count(), Hardware::BASE_CLOCK_RATE); + return Common::Divide128On32(temporal, static_cast<u32>(1000000000)).first; +} + +u64 msToClockCycles(std::chrono::milliseconds ns) { + const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ); + return Common::Divide128On32(temp, 1000).first; +} + +u64 usToClockCycles(std::chrono::microseconds ns) { + const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ); + return Common::Divide128On32(temp, 1000000).first; +} + +u64 nsToClockCycles(std::chrono::nanoseconds ns) { + const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ); + return Common::Divide128On32(temp, 1000000000).first; } u64 CpuCyclesToClockCycles(u64 ticks) { @@ -54,4 +63,22 @@ u64 CpuCyclesToClockCycles(u64 ticks) { return Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first; } +std::chrono::milliseconds CyclesToMs(s64 cycles) { + const u128 temporal = Common::Multiply64Into128(cycles, 1000); + u64 ms = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first; + return std::chrono::milliseconds(ms); +} + +std::chrono::nanoseconds CyclesToNs(s64 cycles) { + const u128 temporal = Common::Multiply64Into128(cycles, 1000000000); + u64 ns = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first; + return std::chrono::nanoseconds(ns); +} + +std::chrono::microseconds CyclesToUs(s64 cycles) { + const u128 temporal = Common::Multiply64Into128(cycles, 1000000); + u64 us = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first; + return std::chrono::microseconds(us); +} + } // namespace Core::Timing diff --git a/src/core/core_timing_util.h b/src/core/core_timing_util.h index addc72b19..e4a046bf9 100644 --- a/src/core/core_timing_util.h +++ b/src/core/core_timing_util.h @@ -6,25 +6,18 @@ #include <chrono> #include "common/common_types.h" -#include "core/hardware_properties.h" namespace Core::Timing { s64 msToCycles(std::chrono::milliseconds ms); s64 usToCycles(std::chrono::microseconds us); s64 nsToCycles(std::chrono::nanoseconds ns); - -inline std::chrono::milliseconds CyclesToMs(s64 cycles) { - return std::chrono::milliseconds(cycles * 1000 / Hardware::BASE_CLOCK_RATE); -} - -inline std::chrono::nanoseconds CyclesToNs(s64 cycles) { - return std::chrono::nanoseconds(cycles * 1000000000 / Hardware::BASE_CLOCK_RATE); -} - -inline std::chrono::microseconds CyclesToUs(s64 cycles) { - return std::chrono::microseconds(cycles * 1000000 / Hardware::BASE_CLOCK_RATE); -} +u64 msToClockCycles(std::chrono::milliseconds ns); +u64 usToClockCycles(std::chrono::microseconds ns); +u64 nsToClockCycles(std::chrono::nanoseconds ns); +std::chrono::milliseconds CyclesToMs(s64 cycles); +std::chrono::nanoseconds CyclesToNs(s64 cycles); +std::chrono::microseconds CyclesToUs(s64 cycles); u64 CpuCyclesToClockCycles(u64 ticks); diff --git a/src/core/cpu_manager.cpp b/src/core/cpu_manager.cpp index 70ddbdcca..983210197 100644 --- a/src/core/cpu_manager.cpp +++ b/src/core/cpu_manager.cpp @@ -2,80 +2,371 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/fiber.h" +#include "common/microprofile.h" +#include "common/thread.h" #include "core/arm/exclusive_monitor.h" #include "core/core.h" -#include "core/core_manager.h" #include "core/core_timing.h" #include "core/cpu_manager.h" #include "core/gdbstub/gdbstub.h" +#include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/physical_core.h" +#include "core/hle/kernel/scheduler.h" +#include "core/hle/kernel/thread.h" +#include "video_core/gpu.h" namespace Core { CpuManager::CpuManager(System& system) : system{system} {} CpuManager::~CpuManager() = default; +void CpuManager::ThreadStart(CpuManager& cpu_manager, std::size_t core) { + cpu_manager.RunThread(core); +} + void CpuManager::Initialize() { - for (std::size_t index = 0; index < core_managers.size(); ++index) { - core_managers[index] = std::make_unique<CoreManager>(system, index); + running_mode = true; + if (is_multicore) { + for (std::size_t core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + core_data[core].host_thread = + std::make_unique<std::thread>(ThreadStart, std::ref(*this), core); + } + } else { + core_data[0].host_thread = std::make_unique<std::thread>(ThreadStart, std::ref(*this), 0); } } void CpuManager::Shutdown() { - for (auto& cpu_core : core_managers) { - cpu_core.reset(); + running_mode = false; + Pause(false); + if (is_multicore) { + for (auto& data : core_data) { + data.host_thread->join(); + data.host_thread.reset(); + } + } else { + core_data[0].host_thread->join(); + core_data[0].host_thread.reset(); + } +} + +std::function<void(void*)> CpuManager::GetGuestThreadStartFunc() { + return GuestThreadFunction; +} + +std::function<void(void*)> CpuManager::GetIdleThreadStartFunc() { + return IdleThreadFunction; +} + +std::function<void(void*)> CpuManager::GetSuspendThreadStartFunc() { + return SuspendThreadFunction; +} + +void CpuManager::GuestThreadFunction(void* cpu_manager_) { + CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_); + if (cpu_manager->is_multicore) { + cpu_manager->MultiCoreRunGuestThread(); + } else { + cpu_manager->SingleCoreRunGuestThread(); + } +} + +void CpuManager::GuestRewindFunction(void* cpu_manager_) { + CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_); + if (cpu_manager->is_multicore) { + cpu_manager->MultiCoreRunGuestLoop(); + } else { + cpu_manager->SingleCoreRunGuestLoop(); + } +} + +void CpuManager::IdleThreadFunction(void* cpu_manager_) { + CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_); + if (cpu_manager->is_multicore) { + cpu_manager->MultiCoreRunIdleThread(); + } else { + cpu_manager->SingleCoreRunIdleThread(); } } -CoreManager& CpuManager::GetCoreManager(std::size_t index) { - return *core_managers.at(index); +void CpuManager::SuspendThreadFunction(void* cpu_manager_) { + CpuManager* cpu_manager = static_cast<CpuManager*>(cpu_manager_); + if (cpu_manager->is_multicore) { + cpu_manager->MultiCoreRunSuspendThread(); + } else { + cpu_manager->SingleCoreRunSuspendThread(); + } } -const CoreManager& CpuManager::GetCoreManager(std::size_t index) const { - return *core_managers.at(index); +void* CpuManager::GetStartFuncParamater() { + return static_cast<void*>(this); } -CoreManager& CpuManager::GetCurrentCoreManager() { - // Otherwise, use single-threaded mode active_core variable - return *core_managers[active_core]; +/////////////////////////////////////////////////////////////////////////////// +/// MultiCore /// +/////////////////////////////////////////////////////////////////////////////// + +void CpuManager::MultiCoreRunGuestThread() { + auto& kernel = system.Kernel(); + { + auto& sched = kernel.CurrentScheduler(); + sched.OnThreadStart(); + } + MultiCoreRunGuestLoop(); } -const CoreManager& CpuManager::GetCurrentCoreManager() const { - // Otherwise, use single-threaded mode active_core variable - return *core_managers[active_core]; +void CpuManager::MultiCoreRunGuestLoop() { + auto& kernel = system.Kernel(); + auto* thread = kernel.CurrentScheduler().GetCurrentThread(); + while (true) { + auto* physical_core = &kernel.CurrentPhysicalCore(); + auto& arm_interface = thread->ArmInterface(); + system.EnterDynarmicProfile(); + while (!physical_core->IsInterrupted()) { + arm_interface.Run(); + physical_core = &kernel.CurrentPhysicalCore(); + } + system.ExitDynarmicProfile(); + arm_interface.ClearExclusiveState(); + auto& scheduler = kernel.CurrentScheduler(); + scheduler.TryDoContextSwitch(); + } } -void CpuManager::RunLoop(bool tight_loop) { - if (GDBStub::IsServerEnabled()) { - GDBStub::HandlePacket(); +void CpuManager::MultiCoreRunIdleThread() { + auto& kernel = system.Kernel(); + while (true) { + auto& physical_core = kernel.CurrentPhysicalCore(); + physical_core.Idle(); + auto& scheduler = kernel.CurrentScheduler(); + scheduler.TryDoContextSwitch(); + } +} - // If the loop is halted and we want to step, use a tiny (1) number of instructions to - // execute. Otherwise, get out of the loop function. - if (GDBStub::GetCpuHaltFlag()) { - if (GDBStub::GetCpuStepFlag()) { - tight_loop = false; - } else { - return; +void CpuManager::MultiCoreRunSuspendThread() { + auto& kernel = system.Kernel(); + { + auto& sched = kernel.CurrentScheduler(); + sched.OnThreadStart(); + } + while (true) { + auto core = kernel.GetCurrentHostThreadID(); + auto& scheduler = kernel.CurrentScheduler(); + Kernel::Thread* current_thread = scheduler.GetCurrentThread(); + Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[core].host_context); + ASSERT(scheduler.ContextSwitchPending()); + ASSERT(core == kernel.GetCurrentHostThreadID()); + scheduler.TryDoContextSwitch(); + } +} + +void CpuManager::MultiCorePause(bool paused) { + if (!paused) { + bool all_not_barrier = false; + while (!all_not_barrier) { + all_not_barrier = true; + for (const auto& data : core_data) { + all_not_barrier &= !data.is_running.load() && data.initialized.load(); + } + } + for (auto& data : core_data) { + data.enter_barrier->Set(); + } + if (paused_state.load()) { + bool all_barrier = false; + while (!all_barrier) { + all_barrier = true; + for (const auto& data : core_data) { + all_barrier &= data.is_paused.load() && data.initialized.load(); + } + } + for (auto& data : core_data) { + data.exit_barrier->Set(); } } + } else { + /// Wait until all cores are paused. + bool all_barrier = false; + while (!all_barrier) { + all_barrier = true; + for (const auto& data : core_data) { + all_barrier &= data.is_paused.load() && data.initialized.load(); + } + } + /// Don't release the barrier } + paused_state = paused; +} - auto& core_timing = system.CoreTiming(); - core_timing.ResetRun(); - bool keep_running{}; - do { - keep_running = false; - for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) { - core_timing.SwitchContext(active_core); - if (core_timing.CanCurrentContextRun()) { - core_managers[active_core]->RunLoop(tight_loop); +/////////////////////////////////////////////////////////////////////////////// +/// SingleCore /// +/////////////////////////////////////////////////////////////////////////////// + +void CpuManager::SingleCoreRunGuestThread() { + auto& kernel = system.Kernel(); + { + auto& sched = kernel.CurrentScheduler(); + sched.OnThreadStart(); + } + SingleCoreRunGuestLoop(); +} + +void CpuManager::SingleCoreRunGuestLoop() { + auto& kernel = system.Kernel(); + auto* thread = kernel.CurrentScheduler().GetCurrentThread(); + while (true) { + auto* physical_core = &kernel.CurrentPhysicalCore(); + auto& arm_interface = thread->ArmInterface(); + system.EnterDynarmicProfile(); + if (!physical_core->IsInterrupted()) { + arm_interface.Run(); + physical_core = &kernel.CurrentPhysicalCore(); + } + system.ExitDynarmicProfile(); + thread->SetPhantomMode(true); + system.CoreTiming().Advance(); + thread->SetPhantomMode(false); + arm_interface.ClearExclusiveState(); + PreemptSingleCore(); + auto& scheduler = kernel.Scheduler(current_core); + scheduler.TryDoContextSwitch(); + } +} + +void CpuManager::SingleCoreRunIdleThread() { + auto& kernel = system.Kernel(); + while (true) { + auto& physical_core = kernel.CurrentPhysicalCore(); + PreemptSingleCore(false); + system.CoreTiming().AddTicks(1000U); + idle_count++; + auto& scheduler = physical_core.Scheduler(); + scheduler.TryDoContextSwitch(); + } +} + +void CpuManager::SingleCoreRunSuspendThread() { + auto& kernel = system.Kernel(); + { + auto& sched = kernel.CurrentScheduler(); + sched.OnThreadStart(); + } + while (true) { + auto core = kernel.GetCurrentHostThreadID(); + auto& scheduler = kernel.CurrentScheduler(); + Kernel::Thread* current_thread = scheduler.GetCurrentThread(); + Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[0].host_context); + ASSERT(scheduler.ContextSwitchPending()); + ASSERT(core == kernel.GetCurrentHostThreadID()); + scheduler.TryDoContextSwitch(); + } +} + +void CpuManager::PreemptSingleCore(bool from_running_enviroment) { + std::size_t old_core = current_core; + auto& scheduler = system.Kernel().Scheduler(old_core); + Kernel::Thread* current_thread = scheduler.GetCurrentThread(); + if (idle_count >= 4 || from_running_enviroment) { + if (!from_running_enviroment) { + system.CoreTiming().Idle(); + idle_count = 0; + } + current_thread->SetPhantomMode(true); + system.CoreTiming().Advance(); + current_thread->SetPhantomMode(false); + } + current_core.store((current_core + 1) % Core::Hardware::NUM_CPU_CORES); + system.CoreTiming().ResetTicks(); + scheduler.Unload(); + auto& next_scheduler = system.Kernel().Scheduler(current_core); + Common::Fiber::YieldTo(current_thread->GetHostContext(), next_scheduler.ControlContext()); + /// May have changed scheduler + auto& current_scheduler = system.Kernel().Scheduler(current_core); + current_scheduler.Reload(); + auto* currrent_thread2 = current_scheduler.GetCurrentThread(); + if (!currrent_thread2->IsIdleThread()) { + idle_count = 0; + } +} + +void CpuManager::SingleCorePause(bool paused) { + if (!paused) { + bool all_not_barrier = false; + while (!all_not_barrier) { + all_not_barrier = !core_data[0].is_running.load() && core_data[0].initialized.load(); + } + core_data[0].enter_barrier->Set(); + if (paused_state.load()) { + bool all_barrier = false; + while (!all_barrier) { + all_barrier = core_data[0].is_paused.load() && core_data[0].initialized.load(); } - keep_running |= core_timing.CanCurrentContextRun(); + core_data[0].exit_barrier->Set(); + } + } else { + /// Wait until all cores are paused. + bool all_barrier = false; + while (!all_barrier) { + all_barrier = core_data[0].is_paused.load() && core_data[0].initialized.load(); } - } while (keep_running); + /// Don't release the barrier + } + paused_state = paused; +} + +void CpuManager::Pause(bool paused) { + if (is_multicore) { + MultiCorePause(paused); + } else { + SingleCorePause(paused); + } +} - if (GDBStub::IsServerEnabled()) { - GDBStub::SetCpuStepFlag(false); +void CpuManager::RunThread(std::size_t core) { + /// Initialization + system.RegisterCoreThread(core); + std::string name; + if (is_multicore) { + name = "yuzu:CPUCore_" + std::to_string(core); + } else { + name = "yuzu:CPUThread"; + } + MicroProfileOnThreadCreate(name.c_str()); + Common::SetCurrentThreadName(name.c_str()); + Common::SetCurrentThreadPriority(Common::ThreadPriority::High); + auto& data = core_data[core]; + data.enter_barrier = std::make_unique<Common::Event>(); + data.exit_barrier = std::make_unique<Common::Event>(); + data.host_context = Common::Fiber::ThreadToFiber(); + data.is_running = false; + data.initialized = true; + const bool sc_sync = !is_async_gpu && !is_multicore; + bool sc_sync_first_use = sc_sync; + /// Running + while (running_mode) { + data.is_running = false; + data.enter_barrier->Wait(); + if (sc_sync_first_use) { + system.GPU().ObtainContext(); + sc_sync_first_use = false; + } + auto& scheduler = system.Kernel().CurrentScheduler(); + Kernel::Thread* current_thread = scheduler.GetCurrentThread(); + data.is_running = true; + Common::Fiber::YieldTo(data.host_context, current_thread->GetHostContext()); + data.is_running = false; + data.is_paused = true; + data.exit_barrier->Wait(); + data.is_paused = false; } + /// Time to cleanup + data.host_context->Exit(); + data.enter_barrier.reset(); + data.exit_barrier.reset(); + data.initialized = false; + + MicroProfileOnThreadExit(); } } // namespace Core diff --git a/src/core/cpu_manager.h b/src/core/cpu_manager.h index 97554d1bb..17420c941 100644 --- a/src/core/cpu_manager.h +++ b/src/core/cpu_manager.h @@ -5,12 +5,22 @@ #pragma once #include <array> +#include <atomic> +#include <functional> #include <memory> +#include <thread> + +#include "common/fiber.h" +#include "common/thread.h" #include "core/hardware_properties.h" +namespace Common { +class Event; +class Fiber; +} // namespace Common + namespace Core { -class CoreManager; class System; class CpuManager { @@ -24,24 +34,74 @@ public: CpuManager& operator=(const CpuManager&) = delete; CpuManager& operator=(CpuManager&&) = delete; + /// Sets if emulation is multicore or single core, must be set before Initialize + void SetMulticore(bool is_multicore) { + this->is_multicore = is_multicore; + } + + /// Sets if emulation is using an asynchronous GPU. + void SetAsyncGpu(bool is_async_gpu) { + this->is_async_gpu = is_async_gpu; + } + void Initialize(); void Shutdown(); - CoreManager& GetCoreManager(std::size_t index); - const CoreManager& GetCoreManager(std::size_t index) const; + void Pause(bool paused); - CoreManager& GetCurrentCoreManager(); - const CoreManager& GetCurrentCoreManager() const; + static std::function<void(void*)> GetGuestThreadStartFunc(); + static std::function<void(void*)> GetIdleThreadStartFunc(); + static std::function<void(void*)> GetSuspendThreadStartFunc(); + void* GetStartFuncParamater(); - std::size_t GetActiveCoreIndex() const { - return active_core; - } + void PreemptSingleCore(bool from_running_enviroment = true); - void RunLoop(bool tight_loop); + std::size_t CurrentCore() const { + return current_core.load(); + } private: - std::array<std::unique_ptr<CoreManager>, Hardware::NUM_CPU_CORES> core_managers; - std::size_t active_core{}; ///< Active core, only used in single thread mode + static void GuestThreadFunction(void* cpu_manager); + static void GuestRewindFunction(void* cpu_manager); + static void IdleThreadFunction(void* cpu_manager); + static void SuspendThreadFunction(void* cpu_manager); + + void MultiCoreRunGuestThread(); + void MultiCoreRunGuestLoop(); + void MultiCoreRunIdleThread(); + void MultiCoreRunSuspendThread(); + void MultiCorePause(bool paused); + + void SingleCoreRunGuestThread(); + void SingleCoreRunGuestLoop(); + void SingleCoreRunIdleThread(); + void SingleCoreRunSuspendThread(); + void SingleCorePause(bool paused); + + static void ThreadStart(CpuManager& cpu_manager, std::size_t core); + + void RunThread(std::size_t core); + + struct CoreData { + std::shared_ptr<Common::Fiber> host_context; + std::unique_ptr<Common::Event> enter_barrier; + std::unique_ptr<Common::Event> exit_barrier; + std::atomic<bool> is_running; + std::atomic<bool> is_paused; + std::atomic<bool> initialized; + std::unique_ptr<std::thread> host_thread; + }; + + std::atomic<bool> running_mode{}; + std::atomic<bool> paused_state{}; + + std::array<CoreData, Core::Hardware::NUM_CPU_CORES> core_data{}; + + bool is_async_gpu{}; + bool is_multicore{}; + std::atomic<std::size_t> current_core{}; + std::size_t idle_count{}; + static constexpr std::size_t max_cycle_runs = 5; System& system; }; diff --git a/src/core/crypto/aes_util.cpp b/src/core/crypto/aes_util.cpp index 4be76bb43..6a9734812 100644 --- a/src/core/crypto/aes_util.cpp +++ b/src/core/crypto/aes_util.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> #include <mbedtls/cipher.h> #include "common/assert.h" #include "common/logging/log.h" @@ -10,8 +11,10 @@ namespace Core::Crypto { namespace { -std::vector<u8> CalculateNintendoTweak(std::size_t sector_id) { - std::vector<u8> out(0x10); +using NintendoTweak = std::array<u8, 16>; + +NintendoTweak CalculateNintendoTweak(std::size_t sector_id) { + NintendoTweak out{}; for (std::size_t i = 0xF; i <= 0xF; --i) { out[i] = sector_id & 0xFF; sector_id >>= 8; @@ -64,13 +67,6 @@ AESCipher<Key, KeySize>::~AESCipher() { } template <typename Key, std::size_t KeySize> -void AESCipher<Key, KeySize>::SetIV(std::vector<u8> iv) { - ASSERT_MSG((mbedtls_cipher_set_iv(&ctx->encryption_context, iv.data(), iv.size()) || - mbedtls_cipher_set_iv(&ctx->decryption_context, iv.data(), iv.size())) == 0, - "Failed to set IV on mbedtls ciphers."); -} - -template <typename Key, std::size_t KeySize> void AESCipher<Key, KeySize>::Transcode(const u8* src, std::size_t size, u8* dest, Op op) const { auto* const context = op == Op::Encrypt ? &ctx->encryption_context : &ctx->decryption_context; @@ -120,10 +116,17 @@ void AESCipher<Key, KeySize>::XTSTranscode(const u8* src, std::size_t size, u8* for (std::size_t i = 0; i < size; i += sector_size) { SetIV(CalculateNintendoTweak(sector_id++)); - Transcode<u8, u8>(src + i, sector_size, dest + i, op); + Transcode(src + i, sector_size, dest + i, op); } } +template <typename Key, std::size_t KeySize> +void AESCipher<Key, KeySize>::SetIVImpl(const u8* data, std::size_t size) { + ASSERT_MSG((mbedtls_cipher_set_iv(&ctx->encryption_context, data, size) || + mbedtls_cipher_set_iv(&ctx->decryption_context, data, size)) == 0, + "Failed to set IV on mbedtls ciphers."); +} + template class AESCipher<Key128>; template class AESCipher<Key256>; } // namespace Core::Crypto diff --git a/src/core/crypto/aes_util.h b/src/core/crypto/aes_util.h index edc4ab910..e2a304186 100644 --- a/src/core/crypto/aes_util.h +++ b/src/core/crypto/aes_util.h @@ -6,7 +6,6 @@ #include <memory> #include <type_traits> -#include <vector> #include "common/common_types.h" #include "core/file_sys/vfs.h" @@ -32,10 +31,12 @@ class AESCipher { public: AESCipher(Key key, Mode mode); - ~AESCipher(); - void SetIV(std::vector<u8> iv); + template <typename ContiguousContainer> + void SetIV(const ContiguousContainer& container) { + SetIVImpl(std::data(container), std::size(container)); + } template <typename Source, typename Dest> void Transcode(const Source* src, std::size_t size, Dest* dest, Op op) const { @@ -59,6 +60,8 @@ public: std::size_t sector_size, Op op); private: + void SetIVImpl(const u8* data, std::size_t size); + std::unique_ptr<CipherContext> ctx; }; } // namespace Core::Crypto diff --git a/src/core/crypto/ctr_encryption_layer.cpp b/src/core/crypto/ctr_encryption_layer.cpp index 902841c77..5c84bb0a4 100644 --- a/src/core/crypto/ctr_encryption_layer.cpp +++ b/src/core/crypto/ctr_encryption_layer.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> #include <cstring> #include "common/assert.h" #include "core/crypto/ctr_encryption_layer.h" @@ -10,8 +11,7 @@ namespace Core::Crypto { CTREncryptionLayer::CTREncryptionLayer(FileSys::VirtualFile base_, Key128 key_, std::size_t base_offset) - : EncryptionLayer(std::move(base_)), base_offset(base_offset), cipher(key_, Mode::CTR), - iv(16, 0) {} + : EncryptionLayer(std::move(base_)), base_offset(base_offset), cipher(key_, Mode::CTR) {} std::size_t CTREncryptionLayer::Read(u8* data, std::size_t length, std::size_t offset) const { if (length == 0) @@ -39,9 +39,8 @@ std::size_t CTREncryptionLayer::Read(u8* data, std::size_t length, std::size_t o return read + Read(data + read, length - read, offset + read); } -void CTREncryptionLayer::SetIV(const std::vector<u8>& iv_) { - const auto length = std::min(iv_.size(), iv.size()); - iv.assign(iv_.cbegin(), iv_.cbegin() + length); +void CTREncryptionLayer::SetIV(const IVData& iv_) { + iv = iv_; } void CTREncryptionLayer::UpdateIV(std::size_t offset) const { diff --git a/src/core/crypto/ctr_encryption_layer.h b/src/core/crypto/ctr_encryption_layer.h index a7bf810f4..a2429f001 100644 --- a/src/core/crypto/ctr_encryption_layer.h +++ b/src/core/crypto/ctr_encryption_layer.h @@ -4,7 +4,8 @@ #pragma once -#include <vector> +#include <array> + #include "core/crypto/aes_util.h" #include "core/crypto/encryption_layer.h" #include "core/crypto/key_manager.h" @@ -14,18 +15,20 @@ namespace Core::Crypto { // Sits on top of a VirtualFile and provides CTR-mode AES decription. class CTREncryptionLayer : public EncryptionLayer { public: + using IVData = std::array<u8, 16>; + CTREncryptionLayer(FileSys::VirtualFile base, Key128 key, std::size_t base_offset); std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override; - void SetIV(const std::vector<u8>& iv); + void SetIV(const IVData& iv); private: std::size_t base_offset; // Must be mutable as operations modify cipher contexts. mutable AESCipher<Key128> cipher; - mutable std::vector<u8> iv; + mutable IVData iv{}; void UpdateIV(std::size_t offset) const; }; diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp index 87e6a1fd3..da15f764a 100644 --- a/src/core/crypto/key_manager.cpp +++ b/src/core/crypto/key_manager.cpp @@ -23,7 +23,6 @@ #include "common/hex_util.h" #include "common/logging/log.h" #include "common/string_util.h" -#include "core/core.h" #include "core/crypto/aes_util.h" #include "core/crypto/key_manager.h" #include "core/crypto/partition_data_manager.h" @@ -36,18 +35,86 @@ #include "core/settings.h" namespace Core::Crypto { +namespace { constexpr u64 CURRENT_CRYPTO_REVISION = 0x5; constexpr u64 FULL_TICKET_SIZE = 0x400; -using namespace Common; +using Common::AsArray; -const std::array<SHA256Hash, 2> eticket_source_hashes{ - "B71DB271DC338DF380AA2C4335EF8873B1AFD408E80B3582D8719FC81C5E511C"_array32, // eticket_rsa_kek_source - "E8965A187D30E57869F562D04383C996DE487BBA5761363D2D4D32391866A85C"_array32, // eticket_rsa_kekek_source +// clang-format off +constexpr std::array eticket_source_hashes{ + AsArray("B71DB271DC338DF380AA2C4335EF8873B1AFD408E80B3582D8719FC81C5E511C"), // eticket_rsa_kek_source + AsArray("E8965A187D30E57869F562D04383C996DE487BBA5761363D2D4D32391866A85C"), // eticket_rsa_kekek_source }; +// clang-format on -const std::map<std::pair<S128KeyType, u64>, std::string> KEYS_VARIABLE_LENGTH{ +constexpr std::array<std::pair<std::string_view, KeyIndex<S128KeyType>>, 30> s128_file_id{{ + {"eticket_rsa_kek", {S128KeyType::ETicketRSAKek, 0, 0}}, + {"eticket_rsa_kek_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::ETicketKek), 0}}, + {"eticket_rsa_kekek_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::ETicketKekek), 0}}, + {"rsa_kek_mask_0", {S128KeyType::RSAKek, static_cast<u64>(RSAKekType::Mask0), 0}}, + {"rsa_kek_seed_3", {S128KeyType::RSAKek, static_cast<u64>(RSAKekType::Seed3), 0}}, + {"rsa_oaep_kek_generation_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::RSAOaepKekGeneration), 0}}, + {"sd_card_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek), 0}}, + {"aes_kek_generation_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration), 0}}, + {"aes_key_generation_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration), 0}}, + {"package2_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Package2), 0}}, + {"master_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Master), 0}}, + {"header_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::HeaderKek), 0}}, + {"key_area_key_application_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey), + static_cast<u64>(KeyAreaKeyType::Application)}}, + {"key_area_key_ocean_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey), + static_cast<u64>(KeyAreaKeyType::Ocean)}}, + {"key_area_key_system_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey), + static_cast<u64>(KeyAreaKeyType::System)}}, + {"titlekek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Titlekek), 0}}, + {"keyblob_mac_key_source", + {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC), 0}}, + {"tsec_key", {S128KeyType::TSEC, 0, 0}}, + {"secure_boot_key", {S128KeyType::SecureBoot, 0, 0}}, + {"sd_seed", {S128KeyType::SDSeed, 0, 0}}, + {"bis_key_0_crypt", {S128KeyType::BIS, 0, static_cast<u64>(BISKeyType::Crypto)}}, + {"bis_key_0_tweak", {S128KeyType::BIS, 0, static_cast<u64>(BISKeyType::Tweak)}}, + {"bis_key_1_crypt", {S128KeyType::BIS, 1, static_cast<u64>(BISKeyType::Crypto)}}, + {"bis_key_1_tweak", {S128KeyType::BIS, 1, static_cast<u64>(BISKeyType::Tweak)}}, + {"bis_key_2_crypt", {S128KeyType::BIS, 2, static_cast<u64>(BISKeyType::Crypto)}}, + {"bis_key_2_tweak", {S128KeyType::BIS, 2, static_cast<u64>(BISKeyType::Tweak)}}, + {"bis_key_3_crypt", {S128KeyType::BIS, 3, static_cast<u64>(BISKeyType::Crypto)}}, + {"bis_key_3_tweak", {S128KeyType::BIS, 3, static_cast<u64>(BISKeyType::Tweak)}}, + {"header_kek", {S128KeyType::HeaderKek, 0, 0}}, + {"sd_card_kek", {S128KeyType::SDKek, 0, 0}}, +}}; + +auto Find128ByName(std::string_view name) { + return std::find_if(s128_file_id.begin(), s128_file_id.end(), + [&name](const auto& pair) { return pair.first == name; }); +} + +constexpr std::array<std::pair<std::string_view, KeyIndex<S256KeyType>>, 6> s256_file_id{{ + {"header_key", {S256KeyType::Header, 0, 0}}, + {"sd_card_save_key_source", {S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::Save), 0}}, + {"sd_card_nca_key_source", {S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::NCA), 0}}, + {"header_key_source", {S256KeyType::HeaderSource, 0, 0}}, + {"sd_card_save_key", {S256KeyType::SDKey, static_cast<u64>(SDKeyType::Save), 0}}, + {"sd_card_nca_key", {S256KeyType::SDKey, static_cast<u64>(SDKeyType::NCA), 0}}, +}}; + +auto Find256ByName(std::string_view name) { + return std::find_if(s256_file_id.begin(), s256_file_id.end(), + [&name](const auto& pair) { return pair.first == name; }); +} + +using KeyArray = std::array<std::pair<std::pair<S128KeyType, u64>, std::string_view>, 7>; +constexpr KeyArray KEYS_VARIABLE_LENGTH{{ {{S128KeyType::Master, 0}, "master_key_"}, {{S128KeyType::Package1, 0}, "package1_key_"}, {{S128KeyType::Package2, 0}, "package2_key_"}, @@ -55,14 +122,13 @@ const std::map<std::pair<S128KeyType, u64>, std::string> KEYS_VARIABLE_LENGTH{ {{S128KeyType::Source, static_cast<u64>(SourceKeyType::Keyblob)}, "keyblob_key_source_"}, {{S128KeyType::Keyblob, 0}, "keyblob_key_"}, {{S128KeyType::KeyblobMAC, 0}, "keyblob_mac_key_"}, -}; +}}; -namespace { template <std::size_t Size> bool IsAllZeroArray(const std::array<u8, Size>& array) { return std::all_of(array.begin(), array.end(), [](const auto& elem) { return elem == 0; }); } -} // namespace +} // Anonymous namespace u64 GetSignatureTypeDataSize(SignatureType type) { switch (type) { @@ -94,13 +160,13 @@ u64 GetSignatureTypePaddingSize(SignatureType type) { } SignatureType Ticket::GetSignatureType() const { - if (auto ticket = std::get_if<RSA4096Ticket>(&data)) { + if (const auto* ticket = std::get_if<RSA4096Ticket>(&data)) { return ticket->sig_type; } - if (auto ticket = std::get_if<RSA2048Ticket>(&data)) { + if (const auto* ticket = std::get_if<RSA2048Ticket>(&data)) { return ticket->sig_type; } - if (auto ticket = std::get_if<ECDSATicket>(&data)) { + if (const auto* ticket = std::get_if<ECDSATicket>(&data)) { return ticket->sig_type; } @@ -108,13 +174,13 @@ SignatureType Ticket::GetSignatureType() const { } TicketData& Ticket::GetData() { - if (auto ticket = std::get_if<RSA4096Ticket>(&data)) { + if (auto* ticket = std::get_if<RSA4096Ticket>(&data)) { return ticket->data; } - if (auto ticket = std::get_if<RSA2048Ticket>(&data)) { + if (auto* ticket = std::get_if<RSA2048Ticket>(&data)) { return ticket->data; } - if (auto ticket = std::get_if<ECDSATicket>(&data)) { + if (auto* ticket = std::get_if<ECDSATicket>(&data)) { return ticket->data; } @@ -122,13 +188,13 @@ TicketData& Ticket::GetData() { } const TicketData& Ticket::GetData() const { - if (auto ticket = std::get_if<RSA4096Ticket>(&data)) { + if (const auto* ticket = std::get_if<RSA4096Ticket>(&data)) { return ticket->data; } - if (auto ticket = std::get_if<RSA2048Ticket>(&data)) { + if (const auto* ticket = std::get_if<RSA2048Ticket>(&data)) { return ticket->data; } - if (auto ticket = std::get_if<ECDSATicket>(&data)) { + if (const auto* ticket = std::get_if<ECDSATicket>(&data)) { return ticket->data; } @@ -231,8 +297,9 @@ void KeyManager::DeriveGeneralPurposeKeys(std::size_t crypto_revision) { } RSAKeyPair<2048> KeyManager::GetETicketRSAKey() const { - if (IsAllZeroArray(eticket_extended_kek) || !HasKey(S128KeyType::ETicketRSAKek)) + if (IsAllZeroArray(eticket_extended_kek) || !HasKey(S128KeyType::ETicketRSAKek)) { return {}; + } const auto eticket_final = GetKey(S128KeyType::ETicketRSAKek); @@ -259,27 +326,30 @@ Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source) } std::optional<Key128> DeriveSDSeed() { - const FileUtil::IOFile save_43(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + - "/system/save/8000000000000043", - "rb+"); - if (!save_43.IsOpen()) - return {}; + const Common::FS::IOFile save_43(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + + "/system/save/8000000000000043", + "rb+"); + if (!save_43.IsOpen()) { + return std::nullopt; + } - const FileUtil::IOFile sd_private( - FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + "/Nintendo/Contents/private", "rb+"); - if (!sd_private.IsOpen()) - return {}; + const Common::FS::IOFile sd_private(Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir) + + "/Nintendo/Contents/private", + "rb+"); + if (!sd_private.IsOpen()) { + return std::nullopt; + } std::array<u8, 0x10> private_seed{}; if (sd_private.ReadBytes(private_seed.data(), private_seed.size()) != private_seed.size()) { - return {}; + return std::nullopt; } std::array<u8, 0x10> buffer{}; std::size_t offset = 0; for (; offset + 0x10 < save_43.GetSize(); ++offset) { if (!save_43.Seek(offset, SEEK_SET)) { - return {}; + return std::nullopt; } save_43.ReadBytes(buffer.data(), buffer.size()); @@ -289,23 +359,26 @@ std::optional<Key128> DeriveSDSeed() { } if (!save_43.Seek(offset + 0x10, SEEK_SET)) { - return {}; + return std::nullopt; } Key128 seed{}; if (save_43.ReadBytes(seed.data(), seed.size()) != seed.size()) { - return {}; + return std::nullopt; } return seed; } Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys) { - if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek))) + if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek))) { return Loader::ResultStatus::ErrorMissingSDKEKSource; - if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration))) + } + if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration))) { return Loader::ResultStatus::ErrorMissingAESKEKGenerationSource; - if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration))) + } + if (!keys.HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration))) { return Loader::ResultStatus::ErrorMissingAESKeyGenerationSource; + } const auto sd_kek_source = keys.GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek)); @@ -318,14 +391,17 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke GenerateKeyEncryptionKey(sd_kek_source, master_00, aes_kek_gen, aes_key_gen); keys.SetKey(S128KeyType::SDKek, sd_kek); - if (!keys.HasKey(S128KeyType::SDSeed)) + if (!keys.HasKey(S128KeyType::SDSeed)) { return Loader::ResultStatus::ErrorMissingSDSeed; + } const auto sd_seed = keys.GetKey(S128KeyType::SDSeed); - if (!keys.HasKey(S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::Save))) + if (!keys.HasKey(S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::Save))) { return Loader::ResultStatus::ErrorMissingSDSaveKeySource; - if (!keys.HasKey(S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::NCA))) + } + if (!keys.HasKey(S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::NCA))) { return Loader::ResultStatus::ErrorMissingSDNCAKeySource; + } std::array<Key256, 2> sd_key_sources{ keys.GetKey(S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::Save)), @@ -334,8 +410,9 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke // Combine sources and seed for (auto& source : sd_key_sources) { - for (std::size_t i = 0; i < source.size(); ++i) - source[i] ^= sd_seed[i & 0xF]; + for (std::size_t i = 0; i < source.size(); ++i) { + source[i] = static_cast<u8>(source[i] ^ sd_seed[i & 0xF]); + } } AESCipher<Key128> cipher(sd_kek, Mode::ECB); @@ -353,9 +430,10 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke return Loader::ResultStatus::Success; } -std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save) { - if (!ticket_save.IsOpen()) +std::vector<Ticket> GetTicketblob(const Common::FS::IOFile& ticket_save) { + if (!ticket_save.IsOpen()) { return {}; + } std::vector<u8> buffer(ticket_save.GetSize()); if (ticket_save.ReadBytes(buffer.data(), buffer.size()) != buffer.size()) { @@ -415,7 +493,7 @@ static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) { offset = i + 1; break; } else if (data[i] != 0x0) { - return {}; + return std::nullopt; } } @@ -425,16 +503,18 @@ static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) { std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket, const RSAKeyPair<2048>& key) { const auto issuer = ticket.GetData().issuer; - if (IsAllZeroArray(issuer)) - return {}; + if (IsAllZeroArray(issuer)) { + return std::nullopt; + } if (issuer[0] != 'R' || issuer[1] != 'o' || issuer[2] != 'o' || issuer[3] != 't') { LOG_INFO(Crypto, "Attempting to parse ticket with non-standard certificate authority."); } Key128 rights_id = ticket.GetData().rights_id; - if (rights_id == Key128{}) - return {}; + if (rights_id == Key128{}) { + return std::nullopt; + } if (!std::any_of(ticket.GetData().title_key_common_pad.begin(), ticket.GetData().title_key_common_pad.end(), [](u8 b) { return b != 0; })) { @@ -466,15 +546,17 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket, std::array<u8, 0xDF> m_2; std::memcpy(m_2.data(), rsa_step.data() + 0x21, m_2.size()); - if (m_0 != 0) - return {}; + if (m_0 != 0) { + return std::nullopt; + } m_1 = m_1 ^ MGF1<0x20>(m_2); m_2 = m_2 ^ MGF1<0xDF>(m_1); const auto offset = FindTicketOffset(m_2); - if (!offset) - return {}; + if (!offset) { + return std::nullopt; + } ASSERT(*offset > 0); Key128 key_temp{}; @@ -485,8 +567,8 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket, KeyManager::KeyManager() { // Initialize keys - const std::string hactool_keys_dir = FileUtil::GetHactoolConfigurationPath(); - const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir); + const std::string hactool_keys_dir = Common::FS::GetHactoolConfigurationPath(); + const std::string yuzu_keys_dir = Common::FS::GetUserPath(Common::FS::UserPath::KeysDir); if (Settings::values.use_dev_keys) { dev_mode = true; AttemptLoadKeyFile(yuzu_keys_dir, hactool_keys_dir, "dev.keys", false); @@ -504,34 +586,39 @@ KeyManager::KeyManager() { } static bool ValidCryptoRevisionString(std::string_view base, size_t begin, size_t length) { - if (base.size() < begin + length) + if (base.size() < begin + length) { return false; + } return std::all_of(base.begin() + begin, base.begin() + begin + length, [](u8 c) { return std::isxdigit(c); }); } void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) { std::ifstream file; - OpenFStream(file, filename, std::ios_base::in); - if (!file.is_open()) + Common::FS::OpenFStream(file, filename, std::ios_base::in); + if (!file.is_open()) { return; + } std::string line; while (std::getline(file, line)) { std::vector<std::string> out; std::stringstream stream(line); std::string item; - while (std::getline(stream, item, '=')) + while (std::getline(stream, item, '=')) { out.push_back(std::move(item)); + } - if (out.size() != 2) + if (out.size() != 2) { continue; + } out[0].erase(std::remove(out[0].begin(), out[0].end(), ' '), out[0].end()); out[1].erase(std::remove(out[1].begin(), out[1].end(), ' '), out[1].end()); - if (out[0].compare(0, 1, "#") == 0) + if (out[0].compare(0, 1, "#") == 0) { continue; + } if (is_title_keys) { auto rights_id_raw = Common::HexStringToArray<16>(out[0]); @@ -541,24 +628,26 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) { s128_keys[{S128KeyType::Titlekey, rights_id[1], rights_id[0]}] = key; } else { out[0] = Common::ToLower(out[0]); - if (s128_file_id.find(out[0]) != s128_file_id.end()) { - const auto index = s128_file_id.at(out[0]); - Key128 key = Common::HexStringToArray<16>(out[1]); + if (const auto iter128 = Find128ByName(out[0]); iter128 != s128_file_id.end()) { + const auto& index = iter128->second; + const Key128 key = Common::HexStringToArray<16>(out[1]); s128_keys[{index.type, index.field1, index.field2}] = key; - } else if (s256_file_id.find(out[0]) != s256_file_id.end()) { - const auto index = s256_file_id.at(out[0]); - Key256 key = Common::HexStringToArray<32>(out[1]); + } else if (const auto iter256 = Find256ByName(out[0]); iter256 != s256_file_id.end()) { + const auto& index = iter256->second; + const Key256 key = Common::HexStringToArray<32>(out[1]); s256_keys[{index.type, index.field1, index.field2}] = key; } else if (out[0].compare(0, 8, "keyblob_") == 0 && out[0].compare(0, 9, "keyblob_k") != 0) { - if (!ValidCryptoRevisionString(out[0], 8, 2)) + if (!ValidCryptoRevisionString(out[0], 8, 2)) { continue; + } const auto index = std::stoul(out[0].substr(8, 2), nullptr, 16); keyblobs[index] = Common::HexStringToArray<0x90>(out[1]); } else if (out[0].compare(0, 18, "encrypted_keyblob_") == 0) { - if (!ValidCryptoRevisionString(out[0], 18, 2)) + if (!ValidCryptoRevisionString(out[0], 18, 2)) { continue; + } const auto index = std::stoul(out[0].substr(18, 2), nullptr, 16); encrypted_keyblobs[index] = Common::HexStringToArray<0xB0>(out[1]); @@ -566,8 +655,9 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) { eticket_extended_kek = Common::HexStringToArray<576>(out[1]); } else { for (const auto& kv : KEYS_VARIABLE_LENGTH) { - if (!ValidCryptoRevisionString(out[0], kv.second.size(), 2)) + if (!ValidCryptoRevisionString(out[0], kv.second.size(), 2)) { continue; + } if (out[0].compare(0, kv.second.size(), kv.second) == 0) { const auto index = std::stoul(out[0].substr(kv.second.size(), 2), nullptr, 16); @@ -602,10 +692,11 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) { void KeyManager::AttemptLoadKeyFile(const std::string& dir1, const std::string& dir2, const std::string& filename, bool title) { - if (FileUtil::Exists(dir1 + DIR_SEP + filename)) + if (Common::FS::Exists(dir1 + DIR_SEP + filename)) { LoadFromFile(dir1 + DIR_SEP + filename, title); - else if (FileUtil::Exists(dir2 + DIR_SEP + filename)) + } else if (Common::FS::Exists(dir2 + DIR_SEP + filename)) { LoadFromFile(dir2 + DIR_SEP + filename, title); + } } bool KeyManager::BaseDeriveNecessary() const { @@ -613,8 +704,9 @@ bool KeyManager::BaseDeriveNecessary() const { return !HasKey(key_type, index1, index2); }; - if (check_key_existence(S256KeyType::Header)) + if (check_key_existence(S256KeyType::Header)) { return true; + } for (size_t i = 0; i < CURRENT_CRYPTO_REVISION; ++i) { if (check_key_existence(S128KeyType::Master, i) || @@ -639,14 +731,16 @@ bool KeyManager::HasKey(S256KeyType id, u64 field1, u64 field2) const { } Key128 KeyManager::GetKey(S128KeyType id, u64 field1, u64 field2) const { - if (!HasKey(id, field1, field2)) + if (!HasKey(id, field1, field2)) { return {}; + } return s128_keys.at({id, field1, field2}); } Key256 KeyManager::GetKey(S256KeyType id, u64 field1, u64 field2) const { - if (!HasKey(id, field1, field2)) + if (!HasKey(id, field1, field2)) { return {}; + } return s256_keys.at({id, field1, field2}); } @@ -668,7 +762,7 @@ Key256 KeyManager::GetBISKey(u8 partition_id) const { template <size_t Size> void KeyManager::WriteKeyToFile(KeyCategory category, std::string_view keyname, const std::array<u8, Size>& key) { - const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir); + const std::string yuzu_keys_dir = Common::FS::GetUserPath(Common::FS::UserPath::KeysDir); std::string filename = "title.keys_autogenerated"; if (category == KeyCategory::Standard) { filename = dev_mode ? "dev.keys_autogenerated" : "prod.keys_autogenerated"; @@ -677,9 +771,9 @@ void KeyManager::WriteKeyToFile(KeyCategory category, std::string_view keyname, } const auto path = yuzu_keys_dir + DIR_SEP + filename; - const auto add_info_text = !FileUtil::Exists(path); - FileUtil::CreateFullPath(path); - FileUtil::IOFile file{path, "a"}; + const auto add_info_text = !Common::FS::Exists(path); + Common::FS::CreateFullPath(path); + Common::FS::IOFile file{path, "a"}; if (!file.IsOpen()) { return; } @@ -695,8 +789,9 @@ void KeyManager::WriteKeyToFile(KeyCategory category, std::string_view keyname, } void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) { - if (s128_keys.find({id, field1, field2}) != s128_keys.end()) + if (s128_keys.find({id, field1, field2}) != s128_keys.end() || key == Key128{}) { return; + } if (id == S128KeyType::Titlekey) { Key128 rights_id; std::memcpy(rights_id.data(), &field2, sizeof(u64)); @@ -711,19 +806,21 @@ void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) { } const auto iter2 = std::find_if( - s128_file_id.begin(), s128_file_id.end(), - [&id, &field1, &field2](const std::pair<std::string, KeyIndex<S128KeyType>> elem) { + s128_file_id.begin(), s128_file_id.end(), [&id, &field1, &field2](const auto& elem) { return std::tie(elem.second.type, elem.second.field1, elem.second.field2) == std::tie(id, field1, field2); }); - if (iter2 != s128_file_id.end()) + if (iter2 != s128_file_id.end()) { WriteKeyToFile(category, iter2->first, key); + } // Variable cases if (id == S128KeyType::KeyArea) { - static constexpr std::array<const char*, 3> kak_names = {"key_area_key_application_{:02X}", - "key_area_key_ocean_{:02X}", - "key_area_key_system_{:02X}"}; + static constexpr std::array<const char*, 3> kak_names = { + "key_area_key_application_{:02X}", + "key_area_key_ocean_{:02X}", + "key_area_key_system_{:02X}", + }; WriteKeyToFile(category, fmt::format(kak_names.at(field2), field1), key); } else if (id == S128KeyType::Master) { WriteKeyToFile(category, fmt::format("master_key_{:02X}", field1), key); @@ -745,43 +842,46 @@ void KeyManager::SetKey(S128KeyType id, Key128 key, u64 field1, u64 field2) { } void KeyManager::SetKey(S256KeyType id, Key256 key, u64 field1, u64 field2) { - if (s256_keys.find({id, field1, field2}) != s256_keys.end()) + if (s256_keys.find({id, field1, field2}) != s256_keys.end() || key == Key256{}) { return; + } const auto iter = std::find_if( - s256_file_id.begin(), s256_file_id.end(), - [&id, &field1, &field2](const std::pair<std::string, KeyIndex<S256KeyType>> elem) { + s256_file_id.begin(), s256_file_id.end(), [&id, &field1, &field2](const auto& elem) { return std::tie(elem.second.type, elem.second.field1, elem.second.field2) == std::tie(id, field1, field2); }); - if (iter != s256_file_id.end()) + if (iter != s256_file_id.end()) { WriteKeyToFile(KeyCategory::Standard, iter->first, key); + } s256_keys[{id, field1, field2}] = key; } bool KeyManager::KeyFileExists(bool title) { - const std::string hactool_keys_dir = FileUtil::GetHactoolConfigurationPath(); - const std::string yuzu_keys_dir = FileUtil::GetUserPath(FileUtil::UserPath::KeysDir); + const std::string hactool_keys_dir = Common::FS::GetHactoolConfigurationPath(); + const std::string yuzu_keys_dir = Common::FS::GetUserPath(Common::FS::UserPath::KeysDir); if (title) { - return FileUtil::Exists(hactool_keys_dir + DIR_SEP + "title.keys") || - FileUtil::Exists(yuzu_keys_dir + DIR_SEP + "title.keys"); + return Common::FS::Exists(hactool_keys_dir + DIR_SEP + "title.keys") || + Common::FS::Exists(yuzu_keys_dir + DIR_SEP + "title.keys"); } if (Settings::values.use_dev_keys) { - return FileUtil::Exists(hactool_keys_dir + DIR_SEP + "dev.keys") || - FileUtil::Exists(yuzu_keys_dir + DIR_SEP + "dev.keys"); + return Common::FS::Exists(hactool_keys_dir + DIR_SEP + "dev.keys") || + Common::FS::Exists(yuzu_keys_dir + DIR_SEP + "dev.keys"); } - return FileUtil::Exists(hactool_keys_dir + DIR_SEP + "prod.keys") || - FileUtil::Exists(yuzu_keys_dir + DIR_SEP + "prod.keys"); + return Common::FS::Exists(hactool_keys_dir + DIR_SEP + "prod.keys") || + Common::FS::Exists(yuzu_keys_dir + DIR_SEP + "prod.keys"); } void KeyManager::DeriveSDSeedLazy() { - if (HasKey(S128KeyType::SDSeed)) + if (HasKey(S128KeyType::SDSeed)) { return; + } const auto res = DeriveSDSeed(); - if (res) + if (res) { SetKey(S128KeyType::SDSeed, *res); + } } static Key128 CalculateCMAC(const u8* source, size_t size, const Key128& key) { @@ -793,11 +893,13 @@ static Key128 CalculateCMAC(const u8* source, size_t size, const Key128& key) { } void KeyManager::DeriveBase() { - if (!BaseDeriveNecessary()) + if (!BaseDeriveNecessary()) { return; + } - if (!HasKey(S128KeyType::SecureBoot) || !HasKey(S128KeyType::TSEC)) + if (!HasKey(S128KeyType::SecureBoot) || !HasKey(S128KeyType::TSEC)) { return; + } const auto has_bis = [this](u64 id) { return HasKey(S128KeyType::BIS, id, static_cast<u64>(BISKeyType::Crypto)) && @@ -814,10 +916,11 @@ void KeyManager::DeriveBase() { static_cast<u64>(BISKeyType::Tweak)); }; - if (has_bis(2) && !has_bis(3)) + if (has_bis(2) && !has_bis(3)) { copy_bis(2, 3); - else if (has_bis(3) && !has_bis(2)) + } else if (has_bis(3) && !has_bis(2)) { copy_bis(3, 2); + } std::bitset<32> revisions(0xFFFFFFFF); for (size_t i = 0; i < revisions.size(); ++i) { @@ -827,15 +930,17 @@ void KeyManager::DeriveBase() { } } - if (!revisions.any()) + if (!revisions.any()) { return; + } const auto sbk = GetKey(S128KeyType::SecureBoot); const auto tsec = GetKey(S128KeyType::TSEC); for (size_t i = 0; i < revisions.size(); ++i) { - if (!revisions[i]) + if (!revisions[i]) { continue; + } // Derive keyblob key const auto key = DeriveKeyblobKey( @@ -844,16 +949,18 @@ void KeyManager::DeriveBase() { SetKey(S128KeyType::Keyblob, key, i); // Derive keyblob MAC key - if (!HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC))) + if (!HasKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC))) { continue; + } const auto mac_key = DeriveKeyblobMACKey( key, GetKey(S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC))); SetKey(S128KeyType::KeyblobMAC, mac_key, i); Key128 cmac = CalculateCMAC(encrypted_keyblobs[i].data() + 0x10, 0xA0, mac_key); - if (std::memcmp(cmac.data(), encrypted_keyblobs[i].data(), cmac.size()) != 0) + if (std::memcmp(cmac.data(), encrypted_keyblobs[i].data(), cmac.size()) != 0) { continue; + } // Decrypt keyblob if (keyblobs[i] == std::array<u8, 0x90>{}) { @@ -877,16 +984,19 @@ void KeyManager::DeriveBase() { revisions.set(); for (size_t i = 0; i < revisions.size(); ++i) { - if (!HasKey(S128KeyType::Master, i)) + if (!HasKey(S128KeyType::Master, i)) { revisions.reset(i); + } } - if (!revisions.any()) + if (!revisions.any()) { return; + } for (size_t i = 0; i < revisions.size(); ++i) { - if (!revisions[i]) + if (!revisions[i]) { continue; + } // Derive general purpose keys DeriveGeneralPurposeKeys(i); @@ -911,21 +1021,24 @@ void KeyManager::DeriveBase() { } } -void KeyManager::DeriveETicket(PartitionDataManager& data) { +void KeyManager::DeriveETicket(PartitionDataManager& data, + const FileSys::ContentProvider& provider) { // ETicket keys - const auto es = Core::System::GetInstance().GetContentProvider().GetEntry( - 0x0100000000000033, FileSys::ContentRecordType::Program); + const auto es = provider.GetEntry(0x0100000000000033, FileSys::ContentRecordType::Program); - if (es == nullptr) + if (es == nullptr) { return; + } const auto exefs = es->GetExeFS(); - if (exefs == nullptr) + if (exefs == nullptr) { return; + } const auto main = exefs->GetFile("main"); - if (main == nullptr) + if (main == nullptr) { return; + } const auto bytes = main->ReadAllBytes(); @@ -935,16 +1048,19 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) { const auto seed3 = data.GetRSAKekSeed3(); const auto mask0 = data.GetRSAKekMask0(); - if (eticket_kek != Key128{}) + if (eticket_kek != Key128{}) { SetKey(S128KeyType::Source, eticket_kek, static_cast<size_t>(SourceKeyType::ETicketKek)); + } if (eticket_kekek != Key128{}) { SetKey(S128KeyType::Source, eticket_kekek, static_cast<size_t>(SourceKeyType::ETicketKekek)); } - if (seed3 != Key128{}) + if (seed3 != Key128{}) { SetKey(S128KeyType::RSAKek, seed3, static_cast<size_t>(RSAKekType::Seed3)); - if (mask0 != Key128{}) + } + if (mask0 != Key128{}) { SetKey(S128KeyType::RSAKek, mask0, static_cast<size_t>(RSAKekType::Mask0)); + } if (eticket_kek == Key128{} || eticket_kekek == Key128{} || seed3 == Key128{} || mask0 == Key128{}) { return; @@ -970,8 +1086,9 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) { AESCipher<Key128> es_kek(temp_kekek, Mode::ECB); es_kek.Transcode(eticket_kek.data(), eticket_kek.size(), eticket_final.data(), Op::Decrypt); - if (eticket_final == Key128{}) + if (eticket_final == Key128{}) { return; + } SetKey(S128KeyType::ETicketRSAKek, eticket_final); @@ -986,18 +1103,20 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) { void KeyManager::PopulateTickets() { const auto rsa_key = GetETicketRSAKey(); - if (rsa_key == RSAKeyPair<2048>{}) + if (rsa_key == RSAKeyPair<2048>{}) { return; + } - if (!common_tickets.empty() && !personal_tickets.empty()) + if (!common_tickets.empty() && !personal_tickets.empty()) { return; + } - const FileUtil::IOFile save1(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + - "/system/save/80000000000000e1", - "rb+"); - const FileUtil::IOFile save2(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + - "/system/save/80000000000000e2", - "rb+"); + const Common::FS::IOFile save1(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + + "/system/save/80000000000000e1", + "rb+"); + const Common::FS::IOFile save2(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + + "/system/save/80000000000000e2", + "rb+"); const auto blob2 = GetTicketblob(save2); auto res = GetTicketblob(save1); @@ -1007,8 +1126,10 @@ void KeyManager::PopulateTickets() { for (std::size_t i = 0; i < res.size(); ++i) { const auto common = i < idx; const auto pair = ParseTicket(res[i], rsa_key); - if (!pair) + if (!pair) { continue; + } + const auto& [rid, key] = *pair; u128 rights_id; std::memcpy(rights_id.data(), rid.data(), rid.size()); @@ -1037,27 +1158,33 @@ void KeyManager::SynthesizeTickets() { } void KeyManager::SetKeyWrapped(S128KeyType id, Key128 key, u64 field1, u64 field2) { - if (key == Key128{}) + if (key == Key128{}) { return; + } SetKey(id, key, field1, field2); } void KeyManager::SetKeyWrapped(S256KeyType id, Key256 key, u64 field1, u64 field2) { - if (key == Key256{}) + if (key == Key256{}) { return; + } + SetKey(id, key, field1, field2); } void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) { - if (!BaseDeriveNecessary()) + if (!BaseDeriveNecessary()) { return; + } - if (!data.HasBoot0()) + if (!data.HasBoot0()) { return; + } for (size_t i = 0; i < encrypted_keyblobs.size(); ++i) { - if (encrypted_keyblobs[i] != std::array<u8, 0xB0>{}) + if (encrypted_keyblobs[i] != std::array<u8, 0xB0>{}) { continue; + } encrypted_keyblobs[i] = data.GetEncryptedKeyblob(i); WriteKeyToFile<0xB0>(KeyCategory::Console, fmt::format("encrypted_keyblob_{:02X}", i), encrypted_keyblobs[i]); @@ -1079,8 +1206,9 @@ void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) { static_cast<u64>(SourceKeyType::Keyblob), i); } - if (data.HasFuses()) + if (data.HasFuses()) { SetKeyWrapped(S128KeyType::SecureBoot, data.GetSecureBootKey()); + } DeriveBase(); @@ -1094,8 +1222,9 @@ void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) { const auto masters = data.GetTZMasterKeys(latest_master); for (size_t i = 0; i < masters.size(); ++i) { - if (masters[i] != Key128{} && !HasKey(S128KeyType::Master, i)) + if (masters[i] != Key128{} && !HasKey(S128KeyType::Master, i)) { SetKey(S128KeyType::Master, masters[i], i); + } } DeriveBase(); @@ -1105,8 +1234,9 @@ void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) { std::array<Key128, 0x20> package2_keys{}; for (size_t i = 0; i < package2_keys.size(); ++i) { - if (HasKey(S128KeyType::Package2, i)) + if (HasKey(S128KeyType::Package2, i)) { package2_keys[i] = GetKey(S128KeyType::Package2, i); + } } data.DecryptPackage2(package2_keys, Package2Type::NormalMain); @@ -1144,12 +1274,15 @@ const std::map<u128, Ticket>& KeyManager::GetPersonalizedTickets() const { bool KeyManager::AddTicketCommon(Ticket raw) { const auto rsa_key = GetETicketRSAKey(); - if (rsa_key == RSAKeyPair<2048>{}) + if (rsa_key == RSAKeyPair<2048>{}) { return false; + } const auto pair = ParseTicket(raw, rsa_key); - if (!pair) + if (!pair) { return false; + } + const auto& [rid, key] = *pair; u128 rights_id; std::memcpy(rights_id.data(), rid.data(), rid.size()); @@ -1160,12 +1293,15 @@ bool KeyManager::AddTicketCommon(Ticket raw) { bool KeyManager::AddTicketPersonalized(Ticket raw) { const auto rsa_key = GetETicketRSAKey(); - if (rsa_key == RSAKeyPair<2048>{}) + if (rsa_key == RSAKeyPair<2048>{}) { return false; + } const auto pair = ParseTicket(raw, rsa_key); - if (!pair) + if (!pair) { return false; + } + const auto& [rid, key] = *pair; u128 rights_id; std::memcpy(rights_id.data(), rid.data(), rid.size()); @@ -1173,57 +1309,4 @@ bool KeyManager::AddTicketPersonalized(Ticket raw) { SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); return true; } - -const boost::container::flat_map<std::string, KeyIndex<S128KeyType>> KeyManager::s128_file_id = { - {"eticket_rsa_kek", {S128KeyType::ETicketRSAKek, 0, 0}}, - {"eticket_rsa_kek_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::ETicketKek), 0}}, - {"eticket_rsa_kekek_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::ETicketKekek), 0}}, - {"rsa_kek_mask_0", {S128KeyType::RSAKek, static_cast<u64>(RSAKekType::Mask0), 0}}, - {"rsa_kek_seed_3", {S128KeyType::RSAKek, static_cast<u64>(RSAKekType::Seed3), 0}}, - {"rsa_oaep_kek_generation_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::RSAOaepKekGeneration), 0}}, - {"sd_card_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::SDKek), 0}}, - {"aes_kek_generation_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKekGeneration), 0}}, - {"aes_key_generation_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::AESKeyGeneration), 0}}, - {"package2_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Package2), 0}}, - {"master_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Master), 0}}, - {"header_kek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::HeaderKek), 0}}, - {"key_area_key_application_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey), - static_cast<u64>(KeyAreaKeyType::Application)}}, - {"key_area_key_ocean_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey), - static_cast<u64>(KeyAreaKeyType::Ocean)}}, - {"key_area_key_system_source", - {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyAreaKey), - static_cast<u64>(KeyAreaKeyType::System)}}, - {"titlekek_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::Titlekek), 0}}, - {"keyblob_mac_key_source", {S128KeyType::Source, static_cast<u64>(SourceKeyType::KeyblobMAC)}}, - {"tsec_key", {S128KeyType::TSEC, 0, 0}}, - {"secure_boot_key", {S128KeyType::SecureBoot, 0, 0}}, - {"sd_seed", {S128KeyType::SDSeed, 0, 0}}, - {"bis_key_0_crypt", {S128KeyType::BIS, 0, static_cast<u64>(BISKeyType::Crypto)}}, - {"bis_key_0_tweak", {S128KeyType::BIS, 0, static_cast<u64>(BISKeyType::Tweak)}}, - {"bis_key_1_crypt", {S128KeyType::BIS, 1, static_cast<u64>(BISKeyType::Crypto)}}, - {"bis_key_1_tweak", {S128KeyType::BIS, 1, static_cast<u64>(BISKeyType::Tweak)}}, - {"bis_key_2_crypt", {S128KeyType::BIS, 2, static_cast<u64>(BISKeyType::Crypto)}}, - {"bis_key_2_tweak", {S128KeyType::BIS, 2, static_cast<u64>(BISKeyType::Tweak)}}, - {"bis_key_3_crypt", {S128KeyType::BIS, 3, static_cast<u64>(BISKeyType::Crypto)}}, - {"bis_key_3_tweak", {S128KeyType::BIS, 3, static_cast<u64>(BISKeyType::Tweak)}}, - {"header_kek", {S128KeyType::HeaderKek, 0, 0}}, - {"sd_card_kek", {S128KeyType::SDKek, 0, 0}}, -}; - -const boost::container::flat_map<std::string, KeyIndex<S256KeyType>> KeyManager::s256_file_id = { - {"header_key", {S256KeyType::Header, 0, 0}}, - {"sd_card_save_key_source", {S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::Save), 0}}, - {"sd_card_nca_key_source", {S256KeyType::SDKeySource, static_cast<u64>(SDKeyType::NCA), 0}}, - {"header_key_source", {S256KeyType::HeaderSource, 0, 0}}, - {"sd_card_save_key", {S256KeyType::SDKey, static_cast<u64>(SDKeyType::Save), 0}}, - {"sd_card_nca_key", {S256KeyType::SDKey, static_cast<u64>(SDKeyType::NCA), 0}}, -}; } // namespace Core::Crypto diff --git a/src/core/crypto/key_manager.h b/src/core/crypto/key_manager.h index 7265c4171..0a7220286 100644 --- a/src/core/crypto/key_manager.h +++ b/src/core/crypto/key_manager.h @@ -10,17 +10,20 @@ #include <string> #include <variant> -#include <boost/container/flat_map.hpp> #include <fmt/format.h> #include "common/common_funcs.h" #include "common/common_types.h" #include "core/crypto/partition_data_manager.h" #include "core/file_sys/vfs_types.h" -namespace FileUtil { +namespace Common::FS { class IOFile; } +namespace FileSys { +class ContentProvider; +} + namespace Loader { enum class ResultStatus : u16; } @@ -223,7 +226,16 @@ bool operator<(const KeyIndex<KeyType>& lhs, const KeyIndex<KeyType>& rhs) { class KeyManager { public: - KeyManager(); + static KeyManager& Instance() { + static KeyManager instance; + return instance; + } + + KeyManager(const KeyManager&) = delete; + KeyManager& operator=(const KeyManager&) = delete; + + KeyManager(KeyManager&&) = delete; + KeyManager& operator=(KeyManager&&) = delete; bool HasKey(S128KeyType id, u64 field1 = 0, u64 field2 = 0) const; bool HasKey(S256KeyType id, u64 field1 = 0, u64 field2 = 0) const; @@ -244,7 +256,7 @@ public: bool BaseDeriveNecessary() const; void DeriveBase(); - void DeriveETicket(PartitionDataManager& data); + void DeriveETicket(PartitionDataManager& data, const FileSys::ContentProvider& provider); void PopulateTickets(); void SynthesizeTickets(); @@ -257,6 +269,8 @@ public: bool AddTicketPersonalized(Ticket raw); private: + KeyManager(); + std::map<KeyIndex<S128KeyType>, Key128> s128_keys; std::map<KeyIndex<S256KeyType>, Key256> s256_keys; @@ -282,9 +296,6 @@ private: void SetKeyWrapped(S128KeyType id, Key128 key, u64 field1 = 0, u64 field2 = 0); void SetKeyWrapped(S256KeyType id, Key256 key, u64 field1 = 0, u64 field2 = 0); - - static const boost::container::flat_map<std::string, KeyIndex<S128KeyType>> s128_file_id; - static const boost::container::flat_map<std::string, KeyIndex<S256KeyType>> s256_file_id; }; Key128 GenerateKeyEncryptionKey(Key128 source, Key128 master, Key128 kek_seed, Key128 key_seed); @@ -297,7 +308,7 @@ std::array<u8, 0x90> DecryptKeyblob(const std::array<u8, 0xB0>& encrypted_keyblo std::optional<Key128> DeriveSDSeed(); Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys); -std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save); +std::vector<Ticket> GetTicketblob(const Common::FS::IOFile& ticket_save); // Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority // (offset 0x140-0x144 is zero) diff --git a/src/core/crypto/partition_data_manager.cpp b/src/core/crypto/partition_data_manager.cpp index d64302f2e..5f1c86a09 100644 --- a/src/core/crypto/partition_data_manager.cpp +++ b/src/core/crypto/partition_data_manager.cpp @@ -26,8 +26,9 @@ #include "core/file_sys/vfs.h" #include "core/file_sys/vfs_offset.h" #include "core/file_sys/vfs_vector.h" +#include "core/loader/loader.h" -using namespace Common; +using Common::AsArray; namespace Core::Crypto { @@ -47,105 +48,123 @@ struct Package2Header { }; static_assert(sizeof(Package2Header) == 0x200, "Package2Header has incorrect size."); -const std::array<SHA256Hash, 0x10> source_hashes{ - "B24BD293259DBC7AC5D63F88E60C59792498E6FC5443402C7FFE87EE8B61A3F0"_array32, // keyblob_mac_key_source - "7944862A3A5C31C6720595EFD302245ABD1B54CCDCF33000557681E65C5664A4"_array32, // master_key_source - "21E2DF100FC9E094DB51B47B9B1D6E94ED379DB8B547955BEF8FE08D8DD35603"_array32, // package2_key_source - "FC02B9D37B42D7A1452E71444F1F700311D1132E301A83B16062E72A78175085"_array32, // aes_kek_generation_source - "FBD10056999EDC7ACDB96098E47E2C3606230270D23281E671F0F389FC5BC585"_array32, // aes_key_generation_source - "C48B619827986C7F4E3081D59DB2B460C84312650E9A8E6B458E53E8CBCA4E87"_array32, // titlekek_source - "04AD66143C726B2A139FB6B21128B46F56C553B2B3887110304298D8D0092D9E"_array32, // key_area_key_application_source - "FD434000C8FF2B26F8E9A9D2D2C12F6BE5773CBB9DC86300E1BD99F8EA33A417"_array32, // key_area_key_ocean_source - "1F17B1FD51AD1C2379B58F152CA4912EC2106441E51722F38700D5937A1162F7"_array32, // key_area_key_system_source - "6B2ED877C2C52334AC51E59ABFA7EC457F4A7D01E46291E9F2EAA45F011D24B7"_array32, // sd_card_kek_source - "D482743563D3EA5DCDC3B74E97C9AC8A342164FA041A1DC80F17F6D31E4BC01C"_array32, // sd_card_save_key_source - "2E751CECF7D93A2B957BD5FFCB082FD038CC2853219DD3092C6DAB9838F5A7CC"_array32, // sd_card_nca_key_source - "1888CAED5551B3EDE01499E87CE0D86827F80820EFB275921055AA4E2ABDFFC2"_array32, // header_kek_source - "8F783E46852DF6BE0BA4E19273C4ADBAEE16380043E1B8C418C4089A8BD64AA6"_array32, // header_key_source - "D1757E52F1AE55FA882EC690BC6F954AC46A83DC22F277F8806BD55577C6EED7"_array32, // rsa_kek_seed3 - "FC02B9D37B42D7A1452E71444F1F700311D1132E301A83B16062E72A78175085"_array32, // rsa_kek_mask0 +// clang-format off +constexpr std::array source_hashes{ + AsArray("B24BD293259DBC7AC5D63F88E60C59792498E6FC5443402C7FFE87EE8B61A3F0"), // keyblob_mac_key_source + AsArray("7944862A3A5C31C6720595EFD302245ABD1B54CCDCF33000557681E65C5664A4"), // master_key_source + AsArray("21E2DF100FC9E094DB51B47B9B1D6E94ED379DB8B547955BEF8FE08D8DD35603"), // package2_key_source + AsArray("FC02B9D37B42D7A1452E71444F1F700311D1132E301A83B16062E72A78175085"), // aes_kek_generation_source + AsArray("FBD10056999EDC7ACDB96098E47E2C3606230270D23281E671F0F389FC5BC585"), // aes_key_generation_source + AsArray("C48B619827986C7F4E3081D59DB2B460C84312650E9A8E6B458E53E8CBCA4E87"), // titlekek_source + AsArray("04AD66143C726B2A139FB6B21128B46F56C553B2B3887110304298D8D0092D9E"), // key_area_key_application_source + AsArray("FD434000C8FF2B26F8E9A9D2D2C12F6BE5773CBB9DC86300E1BD99F8EA33A417"), // key_area_key_ocean_source + AsArray("1F17B1FD51AD1C2379B58F152CA4912EC2106441E51722F38700D5937A1162F7"), // key_area_key_system_source + AsArray("6B2ED877C2C52334AC51E59ABFA7EC457F4A7D01E46291E9F2EAA45F011D24B7"), // sd_card_kek_source + AsArray("D482743563D3EA5DCDC3B74E97C9AC8A342164FA041A1DC80F17F6D31E4BC01C"), // sd_card_save_key_source + AsArray("2E751CECF7D93A2B957BD5FFCB082FD038CC2853219DD3092C6DAB9838F5A7CC"), // sd_card_nca_key_source + AsArray("1888CAED5551B3EDE01499E87CE0D86827F80820EFB275921055AA4E2ABDFFC2"), // header_kek_source + AsArray("8F783E46852DF6BE0BA4E19273C4ADBAEE16380043E1B8C418C4089A8BD64AA6"), // header_key_source + AsArray("D1757E52F1AE55FA882EC690BC6F954AC46A83DC22F277F8806BD55577C6EED7"), // rsa_kek_seed3 + AsArray("FC02B9D37B42D7A1452E71444F1F700311D1132E301A83B16062E72A78175085"), // rsa_kek_mask0 }; - -const std::array<SHA256Hash, 0x20> keyblob_source_hashes{ - "8A06FE274AC491436791FDB388BCDD3AB9943BD4DEF8094418CDAC150FD73786"_array32, // keyblob_key_source_00 - "2D5CAEB2521FEF70B47E17D6D0F11F8CE2C1E442A979AD8035832C4E9FBCCC4B"_array32, // keyblob_key_source_01 - "61C5005E713BAE780641683AF43E5F5C0E03671117F702F401282847D2FC6064"_array32, // keyblob_key_source_02 - "8E9795928E1C4428E1B78F0BE724D7294D6934689C11B190943923B9D5B85903"_array32, // keyblob_key_source_03 - "95FA33AF95AFF9D9B61D164655B32710ED8D615D46C7D6CC3CC70481B686B402"_array32, // keyblob_key_source_04 - "3F5BE7B3C8B1ABD8C10B4B703D44766BA08730562C172A4FE0D6B866B3E2DB3E"_array32, // keyblob_key_source_05 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_06 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_07 - - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_08 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_09 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0A - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0B - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0C - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0D - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0E - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_0F - - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_10 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_11 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_12 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_13 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_14 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_15 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_16 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_17 - - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_18 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_19 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1A - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1B - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1C - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1D - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1E - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // keyblob_key_source_1F +// clang-format on + +// clang-format off +constexpr std::array keyblob_source_hashes{ + AsArray("8A06FE274AC491436791FDB388BCDD3AB9943BD4DEF8094418CDAC150FD73786"), // keyblob_key_source_00 + AsArray("2D5CAEB2521FEF70B47E17D6D0F11F8CE2C1E442A979AD8035832C4E9FBCCC4B"), // keyblob_key_source_01 + AsArray("61C5005E713BAE780641683AF43E5F5C0E03671117F702F401282847D2FC6064"), // keyblob_key_source_02 + AsArray("8E9795928E1C4428E1B78F0BE724D7294D6934689C11B190943923B9D5B85903"), // keyblob_key_source_03 + AsArray("95FA33AF95AFF9D9B61D164655B32710ED8D615D46C7D6CC3CC70481B686B402"), // keyblob_key_source_04 + AsArray("3F5BE7B3C8B1ABD8C10B4B703D44766BA08730562C172A4FE0D6B866B3E2DB3E"), // keyblob_key_source_05 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_06 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_07 + + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_08 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_09 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_0A + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_0B + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_0C + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_0D + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_0E + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_0F + + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_10 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_11 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_12 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_13 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_14 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_15 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_16 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_17 + + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_18 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_19 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_1A + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_1B + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_1C + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_1D + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_1E + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // keyblob_key_source_1F }; - -const std::array<SHA256Hash, 0x20> master_key_hashes{ - "0EE359BE3C864BB0782E1D70A718A0342C551EED28C369754F9C4F691BECF7CA"_array32, // master_key_00 - "4FE707B7E4ABDAF727C894AAF13B1351BFE2AC90D875F73B2E20FA94B9CC661E"_array32, // master_key_01 - "79277C0237A2252EC3DFAC1F7C359C2B3D121E9DB15BB9AB4C2B4408D2F3AE09"_array32, // master_key_02 - "4F36C565D13325F65EE134073C6A578FFCB0008E02D69400836844EAB7432754"_array32, // master_key_03 - "75FF1D95D26113550EE6FCC20ACB58E97EDEB3A2FF52543ED5AEC63BDCC3DA50"_array32, // master_key_04 - "EBE2BCD6704673EC0F88A187BB2AD9F1CC82B718C389425941BDC194DC46B0DD"_array32, // master_key_05 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_06 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_07 - - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_08 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_09 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0A - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0B - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0C - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0D - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0E - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_0F - - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_10 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_11 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_12 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_13 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_14 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_15 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_16 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_17 - - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_18 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_19 - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1A - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1B - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1C - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1D - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1E - "0000000000000000000000000000000000000000000000000000000000000000"_array32, // master_key_1F +// clang-format on + +// clang-format off +constexpr std::array master_key_hashes{ + AsArray("0EE359BE3C864BB0782E1D70A718A0342C551EED28C369754F9C4F691BECF7CA"), // master_key_00 + AsArray("4FE707B7E4ABDAF727C894AAF13B1351BFE2AC90D875F73B2E20FA94B9CC661E"), // master_key_01 + AsArray("79277C0237A2252EC3DFAC1F7C359C2B3D121E9DB15BB9AB4C2B4408D2F3AE09"), // master_key_02 + AsArray("4F36C565D13325F65EE134073C6A578FFCB0008E02D69400836844EAB7432754"), // master_key_03 + AsArray("75FF1D95D26113550EE6FCC20ACB58E97EDEB3A2FF52543ED5AEC63BDCC3DA50"), // master_key_04 + AsArray("EBE2BCD6704673EC0F88A187BB2AD9F1CC82B718C389425941BDC194DC46B0DD"), // master_key_05 + AsArray("9497E6779F5D840F2BBA1DE4E95BA1D6F21EFC94717D5AE5CA37D7EC5BD37A19"), // master_key_06 + AsArray("4EC96B8CB01B8DCE382149443430B2B6EBCB2983348AFA04A25E53609DABEDF6"), // master_key_07 + + AsArray("2998E2E23609BC2675FF062A2D64AF5B1B78DFF463B24119D64A1B64F01B2D51"), // master_key_08 + AsArray("9D486A98067C44B37CF173D3BF577891EB6081FF6B4A166347D9DBBF7025076B"), // master_key_09 + AsArray("4EC5A237A75A083A9C5F6CF615601522A7F822D06BD4BA32612C9CEBBB29BD45"), // master_key_0A + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_0B + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_0C + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_0D + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_0E + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_0F + + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_10 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_11 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_12 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_13 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_14 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_15 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_16 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_17 + + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_18 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_19 + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_1A + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_1B + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_1C + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_1D + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_1E + AsArray("0000000000000000000000000000000000000000000000000000000000000000"), // master_key_1F }; +// clang-format on + +static constexpr u8 CalculateMaxKeyblobSourceHash() { + const auto is_zero = [](const auto& data) { + // TODO: Replace with std::all_of whenever mingw decides to update their + // libraries to include the constexpr variant of it. + for (const auto element : data) { + if (element != 0) { + return false; + } + } + return true; + }; -static u8 CalculateMaxKeyblobSourceHash() { for (s8 i = 0x1F; i >= 0; --i) { - if (keyblob_source_hashes[i] != SHA256Hash{}) + if (!is_zero(keyblob_source_hashes[i])) { return static_cast<u8>(i + 1); + } } return 0; @@ -202,8 +221,8 @@ static std::array<Key128, 0x20> FindEncryptedMasterKeyFromHex(const std::vector< return out; } -FileSys::VirtualFile FindFileInDirWithNames(const FileSys::VirtualDir& dir, - const std::string& name) { +static FileSys::VirtualFile FindFileInDirWithNames(const FileSys::VirtualDir& dir, + const std::string& name) { const auto upper = Common::ToUpper(name); for (const auto& fname : {name, name + ".bin", upper, upper + ".BIN"}) { @@ -345,14 +364,12 @@ FileSys::VirtualFile PartitionDataManager::GetPackage2Raw(Package2Type type) con return package2.at(static_cast<size_t>(type)); } -bool AttemptDecrypt(const std::array<u8, 16>& key, Package2Header& header) { - - const std::vector<u8> iv(header.header_ctr.begin(), header.header_ctr.end()); +static bool AttemptDecrypt(const std::array<u8, 16>& key, Package2Header& header) { Package2Header temp = header; AESCipher<Key128> cipher(key, Mode::CTR); - cipher.SetIV(iv); - cipher.Transcode(&temp.header_ctr, sizeof(Package2Header) - 0x100, &temp.header_ctr, - Op::Decrypt); + cipher.SetIV(header.header_ctr); + cipher.Transcode(&temp.header_ctr, sizeof(Package2Header) - sizeof(Package2Header::signature), + &temp.header_ctr, Op::Decrypt); if (temp.magic == Common::MakeMagic('P', 'K', '2', '1')) { header = temp; return true; @@ -389,7 +406,7 @@ void PartitionDataManager::DecryptPackage2(const std::array<Key128, 0x20>& packa auto c = a->ReadAllBytes(); AESCipher<Key128> cipher(package2_keys[revision], Mode::CTR); - cipher.SetIV({header.section_ctr[1].begin(), header.section_ctr[1].end()}); + cipher.SetIV(header.section_ctr[1]); cipher.Transcode(c.data(), c.size(), c.data(), Op::Decrypt); const auto ini_file = std::make_shared<FileSys::VectorVfsFile>(c); diff --git a/src/core/device_memory.cpp b/src/core/device_memory.cpp new file mode 100644 index 000000000..0c4b440ed --- /dev/null +++ b/src/core/device_memory.cpp @@ -0,0 +1,12 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/device_memory.h" + +namespace Core { + +DeviceMemory::DeviceMemory() : buffer{DramMemoryMap::Size} {} +DeviceMemory::~DeviceMemory() = default; + +} // namespace Core diff --git a/src/core/device_memory.h b/src/core/device_memory.h new file mode 100644 index 000000000..5b1ae28f3 --- /dev/null +++ b/src/core/device_memory.h @@ -0,0 +1,47 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" +#include "common/virtual_buffer.h" + +namespace Core { + +namespace DramMemoryMap { +enum : u64 { + Base = 0x80000000ULL, + Size = 0x100000000ULL, + End = Base + Size, + KernelReserveBase = Base + 0x60000, + SlabHeapBase = KernelReserveBase + 0x85000, + SlapHeapSize = 0xa21000, + SlabHeapEnd = SlabHeapBase + SlapHeapSize, +}; +}; // namespace DramMemoryMap + +class DeviceMemory : NonCopyable { +public: + explicit DeviceMemory(); + ~DeviceMemory(); + + template <typename T> + PAddr GetPhysicalAddr(const T* ptr) const { + return (reinterpret_cast<uintptr_t>(ptr) - reinterpret_cast<uintptr_t>(buffer.data())) + + DramMemoryMap::Base; + } + + u8* GetPointer(PAddr addr) { + return buffer.data() + (addr - DramMemoryMap::Base); + } + + const u8* GetPointer(PAddr addr) const { + return buffer.data() + (addr - DramMemoryMap::Base); + } + +private: + Common::VirtualBuffer<u8> buffer; +}; + +} // namespace Core diff --git a/src/core/file_sys/bis_factory.cpp b/src/core/file_sys/bis_factory.cpp index 0af44f340..7c6304ff0 100644 --- a/src/core/file_sys/bis_factory.cpp +++ b/src/core/file_sys/bis_factory.cpp @@ -4,14 +4,17 @@ #include <fmt/format.h> #include "common/file_util.h" -#include "core/core.h" #include "core/file_sys/bis_factory.h" #include "core/file_sys/mode.h" #include "core/file_sys/registered_cache.h" -#include "core/settings.h" +#include "core/file_sys/vfs.h" namespace FileSys { +constexpr u64 NAND_USER_SIZE = 0x680000000; // 26624 MiB +constexpr u64 NAND_SYSTEM_SIZE = 0xA0000000; // 2560 MiB +constexpr u64 NAND_TOTAL_SIZE = 0x747C00000; // 29820 MiB + BISFactory::BISFactory(VirtualDir nand_root_, VirtualDir load_root_, VirtualDir dump_root_) : nand_root(std::move(nand_root_)), load_root(std::move(load_root_)), dump_root(std::move(dump_root_)), @@ -78,11 +81,11 @@ VirtualDir BISFactory::OpenPartition(BisPartitionId id) const { } } -VirtualFile BISFactory::OpenPartitionStorage(BisPartitionId id) const { - Core::Crypto::KeyManager keys; - Core::Crypto::PartitionDataManager pdm{ - Core::System::GetInstance().GetFilesystem()->OpenDirectory( - FileUtil::GetUserPath(FileUtil::UserPath::SysDataDir), Mode::Read)}; +VirtualFile BISFactory::OpenPartitionStorage(BisPartitionId id, + VirtualFilesystem file_system) const { + auto& keys = Core::Crypto::KeyManager::Instance(); + Core::Crypto::PartitionDataManager pdm{file_system->OpenDirectory( + Common::FS::GetUserPath(Common::FS::UserPath::SysDataDir), Mode::Read)}; keys.PopulateFromPartitionData(pdm); switch (id) { @@ -110,30 +113,29 @@ VirtualDir BISFactory::GetImageDirectory() const { u64 BISFactory::GetSystemNANDFreeSpace() const { const auto sys_dir = GetOrCreateDirectoryRelative(nand_root, "/system"); - if (sys_dir == nullptr) - return 0; + if (sys_dir == nullptr) { + return GetSystemNANDTotalSpace(); + } return GetSystemNANDTotalSpace() - sys_dir->GetSize(); } u64 BISFactory::GetSystemNANDTotalSpace() const { - return static_cast<u64>(Settings::values.nand_system_size); + return NAND_SYSTEM_SIZE; } u64 BISFactory::GetUserNANDFreeSpace() const { - const auto usr_dir = GetOrCreateDirectoryRelative(nand_root, "/user"); - if (usr_dir == nullptr) - return 0; - - return GetUserNANDTotalSpace() - usr_dir->GetSize(); + // For some reason games such as BioShock 1 checks whether this is exactly 0x680000000 bytes. + // Set the free space to be 1 MiB less than the total as a workaround to this issue. + return GetUserNANDTotalSpace() - 0x100000; } u64 BISFactory::GetUserNANDTotalSpace() const { - return static_cast<u64>(Settings::values.nand_user_size); + return NAND_USER_SIZE; } u64 BISFactory::GetFullNANDTotalSpace() const { - return static_cast<u64>(Settings::values.nand_total_size); + return NAND_TOTAL_SIZE; } VirtualDir BISFactory::GetBCATDirectory(u64 title_id) const { diff --git a/src/core/file_sys/bis_factory.h b/src/core/file_sys/bis_factory.h index 8f0451c98..136485881 100644 --- a/src/core/file_sys/bis_factory.h +++ b/src/core/file_sys/bis_factory.h @@ -6,7 +6,8 @@ #include <memory> -#include "core/file_sys/vfs.h" +#include "common/common_types.h" +#include "core/file_sys/vfs_types.h" namespace FileSys { @@ -51,7 +52,7 @@ public: VirtualDir GetModificationDumpRoot(u64 title_id) const; VirtualDir OpenPartition(BisPartitionId id) const; - VirtualFile OpenPartitionStorage(BisPartitionId id) const; + VirtualFile OpenPartitionStorage(BisPartitionId id, VirtualFilesystem file_system) const; VirtualDir GetImageDirectory() const; diff --git a/src/core/file_sys/card_image.cpp b/src/core/file_sys/card_image.cpp index 07d0c8d5d..956da68f7 100644 --- a/src/core/file_sys/card_image.cpp +++ b/src/core/file_sys/card_image.cpp @@ -8,11 +8,11 @@ #include <fmt/ostream.h> #include "common/logging/log.h" +#include "core/crypto/key_manager.h" #include "core/file_sys/card_image.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/nca_metadata.h" #include "core/file_sys/partition_filesystem.h" -#include "core/file_sys/romfs.h" #include "core/file_sys/submission_package.h" #include "core/file_sys/vfs_concat.h" #include "core/file_sys/vfs_offset.h" @@ -31,7 +31,8 @@ constexpr std::array partition_names{ XCI::XCI(VirtualFile file_) : file(std::move(file_)), program_nca_status{Loader::ResultStatus::ErrorXCIMissingProgramNCA}, - partitions(partition_names.size()), partitions_raw(partition_names.size()) { + partitions(partition_names.size()), + partitions_raw(partition_names.size()), keys{Core::Crypto::KeyManager::Instance()} { if (file->ReadObject(&header) != sizeof(GamecardHeader)) { status = Loader::ResultStatus::ErrorBadXCIHeader; return; @@ -178,7 +179,7 @@ u32 XCI::GetSystemUpdateVersion() { return 0; for (const auto& file : update->GetFiles()) { - NCA nca{file, nullptr, 0, keys}; + NCA nca{file, nullptr, 0}; if (nca.GetStatus() != Loader::ResultStatus::Success) continue; @@ -286,7 +287,7 @@ Loader::ResultStatus XCI::AddNCAFromPartition(XCIPartition part) { continue; } - auto nca = std::make_shared<NCA>(file, nullptr, 0, keys); + auto nca = std::make_shared<NCA>(file, nullptr, 0); if (nca->IsUpdate()) { continue; } diff --git a/src/core/file_sys/card_image.h b/src/core/file_sys/card_image.h index c2ee0ea99..2d0a0f285 100644 --- a/src/core/file_sys/card_image.h +++ b/src/core/file_sys/card_image.h @@ -9,9 +9,12 @@ #include <vector> #include "common/common_types.h" #include "common/swap.h" -#include "core/crypto/key_manager.h" #include "core/file_sys/vfs.h" +namespace Core::Crypto { +class KeyManager; +} + namespace Loader { enum class ResultStatus : u16; } @@ -140,6 +143,6 @@ private: u64 update_normal_partition_end; - Core::Crypto::KeyManager keys; + Core::Crypto::KeyManager& keys; }; } // namespace FileSys diff --git a/src/core/file_sys/content_archive.cpp b/src/core/file_sys/content_archive.cpp index b8bbdd1ef..76af47ff9 100644 --- a/src/core/file_sys/content_archive.cpp +++ b/src/core/file_sys/content_archive.cpp @@ -10,10 +10,10 @@ #include "common/logging/log.h" #include "core/crypto/aes_util.h" #include "core/crypto/ctr_encryption_layer.h" +#include "core/crypto/key_manager.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/nca_patch.h" #include "core/file_sys/partition_filesystem.h" -#include "core/file_sys/romfs.h" #include "core/file_sys/vfs_offset.h" #include "core/loader/loader.h" @@ -118,9 +118,9 @@ static bool IsValidNCA(const NCAHeader& header) { return header.magic == Common::MakeMagic('N', 'C', 'A', '3'); } -NCA::NCA(VirtualFile file_, VirtualFile bktr_base_romfs_, u64 bktr_base_ivfc_offset, - Core::Crypto::KeyManager keys_) - : file(std::move(file_)), bktr_base_romfs(std::move(bktr_base_romfs_)), keys(std::move(keys_)) { +NCA::NCA(VirtualFile file_, VirtualFile bktr_base_romfs_, u64 bktr_base_ivfc_offset) + : file(std::move(file_)), + bktr_base_romfs(std::move(bktr_base_romfs_)), keys{Core::Crypto::KeyManager::Instance()} { if (file == nullptr) { status = Loader::ResultStatus::ErrorNullFile; return; @@ -323,7 +323,7 @@ bool NCA::ReadRomFSSection(const NCASectionHeader& section, const NCASectionTabl subsection_buckets.back().entries.push_back({section.bktr.relocation.offset, {0}, ctr_low}); subsection_buckets.back().entries.push_back({size, {0}, 0}); - std::optional<Core::Crypto::Key128> key = {}; + std::optional<Core::Crypto::Key128> key; if (encrypted) { if (has_rights_id) { status = Loader::ResultStatus::Success; @@ -442,18 +442,18 @@ std::optional<Core::Crypto::Key128> NCA::GetTitlekey() { memcpy(rights_id.data(), header.rights_id.data(), 16); if (rights_id == u128{}) { status = Loader::ResultStatus::ErrorInvalidRightsID; - return {}; + return std::nullopt; } auto titlekey = keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]); if (titlekey == Core::Crypto::Key128{}) { status = Loader::ResultStatus::ErrorMissingTitlekey; - return {}; + return std::nullopt; } if (!keys.HasKey(Core::Crypto::S128KeyType::Titlekek, master_key_id)) { status = Loader::ResultStatus::ErrorMissingTitlekek; - return {}; + return std::nullopt; } Core::Crypto::AESCipher<Core::Crypto::Key128> cipher( @@ -477,7 +477,7 @@ VirtualFile NCA::Decrypt(const NCASectionHeader& s_header, VirtualFile in, u64 s case NCASectionCryptoType::BKTR: LOG_TRACE(Crypto, "called with mode=CTR, starting_offset={:016X}", starting_offset); { - std::optional<Core::Crypto::Key128> key = {}; + std::optional<Core::Crypto::Key128> key; if (has_rights_id) { status = Loader::ResultStatus::Success; key = GetTitlekey(); @@ -496,9 +496,10 @@ VirtualFile NCA::Decrypt(const NCASectionHeader& s_header, VirtualFile in, u64 s auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>(std::move(in), *key, starting_offset); - std::vector<u8> iv(16); - for (u8 i = 0; i < 8; ++i) - iv[i] = s_header.raw.section_ctr[0x8 - i - 1]; + Core::Crypto::CTREncryptionLayer::IVData iv{}; + for (std::size_t i = 0; i < 8; ++i) { + iv[i] = s_header.raw.section_ctr[8 - i - 1]; + } out->SetIV(iv); return std::static_pointer_cast<VfsFile>(out); } diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h index e249079b5..69292232a 100644 --- a/src/core/file_sys/content_archive.h +++ b/src/core/file_sys/content_archive.h @@ -99,8 +99,7 @@ inline bool IsDirectoryLogoPartition(const VirtualDir& pfs) { class NCA : public ReadOnlyVfsDirectory { public: explicit NCA(VirtualFile file, VirtualFile bktr_base_romfs = nullptr, - u64 bktr_base_ivfc_offset = 0, - Core::Crypto::KeyManager keys = Core::Crypto::KeyManager()); + u64 bktr_base_ivfc_offset = 0); ~NCA() override; Loader::ResultStatus GetStatus() const; @@ -159,7 +158,7 @@ private: bool encrypted = false; bool is_update = false; - Core::Crypto::KeyManager keys; + Core::Crypto::KeyManager& keys; }; } // namespace FileSys diff --git a/src/core/file_sys/control_metadata.cpp b/src/core/file_sys/control_metadata.cpp index f155a1341..b0a130345 100644 --- a/src/core/file_sys/control_metadata.cpp +++ b/src/core/file_sys/control_metadata.cpp @@ -5,6 +5,7 @@ #include "common/string_util.h" #include "common/swap.h" #include "core/file_sys/control_metadata.h" +#include "core/file_sys/vfs.h" namespace FileSys { @@ -95,6 +96,10 @@ u32 NACP::GetSupportedLanguages() const { return raw.supported_languages; } +u64 NACP::GetDeviceSaveDataSize() const { + return raw.device_save_data_size; +} + std::vector<u8> NACP::GetRawBytes() const { std::vector<u8> out(sizeof(RawNACP)); std::memcpy(out.data(), &raw, sizeof(RawNACP)); diff --git a/src/core/file_sys/control_metadata.h b/src/core/file_sys/control_metadata.h index 2d8c251ac..403c4219a 100644 --- a/src/core/file_sys/control_metadata.h +++ b/src/core/file_sys/control_metadata.h @@ -10,7 +10,7 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" -#include "core/file_sys/vfs.h" +#include "core/file_sys/vfs_types.h" namespace FileSys { @@ -83,7 +83,7 @@ enum class Language : u8 { Italian = 7, Dutch = 8, CanadianFrench = 9, - Portugese = 10, + Portuguese = 10, Russian = 11, Korean = 12, Taiwanese = 13, @@ -113,6 +113,7 @@ public: u32 GetSupportedLanguages() const; std::vector<u8> GetRawBytes() const; bool GetUserAccountSwitchLock() const; + u64 GetDeviceSaveDataSize() const; private: RawNACP raw{}; diff --git a/src/core/file_sys/fsmitm_romfsbuild.cpp b/src/core/file_sys/fsmitm_romfsbuild.cpp index d126ae8dd..c52fafb6f 100644 --- a/src/core/file_sys/fsmitm_romfsbuild.cpp +++ b/src/core/file_sys/fsmitm_romfsbuild.cpp @@ -240,7 +240,7 @@ RomFSBuildContext::RomFSBuildContext(VirtualDir base_, VirtualDir ext_) RomFSBuildContext::~RomFSBuildContext() = default; -std::map<u64, VirtualFile> RomFSBuildContext::Build() { +std::multimap<u64, VirtualFile> RomFSBuildContext::Build() { const u64 dir_hash_table_entry_count = romfs_get_hash_table_count(num_dirs); const u64 file_hash_table_entry_count = romfs_get_hash_table_count(num_files); dir_hash_table_size = 4 * dir_hash_table_entry_count; @@ -266,8 +266,9 @@ std::map<u64, VirtualFile> RomFSBuildContext::Build() { cur_file->offset = file_partition_size; file_partition_size += cur_file->size; cur_file->entry_offset = entry_offset; - entry_offset += sizeof(RomFSFileEntry) + - Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4); + entry_offset += + static_cast<u32>(sizeof(RomFSFileEntry) + + Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4)); prev_file = cur_file; } // Assign deferred parent/sibling ownership. @@ -284,8 +285,9 @@ std::map<u64, VirtualFile> RomFSBuildContext::Build() { for (const auto& it : directories) { cur_dir = it.second; cur_dir->entry_offset = entry_offset; - entry_offset += sizeof(RomFSDirectoryEntry) + - Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4); + entry_offset += + static_cast<u32>(sizeof(RomFSDirectoryEntry) + + Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4)); } // Assign deferred parent/sibling ownership. for (auto it = directories.rbegin(); it->second != root; ++it) { @@ -294,7 +296,7 @@ std::map<u64, VirtualFile> RomFSBuildContext::Build() { cur_dir->parent->child = cur_dir; } - std::map<u64, VirtualFile> out; + std::multimap<u64, VirtualFile> out; // Populate file tables. for (const auto& it : files) { diff --git a/src/core/file_sys/fsmitm_romfsbuild.h b/src/core/file_sys/fsmitm_romfsbuild.h index a62502193..049de180b 100644 --- a/src/core/file_sys/fsmitm_romfsbuild.h +++ b/src/core/file_sys/fsmitm_romfsbuild.h @@ -43,7 +43,7 @@ public: ~RomFSBuildContext(); // This finalizes the context. - std::map<u64, VirtualFile> Build(); + std::multimap<u64, VirtualFile> Build(); private: VirtualDir base; diff --git a/src/core/file_sys/ips_layer.cpp b/src/core/file_sys/ips_layer.cpp index a08a70efd..a6101f1c0 100644 --- a/src/core/file_sys/ips_layer.cpp +++ b/src/core/file_sys/ips_layer.cpp @@ -245,9 +245,11 @@ void IPSwitchCompiler::Parse() { // Read rest of patch while (true) { - if (i + 1 >= lines.size()) + if (i + 1 >= lines.size()) { break; - const auto patch_line = lines[++i]; + } + + const auto& patch_line = lines[++i]; // Start of new patch if (StartsWith(patch_line, "@enabled") || StartsWith(patch_line, "@disabled")) { @@ -297,7 +299,7 @@ void IPSwitchCompiler::Parse() { patch_text->GetName(), offset, Common::HexToString(replace)); } - patch.records.insert_or_assign(offset, std::move(replace)); + patch.records.insert_or_assign(static_cast<u32>(offset), std::move(replace)); } patches.push_back(std::move(patch)); diff --git a/src/core/file_sys/kernel_executable.cpp b/src/core/file_sys/kernel_executable.cpp index 76313679d..ef93ef3ed 100644 --- a/src/core/file_sys/kernel_executable.cpp +++ b/src/core/file_sys/kernel_executable.cpp @@ -2,9 +2,12 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <cstring> + #include "common/string_util.h" #include "core/file_sys/kernel_executable.h" #include "core/file_sys/vfs_offset.h" +#include "core/loader/loader.h" namespace FileSys { diff --git a/src/core/file_sys/kernel_executable.h b/src/core/file_sys/kernel_executable.h index 324a57384..044c554d3 100644 --- a/src/core/file_sys/kernel_executable.h +++ b/src/core/file_sys/kernel_executable.h @@ -4,10 +4,17 @@ #pragma once +#include <array> +#include <vector> + #include "common/common_funcs.h" +#include "common/common_types.h" #include "common/swap.h" #include "core/file_sys/vfs_types.h" -#include "core/loader/loader.h" + +namespace Loader { +enum class ResultStatus : u16; +} namespace FileSys { diff --git a/src/core/file_sys/mode.h b/src/core/file_sys/mode.h index c95205668..2b4f21073 100644 --- a/src/core/file_sys/mode.h +++ b/src/core/file_sys/mode.h @@ -4,6 +4,7 @@ #pragma once +#include "common/common_funcs.h" #include "common/common_types.h" namespace FileSys { @@ -11,13 +12,11 @@ namespace FileSys { enum class Mode : u32 { Read = 1, Write = 2, - ReadWrite = 3, + ReadWrite = Read | Write, Append = 4, - WriteAppend = 6, + WriteAppend = Write | Append, }; -inline u32 operator&(Mode lhs, Mode rhs) { - return static_cast<u32>(lhs) & static_cast<u32>(rhs); -} +DECLARE_ENUM_FLAG_OPERATORS(Mode) } // namespace FileSys diff --git a/src/core/file_sys/nca_metadata.cpp b/src/core/file_sys/nca_metadata.cpp index 93d0df6b9..3596541b2 100644 --- a/src/core/file_sys/nca_metadata.cpp +++ b/src/core/file_sys/nca_metadata.cpp @@ -7,6 +7,7 @@ #include "common/logging/log.h" #include "common/swap.h" #include "core/file_sys/nca_metadata.h" +#include "core/file_sys/vfs.h" namespace FileSys { @@ -107,7 +108,7 @@ std::vector<u8> CNMT::Serialize() const { memcpy(out.data() + sizeof(CNMTHeader), &opt_header, sizeof(OptionalHeader)); } - auto offset = header.table_offset; + u64_le offset = header.table_offset; for (const auto& rec : content_records) { memcpy(out.data() + offset + sizeof(CNMTHeader), &rec, sizeof(ContentRecord)); diff --git a/src/core/file_sys/nca_metadata.h b/src/core/file_sys/nca_metadata.h index 1f82fff0a..53535e5f5 100644 --- a/src/core/file_sys/nca_metadata.h +++ b/src/core/file_sys/nca_metadata.h @@ -10,7 +10,7 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" -#include "core/file_sys/vfs.h" +#include "core/file_sys/vfs_types.h" namespace FileSys { class CNMT; diff --git a/src/core/file_sys/nca_patch.cpp b/src/core/file_sys/nca_patch.cpp index 0090cc6c4..5990a2fd5 100644 --- a/src/core/file_sys/nca_patch.cpp +++ b/src/core/file_sys/nca_patch.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <algorithm> +#include <array> #include <cstddef> #include <cstring> @@ -11,6 +12,49 @@ #include "core/file_sys/nca_patch.h" namespace FileSys { +namespace { +template <bool Subsection, typename BlockType, typename BucketType> +std::pair<std::size_t, std::size_t> SearchBucketEntry(u64 offset, const BlockType& block, + const BucketType& buckets) { + if constexpr (Subsection) { + const auto& last_bucket = buckets[block.number_buckets - 1]; + if (offset >= last_bucket.entries[last_bucket.number_entries].address_patch) { + return {block.number_buckets - 1, last_bucket.number_entries}; + } + } else { + ASSERT_MSG(offset <= block.size, "Offset is out of bounds in BKTR relocation block."); + } + + std::size_t bucket_id = std::count_if( + block.base_offsets.begin() + 1, block.base_offsets.begin() + block.number_buckets, + [&offset](u64 base_offset) { return base_offset <= offset; }); + + const auto& bucket = buckets[bucket_id]; + + if (bucket.number_entries == 1) { + return {bucket_id, 0}; + } + + std::size_t low = 0; + std::size_t mid = 0; + std::size_t high = bucket.number_entries - 1; + while (low <= high) { + mid = (low + high) / 2; + if (bucket.entries[mid].address_patch > offset) { + high = mid - 1; + } else { + if (mid == bucket.number_entries - 1 || + bucket.entries[mid + 1].address_patch > offset) { + return {bucket_id, mid}; + } + + low = mid + 1; + } + } + + UNREACHABLE_MSG("Offset could not be found in BKTR block."); +} +} // Anonymous namespace BKTR::BKTR(VirtualFile base_romfs_, VirtualFile bktr_romfs_, RelocationBlock relocation_, std::vector<RelocationBucket> relocation_buckets_, SubsectionBlock subsection_, @@ -66,7 +110,7 @@ std::size_t BKTR::Read(u8* data, std::size_t length, std::size_t offset) const { Core::Crypto::AESCipher<Core::Crypto::Key128> cipher(key, Core::Crypto::Mode::CTR); // Calculate AES IV - std::vector<u8> iv(16); + std::array<u8, 16> iv{}; auto subsection_ctr = subsection.ctr; auto offset_iv = section_offset + base_offset; for (std::size_t i = 0; i < section_ctr.size(); ++i) @@ -109,46 +153,6 @@ std::size_t BKTR::Read(u8* data, std::size_t length, std::size_t offset) const { return raw_read; } -template <bool Subsection, typename BlockType, typename BucketType> -std::pair<std::size_t, std::size_t> BKTR::SearchBucketEntry(u64 offset, BlockType block, - BucketType buckets) const { - if constexpr (Subsection) { - const auto last_bucket = buckets[block.number_buckets - 1]; - if (offset >= last_bucket.entries[last_bucket.number_entries].address_patch) - return {block.number_buckets - 1, last_bucket.number_entries}; - } else { - ASSERT_MSG(offset <= block.size, "Offset is out of bounds in BKTR relocation block."); - } - - std::size_t bucket_id = std::count_if( - block.base_offsets.begin() + 1, block.base_offsets.begin() + block.number_buckets, - [&offset](u64 base_offset) { return base_offset <= offset; }); - - const auto bucket = buckets[bucket_id]; - - if (bucket.number_entries == 1) - return {bucket_id, 0}; - - std::size_t low = 0; - std::size_t mid = 0; - std::size_t high = bucket.number_entries - 1; - while (low <= high) { - mid = (low + high) / 2; - if (bucket.entries[mid].address_patch > offset) { - high = mid - 1; - } else { - if (mid == bucket.number_entries - 1 || - bucket.entries[mid + 1].address_patch > offset) { - return {bucket_id, mid}; - } - - low = mid + 1; - } - } - - UNREACHABLE_MSG("Offset could not be found in BKTR block."); -} - RelocationEntry BKTR::GetRelocationEntry(u64 offset) const { const auto res = SearchBucketEntry<false>(offset, relocation, relocation_buckets); return relocation_buckets[res.first].entries[res.second]; diff --git a/src/core/file_sys/nca_patch.h b/src/core/file_sys/nca_patch.h index 8e64e8378..60c544f8e 100644 --- a/src/core/file_sys/nca_patch.h +++ b/src/core/file_sys/nca_patch.h @@ -117,10 +117,6 @@ public: bool Rename(std::string_view name) override; private: - template <bool Subsection, typename BlockType, typename BucketType> - std::pair<std::size_t, std::size_t> SearchBucketEntry(u64 offset, BlockType block, - BucketType buckets) const; - RelocationEntry GetRelocationEntry(u64 offset) const; RelocationEntry GetNextRelocationEntry(u64 offset) const; diff --git a/src/core/file_sys/partition_filesystem.cpp b/src/core/file_sys/partition_filesystem.cpp index 846986736..48a2ed4d4 100644 --- a/src/core/file_sys/partition_filesystem.cpp +++ b/src/core/file_sys/partition_filesystem.cpp @@ -21,7 +21,7 @@ bool PartitionFilesystem::Header::HasValidMagicValue() const { magic == Common::MakeMagic('P', 'F', 'S', '0'); } -PartitionFilesystem::PartitionFilesystem(std::shared_ptr<VfsFile> file) { +PartitionFilesystem::PartitionFilesystem(VirtualFile file) { // At least be as large as the header if (file->GetSize() < sizeof(Header)) { status = Loader::ResultStatus::ErrorBadPFSHeader; @@ -89,11 +89,11 @@ std::map<std::string, u64> PartitionFilesystem::GetFileSizes() const { return sizes; } -std::vector<std::shared_ptr<VfsFile>> PartitionFilesystem::GetFiles() const { +std::vector<VirtualFile> PartitionFilesystem::GetFiles() const { return pfs_files; } -std::vector<std::shared_ptr<VfsDirectory>> PartitionFilesystem::GetSubdirectories() const { +std::vector<VirtualDir> PartitionFilesystem::GetSubdirectories() const { return {}; } @@ -101,7 +101,7 @@ std::string PartitionFilesystem::GetName() const { return is_hfs ? "HFS0" : "PFS0"; } -std::shared_ptr<VfsDirectory> PartitionFilesystem::GetParentDirectory() const { +VirtualDir PartitionFilesystem::GetParentDirectory() const { // TODO(DarkLordZach): Add support for nested containers. return nullptr; } diff --git a/src/core/file_sys/partition_filesystem.h b/src/core/file_sys/partition_filesystem.h index 279193b19..0f831148e 100644 --- a/src/core/file_sys/partition_filesystem.h +++ b/src/core/file_sys/partition_filesystem.h @@ -24,7 +24,7 @@ namespace FileSys { */ class PartitionFilesystem : public ReadOnlyVfsDirectory { public: - explicit PartitionFilesystem(std::shared_ptr<VfsFile> file); + explicit PartitionFilesystem(VirtualFile file); ~PartitionFilesystem() override; Loader::ResultStatus GetStatus() const; @@ -32,10 +32,10 @@ public: std::map<std::string, u64> GetFileOffsets() const; std::map<std::string, u64> GetFileSizes() const; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualFile> GetFiles() const override; + std::vector<VirtualDir> GetSubdirectories() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; + VirtualDir GetParentDirectory() const override; void PrintDebugInfo() const; private: diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp index e77e82b8d..e9d1607d0 100644 --- a/src/core/file_sys/patch_manager.cpp +++ b/src/core/file_sys/patch_manager.cpp @@ -10,6 +10,7 @@ #include "common/file_util.h" #include "common/hex_util.h" #include "common/logging/log.h" +#include "common/string_util.h" #include "core/core.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" @@ -26,8 +27,9 @@ #include "core/settings.h" namespace FileSys { +namespace { -constexpr u64 SINGLE_BYTE_MODULUS = 0x100; +constexpr u32 SINGLE_BYTE_MODULUS = 0x100; constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; constexpr std::array<const char*, 14> EXEFS_FILE_NAMES{ @@ -35,20 +37,85 @@ constexpr std::array<const char*, 14> EXEFS_FILE_NAMES{ "subsdk3", "subsdk4", "subsdk5", "subsdk6", "subsdk7", "subsdk8", "subsdk9", }; -std::string FormatTitleVersion(u32 version, TitleVersionFormat format) { +enum class TitleVersionFormat : u8 { + ThreeElements, ///< vX.Y.Z + FourElements, ///< vX.Y.Z.W +}; + +std::string FormatTitleVersion(u32 version, + TitleVersionFormat format = TitleVersionFormat::ThreeElements) { std::array<u8, sizeof(u32)> bytes{}; - bytes[0] = version % SINGLE_BYTE_MODULUS; + bytes[0] = static_cast<u8>(version % SINGLE_BYTE_MODULUS); for (std::size_t i = 1; i < bytes.size(); ++i) { version /= SINGLE_BYTE_MODULUS; - bytes[i] = version % SINGLE_BYTE_MODULUS; + bytes[i] = static_cast<u8>(version % SINGLE_BYTE_MODULUS); } - if (format == TitleVersionFormat::FourElements) + if (format == TitleVersionFormat::FourElements) { return fmt::format("v{}.{}.{}.{}", bytes[3], bytes[2], bytes[1], bytes[0]); + } return fmt::format("v{}.{}.{}", bytes[3], bytes[2], bytes[1]); } -PatchManager::PatchManager(u64 title_id) : title_id(title_id) {} +// Returns a directory with name matching name case-insensitive. Returns nullptr if directory +// doesn't have a directory with name. +VirtualDir FindSubdirectoryCaseless(const VirtualDir dir, std::string_view name) { +#ifdef _WIN32 + return dir->GetSubdirectory(name); +#else + const auto subdirs = dir->GetSubdirectories(); + for (const auto& subdir : subdirs) { + std::string dir_name = Common::ToLower(subdir->GetName()); + if (dir_name == name) { + return subdir; + } + } + + return nullptr; +#endif +} + +std::optional<std::vector<Core::Memory::CheatEntry>> ReadCheatFileFromFolder( + u64 title_id, const PatchManager::BuildID& build_id_, const VirtualDir& base_path, bool upper) { + const auto build_id_raw = Common::HexToString(build_id_, upper); + const auto build_id = build_id_raw.substr(0, sizeof(u64) * 2); + const auto file = base_path->GetFile(fmt::format("{}.txt", build_id)); + + if (file == nullptr) { + LOG_INFO(Common_Filesystem, "No cheats file found for title_id={:016X}, build_id={}", + title_id, build_id); + return std::nullopt; + } + + std::vector<u8> data(file->GetSize()); + if (file->Read(data.data(), data.size()) != data.size()) { + LOG_INFO(Common_Filesystem, "Failed to read cheats file for title_id={:016X}, build_id={}", + title_id, build_id); + return std::nullopt; + } + + const Core::Memory::TextCheatParser parser; + return parser.Parse(std::string_view(reinterpret_cast<const char*>(data.data()), data.size())); +} + +void AppendCommaIfNotEmpty(std::string& to, std::string_view with) { + if (to.empty()) { + to += with; + } else { + to += ", "; + to += with; + } +} + +bool IsDirValidAndNonEmpty(const VirtualDir& dir) { + return dir != nullptr && (!dir->GetFiles().empty() || !dir->GetSubdirectories().empty()); +} +} // Anonymous namespace + +PatchManager::PatchManager(u64 title_id_, + const Service::FileSystem::FileSystemController& fs_controller_, + const ContentProvider& content_provider_) + : title_id{title_id_}, fs_controller{fs_controller_}, content_provider{content_provider_} {} PatchManager::~PatchManager() = default; @@ -64,34 +131,30 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const { if (Settings::values.dump_exefs) { LOG_INFO(Loader, "Dumping ExeFS for title_id={:016X}", title_id); - const auto dump_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationDumpRoot(title_id); + const auto dump_dir = fs_controller.GetModificationDumpRoot(title_id); if (dump_dir != nullptr) { const auto exefs_dir = GetOrCreateDirectoryRelative(dump_dir, "/exefs"); VfsRawCopyD(exefs, exefs_dir); } } - const auto& installed = Core::System::GetInstance().GetContentProvider(); - const auto& disabled = Settings::values.disabled_addons[title_id]; const auto update_disabled = std::find(disabled.cbegin(), disabled.cend(), "Update") != disabled.cend(); // Game Updates const auto update_tid = GetUpdateTitleID(title_id); - const auto update = installed.GetEntry(update_tid, ContentRecordType::Program); + const auto update = content_provider.GetEntry(update_tid, ContentRecordType::Program); if (!update_disabled && update != nullptr && update->GetExeFS() != nullptr && update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { LOG_INFO(Loader, " ExeFS: Update ({}) applied successfully", - FormatTitleVersion(installed.GetEntryVersion(update_tid).value_or(0))); + FormatTitleVersion(content_provider.GetEntryVersion(update_tid).value_or(0))); exefs = update->GetExeFS(); } // LayeredExeFS - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir != nullptr && load_dir->GetSize() > 0) { auto patch_dirs = load_dir->GetSubdirectories(); std::sort( @@ -104,7 +167,7 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const { if (std::find(disabled.begin(), disabled.end(), subdir->GetName()) != disabled.end()) continue; - auto exefs_dir = subdir->GetSubdirectory("exefs"); + auto exefs_dir = FindSubdirectoryCaseless(subdir, "exefs"); if (exefs_dir != nullptr) layers.push_back(std::move(exefs_dir)); } @@ -130,7 +193,7 @@ std::vector<VirtualFile> PatchManager::CollectPatches(const std::vector<VirtualD if (std::find(disabled.cbegin(), disabled.cend(), subdir->GetName()) != disabled.cend()) continue; - auto exefs_dir = subdir->GetSubdirectory("exefs"); + auto exefs_dir = FindSubdirectoryCaseless(subdir, "exefs"); if (exefs_dir != nullptr) { for (const auto& file : exefs_dir->GetFiles()) { if (file->GetExtension() == "ips") { @@ -177,8 +240,7 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st if (Settings::values.dump_nso) { LOG_INFO(Loader, "Dumping NSO for name={}, build_id={}, title_id={:016X}", name, build_id, title_id); - const auto dump_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationDumpRoot(title_id); + const auto dump_dir = fs_controller.GetModificationDumpRoot(title_id); if (dump_dir != nullptr) { const auto nso_dir = GetOrCreateDirectoryRelative(dump_dir, "/nso"); const auto file = nso_dir->CreateFile(fmt::format("{}-{}.nso", name, build_id)); @@ -190,8 +252,7 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st LOG_INFO(Loader, "Patching NSO for name={}, build_id={}", name, build_id); - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir == nullptr) { LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id); return nso; @@ -228,14 +289,13 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st return out; } -bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const { +bool PatchManager::HasNSOPatch(const BuildID& build_id_) const { const auto build_id_raw = Common::HexToString(build_id_); const auto build_id = build_id_raw.substr(0, build_id_raw.find_last_not_of('0') + 1); LOG_INFO(Loader, "Querying NSO patch existence for build_id={}", build_id); - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir == nullptr) { LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id); return false; @@ -248,37 +308,9 @@ bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const { return !CollectPatches(patch_dirs, build_id).empty(); } -namespace { -std::optional<std::vector<Memory::CheatEntry>> ReadCheatFileFromFolder( - const Core::System& system, u64 title_id, const std::array<u8, 0x20>& build_id_, - const VirtualDir& base_path, bool upper) { - const auto build_id_raw = Common::HexToString(build_id_, upper); - const auto build_id = build_id_raw.substr(0, sizeof(u64) * 2); - const auto file = base_path->GetFile(fmt::format("{}.txt", build_id)); - - if (file == nullptr) { - LOG_INFO(Common_Filesystem, "No cheats file found for title_id={:016X}, build_id={}", - title_id, build_id); - return std::nullopt; - } - - std::vector<u8> data(file->GetSize()); - if (file->Read(data.data(), data.size()) != data.size()) { - LOG_INFO(Common_Filesystem, "Failed to read cheats file for title_id={:016X}, build_id={}", - title_id, build_id); - return std::nullopt; - } - - Memory::TextCheatParser parser; - return parser.Parse( - system, std::string_view(reinterpret_cast<const char* const>(data.data()), data.size())); -} - -} // Anonymous namespace - -std::vector<Memory::CheatEntry> PatchManager::CreateCheatList( - const Core::System& system, const std::array<u8, 32>& build_id_) const { - const auto load_dir = system.GetFileSystemController().GetModificationLoadRoot(title_id); +std::vector<Core::Memory::CheatEntry> PatchManager::CreateCheatList( + const BuildID& build_id_) const { + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir == nullptr) { LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id); return {}; @@ -289,22 +321,20 @@ std::vector<Memory::CheatEntry> PatchManager::CreateCheatList( std::sort(patch_dirs.begin(), patch_dirs.end(), [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); }); - std::vector<Memory::CheatEntry> out; + std::vector<Core::Memory::CheatEntry> out; for (const auto& subdir : patch_dirs) { if (std::find(disabled.cbegin(), disabled.cend(), subdir->GetName()) != disabled.cend()) { continue; } - auto cheats_dir = subdir->GetSubdirectory("cheats"); + auto cheats_dir = FindSubdirectoryCaseless(subdir, "cheats"); if (cheats_dir != nullptr) { - auto res = ReadCheatFileFromFolder(system, title_id, build_id_, cheats_dir, true); - if (res.has_value()) { + if (const auto res = ReadCheatFileFromFolder(title_id, build_id_, cheats_dir, true)) { std::copy(res->begin(), res->end(), std::back_inserter(out)); continue; } - res = ReadCheatFileFromFolder(system, title_id, build_id_, cheats_dir, false); - if (res.has_value()) { + if (const auto res = ReadCheatFileFromFolder(title_id, build_id_, cheats_dir, false)) { std::copy(res->begin(), res->end(), std::back_inserter(out)); } } @@ -313,9 +343,9 @@ std::vector<Memory::CheatEntry> PatchManager::CreateCheatList( return out; } -static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type) { - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); +static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type, + const Service::FileSystem::FileSystemController& fs_controller) { + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if ((type != ContentRecordType::Program && type != ContentRecordType::Data) || load_dir == nullptr || load_dir->GetSize() <= 0) { return; @@ -340,11 +370,11 @@ static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType t continue; } - auto romfs_dir = subdir->GetSubdirectory("romfs"); + auto romfs_dir = FindSubdirectoryCaseless(subdir, "romfs"); if (romfs_dir != nullptr) layers.push_back(std::move(romfs_dir)); - auto ext_dir = subdir->GetSubdirectory("romfs_ext"); + auto ext_dir = FindSubdirectoryCaseless(subdir, "romfs_ext"); if (ext_dir != nullptr) layers_ext.push_back(std::move(ext_dir)); } @@ -377,19 +407,19 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content const auto log_string = fmt::format("Patching RomFS for title_id={:016X}, type={:02X}", title_id, static_cast<u8>(type)); - if (type == ContentRecordType::Program || type == ContentRecordType::Data) + if (type == ContentRecordType::Program || type == ContentRecordType::Data) { LOG_INFO(Loader, "{}", log_string); - else + } else { LOG_DEBUG(Loader, "{}", log_string); + } - if (romfs == nullptr) + if (romfs == nullptr) { return romfs; - - const auto& installed = Core::System::GetInstance().GetContentProvider(); + } // Game Updates const auto update_tid = GetUpdateTitleID(title_id); - const auto update = installed.GetEntryRaw(update_tid, type); + const auto update = content_provider.GetEntryRaw(update_tid, type); const auto& disabled = Settings::values.disabled_addons[title_id]; const auto update_disabled = @@ -400,7 +430,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content if (new_nca->GetStatus() == Loader::ResultStatus::Success && new_nca->GetRomFS() != nullptr) { LOG_INFO(Loader, " RomFS: Update ({}) applied successfully", - FormatTitleVersion(installed.GetEntryVersion(update_tid).value_or(0))); + FormatTitleVersion(content_provider.GetEntryVersion(update_tid).value_or(0))); romfs = new_nca->GetRomFS(); } } else if (!update_disabled && update_raw != nullptr) { @@ -413,34 +443,24 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content } // LayeredFS - ApplyLayeredFS(romfs, title_id, type); + ApplyLayeredFS(romfs, title_id, type, fs_controller); return romfs; } -static void AppendCommaIfNotEmpty(std::string& to, const std::string& with) { - if (to.empty()) - to += with; - else - to += ", " + with; -} - -static bool IsDirValidAndNonEmpty(const VirtualDir& dir) { - return dir != nullptr && (!dir->GetFiles().empty() || !dir->GetSubdirectories().empty()); -} - -std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNames( - VirtualFile update_raw) const { - if (title_id == 0) +PatchManager::PatchVersionNames PatchManager::GetPatchVersionNames(VirtualFile update_raw) const { + if (title_id == 0) { return {}; + } + std::map<std::string, std::string, std::less<>> out; - const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto& disabled = Settings::values.disabled_addons[title_id]; // Game Updates const auto update_tid = GetUpdateTitleID(title_id); - PatchManager update{update_tid}; - auto [nacp, discard_icon_file] = update.GetControlMetadata(); + PatchManager update{update_tid, fs_controller, content_provider}; + const auto metadata = update.GetControlMetadata(); + const auto& nacp = metadata.first; const auto update_disabled = std::find(disabled.cbegin(), disabled.cend(), "Update") != disabled.cend(); @@ -449,13 +469,12 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam if (nacp != nullptr) { out.insert_or_assign(update_label, nacp->GetVersionString()); } else { - if (installed.HasEntry(update_tid, ContentRecordType::Program)) { - const auto meta_ver = installed.GetEntryVersion(update_tid); + if (content_provider.HasEntry(update_tid, ContentRecordType::Program)) { + const auto meta_ver = content_provider.GetEntryVersion(update_tid); if (meta_ver.value_or(0) == 0) { out.insert_or_assign(update_label, ""); } else { - out.insert_or_assign( - update_label, FormatTitleVersion(*meta_ver, TitleVersionFormat::ThreeElements)); + out.insert_or_assign(update_label, FormatTitleVersion(*meta_ver)); } } else if (update_raw != nullptr) { out.insert_or_assign(update_label, "PACKED"); @@ -463,13 +482,12 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam } // General Mods (LayeredFS and IPS) - const auto mod_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto mod_dir = fs_controller.GetModificationLoadRoot(title_id); if (mod_dir != nullptr && mod_dir->GetSize() > 0) { for (const auto& mod : mod_dir->GetSubdirectories()) { std::string types; - const auto exefs_dir = mod->GetSubdirectory("exefs"); + const auto exefs_dir = FindSubdirectoryCaseless(mod, "exefs"); if (IsDirValidAndNonEmpty(exefs_dir)) { bool ips = false; bool ipswitch = false; @@ -493,9 +511,9 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam if (layeredfs) AppendCommaIfNotEmpty(types, "LayeredExeFS"); } - if (IsDirValidAndNonEmpty(mod->GetSubdirectory("romfs"))) + if (IsDirValidAndNonEmpty(FindSubdirectoryCaseless(mod, "romfs"))) AppendCommaIfNotEmpty(types, "LayeredFS"); - if (IsDirValidAndNonEmpty(mod->GetSubdirectory("cheats"))) + if (IsDirValidAndNonEmpty(FindSubdirectoryCaseless(mod, "cheats"))) AppendCommaIfNotEmpty(types, "Cheats"); if (types.empty()) @@ -508,13 +526,15 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam } // DLC - const auto dlc_entries = installed.ListEntriesFilter(TitleType::AOC, ContentRecordType::Data); + const auto dlc_entries = + content_provider.ListEntriesFilter(TitleType::AOC, ContentRecordType::Data); std::vector<ContentProviderEntry> dlc_match; dlc_match.reserve(dlc_entries.size()); std::copy_if(dlc_entries.begin(), dlc_entries.end(), std::back_inserter(dlc_match), - [this, &installed](const ContentProviderEntry& entry) { + [this](const ContentProviderEntry& entry) { return (entry.title_id & DLC_BASE_TITLE_ID_MASK) == title_id && - installed.GetEntry(entry)->GetStatus() == Loader::ResultStatus::Success; + content_provider.GetEntry(entry)->GetStatus() == + Loader::ResultStatus::Success; }); if (!dlc_match.empty()) { // Ensure sorted so DLC IDs show in order. @@ -535,49 +555,52 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam } std::optional<u32> PatchManager::GetGameVersion() const { - const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto update_tid = GetUpdateTitleID(title_id); - if (installed.HasEntry(update_tid, ContentRecordType::Program)) { - return installed.GetEntryVersion(update_tid); + if (content_provider.HasEntry(update_tid, ContentRecordType::Program)) { + return content_provider.GetEntryVersion(update_tid); } - return installed.GetEntryVersion(title_id); + return content_provider.GetEntryVersion(title_id); } -std::pair<std::unique_ptr<NACP>, VirtualFile> PatchManager::GetControlMetadata() const { - const auto& installed = Core::System::GetInstance().GetContentProvider(); - - const auto base_control_nca = installed.GetEntry(title_id, ContentRecordType::Control); - if (base_control_nca == nullptr) +PatchManager::Metadata PatchManager::GetControlMetadata() const { + const auto base_control_nca = content_provider.GetEntry(title_id, ContentRecordType::Control); + if (base_control_nca == nullptr) { return {}; + } return ParseControlNCA(*base_control_nca); } -std::pair<std::unique_ptr<NACP>, VirtualFile> PatchManager::ParseControlNCA(const NCA& nca) const { +PatchManager::Metadata PatchManager::ParseControlNCA(const NCA& nca) const { const auto base_romfs = nca.GetRomFS(); - if (base_romfs == nullptr) + if (base_romfs == nullptr) { return {}; + } const auto romfs = PatchRomFS(base_romfs, nca.GetBaseIVFCOffset(), ContentRecordType::Control); - if (romfs == nullptr) + if (romfs == nullptr) { return {}; + } const auto extracted = ExtractRomFS(romfs); - if (extracted == nullptr) + if (extracted == nullptr) { return {}; + } auto nacp_file = extracted->GetFile("control.nacp"); - if (nacp_file == nullptr) + if (nacp_file == nullptr) { nacp_file = extracted->GetFile("Control.nacp"); + } auto nacp = nacp_file == nullptr ? nullptr : std::make_unique<NACP>(nacp_file); VirtualFile icon_file; for (const auto& language : FileSys::LANGUAGE_NAMES) { - icon_file = extracted->GetFile("icon_" + std::string(language) + ".dat"); - if (icon_file != nullptr) + icon_file = extracted->GetFile(std::string("icon_").append(language).append(".dat")); + if (icon_file != nullptr) { break; + } } return {std::move(nacp), icon_file}; diff --git a/src/core/file_sys/patch_manager.h b/src/core/file_sys/patch_manager.h index e857e6e82..fb1853035 100644 --- a/src/core/file_sys/patch_manager.h +++ b/src/core/file_sys/patch_manager.h @@ -6,83 +6,89 @@ #include <map> #include <memory> +#include <optional> #include <string> #include "common/common_types.h" #include "core/file_sys/nca_metadata.h" -#include "core/file_sys/vfs.h" +#include "core/file_sys/vfs_types.h" #include "core/memory/dmnt_cheat_types.h" namespace Core { class System; } +namespace Service::FileSystem { +class FileSystemController; +} + namespace FileSys { +class ContentProvider; class NCA; class NACP; -enum class TitleVersionFormat : u8 { - ThreeElements, ///< vX.Y.Z - FourElements, ///< vX.Y.Z.W -}; - -std::string FormatTitleVersion(u32 version, - TitleVersionFormat format = TitleVersionFormat::ThreeElements); - // A centralized class to manage patches to games. class PatchManager { public: - explicit PatchManager(u64 title_id); + using BuildID = std::array<u8, 0x20>; + using Metadata = std::pair<std::unique_ptr<NACP>, VirtualFile>; + using PatchVersionNames = std::map<std::string, std::string, std::less<>>; + + explicit PatchManager(u64 title_id_, + const Service::FileSystem::FileSystemController& fs_controller_, + const ContentProvider& content_provider_); ~PatchManager(); - u64 GetTitleID() const; + [[nodiscard]] u64 GetTitleID() const; // Currently tracked ExeFS patches: // - Game Updates - VirtualDir PatchExeFS(VirtualDir exefs) const; + [[nodiscard]] VirtualDir PatchExeFS(VirtualDir exefs) const; // Currently tracked NSO patches: // - IPS // - IPSwitch - std::vector<u8> PatchNSO(const std::vector<u8>& nso, const std::string& name) const; + [[nodiscard]] std::vector<u8> PatchNSO(const std::vector<u8>& nso, + const std::string& name) const; // Checks to see if PatchNSO() will have any effect given the NSO's build ID. // Used to prevent expensive copies in NSO loader. - bool HasNSOPatch(const std::array<u8, 0x20>& build_id) const; + [[nodiscard]] bool HasNSOPatch(const BuildID& build_id) const; // Creates a CheatList object with all - std::vector<Memory::CheatEntry> CreateCheatList(const Core::System& system, - const std::array<u8, 0x20>& build_id) const; + [[nodiscard]] std::vector<Core::Memory::CheatEntry> CreateCheatList( + const BuildID& build_id) const; // Currently tracked RomFS patches: // - Game Updates // - LayeredFS - VirtualFile PatchRomFS(VirtualFile base, u64 ivfc_offset, - ContentRecordType type = ContentRecordType::Program, - VirtualFile update_raw = nullptr) const; + [[nodiscard]] VirtualFile PatchRomFS(VirtualFile base, u64 ivfc_offset, + ContentRecordType type = ContentRecordType::Program, + VirtualFile update_raw = nullptr) const; // Returns a vector of pairs between patch names and patch versions. // i.e. Update 3.2.2 will return {"Update", "3.2.2"} - std::map<std::string, std::string, std::less<>> GetPatchVersionNames( - VirtualFile update_raw = nullptr) const; + [[nodiscard]] PatchVersionNames GetPatchVersionNames(VirtualFile update_raw = nullptr) const; // If the game update exists, returns the u32 version field in its Meta-type NCA. If that fails, // it will fallback to the Meta-type NCA of the base game. If that fails, the result will be // std::nullopt - std::optional<u32> GetGameVersion() const; + [[nodiscard]] std::optional<u32> GetGameVersion() const; // Given title_id of the program, attempts to get the control data of the update and parse // it, falling back to the base control data. - std::pair<std::unique_ptr<NACP>, VirtualFile> GetControlMetadata() const; + [[nodiscard]] Metadata GetControlMetadata() const; // Version of GetControlMetadata that takes an arbitrary NCA - std::pair<std::unique_ptr<NACP>, VirtualFile> ParseControlNCA(const NCA& nca) const; + [[nodiscard]] Metadata ParseControlNCA(const NCA& nca) const; private: - std::vector<VirtualFile> CollectPatches(const std::vector<VirtualDir>& patch_dirs, - const std::string& build_id) const; + [[nodiscard]] std::vector<VirtualFile> CollectPatches(const std::vector<VirtualDir>& patch_dirs, + const std::string& build_id) const; u64 title_id; + const Service::FileSystem::FileSystemController& fs_controller; + const ContentProvider& content_provider; }; } // namespace FileSys diff --git a/src/core/file_sys/program_metadata.cpp b/src/core/file_sys/program_metadata.cpp index 1d6c30962..9cf49bf44 100644 --- a/src/core/file_sys/program_metadata.cpp +++ b/src/core/file_sys/program_metadata.cpp @@ -7,6 +7,7 @@ #include "common/logging/log.h" #include "core/file_sys/program_metadata.h" +#include "core/file_sys/vfs.h" #include "core/loader/loader.h" namespace FileSys { @@ -51,6 +52,17 @@ Loader::ResultStatus ProgramMetadata::Load(VirtualFile file) { return Loader::ResultStatus::Success; } +/*static*/ ProgramMetadata ProgramMetadata::GetDefault() { + ProgramMetadata result; + + result.LoadManual( + true /*is_64_bit*/, FileSys::ProgramAddressSpaceType::Is39Bit /*address_space*/, + 0x2c /*main_thread_prio*/, 0 /*main_thread_core*/, 0x00100000 /*main_thread_stack_size*/, + {}, 0xFFFFFFFFFFFFFFFF /*filesystem_permissions*/, {} /*capabilities*/); + + return result; +} + void ProgramMetadata::LoadManual(bool is_64_bit, ProgramAddressSpaceType address_space, s32 main_thread_prio, u32 main_thread_core, u32 main_thread_stack_size, u64 title_id, diff --git a/src/core/file_sys/program_metadata.h b/src/core/file_sys/program_metadata.h index f8759a396..455532567 100644 --- a/src/core/file_sys/program_metadata.h +++ b/src/core/file_sys/program_metadata.h @@ -9,7 +9,7 @@ #include "common/bit_field.h" #include "common/common_types.h" #include "common/swap.h" -#include "core/file_sys/vfs.h" +#include "core/file_sys/vfs_types.h" namespace Loader { enum class ResultStatus : u16; @@ -44,9 +44,13 @@ public: ProgramMetadata(); ~ProgramMetadata(); + /// Gets a default ProgramMetadata configuration, should only be used for homebrew formats where + /// we do not have an NPDM file + static ProgramMetadata GetDefault(); + Loader::ResultStatus Load(VirtualFile file); - // Load from parameters instead of NPDM file, used for KIP + /// Load from parameters instead of NPDM file, used for KIP void LoadManual(bool is_64_bit, ProgramAddressSpaceType address_space, s32 main_thread_prio, u32 main_thread_core, u32 main_thread_stack_size, u64 title_id, u64 filesystem_permissions, KernelCapabilityDescriptors capabilities); diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp index 6e9cf67ef..da01002d5 100644 --- a/src/core/file_sys/registered_cache.cpp +++ b/src/core/file_sys/registered_cache.cpp @@ -257,8 +257,7 @@ std::vector<NcaID> PlaceholderCache::List() const { for (const auto& sdir : dir->GetSubdirectories()) { for (const auto& file : sdir->GetFiles()) { const auto name = file->GetName(); - if (name.length() == 36 && name[32] == '.' && name[33] == 'n' && name[34] == 'c' && - name[35] == 'a') { + if (name.length() == 36 && name.ends_with(".nca")) { out.push_back(Common::HexStringToArray<0x10>(name.substr(0, 32))); } } @@ -344,15 +343,18 @@ VirtualFile RegisteredCache::GetFileAtID(NcaID id) const { static std::optional<NcaID> CheckMapForContentRecord(const std::map<u64, CNMT>& map, u64 title_id, ContentRecordType type) { - if (map.find(title_id) == map.end()) - return {}; - - const auto& cnmt = map.at(title_id); + const auto cmnt_iter = map.find(title_id); + if (cmnt_iter == map.cend()) { + return std::nullopt; + } - const auto iter = std::find_if(cnmt.GetContentRecords().begin(), cnmt.GetContentRecords().end(), + const auto& cnmt = cmnt_iter->second; + const auto& content_records = cnmt.GetContentRecords(); + const auto iter = std::find_if(content_records.cbegin(), content_records.cend(), [type](const ContentRecord& rec) { return rec.type == type; }); - if (iter == cnmt.GetContentRecords().end()) - return {}; + if (iter == content_records.cend()) { + return std::nullopt; + } return std::make_optional(iter->nca_id); } @@ -408,7 +410,7 @@ void RegisteredCache::ProcessFiles(const std::vector<NcaID>& ids) { if (file == nullptr) continue; - const auto nca = std::make_shared<NCA>(parser(file, id), nullptr, 0, keys); + const auto nca = std::make_shared<NCA>(parser(file, id), nullptr, 0); if (nca->GetStatus() != Loader::ResultStatus::Success || nca->GetType() != NCAContentType::Meta) { continue; @@ -467,14 +469,16 @@ VirtualFile RegisteredCache::GetEntryUnparsed(u64 title_id, ContentRecordType ty std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const { const auto meta_iter = meta.find(title_id); - if (meta_iter != meta.end()) + if (meta_iter != meta.cend()) { return meta_iter->second.GetTitleVersion(); + } const auto yuzu_meta_iter = yuzu_meta.find(title_id); - if (yuzu_meta_iter != yuzu_meta.end()) + if (yuzu_meta_iter != yuzu_meta.cend()) { return yuzu_meta_iter->second.GetTitleVersion(); + } - return {}; + return std::nullopt; } VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) const { @@ -486,7 +490,7 @@ std::unique_ptr<NCA> RegisteredCache::GetEntry(u64 title_id, ContentRecordType t const auto raw = GetEntryRaw(title_id, type); if (raw == nullptr) return nullptr; - return std::make_unique<NCA>(raw, nullptr, 0, keys); + return std::make_unique<NCA>(raw, nullptr, 0); } template <typename T> @@ -560,57 +564,139 @@ InstallResult RegisteredCache::InstallEntry(const NSP& nsp, bool overwrite_if_ex return InstallResult::ErrorMetaFailed; } - // Install Metadata File const auto meta_id_raw = (*meta_iter)->GetName().substr(0, 32); const auto meta_id = Common::HexStringToArray<16>(meta_id_raw); - const auto res = RawInstallNCA(**meta_iter, copy, overwrite_if_exists, meta_id); - if (res != InstallResult::Success) - return res; + if ((*meta_iter)->GetSubdirectories().empty()) { + LOG_ERROR(Loader, + "The file you are attempting to install does not contain a section0 within the " + "metadata NCA and is therefore malformed. Verify that the file is valid."); + return InstallResult::ErrorMetaFailed; + } - // Install all the other NCAs const auto section0 = (*meta_iter)->GetSubdirectories()[0]; + + if (section0->GetFiles().empty()) { + LOG_ERROR(Loader, + "The file you are attempting to install does not contain a CNMT within the " + "metadata NCA and is therefore malformed. Verify that the file is valid."); + return InstallResult::ErrorMetaFailed; + } + const auto cnmt_file = section0->GetFiles()[0]; const CNMT cnmt(cnmt_file); + + const auto title_id = cnmt.GetTitleID(); + const auto result = RemoveExistingEntry(title_id); + + // Install Metadata File + const auto res = RawInstallNCA(**meta_iter, copy, overwrite_if_exists, meta_id); + if (res != InstallResult::Success) { + return res; + } + + // Install all the other NCAs for (const auto& record : cnmt.GetContentRecords()) { // Ignore DeltaFragments, they are not useful to us - if (record.type == ContentRecordType::DeltaFragment) + if (record.type == ContentRecordType::DeltaFragment) { continue; + } const auto nca = GetNCAFromNSPForID(nsp, record.nca_id); - if (nca == nullptr) + if (nca == nullptr) { return InstallResult::ErrorCopyFailed; + } const auto res2 = RawInstallNCA(*nca, copy, overwrite_if_exists, record.nca_id); - if (res2 != InstallResult::Success) + if (res2 != InstallResult::Success) { return res2; + } } Refresh(); + if (result) { + return InstallResult::OverwriteExisting; + } return InstallResult::Success; } InstallResult RegisteredCache::InstallEntry(const NCA& nca, TitleType type, bool overwrite_if_exists, const VfsCopyFunction& copy) { - CNMTHeader header{ - nca.GetTitleId(), ///< Title ID - 0, ///< Ignore/Default title version - type, ///< Type - {}, ///< Padding - 0x10, ///< Default table offset - 1, ///< 1 Content Entry - 0, ///< No Meta Entries - {}, ///< Padding + const CNMTHeader header{ + .title_id = nca.GetTitleId(), + .title_version = 0, + .type = type, + .reserved = {}, + .table_offset = 0x10, + .number_content_entries = 1, + .number_meta_entries = 0, + .attributes = 0, + .reserved2 = {}, + .is_committed = 0, + .required_download_system_version = 0, + .reserved3 = {}, }; - OptionalHeader opt_header{0, 0}; + const OptionalHeader opt_header{0, 0}; ContentRecord c_rec{{}, {}, {}, GetCRTypeFromNCAType(nca.GetType()), {}}; const auto& data = nca.GetBaseFile()->ReadBytes(0x100000); mbedtls_sha256_ret(data.data(), data.size(), c_rec.hash.data(), 0); - memcpy(&c_rec.nca_id, &c_rec.hash, 16); + std::memcpy(&c_rec.nca_id, &c_rec.hash, 16); const CNMT new_cnmt(header, opt_header, {c_rec}, {}); - if (!RawInstallYuzuMeta(new_cnmt)) + if (!RawInstallYuzuMeta(new_cnmt)) { return InstallResult::ErrorMetaFailed; + } return RawInstallNCA(nca, copy, overwrite_if_exists, c_rec.nca_id); } +bool RegisteredCache::RemoveExistingEntry(u64 title_id) const { + const auto delete_nca = [this](const NcaID& id) { + const auto path = GetRelativePathFromNcaID(id, false, true, false); + + const bool isFile = dir->GetFileRelative(path) != nullptr; + const bool isDir = dir->GetDirectoryRelative(path) != nullptr; + + if (isFile) { + return dir->DeleteFile(path); + } else if (isDir) { + return dir->DeleteSubdirectoryRecursive(path); + } + + return false; + }; + + // If an entry exists in the registered cache, remove it + if (HasEntry(title_id, ContentRecordType::Meta)) { + LOG_INFO(Loader, + "Previously installed entry (v{}) for title_id={:016X} detected! " + "Attempting to remove...", + GetEntryVersion(title_id).value_or(0), title_id); + + // Get all the ncas associated with the current CNMT and delete them + const auto meta_old_id = + GetNcaIDFromMetadata(title_id, ContentRecordType::Meta).value_or(NcaID{}); + const auto program_id = + GetNcaIDFromMetadata(title_id, ContentRecordType::Program).value_or(NcaID{}); + const auto data_id = + GetNcaIDFromMetadata(title_id, ContentRecordType::Data).value_or(NcaID{}); + const auto control_id = + GetNcaIDFromMetadata(title_id, ContentRecordType::Control).value_or(NcaID{}); + const auto html_id = + GetNcaIDFromMetadata(title_id, ContentRecordType::HtmlDocument).value_or(NcaID{}); + const auto legal_id = + GetNcaIDFromMetadata(title_id, ContentRecordType::LegalInformation).value_or(NcaID{}); + + const auto deleted_meta = delete_nca(meta_old_id); + const auto deleted_program = delete_nca(program_id); + const auto deleted_data = delete_nca(data_id); + const auto deleted_control = delete_nca(control_id); + const auto deleted_html = delete_nca(html_id); + const auto deleted_legal = delete_nca(legal_id); + + return deleted_meta && (deleted_meta || deleted_program || deleted_data || + deleted_control || deleted_html || deleted_legal); + } + + return false; +} + InstallResult RegisteredCache::RawInstallNCA(const NCA& nca, const VfsCopyFunction& copy, bool overwrite_if_exists, std::optional<NcaID> override_id) { @@ -641,12 +727,13 @@ InstallResult RegisteredCache::RawInstallNCA(const NCA& nca, const VfsCopyFuncti LOG_WARNING(Loader, "Overwriting existing NCA..."); VirtualDir c_dir; { c_dir = dir->GetFileRelative(path)->GetContainingDirectory(); } - c_dir->DeleteFile(FileUtil::GetFilename(path)); + c_dir->DeleteFile(Common::FS::GetFilename(path)); } auto out = dir->CreateFileRelative(path); - if (out == nullptr) + if (out == nullptr) { return InstallResult::ErrorCopyFailed; + } return copy(in, out, VFS_RC_LARGE_COPY_BLOCK) ? InstallResult::Success : InstallResult::ErrorCopyFailed; } @@ -848,7 +935,8 @@ VirtualFile ManualContentProvider::GetEntryUnparsed(u64 title_id, ContentRecordT VirtualFile ManualContentProvider::GetEntryRaw(u64 title_id, ContentRecordType type) const { const auto iter = std::find_if(entries.begin(), entries.end(), [title_id, type](const auto& entry) { - const auto [title_type, content_type, e_title_id] = entry.first; + const auto content_type = std::get<1>(entry.first); + const auto e_title_id = std::get<2>(entry.first); return content_type == type && e_title_id == title_id; }); if (iter == entries.end()) @@ -860,7 +948,7 @@ std::unique_ptr<NCA> ManualContentProvider::GetEntry(u64 title_id, ContentRecord const auto res = GetEntryRaw(title_id, type); if (res == nullptr) return nullptr; - return std::make_unique<NCA>(res, nullptr, 0, keys); + return std::make_unique<NCA>(res, nullptr, 0); } std::vector<ContentProviderEntry> ManualContentProvider::ListEntriesFilter( diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h index d1eec240e..5b414b0f0 100644 --- a/src/core/file_sys/registered_cache.h +++ b/src/core/file_sys/registered_cache.h @@ -34,6 +34,7 @@ using VfsCopyFunction = std::function<bool(const VirtualFile&, const VirtualFile enum class InstallResult { Success, + OverwriteExisting, ErrorAlreadyExists, ErrorCopyFailed, ErrorMetaFailed, @@ -88,7 +89,7 @@ public: protected: // A single instance of KeyManager to be used by GetEntry() - Core::Crypto::KeyManager keys; + Core::Crypto::KeyManager& keys = Core::Crypto::KeyManager::Instance(); }; class PlaceholderCache { @@ -132,9 +133,9 @@ public: // Parsing function defines the conversion from raw file to NCA. If there are other steps // besides creating the NCA from the file (e.g. NAX0 on SD Card), that should go in a custom // parsing function. - explicit RegisteredCache(VirtualDir dir, - ContentProviderParsingFunction parsing_function = - [](const VirtualFile& file, const NcaID& id) { return file; }); + explicit RegisteredCache( + VirtualDir dir, ContentProviderParsingFunction parsing_function = + [](const VirtualFile& file, const NcaID& id) { return file; }); ~RegisteredCache() override; void Refresh() override; @@ -168,6 +169,9 @@ public: InstallResult InstallEntry(const NCA& nca, TitleType type, bool overwrite_if_exists = false, const VfsCopyFunction& copy = &VfsRawCopy); + // Removes an existing entry based on title id + bool RemoveExistingEntry(u64 title_id) const; + private: template <typename T> void IterateAllMetadata(std::vector<T>& out, diff --git a/src/core/file_sys/romfs.h b/src/core/file_sys/romfs.h index 2fd07ed04..82e683782 100644 --- a/src/core/file_sys/romfs.h +++ b/src/core/file_sys/romfs.h @@ -4,7 +4,6 @@ #pragma once -#include <array> #include "core/file_sys/vfs.h" namespace FileSys { diff --git a/src/core/file_sys/romfs_factory.cpp b/src/core/file_sys/romfs_factory.cpp index 418a39a7e..987199747 100644 --- a/src/core/file_sys/romfs_factory.cpp +++ b/src/core/file_sys/romfs_factory.cpp @@ -6,7 +6,6 @@ #include "common/assert.h" #include "common/common_types.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/file_sys/card_image.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/nca_metadata.h" @@ -19,7 +18,9 @@ namespace FileSys { -RomFSFactory::RomFSFactory(Loader::AppLoader& app_loader) { +RomFSFactory::RomFSFactory(Loader::AppLoader& app_loader, ContentProvider& provider, + Service::FileSystem::FileSystemController& controller) + : content_provider{provider}, filesystem_controller{controller} { // Load the RomFS from the app if (app_loader.ReadRomFS(file) != Loader::ResultStatus::Success) { LOG_ERROR(Service_FS, "Unable to read RomFS!"); @@ -36,49 +37,50 @@ void RomFSFactory::SetPackedUpdate(VirtualFile update_raw) { } ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess(u64 current_process_title_id) const { - if (!updatable) + if (!updatable) { return MakeResult<VirtualFile>(file); + } - const PatchManager patch_manager(current_process_title_id); + const PatchManager patch_manager{current_process_title_id, filesystem_controller, + content_provider}; return MakeResult<VirtualFile>( patch_manager.PatchRomFS(file, ivfc_offset, ContentRecordType::Program, update_raw)); } ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, ContentRecordType type) const { - std::shared_ptr<NCA> res; - - switch (storage) { - case StorageId::None: - res = Core::System::GetInstance().GetContentProvider().GetEntry(title_id, type); - break; - case StorageId::NandSystem: - res = - Core::System::GetInstance().GetFileSystemController().GetSystemNANDContents()->GetEntry( - title_id, type); - break; - case StorageId::NandUser: - res = Core::System::GetInstance().GetFileSystemController().GetUserNANDContents()->GetEntry( - title_id, type); - break; - case StorageId::SdCard: - res = Core::System::GetInstance().GetFileSystemController().GetSDMCContents()->GetEntry( - title_id, type); - break; - default: - UNIMPLEMENTED_MSG("Unimplemented storage_id={:02X}", static_cast<u8>(storage)); - } - + const std::shared_ptr<NCA> res = GetEntry(title_id, storage, type); if (res == nullptr) { // TODO(DarkLordZach): Find the right error code to use here return RESULT_UNKNOWN; } + const auto romfs = res->GetRomFS(); if (romfs == nullptr) { // TODO(DarkLordZach): Find the right error code to use here return RESULT_UNKNOWN; } + return MakeResult<VirtualFile>(romfs); } +std::shared_ptr<NCA> RomFSFactory::GetEntry(u64 title_id, StorageId storage, + ContentRecordType type) const { + switch (storage) { + case StorageId::None: + return content_provider.GetEntry(title_id, type); + case StorageId::NandSystem: + return filesystem_controller.GetSystemNANDContents()->GetEntry(title_id, type); + case StorageId::NandUser: + return filesystem_controller.GetUserNANDContents()->GetEntry(title_id, type); + case StorageId::SdCard: + return filesystem_controller.GetSDMCContents()->GetEntry(title_id, type); + case StorageId::Host: + case StorageId::GameCard: + default: + UNIMPLEMENTED_MSG("Unimplemented storage_id={:02X}", static_cast<u8>(storage)); + return nullptr; + } +} + } // namespace FileSys diff --git a/src/core/file_sys/romfs_factory.h b/src/core/file_sys/romfs_factory.h index c5d40285c..ec704dfa8 100644 --- a/src/core/file_sys/romfs_factory.h +++ b/src/core/file_sys/romfs_factory.h @@ -13,8 +13,15 @@ namespace Loader { class AppLoader; } // namespace Loader +namespace Service::FileSystem { +class FileSystemController; +} + namespace FileSys { +class ContentProvider; +class NCA; + enum class ContentRecordType : u8; enum class StorageId : u8 { @@ -29,18 +36,26 @@ enum class StorageId : u8 { /// File system interface to the RomFS archive class RomFSFactory { public: - explicit RomFSFactory(Loader::AppLoader& app_loader); + explicit RomFSFactory(Loader::AppLoader& app_loader, ContentProvider& provider, + Service::FileSystem::FileSystemController& controller); ~RomFSFactory(); void SetPackedUpdate(VirtualFile update_raw); - ResultVal<VirtualFile> OpenCurrentProcess(u64 current_process_title_id) const; - ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type) const; + [[nodiscard]] ResultVal<VirtualFile> OpenCurrentProcess(u64 current_process_title_id) const; + [[nodiscard]] ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, + ContentRecordType type) const; private: + [[nodiscard]] std::shared_ptr<NCA> GetEntry(u64 title_id, StorageId storage, + ContentRecordType type) const; + VirtualFile file; VirtualFile update_raw; bool updatable; u64 ivfc_offset; + + ContentProvider& content_provider; + Service::FileSystem::FileSystemController& filesystem_controller; }; } // namespace FileSys diff --git a/src/core/file_sys/savedata_factory.cpp b/src/core/file_sys/savedata_factory.cpp index f3def93ab..ba4efee3a 100644 --- a/src/core/file_sys/savedata_factory.cpp +++ b/src/core/file_sys/savedata_factory.cpp @@ -17,23 +17,23 @@ constexpr char SAVE_DATA_SIZE_FILENAME[] = ".yuzu_save_size"; namespace { -void PrintSaveDataDescriptorWarnings(SaveDataDescriptor meta) { +void PrintSaveDataAttributeWarnings(SaveDataAttribute meta) { if (meta.type == SaveDataType::SystemSaveData || meta.type == SaveDataType::SaveData) { if (meta.zero_1 != 0) { LOG_WARNING(Service_FS, - "Possibly incorrect SaveDataDescriptor, type is " + "Possibly incorrect SaveDataAttribute, type is " "SystemSaveData||SaveData but offset 0x28 is non-zero ({:016X}).", meta.zero_1); } if (meta.zero_2 != 0) { LOG_WARNING(Service_FS, - "Possibly incorrect SaveDataDescriptor, type is " + "Possibly incorrect SaveDataAttribute, type is " "SystemSaveData||SaveData but offset 0x30 is non-zero ({:016X}).", meta.zero_2); } if (meta.zero_3 != 0) { LOG_WARNING(Service_FS, - "Possibly incorrect SaveDataDescriptor, type is " + "Possibly incorrect SaveDataAttribute, type is " "SystemSaveData||SaveData but offset 0x38 is non-zero ({:016X}).", meta.zero_3); } @@ -41,32 +41,32 @@ void PrintSaveDataDescriptorWarnings(SaveDataDescriptor meta) { if (meta.type == SaveDataType::SystemSaveData && meta.title_id != 0) { LOG_WARNING(Service_FS, - "Possibly incorrect SaveDataDescriptor, type is SystemSaveData but title_id is " + "Possibly incorrect SaveDataAttribute, type is SystemSaveData but title_id is " "non-zero ({:016X}).", meta.title_id); } if (meta.type == SaveDataType::DeviceSaveData && meta.user_id != u128{0, 0}) { LOG_WARNING(Service_FS, - "Possibly incorrect SaveDataDescriptor, type is DeviceSaveData but user_id is " + "Possibly incorrect SaveDataAttribute, type is DeviceSaveData but user_id is " "non-zero ({:016X}{:016X})", meta.user_id[1], meta.user_id[0]); } } -bool ShouldSaveDataBeAutomaticallyCreated(SaveDataSpaceId space, const SaveDataDescriptor& desc) { - return desc.type == SaveDataType::CacheStorage || desc.type == SaveDataType::TemporaryStorage || +bool ShouldSaveDataBeAutomaticallyCreated(SaveDataSpaceId space, const SaveDataAttribute& attr) { + return attr.type == SaveDataType::CacheStorage || attr.type == SaveDataType::TemporaryStorage || (space == SaveDataSpaceId::NandUser && ///< Normal Save Data -- Current Title & User - desc.type == SaveDataType::SaveData && desc.title_id == 0 && desc.save_id == 0); + (attr.type == SaveDataType::SaveData || attr.type == SaveDataType::DeviceSaveData) && + attr.title_id == 0 && attr.save_id == 0); } } // Anonymous namespace -std::string SaveDataDescriptor::DebugInfo() const { - return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, " - "save_id={:016X}, " +std::string SaveDataAttribute::DebugInfo() const { + return fmt::format("[title_id={:016X}, user_id={:016X}{:016X}, save_id={:016X}, type={:02X}, " "rank={}, index={}]", - static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id, + title_id, user_id[1], user_id[0], save_id, static_cast<u8>(type), static_cast<u8>(rank), index); } @@ -79,8 +79,8 @@ SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save SaveDataFactory::~SaveDataFactory() = default; ResultVal<VirtualDir> SaveDataFactory::Create(SaveDataSpaceId space, - const SaveDataDescriptor& meta) const { - PrintSaveDataDescriptorWarnings(meta); + const SaveDataAttribute& meta) const { + PrintSaveDataAttributeWarnings(meta); const auto save_directory = GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id); @@ -97,7 +97,7 @@ ResultVal<VirtualDir> SaveDataFactory::Create(SaveDataSpaceId space, } ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, - const SaveDataDescriptor& meta) const { + const SaveDataAttribute& meta) const { const auto save_directory = GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id); @@ -139,8 +139,10 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ u128 user_id, u64 save_id) { // According to switchbrew, if a save is of type SaveData and the title id field is 0, it should // be interpreted as the title id of the current process. - if (type == SaveDataType::SaveData && title_id == 0) { - title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); + if (type == SaveDataType::SaveData || type == SaveDataType::DeviceSaveData) { + if (title_id == 0) { + title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); + } } std::string out = GetSaveDataSpaceIdPath(space); diff --git a/src/core/file_sys/savedata_factory.h b/src/core/file_sys/savedata_factory.h index 991e57aa1..6625bbbd8 100644 --- a/src/core/file_sys/savedata_factory.h +++ b/src/core/file_sys/savedata_factory.h @@ -21,6 +21,7 @@ enum class SaveDataSpaceId : u8 { TemporaryStorage = 3, SdCardUser = 4, ProperSystem = 100, + SafeMode = 101, }; enum class SaveDataType : u8 { @@ -30,28 +31,50 @@ enum class SaveDataType : u8 { DeviceSaveData = 3, TemporaryStorage = 4, CacheStorage = 5, + SystemBcat = 6, }; enum class SaveDataRank : u8 { - Primary, - Secondary, + Primary = 0, + Secondary = 1, }; -struct SaveDataDescriptor { - u64_le title_id; +enum class SaveDataFlags : u32 { + None = (0 << 0), + KeepAfterResettingSystemSaveData = (1 << 0), + KeepAfterRefurbishment = (1 << 1), + KeepAfterResettingSystemSaveDataWithoutUserSaveData = (1 << 2), + NeedsSecureDelete = (1 << 3), +}; + +struct SaveDataAttribute { + u64 title_id; u128 user_id; - u64_le save_id; + u64 save_id; SaveDataType type; SaveDataRank rank; - u16_le index; + u16 index; INSERT_PADDING_BYTES(4); - u64_le zero_1; - u64_le zero_2; - u64_le zero_3; + u64 zero_1; + u64 zero_2; + u64 zero_3; std::string DebugInfo() const; }; -static_assert(sizeof(SaveDataDescriptor) == 0x40, "SaveDataDescriptor has incorrect size."); +static_assert(sizeof(SaveDataAttribute) == 0x40, "SaveDataAttribute has incorrect size."); + +struct SaveDataExtraData { + SaveDataAttribute attr; + u64 owner_id; + s64 timestamp; + SaveDataFlags flags; + INSERT_PADDING_BYTES(4); + s64 available_size; + s64 journal_size; + s64 commit_id; + std::array<u8, 0x190> unused; +}; +static_assert(sizeof(SaveDataExtraData) == 0x200, "SaveDataExtraData has incorrect size."); struct SaveDataSize { u64 normal; @@ -64,8 +87,8 @@ public: explicit SaveDataFactory(VirtualDir dir); ~SaveDataFactory(); - ResultVal<VirtualDir> Create(SaveDataSpaceId space, const SaveDataDescriptor& meta) const; - ResultVal<VirtualDir> Open(SaveDataSpaceId space, const SaveDataDescriptor& meta) const; + ResultVal<VirtualDir> Create(SaveDataSpaceId space, const SaveDataAttribute& meta) const; + ResultVal<VirtualDir> Open(SaveDataSpaceId space, const SaveDataAttribute& meta) const; VirtualDir GetSaveDataSpaceDirectory(SaveDataSpaceId space) const; diff --git a/src/core/file_sys/sdmc_factory.cpp b/src/core/file_sys/sdmc_factory.cpp index 5113a1ca6..cb56d8f2d 100644 --- a/src/core/file_sys/sdmc_factory.cpp +++ b/src/core/file_sys/sdmc_factory.cpp @@ -5,11 +5,13 @@ #include <memory> #include "core/file_sys/registered_cache.h" #include "core/file_sys/sdmc_factory.h" +#include "core/file_sys/vfs.h" #include "core/file_sys/xts_archive.h" -#include "core/settings.h" namespace FileSys { +constexpr u64 SDMC_TOTAL_SIZE = 0x10000000000; // 1 TiB + SDMCFactory::SDMCFactory(VirtualDir dir_) : dir(std::move(dir_)), contents(std::make_unique<RegisteredCache>( GetOrCreateDirectoryRelative(dir, "/Nintendo/Contents/registered"), @@ -46,7 +48,7 @@ u64 SDMCFactory::GetSDMCFreeSpace() const { } u64 SDMCFactory::GetSDMCTotalSpace() const { - return static_cast<u64>(Settings::values.sdmc_size); + return SDMC_TOTAL_SIZE; } } // namespace FileSys diff --git a/src/core/file_sys/sdmc_factory.h b/src/core/file_sys/sdmc_factory.h index 42dc4e08a..2bb92ba93 100644 --- a/src/core/file_sys/sdmc_factory.h +++ b/src/core/file_sys/sdmc_factory.h @@ -5,7 +5,7 @@ #pragma once #include <memory> -#include "core/file_sys/vfs.h" +#include "core/file_sys/vfs_types.h" #include "core/hle/result.h" namespace FileSys { diff --git a/src/core/file_sys/submission_package.cpp b/src/core/file_sys/submission_package.cpp index ef3084681..90641d23b 100644 --- a/src/core/file_sys/submission_package.cpp +++ b/src/core/file_sys/submission_package.cpp @@ -19,42 +19,10 @@ #include "core/loader/loader.h" namespace FileSys { -namespace { -void SetTicketKeys(const std::vector<VirtualFile>& files) { - Core::Crypto::KeyManager keys; - - for (const auto& ticket_file : files) { - if (ticket_file == nullptr) { - continue; - } - - if (ticket_file->GetExtension() != "tik") { - continue; - } - - if (ticket_file->GetSize() < - Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET + sizeof(Core::Crypto::Key128)) { - continue; - } - - Core::Crypto::Key128 key{}; - ticket_file->Read(key.data(), key.size(), Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET); - - // We get the name without the extension in order to create the rights ID. - std::string name_only(ticket_file->GetName()); - name_only.erase(name_only.size() - 4); - - const auto rights_id_raw = Common::HexStringToArray<16>(name_only); - u128 rights_id; - std::memcpy(rights_id.data(), rights_id_raw.data(), sizeof(u128)); - keys.SetKey(Core::Crypto::S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); - } -} -} // Anonymous namespace NSP::NSP(VirtualFile file_) : file(std::move(file_)), status{Loader::ResultStatus::Success}, - pfs(std::make_shared<PartitionFilesystem>(file)) { + pfs(std::make_shared<PartitionFilesystem>(file)), keys{Core::Crypto::KeyManager::Instance()} { if (pfs->GetStatus() != Loader::ResultStatus::Success) { status = pfs->GetStatus(); return; @@ -232,6 +200,35 @@ VirtualDir NSP::GetParentDirectory() const { return file->GetContainingDirectory(); } +void NSP::SetTicketKeys(const std::vector<VirtualFile>& files) { + for (const auto& ticket_file : files) { + if (ticket_file == nullptr) { + continue; + } + + if (ticket_file->GetExtension() != "tik") { + continue; + } + + if (ticket_file->GetSize() < + Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET + sizeof(Core::Crypto::Key128)) { + continue; + } + + Core::Crypto::Key128 key{}; + ticket_file->Read(key.data(), key.size(), Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET); + + // We get the name without the extension in order to create the rights ID. + std::string name_only(ticket_file->GetName()); + name_only.erase(name_only.size() - 4); + + const auto rights_id_raw = Common::HexStringToArray<16>(name_only); + u128 rights_id; + std::memcpy(rights_id.data(), rights_id_raw.data(), sizeof(u128)); + keys.SetKey(Core::Crypto::S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); + } +} + void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) { exefs = pfs; @@ -267,9 +264,9 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) { } const CNMT cnmt(inner_file); - auto& ncas_title = ncas[cnmt.GetTitleID()]; - ncas_title[{cnmt.GetType(), ContentRecordType::Meta}] = nca; + ncas[cnmt.GetTitleID()][{cnmt.GetType(), ContentRecordType::Meta}] = nca; + for (const auto& rec : cnmt.GetContentRecords()) { const auto id_string = Common::HexToString(rec.nca_id, false); auto next_file = pfs->GetFile(fmt::format("{}.nca", id_string)); @@ -285,14 +282,33 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) { continue; } - auto next_nca = std::make_shared<NCA>(std::move(next_file), nullptr, 0, keys); + auto next_nca = std::make_shared<NCA>(std::move(next_file), nullptr, 0); + if (next_nca->GetType() == NCAContentType::Program) { - program_status[cnmt.GetTitleID()] = next_nca->GetStatus(); + program_status[next_nca->GetTitleId()] = next_nca->GetStatus(); } - if (next_nca->GetStatus() == Loader::ResultStatus::Success || - (next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS && - (cnmt.GetTitleID() & 0x800) != 0)) { - ncas_title[{cnmt.GetType(), rec.type}] = std::move(next_nca); + + if (next_nca->GetStatus() != Loader::ResultStatus::Success && + next_nca->GetStatus() != Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { + continue; + } + + // If the last 3 hexadecimal digits of the CNMT TitleID is 0x800 or is missing the + // BKTRBaseRomFS, this is an update NCA. Otherwise, this is a base NCA. + if ((cnmt.GetTitleID() & 0x800) != 0 || + next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { + // If the last 3 hexadecimal digits of the NCA's TitleID is between 0x1 and + // 0x7FF, this is a multi-program update NCA. Otherwise, this is a regular + // update NCA. + if ((next_nca->GetTitleId() & 0x7FF) != 0 && + (next_nca->GetTitleId() & 0x800) == 0) { + ncas[next_nca->GetTitleId()][{cnmt.GetType(), rec.type}] = + std::move(next_nca); + } else { + ncas[cnmt.GetTitleID()][{cnmt.GetType(), rec.type}] = std::move(next_nca); + } + } else { + ncas[next_nca->GetTitleId()][{cnmt.GetType(), rec.type}] = std::move(next_nca); } } diff --git a/src/core/file_sys/submission_package.h b/src/core/file_sys/submission_package.h index ee9b6ce17..c70a11b5b 100644 --- a/src/core/file_sys/submission_package.h +++ b/src/core/file_sys/submission_package.h @@ -10,6 +10,10 @@ #include "common/common_types.h" #include "core/file_sys/vfs.h" +namespace Core::Crypto { +class KeyManager; +} + namespace Loader { enum class ResultStatus : u16; } @@ -59,6 +63,7 @@ public: VirtualDir GetParentDirectory() const override; private: + void SetTicketKeys(const std::vector<VirtualFile>& files); void InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files); void ReadNCAs(const std::vector<VirtualFile>& files); @@ -73,7 +78,7 @@ private: std::map<u64, std::map<std::pair<TitleType, ContentRecordType>, std::shared_ptr<NCA>>> ncas; std::vector<VirtualFile> ticket_files; - Core::Crypto::KeyManager keys; + Core::Crypto::KeyManager& keys; VirtualFile romfs; VirtualDir exefs; diff --git a/src/core/file_sys/system_archive/mii_model.cpp b/src/core/file_sys/system_archive/mii_model.cpp index 6a9add87c..d65c7d234 100644 --- a/src/core/file_sys/system_archive/mii_model.cpp +++ b/src/core/file_sys/system_archive/mii_model.cpp @@ -27,20 +27,14 @@ VirtualDir MiiModel() { auto out = std::make_shared<VectorVfsDirectory>(std::vector<VirtualFile>{}, std::vector<VirtualDir>{}, "data"); - out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_LOW_LINEAR.size()>>( - MiiModelData::TEXTURE_LOW_LINEAR, "NXTextureLowLinear.dat")); - out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_LOW_SRGB.size()>>( - MiiModelData::TEXTURE_LOW_SRGB, "NXTextureLowSRGB.dat")); - out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_MID_LINEAR.size()>>( - MiiModelData::TEXTURE_MID_LINEAR, "NXTextureMidLinear.dat")); - out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_MID_SRGB.size()>>( - MiiModelData::TEXTURE_MID_SRGB, "NXTextureMidSRGB.dat")); - out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_HIGH.size()>>( - MiiModelData::SHAPE_HIGH, "ShapeHigh.dat")); - out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_MID.size()>>( - MiiModelData::SHAPE_MID, "ShapeMid.dat")); - - return std::move(out); + out->AddFile(MakeArrayFile(MiiModelData::TEXTURE_LOW_LINEAR, "NXTextureLowLinear.dat")); + out->AddFile(MakeArrayFile(MiiModelData::TEXTURE_LOW_SRGB, "NXTextureLowSRGB.dat")); + out->AddFile(MakeArrayFile(MiiModelData::TEXTURE_MID_LINEAR, "NXTextureMidLinear.dat")); + out->AddFile(MakeArrayFile(MiiModelData::TEXTURE_MID_SRGB, "NXTextureMidSRGB.dat")); + out->AddFile(MakeArrayFile(MiiModelData::SHAPE_HIGH, "ShapeHigh.dat")); + out->AddFile(MakeArrayFile(MiiModelData::SHAPE_MID, "ShapeMid.dat")); + + return out; } } // namespace FileSys::SystemArchive diff --git a/src/core/file_sys/system_archive/ng_word.cpp b/src/core/file_sys/system_archive/ng_word.cpp index f4443784d..100d3c5db 100644 --- a/src/core/file_sys/system_archive/ng_word.cpp +++ b/src/core/file_sys/system_archive/ng_word.cpp @@ -24,19 +24,18 @@ constexpr std::array<u8, 30> WORD_TXT{ } // namespace NgWord1Data VirtualDir NgWord1() { - std::vector<VirtualFile> files(NgWord1Data::NUMBER_WORD_TXT_FILES); + std::vector<VirtualFile> files; + files.reserve(NgWord1Data::NUMBER_WORD_TXT_FILES); for (std::size_t i = 0; i < files.size(); ++i) { - files[i] = std::make_shared<ArrayVfsFile<NgWord1Data::WORD_TXT.size()>>( - NgWord1Data::WORD_TXT, fmt::format("{}.txt", i)); + files.push_back(MakeArrayFile(NgWord1Data::WORD_TXT, fmt::format("{}.txt", i))); } - files.push_back(std::make_shared<ArrayVfsFile<NgWord1Data::WORD_TXT.size()>>( - NgWord1Data::WORD_TXT, "common.txt")); - files.push_back(std::make_shared<ArrayVfsFile<NgWord1Data::VERSION_DAT.size()>>( - NgWord1Data::VERSION_DAT, "version.dat")); + files.push_back(MakeArrayFile(NgWord1Data::WORD_TXT, "common.txt")); + files.push_back(MakeArrayFile(NgWord1Data::VERSION_DAT, "version.dat")); - return std::make_shared<VectorVfsDirectory>(files, std::vector<VirtualDir>{}, "data"); + return std::make_shared<VectorVfsDirectory>(std::move(files), std::vector<VirtualDir>{}, + "data"); } namespace NgWord2Data { @@ -55,27 +54,22 @@ constexpr std::array<u8, 0x2C> AC_NX_DATA{ } // namespace NgWord2Data VirtualDir NgWord2() { - std::vector<VirtualFile> files(NgWord2Data::NUMBER_AC_NX_FILES * 3); + std::vector<VirtualFile> files; + files.reserve(NgWord2Data::NUMBER_AC_NX_FILES * 3); for (std::size_t i = 0; i < NgWord2Data::NUMBER_AC_NX_FILES; ++i) { - files[3 * i] = std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>( - NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_b1_nx", i)); - files[3 * i + 1] = std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>( - NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_b2_nx", i)); - files[3 * i + 2] = std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>( - NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_not_b_nx", i)); + files.push_back(MakeArrayFile(NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_b1_nx", i))); + files.push_back(MakeArrayFile(NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_b2_nx", i))); + files.push_back(MakeArrayFile(NgWord2Data::AC_NX_DATA, fmt::format("ac_{}_not_b_nx", i))); } - files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>( - NgWord2Data::AC_NX_DATA, "ac_common_b1_nx")); - files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>( - NgWord2Data::AC_NX_DATA, "ac_common_b2_nx")); - files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::AC_NX_DATA.size()>>( - NgWord2Data::AC_NX_DATA, "ac_common_not_b_nx")); - files.push_back(std::make_shared<ArrayVfsFile<NgWord2Data::VERSION_DAT.size()>>( - NgWord2Data::VERSION_DAT, "version.dat")); + files.push_back(MakeArrayFile(NgWord2Data::AC_NX_DATA, "ac_common_b1_nx")); + files.push_back(MakeArrayFile(NgWord2Data::AC_NX_DATA, "ac_common_b2_nx")); + files.push_back(MakeArrayFile(NgWord2Data::AC_NX_DATA, "ac_common_not_b_nx")); + files.push_back(MakeArrayFile(NgWord2Data::VERSION_DAT, "version.dat")); - return std::make_shared<VectorVfsDirectory>(files, std::vector<VirtualDir>{}, "data"); + return std::make_shared<VectorVfsDirectory>(std::move(files), std::vector<VirtualDir>{}, + "data"); } } // namespace FileSys::SystemArchive diff --git a/src/core/file_sys/system_archive/shared_font.cpp b/src/core/file_sys/system_archive/shared_font.cpp index 2c05eb42e..c5cdf7d9b 100644 --- a/src/core/file_sys/system_archive/shared_font.cpp +++ b/src/core/file_sys/system_archive/shared_font.cpp @@ -23,7 +23,7 @@ VirtualFile PackBFTTF(const std::array<u8, Size>& data, const std::string& name) std::vector<u8> bfttf(Size + sizeof(u64)); - u64 offset = 0; + size_t offset = 0; Service::NS::EncryptSharedFont(vec, bfttf, offset); return std::make_shared<VectorVfsFile>(std::move(bfttf), name); } diff --git a/src/core/file_sys/system_archive/system_version.cpp b/src/core/file_sys/system_archive/system_version.cpp index 6e22f97b0..aa313de66 100644 --- a/src/core/file_sys/system_archive/system_version.cpp +++ b/src/core/file_sys/system_archive/system_version.cpp @@ -12,17 +12,17 @@ namespace SystemVersionData { // This section should reflect the best system version to describe yuzu's HLE api. // TODO(DarkLordZach): Update when HLE gets better. -constexpr u8 VERSION_MAJOR = 5; -constexpr u8 VERSION_MINOR = 1; -constexpr u8 VERSION_MICRO = 0; +constexpr u8 VERSION_MAJOR = 10; +constexpr u8 VERSION_MINOR = 0; +constexpr u8 VERSION_MICRO = 2; -constexpr u8 REVISION_MAJOR = 3; +constexpr u8 REVISION_MAJOR = 1; constexpr u8 REVISION_MINOR = 0; constexpr char PLATFORM_STRING[] = "NX"; -constexpr char VERSION_HASH[] = "23f9df53e25709d756e0c76effcb2473bd3447dd"; -constexpr char DISPLAY_VERSION[] = "5.1.0"; -constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 5.1.0-3.0"; +constexpr char VERSION_HASH[] = "f90143fa8bbc061d4f68c35f95f04f8080c0ecdc"; +constexpr char DISPLAY_VERSION[] = "10.0.2"; +constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 10.0.2-1.0"; } // namespace SystemVersionData diff --git a/src/core/file_sys/system_archive/time_zone_binary.cpp b/src/core/file_sys/system_archive/time_zone_binary.cpp index 9806bd197..8fd005012 100644 --- a/src/core/file_sys/system_archive/time_zone_binary.cpp +++ b/src/core/file_sys/system_archive/time_zone_binary.cpp @@ -2,6 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> +#include <vector> + #include "common/swap.h" #include "core/file_sys/system_archive/time_zone_binary.h" #include "core/file_sys/vfs_vector.h" @@ -615,43 +618,49 @@ static constexpr std::array<u8, 9633> LOCATION_NAMES{ 0x0a}; static VirtualFile GenerateDefaultTimeZoneFile() { - struct { + struct TimeZoneInfo { s64_be at; - INSERT_PADDING_BYTES(7); + std::array<u8, 7> padding1; std::array<char, 4> time_zone_chars; - INSERT_PADDING_BYTES(2); + std::array<u8, 2> padding2; std::array<char, 6> time_zone_name; - } time_zone_info{}; + }; - const VirtualFile file{std::make_shared<VectorVfsFile>( - std::vector<u8>(sizeof(Service::Time::TimeZone::TzifHeader) + sizeof(time_zone_info)), + VirtualFile file{std::make_shared<VectorVfsFile>( + std::vector<u8>(sizeof(Service::Time::TimeZone::TzifHeader) + sizeof(TimeZoneInfo)), "GMT")}; - Service::Time::TimeZone::TzifHeader header{}; - header.magic = 0x545a6966; - header.version = 0x32; - header.ttis_gmt_count = 0x1; - header.ttis_std_count = 0x1; - header.time_count = 0x1; - header.type_count = 0x1; - header.char_count = 0x4; + const Service::Time::TimeZone::TzifHeader header{ + .magic = 0x545a6966, + .version = 0x32, + .ttis_gmt_count = 1, + .ttis_std_count = 1, + .time_count = 1, + .type_count = 1, + .char_count = 4, + }; file->WriteObject(header, 0); - time_zone_info.at = 0xf8; - time_zone_info.time_zone_chars = {'G', 'M', 'T', '\0'}; - time_zone_info.time_zone_name = {'\n', 'G', 'M', 'T', '0', '\n'}; + const TimeZoneInfo time_zone_info{ + .at = 0xf8, + .padding1 = {}, + .time_zone_chars = {'G', 'M', 'T', '\0'}, + .padding2 = {}, + .time_zone_name = {'\n', 'G', 'M', 'T', '0', '\n'}, + }; file->WriteObject(time_zone_info, sizeof(Service::Time::TimeZone::TzifHeader)); return file; } VirtualDir TimeZoneBinary() { - const std::vector<VirtualDir> root_dirs{std::make_shared<VectorVfsDirectory>( + std::vector<VirtualDir> root_dirs{std::make_shared<VectorVfsDirectory>( std::vector<VirtualFile>{GenerateDefaultTimeZoneFile()}, std::vector<VirtualDir>{}, "zoneinfo")}; - const std::vector<VirtualFile> root_files{ - std::make_shared<ArrayVfsFile<LOCATION_NAMES.size()>>(LOCATION_NAMES, "binaryList.txt")}; - return std::make_shared<VectorVfsDirectory>(root_files, root_dirs, "data"); + std::vector<VirtualFile> root_files{MakeArrayFile(LOCATION_NAMES, "binaryList.txt")}; + + return std::make_shared<VectorVfsDirectory>(std::move(root_files), std::move(root_dirs), + "data"); } } // namespace FileSys::SystemArchive diff --git a/src/core/file_sys/vfs.cpp b/src/core/file_sys/vfs.cpp index e33327ef0..b2f026b6d 100644 --- a/src/core/file_sys/vfs.cpp +++ b/src/core/file_sys/vfs.cpp @@ -30,7 +30,7 @@ bool VfsFilesystem::IsWritable() const { } VfsEntryType VfsFilesystem::GetEntryType(std::string_view path_) const { - const auto path = FileUtil::SanitizePath(path_); + const auto path = Common::FS::SanitizePath(path_); if (root->GetFileRelative(path) != nullptr) return VfsEntryType::File; if (root->GetDirectoryRelative(path) != nullptr) @@ -40,22 +40,22 @@ VfsEntryType VfsFilesystem::GetEntryType(std::string_view path_) const { } VirtualFile VfsFilesystem::OpenFile(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_); + const auto path = Common::FS::SanitizePath(path_); return root->GetFileRelative(path); } VirtualFile VfsFilesystem::CreateFile(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_); + const auto path = Common::FS::SanitizePath(path_); return root->CreateFileRelative(path); } VirtualFile VfsFilesystem::CopyFile(std::string_view old_path_, std::string_view new_path_) { - const auto old_path = FileUtil::SanitizePath(old_path_); - const auto new_path = FileUtil::SanitizePath(new_path_); + const auto old_path = Common::FS::SanitizePath(old_path_); + const auto new_path = Common::FS::SanitizePath(new_path_); // VfsDirectory impls are only required to implement copy across the current directory. - if (FileUtil::GetParentPath(old_path) == FileUtil::GetParentPath(new_path)) { - if (!root->Copy(FileUtil::GetFilename(old_path), FileUtil::GetFilename(new_path))) + if (Common::FS::GetParentPath(old_path) == Common::FS::GetParentPath(new_path)) { + if (!root->Copy(Common::FS::GetFilename(old_path), Common::FS::GetFilename(new_path))) return nullptr; return OpenFile(new_path, Mode::ReadWrite); } @@ -76,8 +76,8 @@ VirtualFile VfsFilesystem::CopyFile(std::string_view old_path_, std::string_view } VirtualFile VfsFilesystem::MoveFile(std::string_view old_path, std::string_view new_path) { - const auto sanitized_old_path = FileUtil::SanitizePath(old_path); - const auto sanitized_new_path = FileUtil::SanitizePath(new_path); + const auto sanitized_old_path = Common::FS::SanitizePath(old_path); + const auto sanitized_new_path = Common::FS::SanitizePath(new_path); // Again, non-default impls are highly encouraged to provide a more optimized version of this. auto out = CopyFile(sanitized_old_path, sanitized_new_path); @@ -89,26 +89,26 @@ VirtualFile VfsFilesystem::MoveFile(std::string_view old_path, std::string_view } bool VfsFilesystem::DeleteFile(std::string_view path_) { - const auto path = FileUtil::SanitizePath(path_); - auto parent = OpenDirectory(FileUtil::GetParentPath(path), Mode::Write); + const auto path = Common::FS::SanitizePath(path_); + auto parent = OpenDirectory(Common::FS::GetParentPath(path), Mode::Write); if (parent == nullptr) return false; - return parent->DeleteFile(FileUtil::GetFilename(path)); + return parent->DeleteFile(Common::FS::GetFilename(path)); } VirtualDir VfsFilesystem::OpenDirectory(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_); + const auto path = Common::FS::SanitizePath(path_); return root->GetDirectoryRelative(path); } VirtualDir VfsFilesystem::CreateDirectory(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_); + const auto path = Common::FS::SanitizePath(path_); return root->CreateDirectoryRelative(path); } VirtualDir VfsFilesystem::CopyDirectory(std::string_view old_path_, std::string_view new_path_) { - const auto old_path = FileUtil::SanitizePath(old_path_); - const auto new_path = FileUtil::SanitizePath(new_path_); + const auto old_path = Common::FS::SanitizePath(old_path_); + const auto new_path = Common::FS::SanitizePath(new_path_); // Non-default impls are highly encouraged to provide a more optimized version of this. auto old_dir = OpenDirectory(old_path, Mode::Read); @@ -139,8 +139,8 @@ VirtualDir VfsFilesystem::CopyDirectory(std::string_view old_path_, std::string_ } VirtualDir VfsFilesystem::MoveDirectory(std::string_view old_path, std::string_view new_path) { - const auto sanitized_old_path = FileUtil::SanitizePath(old_path); - const auto sanitized_new_path = FileUtil::SanitizePath(new_path); + const auto sanitized_old_path = Common::FS::SanitizePath(old_path); + const auto sanitized_new_path = Common::FS::SanitizePath(new_path); // Non-default impls are highly encouraged to provide a more optimized version of this. auto out = CopyDirectory(sanitized_old_path, sanitized_new_path); @@ -152,28 +152,29 @@ VirtualDir VfsFilesystem::MoveDirectory(std::string_view old_path, std::string_v } bool VfsFilesystem::DeleteDirectory(std::string_view path_) { - const auto path = FileUtil::SanitizePath(path_); - auto parent = OpenDirectory(FileUtil::GetParentPath(path), Mode::Write); + const auto path = Common::FS::SanitizePath(path_); + auto parent = OpenDirectory(Common::FS::GetParentPath(path), Mode::Write); if (parent == nullptr) return false; - return parent->DeleteSubdirectoryRecursive(FileUtil::GetFilename(path)); + return parent->DeleteSubdirectoryRecursive(Common::FS::GetFilename(path)); } VfsFile::~VfsFile() = default; std::string VfsFile::GetExtension() const { - return std::string(FileUtil::GetExtensionFromFilename(GetName())); + return std::string(Common::FS::GetExtensionFromFilename(GetName())); } VfsDirectory::~VfsDirectory() = default; std::optional<u8> VfsFile::ReadByte(std::size_t offset) const { u8 out{}; - std::size_t size = Read(&out, 1, offset); - if (size == 1) + const std::size_t size = Read(&out, sizeof(u8), offset); + if (size == 1) { return out; + } - return {}; + return std::nullopt; } std::vector<u8> VfsFile::ReadBytes(std::size_t size, std::size_t offset) const { @@ -203,7 +204,7 @@ std::string VfsFile::GetFullPath() const { } std::shared_ptr<VfsFile> VfsDirectory::GetFileRelative(std::string_view path) const { - auto vec = FileUtil::SplitPathComponents(path); + auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); if (vec.empty()) { @@ -239,7 +240,7 @@ std::shared_ptr<VfsFile> VfsDirectory::GetFileAbsolute(std::string_view path) co } std::shared_ptr<VfsDirectory> VfsDirectory::GetDirectoryRelative(std::string_view path) const { - auto vec = FileUtil::SplitPathComponents(path); + auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); if (vec.empty()) { @@ -301,7 +302,7 @@ std::size_t VfsDirectory::GetSize() const { } std::shared_ptr<VfsFile> VfsDirectory::CreateFileRelative(std::string_view path) { - auto vec = FileUtil::SplitPathComponents(path); + auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); if (vec.empty()) { @@ -320,7 +321,7 @@ std::shared_ptr<VfsFile> VfsDirectory::CreateFileRelative(std::string_view path) } } - return dir->CreateFileRelative(FileUtil::GetPathWithoutTop(path)); + return dir->CreateFileRelative(Common::FS::GetPathWithoutTop(path)); } std::shared_ptr<VfsFile> VfsDirectory::CreateFileAbsolute(std::string_view path) { @@ -332,7 +333,7 @@ std::shared_ptr<VfsFile> VfsDirectory::CreateFileAbsolute(std::string_view path) } std::shared_ptr<VfsDirectory> VfsDirectory::CreateDirectoryRelative(std::string_view path) { - auto vec = FileUtil::SplitPathComponents(path); + auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); if (vec.empty()) { @@ -351,7 +352,7 @@ std::shared_ptr<VfsDirectory> VfsDirectory::CreateDirectoryRelative(std::string_ } } - return dir->CreateDirectoryRelative(FileUtil::GetPathWithoutTop(path)); + return dir->CreateDirectoryRelative(Common::FS::GetPathWithoutTop(path)); } std::shared_ptr<VfsDirectory> VfsDirectory::CreateDirectoryAbsolute(std::string_view path) { diff --git a/src/core/file_sys/vfs_concat.cpp b/src/core/file_sys/vfs_concat.cpp index 16d801c0c..e0ff70174 100644 --- a/src/core/file_sys/vfs_concat.cpp +++ b/src/core/file_sys/vfs_concat.cpp @@ -11,7 +11,7 @@ namespace FileSys { -static bool VerifyConcatenationMapContinuity(const std::map<u64, VirtualFile>& map) { +static bool VerifyConcatenationMapContinuity(const std::multimap<u64, VirtualFile>& map) { const auto last_valid = --map.end(); for (auto iter = map.begin(); iter != last_valid;) { const auto old = iter++; @@ -27,12 +27,12 @@ ConcatenatedVfsFile::ConcatenatedVfsFile(std::vector<VirtualFile> files_, std::s : name(std::move(name)) { std::size_t next_offset = 0; for (const auto& file : files_) { - files[next_offset] = file; + files.emplace(next_offset, file); next_offset += file->GetSize(); } } -ConcatenatedVfsFile::ConcatenatedVfsFile(std::map<u64, VirtualFile> files_, std::string name) +ConcatenatedVfsFile::ConcatenatedVfsFile(std::multimap<u64, VirtualFile> files_, std::string name) : files(std::move(files_)), name(std::move(name)) { ASSERT(VerifyConcatenationMapContinuity(files)); } @@ -50,7 +50,7 @@ VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(std::vector<VirtualFile> f } VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(u8 filler_byte, - std::map<u64, VirtualFile> files, + std::multimap<u64, VirtualFile> files, std::string name) { if (files.empty()) return nullptr; diff --git a/src/core/file_sys/vfs_concat.h b/src/core/file_sys/vfs_concat.h index c90f9d5d1..7a26343c0 100644 --- a/src/core/file_sys/vfs_concat.h +++ b/src/core/file_sys/vfs_concat.h @@ -15,7 +15,7 @@ namespace FileSys { // read-only. class ConcatenatedVfsFile : public VfsFile { ConcatenatedVfsFile(std::vector<VirtualFile> files, std::string name); - ConcatenatedVfsFile(std::map<u64, VirtualFile> files, std::string name); + ConcatenatedVfsFile(std::multimap<u64, VirtualFile> files, std::string name); public: ~ConcatenatedVfsFile() override; @@ -25,7 +25,7 @@ public: /// Convenience function that turns a map of offsets to files into a concatenated file, filling /// gaps with a given filler byte. - static VirtualFile MakeConcatenatedFile(u8 filler_byte, std::map<u64, VirtualFile> files, + static VirtualFile MakeConcatenatedFile(u8 filler_byte, std::multimap<u64, VirtualFile> files, std::string name); std::string GetName() const override; @@ -40,7 +40,7 @@ public: private: // Maps starting offset to file -- more efficient. - std::map<u64, VirtualFile> files; + std::multimap<u64, VirtualFile> files; std::string name; }; diff --git a/src/core/file_sys/vfs_libzip.cpp b/src/core/file_sys/vfs_libzip.cpp index 11d1978ea..429d7bc8b 100644 --- a/src/core/file_sys/vfs_libzip.cpp +++ b/src/core/file_sys/vfs_libzip.cpp @@ -42,14 +42,14 @@ VirtualDir ExtractZIP(VirtualFile file) { continue; if (name.back() != '/') { - std::unique_ptr<zip_file_t, decltype(&zip_fclose)> file{ + std::unique_ptr<zip_file_t, decltype(&zip_fclose)> file2{ zip_fopen_index(zip.get(), i, 0), zip_fclose}; std::vector<u8> buf(stat.size); - if (zip_fread(file.get(), buf.data(), buf.size()) != buf.size()) + if (zip_fread(file2.get(), buf.data(), buf.size()) != s64(buf.size())) return nullptr; - const auto parts = FileUtil::SplitPathComponents(stat.name); + const auto parts = Common::FS::SplitPathComponents(stat.name); const auto new_file = std::make_shared<VectorVfsFile>(buf, parts.back()); std::shared_ptr<VectorVfsDirectory> dtrv = out; diff --git a/src/core/file_sys/vfs_offset.cpp b/src/core/file_sys/vfs_offset.cpp index c96f88488..7714d3de5 100644 --- a/src/core/file_sys/vfs_offset.cpp +++ b/src/core/file_sys/vfs_offset.cpp @@ -58,10 +58,11 @@ std::size_t OffsetVfsFile::Write(const u8* data, std::size_t length, std::size_t } std::optional<u8> OffsetVfsFile::ReadByte(std::size_t r_offset) const { - if (r_offset < size) - return file->ReadByte(offset + r_offset); + if (r_offset >= size) { + return std::nullopt; + } - return {}; + return file->ReadByte(offset + r_offset); } std::vector<u8> OffsetVfsFile::ReadBytes(std::size_t r_size, std::size_t r_offset) const { diff --git a/src/core/file_sys/vfs_real.cpp b/src/core/file_sys/vfs_real.cpp index e21300a7c..488687ba9 100644 --- a/src/core/file_sys/vfs_real.cpp +++ b/src/core/file_sys/vfs_real.cpp @@ -14,24 +14,28 @@ namespace FileSys { +namespace FS = Common::FS; + static std::string ModeFlagsToString(Mode mode) { std::string mode_str; // Calculate the correct open mode for the file. - if (mode & Mode::Read && mode & Mode::Write) { - if (mode & Mode::Append) + if (True(mode & Mode::Read) && True(mode & Mode::Write)) { + if (True(mode & Mode::Append)) { mode_str = "a+"; - else + } else { mode_str = "r+"; + } } else { - if (mode & Mode::Read) + if (True(mode & Mode::Read)) { mode_str = "r"; - else if (mode & Mode::Append) + } else if (True(mode & Mode::Append)) { mode_str = "a"; - else if (mode & Mode::Write) + } else if (True(mode & Mode::Write)) { mode_str = "w"; - else + } else { UNREACHABLE_MSG("Invalid file open mode: {:02X}", static_cast<u8>(mode)); + } } mode_str += "b"; @@ -55,102 +59,118 @@ bool RealVfsFilesystem::IsWritable() const { } VfsEntryType RealVfsFilesystem::GetEntryType(std::string_view path_) const { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - if (!FileUtil::Exists(path)) + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); + if (!FS::Exists(path)) { return VfsEntryType::None; - if (FileUtil::IsDirectory(path)) + } + if (FS::IsDirectory(path)) { return VfsEntryType::Directory; + } return VfsEntryType::File; } VirtualFile RealVfsFilesystem::OpenFile(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - if (cache.find(path) != cache.end()) { - auto weak = cache[path]; + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); + + if (const auto weak_iter = cache.find(path); weak_iter != cache.cend()) { + const auto& weak = weak_iter->second; + if (!weak.expired()) { return std::shared_ptr<RealVfsFile>(new RealVfsFile(*this, weak.lock(), path, perms)); } } - if (!FileUtil::Exists(path) && (perms & Mode::WriteAppend) != 0) - FileUtil::CreateEmptyFile(path); + if (!FS::Exists(path) && True(perms & Mode::WriteAppend)) { + FS::CreateEmptyFile(path); + } - auto backing = std::make_shared<FileUtil::IOFile>(path, ModeFlagsToString(perms).c_str()); - cache[path] = backing; + auto backing = std::make_shared<FS::IOFile>(path, ModeFlagsToString(perms).c_str()); + cache.insert_or_assign(path, backing); // Cannot use make_shared as RealVfsFile constructor is private return std::shared_ptr<RealVfsFile>(new RealVfsFile(*this, backing, path, perms)); } VirtualFile RealVfsFilesystem::CreateFile(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - const auto path_fwd = FileUtil::SanitizePath(path, FileUtil::DirectorySeparator::ForwardSlash); - if (!FileUtil::Exists(path)) { - FileUtil::CreateFullPath(path_fwd); - if (!FileUtil::CreateEmptyFile(path)) + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); + const auto path_fwd = FS::SanitizePath(path, FS::DirectorySeparator::ForwardSlash); + if (!FS::Exists(path)) { + FS::CreateFullPath(path_fwd); + if (!FS::CreateEmptyFile(path)) { return nullptr; + } } return OpenFile(path, perms); } VirtualFile RealVfsFilesystem::CopyFile(std::string_view old_path_, std::string_view new_path_) { - const auto old_path = - FileUtil::SanitizePath(old_path_, FileUtil::DirectorySeparator::PlatformDefault); - const auto new_path = - FileUtil::SanitizePath(new_path_, FileUtil::DirectorySeparator::PlatformDefault); + const auto old_path = FS::SanitizePath(old_path_, FS::DirectorySeparator::PlatformDefault); + const auto new_path = FS::SanitizePath(new_path_, FS::DirectorySeparator::PlatformDefault); - if (!FileUtil::Exists(old_path) || FileUtil::Exists(new_path) || - FileUtil::IsDirectory(old_path) || !FileUtil::Copy(old_path, new_path)) + if (!FS::Exists(old_path) || FS::Exists(new_path) || FS::IsDirectory(old_path) || + !FS::Copy(old_path, new_path)) { return nullptr; + } return OpenFile(new_path, Mode::ReadWrite); } VirtualFile RealVfsFilesystem::MoveFile(std::string_view old_path_, std::string_view new_path_) { - const auto old_path = - FileUtil::SanitizePath(old_path_, FileUtil::DirectorySeparator::PlatformDefault); - const auto new_path = - FileUtil::SanitizePath(new_path_, FileUtil::DirectorySeparator::PlatformDefault); + const auto old_path = FS::SanitizePath(old_path_, FS::DirectorySeparator::PlatformDefault); + const auto new_path = FS::SanitizePath(new_path_, FS::DirectorySeparator::PlatformDefault); + const auto cached_file_iter = cache.find(old_path); - if (!FileUtil::Exists(old_path) || FileUtil::Exists(new_path) || - FileUtil::IsDirectory(old_path) || !FileUtil::Rename(old_path, new_path)) - return nullptr; + if (cached_file_iter != cache.cend()) { + auto file = cached_file_iter->second.lock(); - if (cache.find(old_path) != cache.end()) { - auto cached = cache[old_path]; - if (!cached.expired()) { - auto file = cached.lock(); - file->Open(new_path, "r+b"); - cache.erase(old_path); - cache[new_path] = file; + if (!cached_file_iter->second.expired()) { + file->Close(); } + + if (!FS::Exists(old_path) || FS::Exists(new_path) || FS::IsDirectory(old_path) || + !FS::Rename(old_path, new_path)) { + return nullptr; + } + + cache.erase(old_path); + file->Open(new_path, "r+b"); + cache.insert_or_assign(new_path, std::move(file)); + } else { + UNREACHABLE(); + return nullptr; } + return OpenFile(new_path, Mode::ReadWrite); } bool RealVfsFilesystem::DeleteFile(std::string_view path_) { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - if (cache.find(path) != cache.end()) { - if (!cache[path].expired()) - cache[path].lock()->Close(); + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); + const auto cached_iter = cache.find(path); + + if (cached_iter != cache.cend()) { + if (!cached_iter->second.expired()) { + cached_iter->second.lock()->Close(); + } cache.erase(path); } - return FileUtil::Delete(path); + + return FS::Delete(path); } VirtualDir RealVfsFilesystem::OpenDirectory(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); // Cannot use make_shared as RealVfsDirectory constructor is private return std::shared_ptr<RealVfsDirectory>(new RealVfsDirectory(*this, path, perms)); } VirtualDir RealVfsFilesystem::CreateDirectory(std::string_view path_, Mode perms) { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - const auto path_fwd = FileUtil::SanitizePath(path, FileUtil::DirectorySeparator::ForwardSlash); - if (!FileUtil::Exists(path)) { - FileUtil::CreateFullPath(path_fwd); - if (!FileUtil::CreateDir(path)) + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); + const auto path_fwd = FS::SanitizePath(path, FS::DirectorySeparator::ForwardSlash); + if (!FS::Exists(path)) { + FS::CreateFullPath(path_fwd); + if (!FS::CreateDir(path)) { return nullptr; + } } // Cannot use make_shared as RealVfsDirectory constructor is private return std::shared_ptr<RealVfsDirectory>(new RealVfsDirectory(*this, path, perms)); @@ -158,67 +178,75 @@ VirtualDir RealVfsFilesystem::CreateDirectory(std::string_view path_, Mode perms VirtualDir RealVfsFilesystem::CopyDirectory(std::string_view old_path_, std::string_view new_path_) { - const auto old_path = - FileUtil::SanitizePath(old_path_, FileUtil::DirectorySeparator::PlatformDefault); - const auto new_path = - FileUtil::SanitizePath(new_path_, FileUtil::DirectorySeparator::PlatformDefault); - if (!FileUtil::Exists(old_path) || FileUtil::Exists(new_path) || - !FileUtil::IsDirectory(old_path)) + const auto old_path = FS::SanitizePath(old_path_, FS::DirectorySeparator::PlatformDefault); + const auto new_path = FS::SanitizePath(new_path_, FS::DirectorySeparator::PlatformDefault); + if (!FS::Exists(old_path) || FS::Exists(new_path) || !FS::IsDirectory(old_path)) { return nullptr; - FileUtil::CopyDir(old_path, new_path); + } + FS::CopyDir(old_path, new_path); return OpenDirectory(new_path, Mode::ReadWrite); } VirtualDir RealVfsFilesystem::MoveDirectory(std::string_view old_path_, std::string_view new_path_) { - const auto old_path = - FileUtil::SanitizePath(old_path_, FileUtil::DirectorySeparator::PlatformDefault); - const auto new_path = - FileUtil::SanitizePath(new_path_, FileUtil::DirectorySeparator::PlatformDefault); - if (!FileUtil::Exists(old_path) || FileUtil::Exists(new_path) || - FileUtil::IsDirectory(old_path) || !FileUtil::Rename(old_path, new_path)) + const auto old_path = FS::SanitizePath(old_path_, FS::DirectorySeparator::PlatformDefault); + const auto new_path = FS::SanitizePath(new_path_, FS::DirectorySeparator::PlatformDefault); + + if (!FS::Exists(old_path) || FS::Exists(new_path) || FS::IsDirectory(old_path) || + !FS::Rename(old_path, new_path)) { return nullptr; + } for (auto& kv : cache) { - // Path in cache starts with old_path - if (kv.first.rfind(old_path, 0) == 0) { - const auto file_old_path = - FileUtil::SanitizePath(kv.first, FileUtil::DirectorySeparator::PlatformDefault); - const auto file_new_path = - FileUtil::SanitizePath(new_path + DIR_SEP + kv.first.substr(old_path.size()), - FileUtil::DirectorySeparator::PlatformDefault); - auto cached = cache[file_old_path]; - if (!cached.expired()) { - auto file = cached.lock(); - file->Open(file_new_path, "r+b"); - cache.erase(file_old_path); - cache[file_new_path] = file; - } + // If the path in the cache doesn't start with old_path, then bail on this file. + if (kv.first.rfind(old_path, 0) != 0) { + continue; + } + + const auto file_old_path = + FS::SanitizePath(kv.first, FS::DirectorySeparator::PlatformDefault); + auto file_new_path = FS::SanitizePath(new_path + DIR_SEP + kv.first.substr(old_path.size()), + FS::DirectorySeparator::PlatformDefault); + const auto& cached = cache[file_old_path]; + + if (cached.expired()) { + continue; } + + auto file = cached.lock(); + file->Open(file_new_path, "r+b"); + cache.erase(file_old_path); + cache.insert_or_assign(std::move(file_new_path), std::move(file)); } return OpenDirectory(new_path, Mode::ReadWrite); } bool RealVfsFilesystem::DeleteDirectory(std::string_view path_) { - const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); + const auto path = FS::SanitizePath(path_, FS::DirectorySeparator::PlatformDefault); + for (auto& kv : cache) { - // Path in cache starts with old_path - if (kv.first.rfind(path, 0) == 0) { - if (!cache[kv.first].expired()) - cache[kv.first].lock()->Close(); - cache.erase(kv.first); + // If the path in the cache doesn't start with path, then bail on this file. + if (kv.first.rfind(path, 0) != 0) { + continue; + } + + const auto& entry = cache[kv.first]; + if (!entry.expired()) { + entry.lock()->Close(); } + + cache.erase(kv.first); } - return FileUtil::DeleteDirRecursively(path); + + return FS::DeleteDirRecursively(path); } -RealVfsFile::RealVfsFile(RealVfsFilesystem& base_, std::shared_ptr<FileUtil::IOFile> backing_, +RealVfsFile::RealVfsFile(RealVfsFilesystem& base_, std::shared_ptr<FS::IOFile> backing_, const std::string& path_, Mode perms_) - : base(base_), backing(std::move(backing_)), path(path_), - parent_path(FileUtil::GetParentPath(path_)), - path_components(FileUtil::SplitPathComponents(path_)), - parent_components(FileUtil::SliceVector(path_components, 0, path_components.size() - 1)), + : base(base_), backing(std::move(backing_)), path(path_), parent_path(FS::GetParentPath(path_)), + path_components(FS::SplitPathComponents(path_)), + parent_components(FS::SliceVector(path_components, 0, path_components.size() - 1)), perms(perms_) {} RealVfsFile::~RealVfsFile() = default; @@ -240,22 +268,24 @@ std::shared_ptr<VfsDirectory> RealVfsFile::GetContainingDirectory() const { } bool RealVfsFile::IsWritable() const { - return (perms & Mode::WriteAppend) != 0; + return True(perms & Mode::WriteAppend); } bool RealVfsFile::IsReadable() const { - return (perms & Mode::ReadWrite) != 0; + return True(perms & Mode::ReadWrite); } std::size_t RealVfsFile::Read(u8* data, std::size_t length, std::size_t offset) const { - if (!backing->Seek(offset, SEEK_SET)) + if (!backing->Seek(static_cast<s64>(offset), SEEK_SET)) { return 0; + } return backing->ReadBytes(data, length); } std::size_t RealVfsFile::Write(const u8* data, std::size_t length, std::size_t offset) { - if (!backing->Seek(offset, SEEK_SET)) + if (!backing->Seek(static_cast<s64>(offset), SEEK_SET)) { return 0; + } return backing->WriteBytes(data, length); } @@ -272,16 +302,18 @@ bool RealVfsFile::Close() { template <> std::vector<VirtualFile> RealVfsDirectory::IterateEntries<RealVfsFile, VfsFile>() const { - if (perms == Mode::Append) + if (perms == Mode::Append) { return {}; + } std::vector<VirtualFile> out; - FileUtil::ForeachDirectoryEntry( + FS::ForeachDirectoryEntry( nullptr, path, [&out, this](u64* entries_out, const std::string& directory, const std::string& filename) { const std::string full_path = directory + DIR_SEP + filename; - if (!FileUtil::IsDirectory(full_path)) + if (!FS::IsDirectory(full_path)) { out.emplace_back(base.OpenFile(full_path, perms)); + } return true; }); @@ -290,16 +322,18 @@ std::vector<VirtualFile> RealVfsDirectory::IterateEntries<RealVfsFile, VfsFile>( template <> std::vector<VirtualDir> RealVfsDirectory::IterateEntries<RealVfsDirectory, VfsDirectory>() const { - if (perms == Mode::Append) + if (perms == Mode::Append) { return {}; + } std::vector<VirtualDir> out; - FileUtil::ForeachDirectoryEntry( + FS::ForeachDirectoryEntry( nullptr, path, [&out, this](u64* entries_out, const std::string& directory, const std::string& filename) { const std::string full_path = directory + DIR_SEP + filename; - if (FileUtil::IsDirectory(full_path)) + if (FS::IsDirectory(full_path)) { out.emplace_back(base.OpenDirectory(full_path, perms)); + } return true; }); @@ -307,28 +341,30 @@ std::vector<VirtualDir> RealVfsDirectory::IterateEntries<RealVfsDirectory, VfsDi } RealVfsDirectory::RealVfsDirectory(RealVfsFilesystem& base_, const std::string& path_, Mode perms_) - : base(base_), path(FileUtil::RemoveTrailingSlash(path_)), - parent_path(FileUtil::GetParentPath(path)), - path_components(FileUtil::SplitPathComponents(path)), - parent_components(FileUtil::SliceVector(path_components, 0, path_components.size() - 1)), + : base(base_), path(FS::RemoveTrailingSlash(path_)), parent_path(FS::GetParentPath(path)), + path_components(FS::SplitPathComponents(path)), + parent_components(FS::SliceVector(path_components, 0, path_components.size() - 1)), perms(perms_) { - if (!FileUtil::Exists(path) && perms & Mode::WriteAppend) - FileUtil::CreateDir(path); + if (!FS::Exists(path) && True(perms & Mode::WriteAppend)) { + FS::CreateDir(path); + } } RealVfsDirectory::~RealVfsDirectory() = default; std::shared_ptr<VfsFile> RealVfsDirectory::GetFileRelative(std::string_view path) const { - const auto full_path = FileUtil::SanitizePath(this->path + DIR_SEP + std::string(path)); - if (!FileUtil::Exists(full_path) || FileUtil::IsDirectory(full_path)) + const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); + if (!FS::Exists(full_path) || FS::IsDirectory(full_path)) { return nullptr; + } return base.OpenFile(full_path, perms); } std::shared_ptr<VfsDirectory> RealVfsDirectory::GetDirectoryRelative(std::string_view path) const { - const auto full_path = FileUtil::SanitizePath(this->path + DIR_SEP + std::string(path)); - if (!FileUtil::Exists(full_path) || !FileUtil::IsDirectory(full_path)) + const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); + if (!FS::Exists(full_path) || !FS::IsDirectory(full_path)) { return nullptr; + } return base.OpenDirectory(full_path, perms); } @@ -341,17 +377,17 @@ std::shared_ptr<VfsDirectory> RealVfsDirectory::GetSubdirectory(std::string_view } std::shared_ptr<VfsFile> RealVfsDirectory::CreateFileRelative(std::string_view path) { - const auto full_path = FileUtil::SanitizePath(this->path + DIR_SEP + std::string(path)); + const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); return base.CreateFile(full_path, perms); } std::shared_ptr<VfsDirectory> RealVfsDirectory::CreateDirectoryRelative(std::string_view path) { - const auto full_path = FileUtil::SanitizePath(this->path + DIR_SEP + std::string(path)); + const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); return base.CreateDirectory(full_path, perms); } bool RealVfsDirectory::DeleteSubdirectoryRecursive(std::string_view name) { - auto full_path = FileUtil::SanitizePath(this->path + DIR_SEP + std::string(name)); + const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(name)); return base.DeleteDirectory(full_path); } @@ -364,11 +400,11 @@ std::vector<std::shared_ptr<VfsDirectory>> RealVfsDirectory::GetSubdirectories() } bool RealVfsDirectory::IsWritable() const { - return (perms & Mode::WriteAppend) != 0; + return True(perms & Mode::WriteAppend); } bool RealVfsDirectory::IsReadable() const { - return (perms & Mode::ReadWrite) != 0; + return True(perms & Mode::ReadWrite); } std::string RealVfsDirectory::GetName() const { @@ -376,8 +412,9 @@ std::string RealVfsDirectory::GetName() const { } std::shared_ptr<VfsDirectory> RealVfsDirectory::GetParentDirectory() const { - if (path_components.size() <= 1) + if (path_components.size() <= 1) { return nullptr; + } return base.OpenDirectory(parent_path, perms); } @@ -414,16 +451,17 @@ std::string RealVfsDirectory::GetFullPath() const { } std::map<std::string, VfsEntryType, std::less<>> RealVfsDirectory::GetEntries() const { - if (perms == Mode::Append) + if (perms == Mode::Append) { return {}; + } std::map<std::string, VfsEntryType, std::less<>> out; - FileUtil::ForeachDirectoryEntry( + FS::ForeachDirectoryEntry( nullptr, path, [&out](u64* entries_out, const std::string& directory, const std::string& filename) { const std::string full_path = directory + DIR_SEP + filename; - out.emplace(filename, FileUtil::IsDirectory(full_path) ? VfsEntryType::Directory - : VfsEntryType::File); + out.emplace(filename, + FS::IsDirectory(full_path) ? VfsEntryType::Directory : VfsEntryType::File); return true; }); diff --git a/src/core/file_sys/vfs_real.h b/src/core/file_sys/vfs_real.h index a0a857a31..0b537b22c 100644 --- a/src/core/file_sys/vfs_real.h +++ b/src/core/file_sys/vfs_real.h @@ -9,7 +9,7 @@ #include "core/file_sys/mode.h" #include "core/file_sys/vfs.h" -namespace FileUtil { +namespace Common::FS { class IOFile; } @@ -36,7 +36,7 @@ public: bool DeleteDirectory(std::string_view path) override; private: - boost::container::flat_map<std::string, std::weak_ptr<FileUtil::IOFile>> cache; + boost::container::flat_map<std::string, std::weak_ptr<Common::FS::IOFile>> cache; }; // An implmentation of VfsFile that represents a file on the user's computer. @@ -58,13 +58,13 @@ public: bool Rename(std::string_view name) override; private: - RealVfsFile(RealVfsFilesystem& base, std::shared_ptr<FileUtil::IOFile> backing, + RealVfsFile(RealVfsFilesystem& base, std::shared_ptr<Common::FS::IOFile> backing, const std::string& path, Mode perms = Mode::Read); bool Close(); RealVfsFilesystem& base; - std::shared_ptr<FileUtil::IOFile> backing; + std::shared_ptr<Common::FS::IOFile> backing; std::string path; std::string parent_path; std::vector<std::string> path_components; diff --git a/src/core/file_sys/vfs_static.h b/src/core/file_sys/vfs_static.h index 9f5a90b1b..8b27c30fa 100644 --- a/src/core/file_sys/vfs_static.h +++ b/src/core/file_sys/vfs_static.h @@ -54,9 +54,11 @@ public: } std::optional<u8> ReadByte(std::size_t offset) const override { - if (offset < size) - return value; - return {}; + if (offset >= size) { + return std::nullopt; + } + + return value; } std::vector<u8> ReadBytes(std::size_t length, std::size_t offset) const override { diff --git a/src/core/file_sys/vfs_vector.h b/src/core/file_sys/vfs_vector.h index ac36cb2ee..95d3da2f2 100644 --- a/src/core/file_sys/vfs_vector.h +++ b/src/core/file_sys/vfs_vector.h @@ -4,7 +4,11 @@ #pragma once +#include <array> #include <cstring> +#include <memory> +#include <string> +#include <vector> #include "core/file_sys/vfs.h" namespace FileSys { @@ -13,7 +17,8 @@ namespace FileSys { template <std::size_t size> class ArrayVfsFile : public VfsFile { public: - ArrayVfsFile(std::array<u8, size> data, std::string name = "", VirtualDir parent = nullptr) + explicit ArrayVfsFile(const std::array<u8, size>& data, std::string name = "", + VirtualDir parent = nullptr) : data(data), name(std::move(name)), parent(std::move(parent)) {} std::string GetName() const override { @@ -61,6 +66,12 @@ private: VirtualDir parent; }; +template <std::size_t Size, typename... Args> +std::shared_ptr<ArrayVfsFile<Size>> MakeArrayFile(const std::array<u8, Size>& data, + Args&&... args) { + return std::make_shared<ArrayVfsFile<Size>>(data, std::forward<Args>(args)...); +} + // An implementation of VfsFile that is backed by a vector optionally supplied upon construction class VectorVfsFile : public VfsFile { public: diff --git a/src/core/file_sys/xts_archive.cpp b/src/core/file_sys/xts_archive.cpp index 86e06ccb9..24c58e7ae 100644 --- a/src/core/file_sys/xts_archive.cpp +++ b/src/core/file_sys/xts_archive.cpp @@ -15,8 +15,9 @@ #include "common/hex_util.h" #include "common/string_util.h" #include "core/crypto/aes_util.h" +#include "core/crypto/key_manager.h" #include "core/crypto/xts_encryption_layer.h" -#include "core/file_sys/partition_filesystem.h" +#include "core/file_sys/content_archive.h" #include "core/file_sys/vfs_offset.h" #include "core/file_sys/xts_archive.h" #include "core/loader/loader.h" @@ -43,8 +44,10 @@ static bool CalculateHMAC256(Destination* out, const SourceKey* key, std::size_t return true; } -NAX::NAX(VirtualFile file_) : header(std::make_unique<NAXHeader>()), file(std::move(file_)) { - std::string path = FileUtil::SanitizePath(file->GetFullPath()); +NAX::NAX(VirtualFile file_) + : header(std::make_unique<NAXHeader>()), + file(std::move(file_)), keys{Core::Crypto::KeyManager::Instance()} { + std::string path = Common::FS::SanitizePath(file->GetFullPath()); static const std::regex nax_path_regex("/registered/(000000[0-9A-F]{2})/([0-9A-F]{32})\\.nca", std::regex_constants::ECMAScript | std::regex_constants::icase); @@ -60,7 +63,8 @@ NAX::NAX(VirtualFile file_) : header(std::make_unique<NAXHeader>()), file(std::m } NAX::NAX(VirtualFile file_, std::array<u8, 0x10> nca_id) - : header(std::make_unique<NAXHeader>()), file(std::move(file_)) { + : header(std::make_unique<NAXHeader>()), + file(std::move(file_)), keys{Core::Crypto::KeyManager::Instance()} { Core::Crypto::SHA256Hash hash{}; mbedtls_sha256_ret(nca_id.data(), nca_id.size(), hash.data(), 0); status = Parse(fmt::format("/registered/000000{:02X}/{}.nca", hash[0], @@ -70,14 +74,18 @@ NAX::NAX(VirtualFile file_, std::array<u8, 0x10> nca_id) NAX::~NAX() = default; Loader::ResultStatus NAX::Parse(std::string_view path) { - if (file->ReadObject(header.get()) != sizeof(NAXHeader)) + if (file == nullptr) { + return Loader::ResultStatus::ErrorNullFile; + } + if (file->ReadObject(header.get()) != sizeof(NAXHeader)) { return Loader::ResultStatus::ErrorBadNAXHeader; - - if (header->magic != Common::MakeMagic('N', 'A', 'X', '0')) + } + if (header->magic != Common::MakeMagic('N', 'A', 'X', '0')) { return Loader::ResultStatus::ErrorBadNAXHeader; - - if (file->GetSize() < NAX_HEADER_PADDING_SIZE + header->file_size) + } + if (file->GetSize() < NAX_HEADER_PADDING_SIZE + header->file_size) { return Loader::ResultStatus::ErrorIncorrectNAXFileSize; + } keys.DeriveSDSeedLazy(); std::array<Core::Crypto::Key256, 2> sd_keys{}; diff --git a/src/core/file_sys/xts_archive.h b/src/core/file_sys/xts_archive.h index 7704dee90..c472e226e 100644 --- a/src/core/file_sys/xts_archive.h +++ b/src/core/file_sys/xts_archive.h @@ -9,12 +9,16 @@ #include "common/common_types.h" #include "common/swap.h" #include "core/crypto/key_manager.h" -#include "core/file_sys/content_archive.h" #include "core/file_sys/vfs.h" -#include "core/loader/loader.h" + +namespace Loader { +enum class ResultStatus : u16; +} namespace FileSys { +class NCA; + struct NAXHeader { std::array<u8, 0x20> hmac; u64_le magic; @@ -62,6 +66,6 @@ private: VirtualFile dec_file; - Core::Crypto::KeyManager keys; + Core::Crypto::KeyManager& keys; }; } // namespace FileSys diff --git a/src/core/frontend/applets/controller.cpp b/src/core/frontend/applets/controller.cpp new file mode 100644 index 000000000..03bbedf8b --- /dev/null +++ b/src/core/frontend/applets/controller.cpp @@ -0,0 +1,81 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "common/logging/log.h" +#include "core/frontend/applets/controller.h" +#include "core/hle/service/hid/controllers/npad.h" +#include "core/hle/service/hid/hid.h" +#include "core/hle/service/sm/sm.h" + +namespace Core::Frontend { + +ControllerApplet::~ControllerApplet() = default; + +DefaultControllerApplet::DefaultControllerApplet(Service::SM::ServiceManager& service_manager_) + : service_manager{service_manager_} {} + +DefaultControllerApplet::~DefaultControllerApplet() = default; + +void DefaultControllerApplet::ReconfigureControllers(std::function<void()> callback, + const ControllerParameters& parameters) const { + LOG_INFO(Service_HID, "called, deducing the best configuration based on the given parameters!"); + + auto& npad = + service_manager.GetService<Service::HID::Hid>("hid") + ->GetAppletResource() + ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad); + + auto& players = Settings::values.players.GetValue(); + + const std::size_t min_supported_players = + parameters.enable_single_mode ? 1 : parameters.min_players; + + // Disconnect Handheld first. + npad.DisconnectNpadAtIndex(8); + + // Deduce the best configuration based on the input parameters. + for (std::size_t index = 0; index < players.size() - 2; ++index) { + // First, disconnect all controllers regardless of the value of keep_controllers_connected. + // This makes it easy to connect the desired controllers. + npad.DisconnectNpadAtIndex(index); + + // Only connect the minimum number of required players. + if (index >= min_supported_players) { + continue; + } + + // Connect controllers based on the following priority list from highest to lowest priority: + // Pro Controller -> Dual Joycons -> Left Joycon/Right Joycon -> Handheld + if (parameters.allow_pro_controller) { + npad.AddNewControllerAt( + npad.MapSettingsTypeToNPad(Settings::ControllerType::ProController), index); + } else if (parameters.allow_dual_joycons) { + npad.AddNewControllerAt( + npad.MapSettingsTypeToNPad(Settings::ControllerType::DualJoyconDetached), index); + } else if (parameters.allow_left_joycon && parameters.allow_right_joycon) { + // Assign left joycons to even player indices and right joycons to odd player indices. + // We do this since Captain Toad Treasure Tracker expects a left joycon for Player 1 and + // a right Joycon for Player 2 in 2 Player Assist mode. + if (index % 2 == 0) { + npad.AddNewControllerAt( + npad.MapSettingsTypeToNPad(Settings::ControllerType::LeftJoycon), index); + } else { + npad.AddNewControllerAt( + npad.MapSettingsTypeToNPad(Settings::ControllerType::RightJoycon), index); + } + } else if (index == 0 && parameters.enable_single_mode && parameters.allow_handheld && + !Settings::values.use_docked_mode.GetValue()) { + // We should *never* reach here under any normal circumstances. + npad.AddNewControllerAt(npad.MapSettingsTypeToNPad(Settings::ControllerType::Handheld), + index); + } else { + UNREACHABLE_MSG("Unable to add a new controller based on the given parameters!"); + } + } + + callback(); +} + +} // namespace Core::Frontend diff --git a/src/core/frontend/applets/controller.h b/src/core/frontend/applets/controller.h new file mode 100644 index 000000000..dff71d8d9 --- /dev/null +++ b/src/core/frontend/applets/controller.h @@ -0,0 +1,56 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <functional> + +#include "common/common_types.h" + +namespace Service::SM { +class ServiceManager; +} + +namespace Core::Frontend { + +using BorderColor = std::array<u8, 4>; +using ExplainText = std::array<char, 0x81>; + +struct ControllerParameters { + s8 min_players{}; + s8 max_players{}; + bool keep_controllers_connected{}; + bool enable_single_mode{}; + bool enable_border_color{}; + std::vector<BorderColor> border_colors{}; + bool enable_explain_text{}; + std::vector<ExplainText> explain_text{}; + bool allow_pro_controller{}; + bool allow_handheld{}; + bool allow_dual_joycons{}; + bool allow_left_joycon{}; + bool allow_right_joycon{}; +}; + +class ControllerApplet { +public: + virtual ~ControllerApplet(); + + virtual void ReconfigureControllers(std::function<void()> callback, + const ControllerParameters& parameters) const = 0; +}; + +class DefaultControllerApplet final : public ControllerApplet { +public: + explicit DefaultControllerApplet(Service::SM::ServiceManager& service_manager_); + ~DefaultControllerApplet() override; + + void ReconfigureControllers(std::function<void()> callback, + const ControllerParameters& parameters) const override; + +private: + Service::SM::ServiceManager& service_manager; +}; + +} // namespace Core::Frontend diff --git a/src/core/frontend/emu_window.cpp b/src/core/frontend/emu_window.cpp index eda466a5d..8c1193894 100644 --- a/src/core/frontend/emu_window.cpp +++ b/src/core/frontend/emu_window.cpp @@ -46,7 +46,7 @@ private: EmuWindow::EmuWindow() { // TODO: Find a better place to set this. config.min_client_area_size = - std::make_pair(Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height); + std::make_pair(Layout::MinimumSize::Width, Layout::MinimumSize::Height); active_config = config; touch_state = std::make_shared<TouchState>(); Input::RegisterFactory<Input::TouchDevice>("emu_window", touch_state); @@ -84,10 +84,12 @@ void EmuWindow::TouchPressed(unsigned framebuffer_x, unsigned framebuffer_y) { return; std::lock_guard guard{touch_state->mutex}; - touch_state->touch_x = static_cast<float>(framebuffer_x - framebuffer_layout.screen.left) / - (framebuffer_layout.screen.right - framebuffer_layout.screen.left); - touch_state->touch_y = static_cast<float>(framebuffer_y - framebuffer_layout.screen.top) / - (framebuffer_layout.screen.bottom - framebuffer_layout.screen.top); + touch_state->touch_x = + static_cast<float>(framebuffer_x - framebuffer_layout.screen.left) / + static_cast<float>(framebuffer_layout.screen.right - framebuffer_layout.screen.left); + touch_state->touch_y = + static_cast<float>(framebuffer_y - framebuffer_layout.screen.top) / + static_cast<float>(framebuffer_layout.screen.bottom - framebuffer_layout.screen.top); touch_state->touch_pressed = true; } diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h index 13aa14934..3e8780243 100644 --- a/src/core/frontend/emu_window.h +++ b/src/core/frontend/emu_window.h @@ -39,7 +39,7 @@ public: class Scoped { public: - explicit Scoped(GraphicsContext& context_) : context(context_) { + [[nodiscard]] explicit Scoped(GraphicsContext& context_) : context(context_) { context.MakeCurrent(); } ~Scoped() { @@ -52,7 +52,7 @@ public: /// Calls MakeCurrent on the context and calls DoneCurrent when the scope for the returned value /// ends - Scoped Acquire() { + [[nodiscard]] Scoped Acquire() { return Scoped{*this}; } }; diff --git a/src/core/frontend/framebuffer_layout.cpp b/src/core/frontend/framebuffer_layout.cpp index 68a0e0906..b9a270a55 100644 --- a/src/core/frontend/framebuffer_layout.cpp +++ b/src/core/frontend/framebuffer_layout.cpp @@ -14,8 +14,8 @@ namespace Layout { template <class T> static Common::Rectangle<T> MaxRectangle(Common::Rectangle<T> window_area, float screen_aspect_ratio) { - float scale = std::min(static_cast<float>(window_area.GetWidth()), - window_area.GetHeight() / screen_aspect_ratio); + const float scale = std::min(static_cast<float>(window_area.GetWidth()), + static_cast<float>(window_area.GetHeight()) / screen_aspect_ratio); return Common::Rectangle<T>{0, 0, static_cast<T>(std::round(scale)), static_cast<T>(std::round(scale * screen_aspect_ratio))}; } @@ -25,11 +25,11 @@ FramebufferLayout DefaultFrameLayout(u32 width, u32 height) { ASSERT(height > 0); // The drawing code needs at least somewhat valid values for both screens // so just calculate them both even if the other isn't showing. - FramebufferLayout res{width, height}; + FramebufferLayout res{width, height, false, {}}; - const float window_aspect_ratio = static_cast<float>(height) / width; + const float window_aspect_ratio = static_cast<float>(height) / static_cast<float>(width); const float emulation_aspect_ratio = EmulationAspectRatio( - static_cast<AspectRatio>(Settings::values.aspect_ratio), window_aspect_ratio); + static_cast<AspectRatio>(Settings::values.aspect_ratio.GetValue()), window_aspect_ratio); const Common::Rectangle<u32> screen_window_area{0, 0, width, height}; Common::Rectangle<u32> screen = MaxRectangle(screen_window_area, emulation_aspect_ratio); @@ -47,7 +47,7 @@ FramebufferLayout DefaultFrameLayout(u32 width, u32 height) { FramebufferLayout FrameLayoutFromResolutionScale(u32 res_scale) { u32 width, height; - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { width = ScreenDocked::Width * res_scale; height = ScreenDocked::Height * res_scale; } else { diff --git a/src/core/frontend/framebuffer_layout.h b/src/core/frontend/framebuffer_layout.h index 15ecfb13d..e2e3bbbb3 100644 --- a/src/core/frontend/framebuffer_layout.h +++ b/src/core/frontend/framebuffer_layout.h @@ -4,10 +4,16 @@ #pragma once +#include "common/common_types.h" #include "common/math_util.h" namespace Layout { +namespace MinimumSize { +constexpr u32 Width = 640; +constexpr u32 Height = 360; +} // namespace MinimumSize + namespace ScreenUndocked { constexpr u32 Width = 1280; constexpr u32 Height = 720; diff --git a/src/core/frontend/input.h b/src/core/frontend/input.h index 2b098b7c6..11c2e96ca 100644 --- a/src/core/frontend/input.h +++ b/src/core/frontend/input.h @@ -30,7 +30,12 @@ public: virtual StatusType GetStatus() const { return {}; } - virtual bool GetAnalogDirectionStatus(AnalogDirection direction) const { + virtual bool GetAnalogDirectionStatus([[maybe_unused]] AnalogDirection direction) const { + return {}; + } + virtual bool SetRumblePlay([[maybe_unused]] f32 amp_low, [[maybe_unused]] f32 freq_low, + [[maybe_unused]] f32 amp_high, + [[maybe_unused]] f32 freq_high) const { return {}; } }; @@ -119,11 +124,18 @@ using ButtonDevice = InputDevice<bool>; using AnalogDevice = InputDevice<std::tuple<float, float>>; /** - * A motion device is an input device that returns a tuple of accelerometer state vector and - * gyroscope state vector. + * A vibration device is an input device that returns an unsigned byte as status. + * It represents whether the vibration device supports vibration or not. + * If the status returns 1, it supports vibration. Otherwise, it does not support vibration. + */ +using VibrationDevice = InputDevice<u8>; + +/** + * A motion status is an object that returns a tuple of accelerometer state vector, + * gyroscope state vector, rotation state vector and orientation state matrix. * * For both vectors: - * x+ is the same direction as LEFT on D-pad. + * x+ is the same direction as RIGHT on D-pad. * y+ is normal to the touch screen, pointing outward. * z+ is the same direction as UP on D-pad. * @@ -133,8 +145,22 @@ using AnalogDevice = InputDevice<std::tuple<float, float>>; * For gyroscope state vector: * Orientation is determined by right-hand rule. * Units: deg/sec + * + * For rotation state vector + * Units: rotations + * + * For orientation state matrix + * x vector + * y vector + * z vector + */ +using MotionStatus = std::tuple<Common::Vec3<float>, Common::Vec3<float>, Common::Vec3<float>, + std::array<Common::Vec3f, 3>>; + +/** + * A motion device is an input device that returns a motion status object */ -using MotionDevice = InputDevice<std::tuple<Common::Vec3<float>, Common::Vec3<float>>>; +using MotionDevice = InputDevice<MotionStatus>; /** * A touch device is an input device that returns a tuple of two floats and a bool. The floats are diff --git a/src/core/gdbstub/gdbstub.cpp b/src/core/gdbstub/gdbstub.cpp index 6d15aeed9..97ee65464 100644 --- a/src/core/gdbstub/gdbstub.cpp +++ b/src/core/gdbstub/gdbstub.cpp @@ -35,11 +35,10 @@ #include "common/swap.h" #include "core/arm/arm_interface.h" #include "core/core.h" -#include "core/core_manager.h" #include "core/gdbstub/gdbstub.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/vm_manager.h" #include "core/loader/loader.h" #include "core/memory.h" @@ -292,11 +291,11 @@ static void FpuWrite(std::size_t id, u128 val, Kernel::Thread* thread = nullptr) */ static u8 HexCharToValue(u8 hex) { if (hex >= '0' && hex <= '9') { - return hex - '0'; + return static_cast<u8>(hex - '0'); } else if (hex >= 'a' && hex <= 'f') { - return hex - 'a' + 0xA; + return static_cast<u8>(hex - 'a' + 0xA); } else if (hex >= 'A' && hex <= 'F') { - return hex - 'A' + 0xA; + return static_cast<u8>(hex - 'A' + 0xA); } LOG_ERROR(Debug_GDBStub, "Invalid nibble: {} ({:02X})", hex, hex); @@ -311,9 +310,9 @@ static u8 HexCharToValue(u8 hex) { static u8 NibbleToHex(u8 n) { n &= 0xF; if (n < 0xA) { - return '0' + n; + return static_cast<u8>('0' + n); } else { - return 'a' + n - 0xA; + return static_cast<u8>('a' + n - 0xA); } } @@ -356,8 +355,8 @@ static u64 HexToLong(const u8* src, std::size_t len) { */ static void MemToGdbHex(u8* dest, const u8* src, std::size_t len) { while (len-- > 0) { - u8 tmp = *src++; - *dest++ = NibbleToHex(tmp >> 4); + const u8 tmp = *src++; + *dest++ = NibbleToHex(static_cast<u8>(tmp >> 4)); *dest++ = NibbleToHex(tmp); } } @@ -371,7 +370,7 @@ static void MemToGdbHex(u8* dest, const u8* src, std::size_t len) { */ static void GdbHexToMem(u8* dest, const u8* src, std::size_t len) { while (len-- > 0) { - *dest++ = (HexCharToValue(src[0]) << 4) | HexCharToValue(src[1]); + *dest++ = static_cast<u8>((HexCharToValue(src[0]) << 4) | HexCharToValue(src[1])); src += 2; } } @@ -603,22 +602,22 @@ static void SendReply(const char* reply) { memcpy(command_buffer + 1, reply, command_length); - u8 checksum = CalculateChecksum(command_buffer, command_length + 1); + const u8 checksum = CalculateChecksum(command_buffer, command_length + 1); command_buffer[0] = GDB_STUB_START; command_buffer[command_length + 1] = GDB_STUB_END; - command_buffer[command_length + 2] = NibbleToHex(checksum >> 4); + command_buffer[command_length + 2] = NibbleToHex(static_cast<u8>(checksum >> 4)); command_buffer[command_length + 3] = NibbleToHex(checksum); u8* ptr = command_buffer; u32 left = command_length + 4; while (left > 0) { - int sent_size = send(gdbserver_socket, reinterpret_cast<char*>(ptr), left, 0); + const auto sent_size = send(gdbserver_socket, reinterpret_cast<char*>(ptr), left, 0); if (sent_size < 0) { LOG_ERROR(Debug_GDBStub, "gdb: send failed"); return Shutdown(); } - left -= sent_size; + left -= static_cast<u32>(sent_size); ptr += sent_size; } } @@ -643,7 +642,7 @@ static void HandleQuery() { SendReply(target_xml); } else if (strncmp(query, "Offsets", strlen("Offsets")) == 0) { const VAddr base_address = - Core::System::GetInstance().CurrentProcess()->VMManager().GetCodeRegionBaseAddress(); + Core::System::GetInstance().CurrentProcess()->PageTable().GetCodeRegionStart(); std::string buffer = fmt::format("TextSeg={:0x}", base_address); SendReply(buffer.c_str()); } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) { @@ -778,10 +777,10 @@ static void ReadCommand() { command_buffer[command_length++] = c; } - u8 checksum_received = HexCharToValue(ReadByte()) << 4; - checksum_received |= HexCharToValue(ReadByte()); + auto checksum_received = static_cast<u32>(HexCharToValue(ReadByte()) << 4); + checksum_received |= static_cast<u32>(HexCharToValue(ReadByte())); - u8 checksum_calculated = CalculateChecksum(command_buffer, command_length); + const u32 checksum_calculated = CalculateChecksum(command_buffer, command_length); if (checksum_received != checksum_calculated) { LOG_ERROR(Debug_GDBStub, @@ -1389,10 +1388,9 @@ void SendTrap(Kernel::Thread* thread, int trap) { return; } - if (!halt_loop || current_thread == thread) { - current_thread = thread; - SendSignal(thread, trap); - } + current_thread = thread; + SendSignal(thread, trap); + halt_loop = true; send_trap = false; } diff --git a/src/core/hardware_interrupt_manager.cpp b/src/core/hardware_interrupt_manager.cpp index c629d9fa1..645f26e91 100644 --- a/src/core/hardware_interrupt_manager.cpp +++ b/src/core/hardware_interrupt_manager.cpp @@ -11,19 +11,20 @@ namespace Core::Hardware { InterruptManager::InterruptManager(Core::System& system_in) : system(system_in) { - gpu_interrupt_event = Core::Timing::CreateEvent("GPUInterrupt", [this](u64 message, s64) { - auto nvdrv = system.ServiceManager().GetService<Service::Nvidia::NVDRV>("nvdrv"); - const u32 syncpt = static_cast<u32>(message >> 32); - const u32 value = static_cast<u32>(message); - nvdrv->SignalGPUInterruptSyncpt(syncpt, value); - }); + gpu_interrupt_event = Core::Timing::CreateEvent( + "GPUInterrupt", [this](std::uintptr_t message, std::chrono::nanoseconds) { + auto nvdrv = system.ServiceManager().GetService<Service::Nvidia::NVDRV>("nvdrv"); + const u32 syncpt = static_cast<u32>(message >> 32); + const u32 value = static_cast<u32>(message); + nvdrv->SignalGPUInterruptSyncpt(syncpt, value); + }); } InterruptManager::~InterruptManager() = default; void InterruptManager::GPUInterruptSyncpt(const u32 syncpoint_id, const u32 value) { const u64 msg = (static_cast<u64>(syncpoint_id) << 32ULL) | value; - system.CoreTiming().ScheduleEvent(10, gpu_interrupt_event, msg); + system.CoreTiming().ScheduleEvent(std::chrono::nanoseconds{10}, gpu_interrupt_event, msg); } } // namespace Core::Hardware diff --git a/src/core/hardware_properties.h b/src/core/hardware_properties.h index b04e046ed..456b41e1b 100644 --- a/src/core/hardware_properties.h +++ b/src/core/hardware_properties.h @@ -42,6 +42,10 @@ struct EmuThreadHandle { constexpr u32 invalid_handle = 0xFFFFFFFF; return {invalid_handle, invalid_handle}; } + + bool IsInvalid() const { + return (*this) == InvalidHandle(); + } }; } // namespace Core diff --git a/src/core/hle/ipc_helpers.h b/src/core/hle/ipc_helpers.h index 0dc6a4a43..d57776ce9 100644 --- a/src/core/hle/ipc_helpers.h +++ b/src/core/hle/ipc_helpers.h @@ -12,7 +12,6 @@ #include <utility> #include "common/assert.h" #include "common/common_types.h" -#include "core/core.h" #include "core/hle/ipc.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" @@ -38,10 +37,11 @@ public: explicit RequestHelperBase(Kernel::HLERequestContext& context) : context(&context), cmdbuf(context.CommandBuffer()) {} - void Skip(unsigned size_in_words, bool set_to_null) { - if (set_to_null) + void Skip(u32 size_in_words, bool set_to_null) { + if (set_to_null) { memset(cmdbuf + index, 0, size_in_words * sizeof(u32)); - index += size_in_words; + } + index += static_cast<ptrdiff_t>(size_in_words); } /** @@ -49,15 +49,15 @@ public: */ void AlignWithPadding() { if (index & 3) { - Skip(4 - (index & 3), true); + Skip(static_cast<u32>(4 - (index & 3)), true); } } - unsigned GetCurrentOffset() const { - return static_cast<unsigned>(index); + u32 GetCurrentOffset() const { + return static_cast<u32>(index); } - void SetCurrentOffset(unsigned offset) { + void SetCurrentOffset(u32 offset) { index = static_cast<ptrdiff_t>(offset); } }; @@ -72,14 +72,12 @@ public: AlwaysMoveHandles = 1, }; - explicit ResponseBuilder(u32* command_buffer) : RequestHelperBase(command_buffer) {} - explicit ResponseBuilder(Kernel::HLERequestContext& context, u32 normal_params_size, u32 num_handles_to_copy = 0, u32 num_objects_to_move = 0, Flags flags = Flags::None) - : RequestHelperBase(context), normal_params_size(normal_params_size), - num_handles_to_copy(num_handles_to_copy), num_objects_to_move(num_objects_to_move) { + num_handles_to_copy(num_handles_to_copy), + num_objects_to_move(num_objects_to_move), kernel{context.kernel} { memset(cmdbuf, 0, sizeof(u32) * IPC::COMMAND_BUFFER_LENGTH); @@ -89,7 +87,7 @@ public: // The entire size of the raw data section in u32 units, including the 16 bytes of mandatory // padding. - u32 raw_data_size = sizeof(IPC::DataPayloadHeader) / 4 + 4 + normal_params_size; + u64 raw_data_size = sizeof(IPC::DataPayloadHeader) / 4 + 4 + normal_params_size; u32 num_handles_to_move{}; u32 num_domain_objects{}; @@ -105,7 +103,7 @@ public: raw_data_size += sizeof(DomainMessageHeader) / 4 + num_domain_objects; } - header.data_size.Assign(raw_data_size); + header.data_size.Assign(static_cast<u32>(raw_data_size)); if (num_handles_to_copy || num_handles_to_move) { header.enable_handle_descriptor.Assign(1); } @@ -139,7 +137,6 @@ public: if (context->Session()->IsDomain()) { context->AddDomainObject(std::move(iface)); } else { - auto& kernel = Core::System::GetInstance().Kernel(); auto [client, server] = Kernel::Session::Create(kernel, iface->GetServiceName()); context->AddMoveObject(std::move(client)); iface->ClientConnected(std::move(server)); @@ -213,6 +210,7 @@ private: u32 num_handles_to_copy{}; u32 num_objects_to_move{}; ///< Domain objects or move handles, context dependent std::ptrdiff_t datapayload_index{}; + Kernel::KernelCore& kernel; }; /// Push /// @@ -229,6 +227,8 @@ inline void ResponseBuilder::Push(u32 value) { template <typename T> void ResponseBuilder::PushRaw(const T& value) { + static_assert(std::is_trivially_copyable_v<T>, + "It's undefined behavior to use memcpy with non-trivially copyable objects"); std::memcpy(cmdbuf + index, &value, sizeof(T)); index += (sizeof(T) + 3) / 4; // round up to word length } @@ -384,6 +384,8 @@ inline s32 RequestParser::Pop() { template <typename T> void RequestParser::PopRaw(T& value) { + static_assert(std::is_trivially_copyable_v<T>, + "It's undefined behavior to use memcpy with non-trivially copyable objects"); std::memcpy(&value, cmdbuf + index, sizeof(T)); index += (sizeof(T) + 3) / 4; // round up to word length } diff --git a/src/core/hle/kernel/address_arbiter.cpp b/src/core/hle/kernel/address_arbiter.cpp index 8475b698c..b882eaa0f 100644 --- a/src/core/hle/kernel/address_arbiter.cpp +++ b/src/core/hle/kernel/address_arbiter.cpp @@ -7,11 +7,15 @@ #include "common/assert.h" #include "common/common_types.h" +#include "core/arm/exclusive_monitor.h" #include "core/core.h" #include "core/hle/kernel/address_arbiter.h" #include "core/hle/kernel/errors.h" +#include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" #include "core/hle/result.h" #include "core/memory.h" @@ -29,12 +33,10 @@ void AddressArbiter::WakeThreads(const std::vector<std::shared_ptr<Thread>>& wai // Signal the waiting threads. for (std::size_t i = 0; i < last; i++) { - ASSERT(waiting_threads[i]->GetStatus() == ThreadStatus::WaitArb); - waiting_threads[i]->SetWaitSynchronizationResult(RESULT_SUCCESS); + waiting_threads[i]->SetSynchronizationResults(nullptr, RESULT_SUCCESS); RemoveThread(waiting_threads[i]); - waiting_threads[i]->SetArbiterWaitAddress(0); + waiting_threads[i]->WaitForArbitration(false); waiting_threads[i]->ResumeFromWait(); - system.PrepareReschedule(waiting_threads[i]->GetProcessorID()); } } @@ -56,6 +58,7 @@ ResultCode AddressArbiter::SignalToAddress(VAddr address, SignalType type, s32 v } ResultCode AddressArbiter::SignalToAddressOnly(VAddr address, s32 num_to_wake) { + SchedulerLock lock(system.Kernel()); const std::vector<std::shared_ptr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address); WakeThreads(waiting_threads, num_to_wake); @@ -64,6 +67,7 @@ ResultCode AddressArbiter::SignalToAddressOnly(VAddr address, s32 num_to_wake) { ResultCode AddressArbiter::IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake) { + SchedulerLock lock(system.Kernel()); auto& memory = system.Memory(); // Ensure that we can write to the address. @@ -71,16 +75,24 @@ ResultCode AddressArbiter::IncrementAndSignalToAddressIfEqual(VAddr address, s32 return ERR_INVALID_ADDRESS_STATE; } - if (static_cast<s32>(memory.Read32(address)) != value) { - return ERR_INVALID_STATE; - } + const std::size_t current_core = system.CurrentCoreIndex(); + auto& monitor = system.Monitor(); + u32 current_value; + do { + current_value = monitor.ExclusiveRead32(current_core, address); + + if (current_value != static_cast<u32>(value)) { + return ERR_INVALID_STATE; + } + current_value++; + } while (!monitor.ExclusiveWrite32(current_core, address, current_value)); - memory.Write32(address, static_cast<u32>(value + 1)); return SignalToAddressOnly(address, num_to_wake); } ResultCode AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake) { + SchedulerLock lock(system.Kernel()); auto& memory = system.Memory(); // Ensure that we can write to the address. @@ -92,29 +104,33 @@ ResultCode AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr a const std::vector<std::shared_ptr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address); - // Determine the modified value depending on the waiting count. + const std::size_t current_core = system.CurrentCoreIndex(); + auto& monitor = system.Monitor(); s32 updated_value; - if (num_to_wake <= 0) { - if (waiting_threads.empty()) { - updated_value = value + 1; - } else { - updated_value = value - 1; + do { + updated_value = monitor.ExclusiveRead32(current_core, address); + + if (updated_value != value) { + return ERR_INVALID_STATE; } - } else { - if (waiting_threads.empty()) { - updated_value = value + 1; - } else if (waiting_threads.size() <= static_cast<u32>(num_to_wake)) { - updated_value = value - 1; + // Determine the modified value depending on the waiting count. + if (num_to_wake <= 0) { + if (waiting_threads.empty()) { + updated_value = value + 1; + } else { + updated_value = value - 1; + } } else { - updated_value = value; + if (waiting_threads.empty()) { + updated_value = value + 1; + } else if (waiting_threads.size() <= static_cast<u32>(num_to_wake)) { + updated_value = value - 1; + } else { + updated_value = value; + } } - } + } while (!monitor.ExclusiveWrite32(current_core, address, updated_value)); - if (static_cast<s32>(memory.Read32(address)) != value) { - return ERR_INVALID_STATE; - } - - memory.Write32(address, static_cast<u32>(updated_value)); WakeThreads(waiting_threads, num_to_wake); return RESULT_SUCCESS; } @@ -136,60 +152,127 @@ ResultCode AddressArbiter::WaitForAddress(VAddr address, ArbitrationType type, s ResultCode AddressArbiter::WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout, bool should_decrement) { auto& memory = system.Memory(); + auto& kernel = system.Kernel(); + Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); - // Ensure that we can read the address. - if (!memory.IsValidVirtualAddress(address)) { - return ERR_INVALID_ADDRESS_STATE; - } + Handle event_handle = InvalidHandle; + { + SchedulerLockAndSleep lock(kernel, event_handle, current_thread, timeout); + + if (current_thread->IsPendingTermination()) { + lock.CancelSleep(); + return ERR_THREAD_TERMINATING; + } + + // Ensure that we can read the address. + if (!memory.IsValidVirtualAddress(address)) { + lock.CancelSleep(); + return ERR_INVALID_ADDRESS_STATE; + } + + s32 current_value = static_cast<s32>(memory.Read32(address)); + if (current_value >= value) { + lock.CancelSleep(); + return ERR_INVALID_STATE; + } + + current_thread->SetSynchronizationResults(nullptr, RESULT_TIMEOUT); + + s32 decrement_value; + + const std::size_t current_core = system.CurrentCoreIndex(); + auto& monitor = system.Monitor(); + do { + current_value = static_cast<s32>(monitor.ExclusiveRead32(current_core, address)); + if (should_decrement) { + decrement_value = current_value - 1; + } else { + decrement_value = current_value; + } + } while ( + !monitor.ExclusiveWrite32(current_core, address, static_cast<u32>(decrement_value))); + + // Short-circuit without rescheduling, if timeout is zero. + if (timeout == 0) { + lock.CancelSleep(); + return RESULT_TIMEOUT; + } - const s32 cur_value = static_cast<s32>(memory.Read32(address)); - if (cur_value >= value) { - return ERR_INVALID_STATE; + current_thread->SetArbiterWaitAddress(address); + InsertThread(SharedFrom(current_thread)); + current_thread->SetStatus(ThreadStatus::WaitArb); + current_thread->WaitForArbitration(true); } - if (should_decrement) { - memory.Write32(address, static_cast<u32>(cur_value - 1)); + if (event_handle != InvalidHandle) { + auto& time_manager = kernel.TimeManager(); + time_manager.UnscheduleTimeEvent(event_handle); } - // Short-circuit without rescheduling, if timeout is zero. - if (timeout == 0) { - return RESULT_TIMEOUT; + { + SchedulerLock lock(kernel); + if (current_thread->IsWaitingForArbitration()) { + RemoveThread(SharedFrom(current_thread)); + current_thread->WaitForArbitration(false); + } } - return WaitForAddressImpl(address, timeout); + return current_thread->GetSignalingResult(); } ResultCode AddressArbiter::WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout) { auto& memory = system.Memory(); + auto& kernel = system.Kernel(); + Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); - // Ensure that we can read the address. - if (!memory.IsValidVirtualAddress(address)) { - return ERR_INVALID_ADDRESS_STATE; - } + Handle event_handle = InvalidHandle; + { + SchedulerLockAndSleep lock(kernel, event_handle, current_thread, timeout); + + if (current_thread->IsPendingTermination()) { + lock.CancelSleep(); + return ERR_THREAD_TERMINATING; + } + + // Ensure that we can read the address. + if (!memory.IsValidVirtualAddress(address)) { + lock.CancelSleep(); + return ERR_INVALID_ADDRESS_STATE; + } - // Only wait for the address if equal. - if (static_cast<s32>(memory.Read32(address)) != value) { - return ERR_INVALID_STATE; + s32 current_value = static_cast<s32>(memory.Read32(address)); + if (current_value != value) { + lock.CancelSleep(); + return ERR_INVALID_STATE; + } + + // Short-circuit without rescheduling, if timeout is zero. + if (timeout == 0) { + lock.CancelSleep(); + return RESULT_TIMEOUT; + } + + current_thread->SetSynchronizationResults(nullptr, RESULT_TIMEOUT); + current_thread->SetArbiterWaitAddress(address); + InsertThread(SharedFrom(current_thread)); + current_thread->SetStatus(ThreadStatus::WaitArb); + current_thread->WaitForArbitration(true); } - // Short-circuit without rescheduling if timeout is zero. - if (timeout == 0) { - return RESULT_TIMEOUT; + if (event_handle != InvalidHandle) { + auto& time_manager = kernel.TimeManager(); + time_manager.UnscheduleTimeEvent(event_handle); } - return WaitForAddressImpl(address, timeout); -} + { + SchedulerLock lock(kernel); + if (current_thread->IsWaitingForArbitration()) { + RemoveThread(SharedFrom(current_thread)); + current_thread->WaitForArbitration(false); + } + } -ResultCode AddressArbiter::WaitForAddressImpl(VAddr address, s64 timeout) { - Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); - current_thread->SetArbiterWaitAddress(address); - InsertThread(SharedFrom(current_thread)); - current_thread->SetStatus(ThreadStatus::WaitArb); - current_thread->InvalidateWakeupCallback(); - current_thread->WakeAfterDelay(timeout); - - system.PrepareReschedule(current_thread->GetProcessorID()); - return RESULT_TIMEOUT; + return current_thread->GetSignalingResult(); } void AddressArbiter::HandleWakeupThread(std::shared_ptr<Thread> thread) { @@ -221,9 +304,9 @@ void AddressArbiter::RemoveThread(std::shared_ptr<Thread> thread) { const auto iter = std::find_if(thread_list.cbegin(), thread_list.cend(), [&thread](const auto& entry) { return thread == entry; }); - ASSERT(iter != thread_list.cend()); - - thread_list.erase(iter); + if (iter != thread_list.cend()) { + thread_list.erase(iter); + } } std::vector<std::shared_ptr<Thread>> AddressArbiter::GetThreadsWaitingOnAddress( diff --git a/src/core/hle/kernel/address_arbiter.h b/src/core/hle/kernel/address_arbiter.h index f958eee5a..0b05d533c 100644 --- a/src/core/hle/kernel/address_arbiter.h +++ b/src/core/hle/kernel/address_arbiter.h @@ -73,9 +73,6 @@ private: /// Waits on an address if the value passed is equal to the argument value. ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout); - // Waits on the given address with a timeout in nanoseconds - ResultCode WaitForAddressImpl(VAddr address, s64 timeout); - /// Wake up num_to_wake (or all) threads in a vector. void WakeThreads(const std::vector<std::shared_ptr<Thread>>& waiting_threads, s32 num_to_wake); diff --git a/src/core/hle/kernel/client_port.cpp b/src/core/hle/kernel/client_port.cpp index 5498fd313..8aff2227a 100644 --- a/src/core/hle/kernel/client_port.cpp +++ b/src/core/hle/kernel/client_port.cpp @@ -34,7 +34,7 @@ ResultVal<std::shared_ptr<ClientSession>> ClientPort::Connect() { } // Wake the threads waiting on the ServerPort - server_port->WakeupAllWaitingThreads(); + server_port->Signal(); return MakeResult(std::move(client)); } diff --git a/src/core/hle/kernel/client_session.cpp b/src/core/hle/kernel/client_session.cpp index 6d66276bc..be9eba519 100644 --- a/src/core/hle/kernel/client_session.cpp +++ b/src/core/hle/kernel/client_session.cpp @@ -47,14 +47,16 @@ ResultVal<std::shared_ptr<ClientSession>> ClientSession::Create(KernelCore& kern return MakeResult(std::move(client_session)); } -ResultCode ClientSession::SendSyncRequest(std::shared_ptr<Thread> thread, Memory::Memory& memory) { +ResultCode ClientSession::SendSyncRequest(std::shared_ptr<Thread> thread, + Core::Memory::Memory& memory, + Core::Timing::CoreTiming& core_timing) { // Keep ServerSession alive until we're done working with it. if (!parent->Server()) { return ERR_SESSION_CLOSED_BY_REMOTE; } // Signal the server session that new data is available - return parent->Server()->HandleSyncRequest(std::move(thread), memory); + return parent->Server()->HandleSyncRequest(std::move(thread), memory, core_timing); } } // namespace Kernel diff --git a/src/core/hle/kernel/client_session.h b/src/core/hle/kernel/client_session.h index d15b09554..e5e0690c2 100644 --- a/src/core/hle/kernel/client_session.h +++ b/src/core/hle/kernel/client_session.h @@ -12,10 +12,14 @@ union ResultCode; -namespace Memory { +namespace Core::Memory { class Memory; } +namespace Core::Timing { +class CoreTiming; +} + namespace Kernel { class KernelCore; @@ -42,7 +46,8 @@ public: return HANDLE_TYPE; } - ResultCode SendSyncRequest(std::shared_ptr<Thread> thread, Memory::Memory& memory); + ResultCode SendSyncRequest(std::shared_ptr<Thread> thread, Core::Memory::Memory& memory, + Core::Timing::CoreTiming& core_timing); bool ShouldWait(const Thread* thread) const override; diff --git a/src/core/hle/kernel/errors.h b/src/core/hle/kernel/errors.h index 8097b3863..d4e5d88cf 100644 --- a/src/core/hle/kernel/errors.h +++ b/src/core/hle/kernel/errors.h @@ -12,8 +12,10 @@ namespace Kernel { constexpr ResultCode ERR_MAX_CONNECTIONS_REACHED{ErrorModule::Kernel, 7}; constexpr ResultCode ERR_INVALID_CAPABILITY_DESCRIPTOR{ErrorModule::Kernel, 14}; +constexpr ResultCode ERR_THREAD_TERMINATING{ErrorModule::Kernel, 59}; constexpr ResultCode ERR_INVALID_SIZE{ErrorModule::Kernel, 101}; constexpr ResultCode ERR_INVALID_ADDRESS{ErrorModule::Kernel, 102}; +constexpr ResultCode ERR_OUT_OF_RESOURCES{ErrorModule::Kernel, 103}; constexpr ResultCode ERR_OUT_OF_MEMORY{ErrorModule::Kernel, 104}; constexpr ResultCode ERR_HANDLE_TABLE_FULL{ErrorModule::Kernel, 105}; constexpr ResultCode ERR_INVALID_ADDRESS_STATE{ErrorModule::Kernel, 106}; diff --git a/src/core/hle/kernel/handle_table.cpp b/src/core/hle/kernel/handle_table.cpp index e441a27fc..3e745c18b 100644 --- a/src/core/hle/kernel/handle_table.cpp +++ b/src/core/hle/kernel/handle_table.cpp @@ -8,7 +8,9 @@ #include "core/core.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" +#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" namespace Kernel { @@ -22,7 +24,7 @@ constexpr u16 GetGeneration(Handle handle) { } } // Anonymous namespace -HandleTable::HandleTable() { +HandleTable::HandleTable(KernelCore& kernel) : kernel{kernel} { Clear(); } @@ -30,6 +32,7 @@ HandleTable::~HandleTable() = default; ResultCode HandleTable::SetSize(s32 handle_table_size) { if (static_cast<u32>(handle_table_size) > MAX_COUNT) { + LOG_ERROR(Kernel, "Handle table size {} is greater than {}", handle_table_size, MAX_COUNT); return ERR_OUT_OF_MEMORY; } @@ -80,6 +83,7 @@ ResultVal<Handle> HandleTable::Duplicate(Handle handle) { ResultCode HandleTable::Close(Handle handle) { if (!IsValid(handle)) { + LOG_ERROR(Kernel, "Handle is not valid! handle={:08X}", handle); return ERR_INVALID_HANDLE; } @@ -101,9 +105,9 @@ bool HandleTable::IsValid(Handle handle) const { std::shared_ptr<Object> HandleTable::GetGeneric(Handle handle) const { if (handle == CurrentThread) { - return SharedFrom(GetCurrentThread()); + return SharedFrom(kernel.CurrentScheduler().GetCurrentThread()); } else if (handle == CurrentProcess) { - return SharedFrom(Core::System::GetInstance().CurrentProcess()); + return SharedFrom(kernel.CurrentProcess()); } if (!IsValid(handle)) { @@ -114,7 +118,7 @@ std::shared_ptr<Object> HandleTable::GetGeneric(Handle handle) const { void HandleTable::Clear() { for (u16 i = 0; i < table_size; ++i) { - generations[i] = i + 1; + generations[i] = static_cast<u16>(i + 1); objects[i] = nullptr; } next_free_slot = 0; diff --git a/src/core/hle/kernel/handle_table.h b/src/core/hle/kernel/handle_table.h index 8029660ed..c9dab8cdd 100644 --- a/src/core/hle/kernel/handle_table.h +++ b/src/core/hle/kernel/handle_table.h @@ -14,6 +14,8 @@ namespace Kernel { +class KernelCore; + enum KernelHandle : Handle { InvalidHandle = 0, CurrentThread = 0xFFFF8000, @@ -48,7 +50,7 @@ public: /// This is the maximum limit of handles allowed per process in Horizon static constexpr std::size_t MAX_COUNT = 1024; - HandleTable(); + explicit HandleTable(KernelCore& kernel); ~HandleTable(); /** @@ -134,6 +136,9 @@ private: /// Head of the free slots linked list. u16 next_free_slot = 0; + + /// Underlying kernel instance that this handle table operates under. + KernelCore& kernel; }; } // namespace Kernel diff --git a/src/core/hle/kernel/hle_ipc.cpp b/src/core/hle/kernel/hle_ipc.cpp index c558a2f33..81f85643b 100644 --- a/src/core/hle/kernel/hle_ipc.cpp +++ b/src/core/hle/kernel/hle_ipc.cpp @@ -13,16 +13,18 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/readable_event.h" +#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/server_session.h" #include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" #include "core/hle/kernel/writable_event.h" #include "core/memory.h" @@ -47,31 +49,33 @@ std::shared_ptr<WritableEvent> HLERequestContext::SleepClientThread( const std::string& reason, u64 timeout, WakeupCallback&& callback, std::shared_ptr<WritableEvent> writable_event) { // Put the client thread to sleep until the wait event is signaled or the timeout expires. - thread->SetWakeupCallback( - [context = *this, callback](ThreadWakeupReason reason, std::shared_ptr<Thread> thread, - std::shared_ptr<SynchronizationObject> object, - std::size_t index) mutable -> bool { - ASSERT(thread->GetStatus() == ThreadStatus::WaitHLEEvent); - callback(thread, context, reason); - context.WriteToOutgoingCommandBuffer(*thread); - return true; - }); - - auto& kernel = Core::System::GetInstance().Kernel(); + if (!writable_event) { // Create event if not provided const auto pair = WritableEvent::CreateEventPair(kernel, "HLE Pause Event: " + reason); writable_event = pair.writable; } - const auto readable_event{writable_event->GetReadableEvent()}; - writable_event->Clear(); - thread->SetStatus(ThreadStatus::WaitHLEEvent); - thread->SetSynchronizationObjects({readable_event}); - readable_event->AddWaitingThread(thread); - - if (timeout > 0) { - thread->WakeAfterDelay(timeout); + { + Handle event_handle = InvalidHandle; + SchedulerLockAndSleep lock(kernel, event_handle, thread.get(), timeout); + thread->SetHLECallback( + [context = *this, callback](std::shared_ptr<Thread> thread) mutable -> bool { + ThreadWakeupReason reason = thread->GetSignalingResult() == RESULT_TIMEOUT + ? ThreadWakeupReason::Timeout + : ThreadWakeupReason::Signal; + callback(thread, context, reason); + context.WriteToOutgoingCommandBuffer(*thread); + return true; + }); + const auto readable_event{writable_event->GetReadableEvent()}; + writable_event->Clear(); + thread->SetHLESyncObject(readable_event.get()); + thread->SetStatus(ThreadStatus::WaitHLEEvent); + thread->SetSynchronizationResults(nullptr, RESULT_TIMEOUT); + readable_event->AddWaitingThread(thread); + lock.Release(); + thread->SetHLETimeEvent(event_handle); } is_thread_waiting = true; @@ -79,9 +83,11 @@ std::shared_ptr<WritableEvent> HLERequestContext::SleepClientThread( return writable_event; } -HLERequestContext::HLERequestContext(std::shared_ptr<Kernel::ServerSession> server_session, +HLERequestContext::HLERequestContext(KernelCore& kernel, Core::Memory::Memory& memory, + std::shared_ptr<ServerSession> server_session, std::shared_ptr<Thread> thread) - : server_session(std::move(server_session)), thread(std::move(thread)) { + : server_session(std::move(server_session)), + thread(std::move(thread)), kernel{kernel}, memory{memory} { cmd_buf[0] = 0; } @@ -216,7 +222,6 @@ ResultCode HLERequestContext::PopulateFromIncomingCommandBuffer(const HandleTabl ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(Thread& thread) { auto& owner_process = *thread.GetOwnerProcess(); auto& handle_table = owner_process.GetHandleTable(); - auto& memory = Core::System::GetInstance().Memory(); std::array<u32, IPC::COMMAND_BUFFER_LENGTH> dst_cmdbuf; memory.ReadBlock(owner_process, thread.GetTLSAddress(), dst_cmdbuf.data(), @@ -282,20 +287,21 @@ ResultCode HLERequestContext::WriteToOutgoingCommandBuffer(Thread& thread) { return RESULT_SUCCESS; } -std::vector<u8> HLERequestContext::ReadBuffer(int buffer_index) const { - std::vector<u8> buffer; +std::vector<u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) const { + std::vector<u8> buffer{}; const bool is_buffer_a{BufferDescriptorA().size() > buffer_index && BufferDescriptorA()[buffer_index].Size()}; - auto& memory = Core::System::GetInstance().Memory(); if (is_buffer_a) { - ASSERT_MSG(BufferDescriptorA().size() > buffer_index, - "BufferDescriptorA invalid buffer_index {}", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorA().size() > buffer_index, { return buffer; }, + "BufferDescriptorA invalid buffer_index {}", buffer_index); buffer.resize(BufferDescriptorA()[buffer_index].Size()); memory.ReadBlock(BufferDescriptorA()[buffer_index].Address(), buffer.data(), buffer.size()); } else { - ASSERT_MSG(BufferDescriptorX().size() > buffer_index, - "BufferDescriptorX invalid buffer_index {}", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorX().size() > buffer_index, { return buffer; }, + "BufferDescriptorX invalid buffer_index {}", buffer_index); buffer.resize(BufferDescriptorX()[buffer_index].Size()); memory.ReadBlock(BufferDescriptorX()[buffer_index].Address(), buffer.data(), buffer.size()); } @@ -304,7 +310,7 @@ std::vector<u8> HLERequestContext::ReadBuffer(int buffer_index) const { } std::size_t HLERequestContext::WriteBuffer(const void* buffer, std::size_t size, - int buffer_index) const { + std::size_t buffer_index) const { if (size == 0) { LOG_WARNING(Core, "skip empty buffer write"); return 0; @@ -319,54 +325,54 @@ std::size_t HLERequestContext::WriteBuffer(const void* buffer, std::size_t size, size = buffer_size; // TODO(bunnei): This needs to be HW tested } - auto& memory = Core::System::GetInstance().Memory(); if (is_buffer_b) { - ASSERT_MSG(BufferDescriptorB().size() > buffer_index, - "BufferDescriptorB invalid buffer_index {}", buffer_index); - ASSERT_MSG(BufferDescriptorB()[buffer_index].Size() >= size, - "BufferDescriptorB buffer_index {} is not large enough", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorB().size() > buffer_index && + BufferDescriptorB()[buffer_index].Size() >= size, + { return 0; }, "BufferDescriptorB is invalid, index={}, size={}", buffer_index, size); memory.WriteBlock(BufferDescriptorB()[buffer_index].Address(), buffer, size); } else { - ASSERT_MSG(BufferDescriptorC().size() > buffer_index, - "BufferDescriptorC invalid buffer_index {}", buffer_index); - ASSERT_MSG(BufferDescriptorC()[buffer_index].Size() >= size, - "BufferDescriptorC buffer_index {} is not large enough", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorC().size() > buffer_index && + BufferDescriptorC()[buffer_index].Size() >= size, + { return 0; }, "BufferDescriptorC is invalid, index={}, size={}", buffer_index, size); memory.WriteBlock(BufferDescriptorC()[buffer_index].Address(), buffer, size); } return size; } -std::size_t HLERequestContext::GetReadBufferSize(int buffer_index) const { +std::size_t HLERequestContext::GetReadBufferSize(std::size_t buffer_index) const { const bool is_buffer_a{BufferDescriptorA().size() > buffer_index && BufferDescriptorA()[buffer_index].Size()}; if (is_buffer_a) { - ASSERT_MSG(BufferDescriptorA().size() > buffer_index, - "BufferDescriptorA invalid buffer_index {}", buffer_index); - ASSERT_MSG(BufferDescriptorA()[buffer_index].Size() > 0, - "BufferDescriptorA buffer_index {} is empty", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorA().size() > buffer_index, { return 0; }, + "BufferDescriptorA invalid buffer_index {}", buffer_index); return BufferDescriptorA()[buffer_index].Size(); } else { - ASSERT_MSG(BufferDescriptorX().size() > buffer_index, - "BufferDescriptorX invalid buffer_index {}", buffer_index); - ASSERT_MSG(BufferDescriptorX()[buffer_index].Size() > 0, - "BufferDescriptorX buffer_index {} is empty", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorX().size() > buffer_index, { return 0; }, + "BufferDescriptorX invalid buffer_index {}", buffer_index); return BufferDescriptorX()[buffer_index].Size(); } } -std::size_t HLERequestContext::GetWriteBufferSize(int buffer_index) const { +std::size_t HLERequestContext::GetWriteBufferSize(std::size_t buffer_index) const { const bool is_buffer_b{BufferDescriptorB().size() > buffer_index && BufferDescriptorB()[buffer_index].Size()}; if (is_buffer_b) { - ASSERT_MSG(BufferDescriptorB().size() > buffer_index, - "BufferDescriptorB invalid buffer_index {}", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorB().size() > buffer_index, { return 0; }, + "BufferDescriptorB invalid buffer_index {}", buffer_index); return BufferDescriptorB()[buffer_index].Size(); } else { - ASSERT_MSG(BufferDescriptorC().size() > buffer_index, - "BufferDescriptorC invalid buffer_index {}", buffer_index); + ASSERT_OR_EXECUTE_MSG( + BufferDescriptorC().size() > buffer_index, { return 0; }, + "BufferDescriptorC invalid buffer_index {}", buffer_index); return BufferDescriptorC()[buffer_index].Size(); } + return 0; } std::string HLERequestContext::Description() const { diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h index 050ad8fd7..c31a65476 100644 --- a/src/core/hle/kernel/hle_ipc.h +++ b/src/core/hle/kernel/hle_ipc.h @@ -13,12 +13,21 @@ #include <vector> #include <boost/container/small_vector.hpp> #include "common/common_types.h" +#include "common/concepts.h" #include "common/swap.h" #include "core/hle/ipc.h" #include "core/hle/kernel/object.h" union ResultCode; +namespace Core::Memory { +class Memory; +} + +namespace IPC { +class ResponseBuilder; +} + namespace Service { class ServiceFrameworkBase; } @@ -28,6 +37,7 @@ namespace Kernel { class Domain; class HandleTable; class HLERequestContext; +class KernelCore; class Process; class ServerSession; class Thread; @@ -98,7 +108,8 @@ protected: */ class HLERequestContext { public: - explicit HLERequestContext(std::shared_ptr<ServerSession> session, + explicit HLERequestContext(KernelCore& kernel, Core::Memory::Memory& memory, + std::shared_ptr<ServerSession> session, std::shared_ptr<Thread> thread); ~HLERequestContext(); @@ -179,36 +190,39 @@ public: } /// Helper function to read a buffer using the appropriate buffer descriptor - std::vector<u8> ReadBuffer(int buffer_index = 0) const; + std::vector<u8> ReadBuffer(std::size_t buffer_index = 0) const; /// Helper function to write a buffer using the appropriate buffer descriptor - std::size_t WriteBuffer(const void* buffer, std::size_t size, int buffer_index = 0) const; + std::size_t WriteBuffer(const void* buffer, std::size_t size, + std::size_t buffer_index = 0) const; /* Helper function to write a buffer using the appropriate buffer descriptor * - * @tparam ContiguousContainer an arbitrary container that satisfies the - * ContiguousContainer concept in the C++ standard library. + * @tparam T an arbitrary container that satisfies the + * ContiguousContainer concept in the C++ standard library or a trivially copyable type. * - * @param container The container to write the data of into a buffer. + * @param data The container/data to write into a buffer. * @param buffer_index The buffer in particular to write to. */ - template <typename ContiguousContainer, - typename = std::enable_if_t<!std::is_pointer_v<ContiguousContainer>>> - std::size_t WriteBuffer(const ContiguousContainer& container, int buffer_index = 0) const { - using ContiguousType = typename ContiguousContainer::value_type; - - static_assert(std::is_trivially_copyable_v<ContiguousType>, - "Container to WriteBuffer must contain trivially copyable objects"); - - return WriteBuffer(std::data(container), std::size(container) * sizeof(ContiguousType), - buffer_index); + template <typename T, typename = std::enable_if_t<!std::is_pointer_v<T>>> + std::size_t WriteBuffer(const T& data, std::size_t buffer_index = 0) const { + if constexpr (Common::IsSTLContainer<T>) { + using ContiguousType = typename T::value_type; + static_assert(std::is_trivially_copyable_v<ContiguousType>, + "Container to WriteBuffer must contain trivially copyable objects"); + return WriteBuffer(std::data(data), std::size(data) * sizeof(ContiguousType), + buffer_index); + } else { + static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable"); + return WriteBuffer(&data, sizeof(T), buffer_index); + } } /// Helper function to get the size of the input buffer - std::size_t GetReadBufferSize(int buffer_index = 0) const; + std::size_t GetReadBufferSize(std::size_t buffer_index = 0) const; /// Helper function to get the size of the output buffer - std::size_t GetWriteBufferSize(int buffer_index = 0) const; + std::size_t GetWriteBufferSize(std::size_t buffer_index = 0) const; template <typename T> std::shared_ptr<T> GetCopyObject(std::size_t index) { @@ -277,6 +291,8 @@ public: } private: + friend class IPC::ResponseBuilder; + void ParseCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf, bool incoming); std::array<u32, IPC::COMMAND_BUFFER_LENGTH> cmd_buf; @@ -303,6 +319,9 @@ private: std::vector<std::shared_ptr<SessionRequestHandler>> domain_request_handlers; bool is_thread_waiting{}; + + KernelCore& kernel; + Core::Memory::Memory& memory; }; } // namespace Kernel diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index 014d647cf..bb3e312a7 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -2,31 +2,40 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> #include <atomic> #include <bitset> #include <functional> #include <memory> -#include <mutex> #include <thread> #include <unordered_map> #include <utility> #include "common/assert.h" #include "common/logging/log.h" +#include "common/microprofile.h" +#include "common/thread.h" #include "core/arm/arm_interface.h" +#include "core/arm/cpu_interrupt_handler.h" #include "core/arm/exclusive_monitor.h" #include "core/core.h" #include "core/core_timing.h" #include "core/core_timing_util.h" +#include "core/cpu_manager.h" +#include "core/device_memory.h" #include "core/hardware_properties.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/memory_layout.h" +#include "core/hle/kernel/memory/memory_manager.h" +#include "core/hle/kernel/memory/slab_heap.h" #include "core/hle/kernel/physical_core.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/resource_limit.h" #include "core/hle/kernel/scheduler.h" +#include "core/hle/kernel/shared_memory.h" #include "core/hle/kernel/synchronization.h" #include "core/hle/kernel/thread.h" #include "core/hle/kernel/time_manager.h" @@ -34,84 +43,29 @@ #include "core/hle/result.h" #include "core/memory.h" -namespace Kernel { - -/** - * Callback that will wake up the thread it was scheduled for - * @param thread_handle The handle of the thread that's been awoken - * @param cycles_late The number of CPU cycles that have passed since the desired wakeup time - */ -static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] s64 cycles_late) { - const auto proper_handle = static_cast<Handle>(thread_handle); - const auto& system = Core::System::GetInstance(); - - // Lock the global kernel mutex when we enter the kernel HLE. - std::lock_guard lock{HLE::g_hle_lock}; - - std::shared_ptr<Thread> thread = - system.Kernel().RetrieveThreadFromGlobalHandleTable(proper_handle); - if (thread == nullptr) { - LOG_CRITICAL(Kernel, "Callback fired for invalid thread {:08X}", proper_handle); - return; - } - - bool resume = true; - - if (thread->GetStatus() == ThreadStatus::WaitSynch || - thread->GetStatus() == ThreadStatus::WaitHLEEvent) { - // Remove the thread from each of its waiting objects' waitlists - for (const auto& object : thread->GetSynchronizationObjects()) { - object->RemoveWaitingThread(thread); - } - thread->ClearSynchronizationObjects(); - - // Invoke the wakeup callback before clearing the wait objects - if (thread->HasWakeupCallback()) { - resume = thread->InvokeWakeupCallback(ThreadWakeupReason::Timeout, thread, nullptr, 0); - } - } else if (thread->GetStatus() == ThreadStatus::WaitMutex || - thread->GetStatus() == ThreadStatus::WaitCondVar) { - thread->SetMutexWaitAddress(0); - thread->SetWaitHandle(0); - if (thread->GetStatus() == ThreadStatus::WaitCondVar) { - thread->GetOwnerProcess()->RemoveConditionVariableThread(thread); - thread->SetCondVarWaitAddress(0); - } - - auto* const lock_owner = thread->GetLockOwner(); - // Threads waking up by timeout from WaitProcessWideKey do not perform priority inheritance - // and don't have a lock owner unless SignalProcessWideKey was called first and the thread - // wasn't awakened due to the mutex already being acquired. - if (lock_owner != nullptr) { - lock_owner->RemoveMutexWaiter(thread); - } - } - - if (thread->GetStatus() == ThreadStatus::WaitArb) { - auto& address_arbiter = thread->GetOwnerProcess()->GetAddressArbiter(); - address_arbiter.HandleWakeupThread(thread); - } +MICROPROFILE_DEFINE(Kernel_SVC, "Kernel", "SVC", MP_RGB(70, 200, 70)); - if (resume) { - if (thread->GetStatus() == ThreadStatus::WaitCondVar || - thread->GetStatus() == ThreadStatus::WaitArb) { - thread->SetWaitSynchronizationResult(RESULT_TIMEOUT); - } - thread->ResumeFromWait(); - } -} +namespace Kernel { struct KernelCore::Impl { explicit Impl(Core::System& system, KernelCore& kernel) - : global_scheduler{kernel}, synchronization{system}, time_manager{system}, system{system} {} + : global_scheduler{kernel}, synchronization{system}, time_manager{system}, + global_handle_table{kernel}, system{system} {} + + void SetMulticore(bool is_multicore) { + this->is_multicore = is_multicore; + } void Initialize(KernelCore& kernel) { Shutdown(); + RegisterHostThread(); InitializePhysicalCores(); InitializeSystemResourceLimit(kernel); - InitializeThreads(); - InitializePreemption(); + InitializeMemoryLayout(); + InitializePreemption(kernel); + InitializeSchedulers(); + InitializeSuspendThreads(); } void Shutdown() { @@ -120,13 +74,24 @@ struct KernelCore::Impl { next_user_process_id = Process::ProcessIDMin; next_thread_id = 1; + for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { + if (suspend_threads[i]) { + suspend_threads[i].reset(); + } + } + + for (std::size_t i = 0; i < cores.size(); i++) { + cores[i].Shutdown(); + schedulers[i].reset(); + } + cores.clear(); + process_list.clear(); current_process = nullptr; system_resource_limit = nullptr; global_handle_table.Clear(); - thread_wakeup_event_type = nullptr; preemption_event = nullptr; global_scheduler.Shutdown(); @@ -139,13 +104,25 @@ struct KernelCore::Impl { cores.clear(); exclusive_monitor.reset(); + + num_host_threads = 0; + std::fill(register_host_thread_keys.begin(), register_host_thread_keys.end(), + std::thread::id{}); + std::fill(register_host_thread_values.begin(), register_host_thread_values.end(), 0); } void InitializePhysicalCores() { exclusive_monitor = Core::MakeExclusiveMonitor(system.Memory(), Core::Hardware::NUM_CPU_CORES); for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { - cores.emplace_back(system, i, *exclusive_monitor); + schedulers[i] = std::make_unique<Kernel::Scheduler>(system, i); + cores.emplace_back(system, i, *schedulers[i], interrupts[i]); + } + } + + void InitializeSchedulers() { + for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { + cores[i].Scheduler().Initialize(); } } @@ -154,71 +131,105 @@ struct KernelCore::Impl { system_resource_limit = ResourceLimit::Create(kernel); // If setting the default system values fails, then something seriously wrong has occurred. - ASSERT(system_resource_limit->SetLimitValue(ResourceType::PhysicalMemory, 0x200000000) + ASSERT(system_resource_limit->SetLimitValue(ResourceType::PhysicalMemory, 0x100000000) .IsSuccess()); ASSERT(system_resource_limit->SetLimitValue(ResourceType::Threads, 800).IsSuccess()); ASSERT(system_resource_limit->SetLimitValue(ResourceType::Events, 700).IsSuccess()); ASSERT(system_resource_limit->SetLimitValue(ResourceType::TransferMemory, 200).IsSuccess()); ASSERT(system_resource_limit->SetLimitValue(ResourceType::Sessions, 900).IsSuccess()); - } - void InitializeThreads() { - thread_wakeup_event_type = - Core::Timing::CreateEvent("ThreadWakeupCallback", ThreadWakeupCallback); + if (!system_resource_limit->Reserve(ResourceType::PhysicalMemory, 0) || + !system_resource_limit->Reserve(ResourceType::PhysicalMemory, 0x60000)) { + UNREACHABLE(); + } } - void InitializePreemption() { - preemption_event = - Core::Timing::CreateEvent("PreemptionCallback", [this](u64 userdata, s64 cycles_late) { - global_scheduler.PreemptThreads(); - s64 time_interval = Core::Timing::msToCycles(std::chrono::milliseconds(10)); + void InitializePreemption(KernelCore& kernel) { + preemption_event = Core::Timing::CreateEvent( + "PreemptionCallback", [this, &kernel](std::uintptr_t, std::chrono::nanoseconds) { + { + SchedulerLock lock(kernel); + global_scheduler.PreemptThreads(); + } + const auto time_interval = std::chrono::nanoseconds{ + Core::Timing::msToCycles(std::chrono::milliseconds(10))}; system.CoreTiming().ScheduleEvent(time_interval, preemption_event); }); - s64 time_interval = Core::Timing::msToCycles(std::chrono::milliseconds(10)); + const auto time_interval = + std::chrono::nanoseconds{Core::Timing::msToCycles(std::chrono::milliseconds(10))}; system.CoreTiming().ScheduleEvent(time_interval, preemption_event); } + void InitializeSuspendThreads() { + for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { + std::string name = "Suspend Thread Id:" + std::to_string(i); + std::function<void(void*)> init_func = Core::CpuManager::GetSuspendThreadStartFunc(); + void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater(); + const auto type = + static_cast<ThreadType>(THREADTYPE_KERNEL | THREADTYPE_HLE | THREADTYPE_SUSPEND); + auto thread_res = + Thread::Create(system, type, std::move(name), 0, 0, 0, static_cast<u32>(i), 0, + nullptr, std::move(init_func), init_func_parameter); + + suspend_threads[i] = std::move(thread_res).Unwrap(); + } + } + void MakeCurrentProcess(Process* process) { current_process = process; - if (process == nullptr) { return; } - - for (auto& core : cores) { - core.SetIs64Bit(process->Is64BitProcess()); + const u32 core_id = GetCurrentHostThreadID(); + if (core_id < Core::Hardware::NUM_CPU_CORES) { + system.Memory().SetCurrentPageTable(*process, core_id); } - - system.Memory().SetCurrentPageTable(*process); } void RegisterCoreThread(std::size_t core_id) { - std::unique_lock lock{register_thread_mutex}; const std::thread::id this_id = std::this_thread::get_id(); - const auto it = host_thread_ids.find(this_id); + if (!is_multicore) { + single_core_thread_id = this_id; + } + const auto end = + register_host_thread_keys.begin() + static_cast<ptrdiff_t>(num_host_threads); + const auto it = std::find(register_host_thread_keys.begin(), end, this_id); ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); - ASSERT(it == host_thread_ids.end()); - ASSERT(!registered_core_threads[core_id]); - host_thread_ids[this_id] = static_cast<u32>(core_id); - registered_core_threads.set(core_id); + ASSERT(it == end); + InsertHostThread(static_cast<u32>(core_id)); } void RegisterHostThread() { - std::unique_lock lock{register_thread_mutex}; const std::thread::id this_id = std::this_thread::get_id(); - const auto it = host_thread_ids.find(this_id); - ASSERT(it == host_thread_ids.end()); - host_thread_ids[this_id] = registered_thread_ids++; + const auto end = + register_host_thread_keys.begin() + static_cast<ptrdiff_t>(num_host_threads); + const auto it = std::find(register_host_thread_keys.begin(), end, this_id); + if (it == end) { + InsertHostThread(registered_thread_ids++); + } + } + + void InsertHostThread(u32 value) { + const size_t index = num_host_threads++; + ASSERT_MSG(index < NUM_REGISTRABLE_HOST_THREADS, "Too many host threads"); + register_host_thread_values[index] = value; + register_host_thread_keys[index] = std::this_thread::get_id(); } - u32 GetCurrentHostThreadID() const { + [[nodiscard]] u32 GetCurrentHostThreadID() const { const std::thread::id this_id = std::this_thread::get_id(); - const auto it = host_thread_ids.find(this_id); - if (it == host_thread_ids.end()) { + if (!is_multicore && single_core_thread_id == this_id) { + return static_cast<u32>(system.GetCpuManager().CurrentCore()); + } + const auto end = + register_host_thread_keys.begin() + static_cast<ptrdiff_t>(num_host_threads); + const auto it = std::find(register_host_thread_keys.begin(), end, this_id); + if (it == end) { return Core::INVALID_HOST_THREAD_ID; } - return it->second; + return register_host_thread_values[static_cast<size_t>( + std::distance(register_host_thread_keys.begin(), it))]; } Core::EmuThreadHandle GetCurrentEmuThreadID() const { @@ -229,7 +240,7 @@ struct KernelCore::Impl { } const Kernel::Scheduler& sched = cores[result.host_handle].Scheduler(); const Kernel::Thread* current = sched.GetCurrentThread(); - if (current != nullptr) { + if (current != nullptr && !current->IsPhantomMode()) { result.guest_handle = current->GetGlobalHandle(); } else { result.guest_handle = InvalidHandle; @@ -237,6 +248,57 @@ struct KernelCore::Impl { return result; } + void InitializeMemoryLayout() { + // Initialize memory layout + constexpr Memory::MemoryLayout layout{Memory::MemoryLayout::GetDefaultLayout()}; + constexpr std::size_t hid_size{0x40000}; + constexpr std::size_t font_size{0x1100000}; + constexpr std::size_t irs_size{0x8000}; + constexpr std::size_t time_size{0x1000}; + constexpr PAddr hid_addr{layout.System().StartAddress()}; + constexpr PAddr font_pa{layout.System().StartAddress() + hid_size}; + constexpr PAddr irs_addr{layout.System().StartAddress() + hid_size + font_size}; + constexpr PAddr time_addr{layout.System().StartAddress() + hid_size + font_size + irs_size}; + + // Initialize memory manager + memory_manager = std::make_unique<Memory::MemoryManager>(); + memory_manager->InitializeManager(Memory::MemoryManager::Pool::Application, + layout.Application().StartAddress(), + layout.Application().EndAddress()); + memory_manager->InitializeManager(Memory::MemoryManager::Pool::Applet, + layout.Applet().StartAddress(), + layout.Applet().EndAddress()); + memory_manager->InitializeManager(Memory::MemoryManager::Pool::System, + layout.System().StartAddress(), + layout.System().EndAddress()); + + hid_shared_mem = Kernel::SharedMemory::Create( + system.Kernel(), system.DeviceMemory(), nullptr, + {hid_addr, hid_size / Memory::PageSize}, Memory::MemoryPermission::None, + Memory::MemoryPermission::Read, hid_addr, hid_size, "HID:SharedMemory"); + font_shared_mem = Kernel::SharedMemory::Create( + system.Kernel(), system.DeviceMemory(), nullptr, + {font_pa, font_size / Memory::PageSize}, Memory::MemoryPermission::None, + Memory::MemoryPermission::Read, font_pa, font_size, "Font:SharedMemory"); + irs_shared_mem = Kernel::SharedMemory::Create( + system.Kernel(), system.DeviceMemory(), nullptr, + {irs_addr, irs_size / Memory::PageSize}, Memory::MemoryPermission::None, + Memory::MemoryPermission::Read, irs_addr, irs_size, "IRS:SharedMemory"); + time_shared_mem = Kernel::SharedMemory::Create( + system.Kernel(), system.DeviceMemory(), nullptr, + {time_addr, time_size / Memory::PageSize}, Memory::MemoryPermission::None, + Memory::MemoryPermission::Read, time_addr, time_size, "Time:SharedMemory"); + + // Allocate slab heaps + user_slab_heap_pages = std::make_unique<Memory::SlabHeap<Memory::Page>>(); + + // Initialize slab heaps + constexpr u64 user_slab_heap_size{0x3de000}; + user_slab_heap_pages->Initialize( + system.DeviceMemory().GetPointer(Core::DramMemoryMap::SlabHeapBase), + user_slab_heap_size); + } + std::atomic<u32> next_object_id{0}; std::atomic<u64> next_kernel_process_id{Process::InitialKIPIDMin}; std::atomic<u64> next_user_process_id{Process::ProcessIDMin}; @@ -251,12 +313,11 @@ struct KernelCore::Impl { std::shared_ptr<ResourceLimit> system_resource_limit; - std::shared_ptr<Core::Timing::EventType> thread_wakeup_event_type; std::shared_ptr<Core::Timing::EventType> preemption_event; // This is the kernel's handle table or supervisor handle table which // stores all the objects in place. - Kernel::HandleTable global_handle_table; + HandleTable global_handle_table; /// Map of named ports managed by the kernel, which can be retrieved using /// the ConnectToPort SVC. @@ -266,10 +327,33 @@ struct KernelCore::Impl { std::vector<Kernel::PhysicalCore> cores; // 0-3 IDs represent core threads, >3 represent others - std::unordered_map<std::thread::id, u32> host_thread_ids; - u32 registered_thread_ids{Core::Hardware::NUM_CPU_CORES}; - std::bitset<Core::Hardware::NUM_CPU_CORES> registered_core_threads; - std::mutex register_thread_mutex; + std::atomic<u32> registered_thread_ids{Core::Hardware::NUM_CPU_CORES}; + + // Number of host threads is a relatively high number to avoid overflowing + static constexpr size_t NUM_REGISTRABLE_HOST_THREADS = 64; + std::atomic<size_t> num_host_threads{0}; + std::array<std::atomic<std::thread::id>, NUM_REGISTRABLE_HOST_THREADS> + register_host_thread_keys{}; + std::array<std::atomic<u32>, NUM_REGISTRABLE_HOST_THREADS> register_host_thread_values{}; + + // Kernel memory management + std::unique_ptr<Memory::MemoryManager> memory_manager; + std::unique_ptr<Memory::SlabHeap<Memory::Page>> user_slab_heap_pages; + + // Shared memory for services + std::shared_ptr<Kernel::SharedMemory> hid_shared_mem; + std::shared_ptr<Kernel::SharedMemory> font_shared_mem; + std::shared_ptr<Kernel::SharedMemory> irs_shared_mem; + std::shared_ptr<Kernel::SharedMemory> time_shared_mem; + + std::array<std::shared_ptr<Thread>, Core::Hardware::NUM_CPU_CORES> suspend_threads{}; + std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES> interrupts{}; + std::array<std::unique_ptr<Kernel::Scheduler>, Core::Hardware::NUM_CPU_CORES> schedulers{}; + + bool is_multicore{}; + std::thread::id single_core_thread_id{}; + + std::array<u64, Core::Hardware::NUM_CPU_CORES> svc_ticks{}; // System context Core::System& system; @@ -280,6 +364,10 @@ KernelCore::~KernelCore() { Shutdown(); } +void KernelCore::SetMulticore(bool is_multicore) { + impl->SetMulticore(is_multicore); +} + void KernelCore::Initialize() { impl->Initialize(*this); } @@ -325,11 +413,11 @@ const Kernel::GlobalScheduler& KernelCore::GlobalScheduler() const { } Kernel::Scheduler& KernelCore::Scheduler(std::size_t id) { - return impl->cores[id].Scheduler(); + return *impl->schedulers[id]; } const Kernel::Scheduler& KernelCore::Scheduler(std::size_t id) const { - return impl->cores[id].Scheduler(); + return *impl->schedulers[id]; } Kernel::PhysicalCore& KernelCore::PhysicalCore(std::size_t id) { @@ -340,6 +428,39 @@ const Kernel::PhysicalCore& KernelCore::PhysicalCore(std::size_t id) const { return impl->cores[id]; } +Kernel::PhysicalCore& KernelCore::CurrentPhysicalCore() { + u32 core_id = impl->GetCurrentHostThreadID(); + ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); + return impl->cores[core_id]; +} + +const Kernel::PhysicalCore& KernelCore::CurrentPhysicalCore() const { + u32 core_id = impl->GetCurrentHostThreadID(); + ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); + return impl->cores[core_id]; +} + +Kernel::Scheduler& KernelCore::CurrentScheduler() { + u32 core_id = impl->GetCurrentHostThreadID(); + ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); + return *impl->schedulers[core_id]; +} + +const Kernel::Scheduler& KernelCore::CurrentScheduler() const { + u32 core_id = impl->GetCurrentHostThreadID(); + ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); + return *impl->schedulers[core_id]; +} + +std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>& KernelCore::Interrupts() { + return impl->interrupts; +} + +const std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>& KernelCore::Interrupts() + const { + return impl->interrupts; +} + Kernel::Synchronization& KernelCore::Synchronization() { return impl->synchronization; } @@ -365,15 +486,17 @@ const Core::ExclusiveMonitor& KernelCore::GetExclusiveMonitor() const { } void KernelCore::InvalidateAllInstructionCaches() { - for (std::size_t i = 0; i < impl->global_scheduler.CpuCoresCount(); i++) { - PhysicalCore(i).ArmInterface().ClearInstructionCache(); + auto& threads = GlobalScheduler().GetThreadList(); + for (auto& thread : threads) { + if (!thread->IsHLEThread()) { + auto& arm_interface = thread->ArmInterface(); + arm_interface.ClearInstructionCache(); + } } } void KernelCore::PrepareReschedule(std::size_t id) { - if (id < impl->global_scheduler.CpuCoresCount()) { - impl->cores[id].Stop(); - } + // TODO: Reimplement, this } void KernelCore::AddNamedPort(std::string name, std::shared_ptr<ClientPort> port) { @@ -409,10 +532,6 @@ u64 KernelCore::CreateNewUserProcessID() { return impl->next_user_process_id++; } -const std::shared_ptr<Core::Timing::EventType>& KernelCore::ThreadWakeupCallbackEventType() const { - return impl->thread_wakeup_event_type; -} - Kernel::HandleTable& KernelCore::GlobalHandleTable() { return impl->global_handle_table; } @@ -437,4 +556,82 @@ Core::EmuThreadHandle KernelCore::GetCurrentEmuThreadID() const { return impl->GetCurrentEmuThreadID(); } +Memory::MemoryManager& KernelCore::MemoryManager() { + return *impl->memory_manager; +} + +const Memory::MemoryManager& KernelCore::MemoryManager() const { + return *impl->memory_manager; +} + +Memory::SlabHeap<Memory::Page>& KernelCore::GetUserSlabHeapPages() { + return *impl->user_slab_heap_pages; +} + +const Memory::SlabHeap<Memory::Page>& KernelCore::GetUserSlabHeapPages() const { + return *impl->user_slab_heap_pages; +} + +Kernel::SharedMemory& KernelCore::GetHidSharedMem() { + return *impl->hid_shared_mem; +} + +const Kernel::SharedMemory& KernelCore::GetHidSharedMem() const { + return *impl->hid_shared_mem; +} + +Kernel::SharedMemory& KernelCore::GetFontSharedMem() { + return *impl->font_shared_mem; +} + +const Kernel::SharedMemory& KernelCore::GetFontSharedMem() const { + return *impl->font_shared_mem; +} + +Kernel::SharedMemory& KernelCore::GetIrsSharedMem() { + return *impl->irs_shared_mem; +} + +const Kernel::SharedMemory& KernelCore::GetIrsSharedMem() const { + return *impl->irs_shared_mem; +} + +Kernel::SharedMemory& KernelCore::GetTimeSharedMem() { + return *impl->time_shared_mem; +} + +const Kernel::SharedMemory& KernelCore::GetTimeSharedMem() const { + return *impl->time_shared_mem; +} + +void KernelCore::Suspend(bool in_suspention) { + const bool should_suspend = exception_exited || in_suspention; + { + SchedulerLock lock(*this); + ThreadStatus status = should_suspend ? ThreadStatus::Ready : ThreadStatus::WaitSleep; + for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { + impl->suspend_threads[i]->SetStatus(status); + } + } +} + +bool KernelCore::IsMulticore() const { + return impl->is_multicore; +} + +void KernelCore::ExceptionalExit() { + exception_exited = true; + Suspend(true); +} + +void KernelCore::EnterSVCProfile() { + std::size_t core = impl->GetCurrentHostThreadID(); + impl->svc_ticks[core] = MicroProfileEnter(MICROPROFILE_TOKEN(Kernel_SVC)); +} + +void KernelCore::ExitSVCProfile() { + std::size_t core = impl->GetCurrentHostThreadID(); + MicroProfileLeave(MICROPROFILE_TOKEN(Kernel_SVC), impl->svc_ticks[core]); +} + } // namespace Kernel diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h index c4f78ab71..16285c3f0 100644 --- a/src/core/hle/kernel/kernel.h +++ b/src/core/hle/kernel/kernel.h @@ -4,14 +4,18 @@ #pragma once +#include <array> #include <memory> #include <string> #include <unordered_map> #include <vector> +#include "core/arm/cpu_interrupt_handler.h" +#include "core/hardware_properties.h" +#include "core/hle/kernel/memory/memory_types.h" #include "core/hle/kernel/object.h" namespace Core { -struct EmuThreadHandle; +class CPUInterruptHandler; class ExclusiveMonitor; class System; } // namespace Core @@ -23,6 +27,12 @@ struct EventType; namespace Kernel { +namespace Memory { +class MemoryManager; +template <typename T> +class SlabHeap; +} // namespace Memory + class AddressArbiter; class ClientPort; class GlobalScheduler; @@ -31,6 +41,7 @@ class PhysicalCore; class Process; class ResourceLimit; class Scheduler; +class SharedMemory; class Synchronization; class Thread; class TimeManager; @@ -57,6 +68,9 @@ public: KernelCore(KernelCore&&) = delete; KernelCore& operator=(KernelCore&&) = delete; + /// Sets if emulation is multicore or single core, must be set before Initialize + void SetMulticore(bool is_multicore); + /// Resets the kernel to a clean slate for use. void Initialize(); @@ -102,6 +116,18 @@ public: /// Gets the an instance of the respective physical CPU core. const Kernel::PhysicalCore& PhysicalCore(std::size_t id) const; + /// Gets the sole instance of the Scheduler at the current running core. + Kernel::Scheduler& CurrentScheduler(); + + /// Gets the sole instance of the Scheduler at the current running core. + const Kernel::Scheduler& CurrentScheduler() const; + + /// Gets the an instance of the current physical CPU core. + Kernel::PhysicalCore& CurrentPhysicalCore(); + + /// Gets the an instance of the current physical CPU core. + const Kernel::PhysicalCore& CurrentPhysicalCore() const; + /// Gets the an instance of the Synchronization Interface. Kernel::Synchronization& Synchronization(); @@ -121,6 +147,10 @@ public: const Core::ExclusiveMonitor& GetExclusiveMonitor() const; + std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>& Interrupts(); + + const std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>& Interrupts() const; + void InvalidateAllInstructionCaches(); /// Adds a port to the named port table @@ -147,6 +177,54 @@ public: /// Register the current thread as a non CPU core thread. void RegisterHostThread(); + /// Gets the virtual memory manager for the kernel. + Memory::MemoryManager& MemoryManager(); + + /// Gets the virtual memory manager for the kernel. + const Memory::MemoryManager& MemoryManager() const; + + /// Gets the slab heap allocated for user space pages. + Memory::SlabHeap<Memory::Page>& GetUserSlabHeapPages(); + + /// Gets the slab heap allocated for user space pages. + const Memory::SlabHeap<Memory::Page>& GetUserSlabHeapPages() const; + + /// Gets the shared memory object for HID services. + Kernel::SharedMemory& GetHidSharedMem(); + + /// Gets the shared memory object for HID services. + const Kernel::SharedMemory& GetHidSharedMem() const; + + /// Gets the shared memory object for font services. + Kernel::SharedMemory& GetFontSharedMem(); + + /// Gets the shared memory object for font services. + const Kernel::SharedMemory& GetFontSharedMem() const; + + /// Gets the shared memory object for IRS services. + Kernel::SharedMemory& GetIrsSharedMem(); + + /// Gets the shared memory object for IRS services. + const Kernel::SharedMemory& GetIrsSharedMem() const; + + /// Gets the shared memory object for Time services. + Kernel::SharedMemory& GetTimeSharedMem(); + + /// Gets the shared memory object for Time services. + const Kernel::SharedMemory& GetTimeSharedMem() const; + + /// Suspend/unsuspend the OS. + void Suspend(bool in_suspention); + + /// Exceptional exit the OS. + void ExceptionalExit(); + + bool IsMulticore() const; + + void EnterSVCProfile(); + + void ExitSVCProfile(); + private: friend class Object; friend class Process; @@ -164,9 +242,6 @@ private: /// Creates a new thread ID, incrementing the internal thread ID counter. u64 CreateNewThreadID(); - /// Retrieves the event type used for thread wakeup callbacks. - const std::shared_ptr<Core::Timing::EventType>& ThreadWakeupCallbackEventType() const; - /// Provides a reference to the global handle table. Kernel::HandleTable& GlobalHandleTable(); @@ -175,6 +250,7 @@ private: struct Impl; std::unique_ptr<Impl> impl; + bool exception_exited{}; }; } // namespace Kernel diff --git a/src/core/hle/kernel/memory/address_space_info.cpp b/src/core/hle/kernel/memory/address_space_info.cpp new file mode 100644 index 000000000..e4288cab4 --- /dev/null +++ b/src/core/hle/kernel/memory/address_space_info.cpp @@ -0,0 +1,117 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#include <array> + +#include "common/assert.h" +#include "core/hle/kernel/memory/address_space_info.h" + +namespace Kernel::Memory { + +namespace { + +enum : u64 { + Size_1_MB = 0x100000, + Size_2_MB = 2 * Size_1_MB, + Size_128_MB = 128 * Size_1_MB, + Size_1_GB = 0x40000000, + Size_2_GB = 2 * Size_1_GB, + Size_4_GB = 4 * Size_1_GB, + Size_6_GB = 6 * Size_1_GB, + Size_64_GB = 64 * Size_1_GB, + Size_512_GB = 512 * Size_1_GB, + Invalid = std::numeric_limits<u64>::max(), +}; + +// clang-format off +constexpr std::array<AddressSpaceInfo, 13> AddressSpaceInfos{{ + { .bit_width = 32, .address = Size_2_MB , .size = Size_1_GB - Size_2_MB , .type = AddressSpaceInfo::Type::Is32Bit, }, + { .bit_width = 32, .address = Size_1_GB , .size = Size_4_GB - Size_1_GB , .type = AddressSpaceInfo::Type::Small64Bit, }, + { .bit_width = 32, .address = Invalid , .size = Size_1_GB , .type = AddressSpaceInfo::Type::Heap, }, + { .bit_width = 32, .address = Invalid , .size = Size_1_GB , .type = AddressSpaceInfo::Type::Alias, }, + { .bit_width = 36, .address = Size_128_MB, .size = Size_2_GB - Size_128_MB, .type = AddressSpaceInfo::Type::Is32Bit, }, + { .bit_width = 36, .address = Size_2_GB , .size = Size_64_GB - Size_2_GB , .type = AddressSpaceInfo::Type::Small64Bit, }, + { .bit_width = 36, .address = Invalid , .size = Size_6_GB , .type = AddressSpaceInfo::Type::Heap, }, + { .bit_width = 36, .address = Invalid , .size = Size_6_GB , .type = AddressSpaceInfo::Type::Alias, }, + { .bit_width = 39, .address = Size_128_MB, .size = Size_512_GB - Size_128_MB, .type = AddressSpaceInfo::Type::Large64Bit, }, + { .bit_width = 39, .address = Invalid , .size = Size_64_GB , .type = AddressSpaceInfo::Type::Is32Bit }, + { .bit_width = 39, .address = Invalid , .size = Size_6_GB , .type = AddressSpaceInfo::Type::Heap, }, + { .bit_width = 39, .address = Invalid , .size = Size_64_GB , .type = AddressSpaceInfo::Type::Alias, }, + { .bit_width = 39, .address = Invalid , .size = Size_2_GB , .type = AddressSpaceInfo::Type::Stack, }, +}}; +// clang-format on + +constexpr bool IsAllowedIndexForAddress(std::size_t index) { + return index < AddressSpaceInfos.size() && AddressSpaceInfos[index].address != Invalid; +} + +using IndexArray = std::array<std::size_t, static_cast<std::size_t>(AddressSpaceInfo::Type::Count)>; + +constexpr IndexArray AddressSpaceIndices32Bit{ + 0, 1, 0, 2, 0, 3, +}; + +constexpr IndexArray AddressSpaceIndices36Bit{ + 4, 5, 4, 6, 4, 7, +}; + +constexpr IndexArray AddressSpaceIndices39Bit{ + 9, 8, 8, 10, 12, 11, +}; + +constexpr bool IsAllowed32BitType(AddressSpaceInfo::Type type) { + return type < AddressSpaceInfo::Type::Count && type != AddressSpaceInfo::Type::Large64Bit && + type != AddressSpaceInfo::Type::Stack; +} + +constexpr bool IsAllowed36BitType(AddressSpaceInfo::Type type) { + return type < AddressSpaceInfo::Type::Count && type != AddressSpaceInfo::Type::Large64Bit && + type != AddressSpaceInfo::Type::Stack; +} + +constexpr bool IsAllowed39BitType(AddressSpaceInfo::Type type) { + return type < AddressSpaceInfo::Type::Count && type != AddressSpaceInfo::Type::Small64Bit; +} + +} // namespace + +u64 AddressSpaceInfo::GetAddressSpaceStart(std::size_t width, Type type) { + const std::size_t index{static_cast<std::size_t>(type)}; + switch (width) { + case 32: + ASSERT(IsAllowed32BitType(type)); + ASSERT(IsAllowedIndexForAddress(AddressSpaceIndices32Bit[index])); + return AddressSpaceInfos[AddressSpaceIndices32Bit[index]].address; + case 36: + ASSERT(IsAllowed36BitType(type)); + ASSERT(IsAllowedIndexForAddress(AddressSpaceIndices36Bit[index])); + return AddressSpaceInfos[AddressSpaceIndices36Bit[index]].address; + case 39: + ASSERT(IsAllowed39BitType(type)); + ASSERT(IsAllowedIndexForAddress(AddressSpaceIndices39Bit[index])); + return AddressSpaceInfos[AddressSpaceIndices39Bit[index]].address; + } + UNREACHABLE(); +} + +std::size_t AddressSpaceInfo::GetAddressSpaceSize(std::size_t width, Type type) { + const std::size_t index{static_cast<std::size_t>(type)}; + switch (width) { + case 32: + ASSERT(IsAllowed32BitType(type)); + return AddressSpaceInfos[AddressSpaceIndices32Bit[index]].size; + case 36: + ASSERT(IsAllowed36BitType(type)); + return AddressSpaceInfos[AddressSpaceIndices36Bit[index]].size; + case 39: + ASSERT(IsAllowed39BitType(type)); + return AddressSpaceInfos[AddressSpaceIndices39Bit[index]].size; + } + UNREACHABLE(); +} + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/address_space_info.h b/src/core/hle/kernel/memory/address_space_info.h new file mode 100644 index 000000000..a4e6e91e5 --- /dev/null +++ b/src/core/hle/kernel/memory/address_space_info.h @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include "common/common_types.h" + +namespace Kernel::Memory { + +struct AddressSpaceInfo final { + enum class Type : u32 { + Is32Bit = 0, + Small64Bit = 1, + Large64Bit = 2, + Heap = 3, + Stack = 4, + Alias = 5, + Count, + }; + + static u64 GetAddressSpaceStart(std::size_t width, Type type); + static std::size_t GetAddressSpaceSize(std::size_t width, Type type); + + const std::size_t bit_width{}; + const std::size_t address{}; + const std::size_t size{}; + const Type type{}; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_block.h b/src/core/hle/kernel/memory/memory_block.h new file mode 100644 index 000000000..9d7839d08 --- /dev/null +++ b/src/core/hle/kernel/memory/memory_block.h @@ -0,0 +1,335 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/common_types.h" +#include "core/hle/kernel/memory/memory_types.h" +#include "core/hle/kernel/svc_types.h" + +namespace Kernel::Memory { + +enum class MemoryState : u32 { + None = 0, + Mask = 0xFF, + All = ~None, + + FlagCanReprotect = (1 << 8), + FlagCanDebug = (1 << 9), + FlagCanUseIpc = (1 << 10), + FlagCanUseNonDeviceIpc = (1 << 11), + FlagCanUseNonSecureIpc = (1 << 12), + FlagMapped = (1 << 13), + FlagCode = (1 << 14), + FlagCanAlias = (1 << 15), + FlagCanCodeAlias = (1 << 16), + FlagCanTransfer = (1 << 17), + FlagCanQueryPhysical = (1 << 18), + FlagCanDeviceMap = (1 << 19), + FlagCanAlignedDeviceMap = (1 << 20), + FlagCanIpcUserBuffer = (1 << 21), + FlagReferenceCounted = (1 << 22), + FlagCanMapProcess = (1 << 23), + FlagCanChangeAttribute = (1 << 24), + FlagCanCodeMemory = (1 << 25), + + FlagsData = FlagCanReprotect | FlagCanUseIpc | FlagCanUseNonDeviceIpc | FlagCanUseNonSecureIpc | + FlagMapped | FlagCanAlias | FlagCanTransfer | FlagCanQueryPhysical | + FlagCanDeviceMap | FlagCanAlignedDeviceMap | FlagCanIpcUserBuffer | + FlagReferenceCounted | FlagCanChangeAttribute, + + FlagsCode = FlagCanDebug | FlagCanUseIpc | FlagCanUseNonDeviceIpc | FlagCanUseNonSecureIpc | + FlagMapped | FlagCode | FlagCanQueryPhysical | FlagCanDeviceMap | + FlagCanAlignedDeviceMap | FlagReferenceCounted, + + FlagsMisc = FlagMapped | FlagReferenceCounted | FlagCanQueryPhysical | FlagCanDeviceMap, + + Free = static_cast<u32>(Svc::MemoryState::Free), + Io = static_cast<u32>(Svc::MemoryState::Io) | FlagMapped, + Static = static_cast<u32>(Svc::MemoryState::Static) | FlagMapped | FlagCanQueryPhysical, + Code = static_cast<u32>(Svc::MemoryState::Code) | FlagsCode | FlagCanMapProcess, + CodeData = static_cast<u32>(Svc::MemoryState::CodeData) | FlagsData | FlagCanMapProcess | + FlagCanCodeMemory, + Shared = static_cast<u32>(Svc::MemoryState::Shared) | FlagMapped | FlagReferenceCounted, + Normal = static_cast<u32>(Svc::MemoryState::Normal) | FlagsData | FlagCanCodeMemory, + + AliasCode = static_cast<u32>(Svc::MemoryState::AliasCode) | FlagsCode | FlagCanMapProcess | + FlagCanCodeAlias, + AliasCodeData = static_cast<u32>(Svc::MemoryState::AliasCodeData) | FlagsData | + FlagCanMapProcess | FlagCanCodeAlias | FlagCanCodeMemory, + + Ipc = static_cast<u32>(Svc::MemoryState::Ipc) | FlagsMisc | FlagCanAlignedDeviceMap | + FlagCanUseIpc | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + + Stack = static_cast<u32>(Svc::MemoryState::Stack) | FlagsMisc | FlagCanAlignedDeviceMap | + FlagCanUseIpc | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + + ThreadLocal = + static_cast<u32>(Svc::MemoryState::ThreadLocal) | FlagMapped | FlagReferenceCounted, + + Transfered = static_cast<u32>(Svc::MemoryState::Transfered) | FlagsMisc | + FlagCanAlignedDeviceMap | FlagCanChangeAttribute | FlagCanUseIpc | + FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + + SharedTransfered = static_cast<u32>(Svc::MemoryState::SharedTransfered) | FlagsMisc | + FlagCanAlignedDeviceMap | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + + SharedCode = static_cast<u32>(Svc::MemoryState::SharedCode) | FlagMapped | + FlagReferenceCounted | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + + Inaccessible = static_cast<u32>(Svc::MemoryState::Inaccessible), + + NonSecureIpc = static_cast<u32>(Svc::MemoryState::NonSecureIpc) | FlagsMisc | + FlagCanAlignedDeviceMap | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + + NonDeviceIpc = + static_cast<u32>(Svc::MemoryState::NonDeviceIpc) | FlagsMisc | FlagCanUseNonDeviceIpc, + + Kernel = static_cast<u32>(Svc::MemoryState::Kernel) | FlagMapped, + + GeneratedCode = static_cast<u32>(Svc::MemoryState::GeneratedCode) | FlagMapped | + FlagReferenceCounted | FlagCanDebug, + CodeOut = static_cast<u32>(Svc::MemoryState::CodeOut) | FlagMapped | FlagReferenceCounted, +}; +DECLARE_ENUM_FLAG_OPERATORS(MemoryState); + +static_assert(static_cast<u32>(MemoryState::Free) == 0x00000000); +static_assert(static_cast<u32>(MemoryState::Io) == 0x00002001); +static_assert(static_cast<u32>(MemoryState::Static) == 0x00042002); +static_assert(static_cast<u32>(MemoryState::Code) == 0x00DC7E03); +static_assert(static_cast<u32>(MemoryState::CodeData) == 0x03FEBD04); +static_assert(static_cast<u32>(MemoryState::Normal) == 0x037EBD05); +static_assert(static_cast<u32>(MemoryState::Shared) == 0x00402006); +static_assert(static_cast<u32>(MemoryState::AliasCode) == 0x00DD7E08); +static_assert(static_cast<u32>(MemoryState::AliasCodeData) == 0x03FFBD09); +static_assert(static_cast<u32>(MemoryState::Ipc) == 0x005C3C0A); +static_assert(static_cast<u32>(MemoryState::Stack) == 0x005C3C0B); +static_assert(static_cast<u32>(MemoryState::ThreadLocal) == 0x0040200C); +static_assert(static_cast<u32>(MemoryState::Transfered) == 0x015C3C0D); +static_assert(static_cast<u32>(MemoryState::SharedTransfered) == 0x005C380E); +static_assert(static_cast<u32>(MemoryState::SharedCode) == 0x0040380F); +static_assert(static_cast<u32>(MemoryState::Inaccessible) == 0x00000010); +static_assert(static_cast<u32>(MemoryState::NonSecureIpc) == 0x005C3811); +static_assert(static_cast<u32>(MemoryState::NonDeviceIpc) == 0x004C2812); +static_assert(static_cast<u32>(MemoryState::Kernel) == 0x00002013); +static_assert(static_cast<u32>(MemoryState::GeneratedCode) == 0x00402214); +static_assert(static_cast<u32>(MemoryState::CodeOut) == 0x00402015); + +enum class MemoryPermission : u8 { + None = 0, + Mask = static_cast<u8>(~None), + + Read = 1 << 0, + Write = 1 << 1, + Execute = 1 << 2, + + ReadAndWrite = Read | Write, + ReadAndExecute = Read | Execute, + + UserMask = static_cast<u8>(Svc::MemoryPermission::Read | Svc::MemoryPermission::Write | + Svc::MemoryPermission::Execute), +}; +DECLARE_ENUM_FLAG_OPERATORS(MemoryPermission); + +enum class MemoryAttribute : u8 { + None = 0x00, + Mask = 0x7F, + All = Mask, + DontCareMask = 0x80, + + Locked = static_cast<u8>(Svc::MemoryAttribute::Locked), + IpcLocked = static_cast<u8>(Svc::MemoryAttribute::IpcLocked), + DeviceShared = static_cast<u8>(Svc::MemoryAttribute::DeviceShared), + Uncached = static_cast<u8>(Svc::MemoryAttribute::Uncached), + + IpcAndDeviceMapped = IpcLocked | DeviceShared, + LockedAndIpcLocked = Locked | IpcLocked, + DeviceSharedAndUncached = DeviceShared | Uncached +}; +DECLARE_ENUM_FLAG_OPERATORS(MemoryAttribute); + +static_assert((static_cast<u8>(MemoryAttribute::Mask) & + static_cast<u8>(MemoryAttribute::DontCareMask)) == 0); + +struct MemoryInfo { + VAddr addr{}; + std::size_t size{}; + MemoryState state{}; + MemoryPermission perm{}; + MemoryAttribute attribute{}; + MemoryPermission original_perm{}; + u16 ipc_lock_count{}; + u16 device_use_count{}; + + constexpr Svc::MemoryInfo GetSvcMemoryInfo() const { + return { + addr, + size, + static_cast<Svc::MemoryState>(state & MemoryState::Mask), + static_cast<Svc::MemoryAttribute>(attribute & MemoryAttribute::Mask), + static_cast<Svc::MemoryPermission>(perm & MemoryPermission::UserMask), + ipc_lock_count, + device_use_count, + }; + } + + constexpr VAddr GetAddress() const { + return addr; + } + constexpr std::size_t GetSize() const { + return size; + } + constexpr std::size_t GetNumPages() const { + return GetSize() / PageSize; + } + constexpr VAddr GetEndAddress() const { + return GetAddress() + GetSize(); + } + constexpr VAddr GetLastAddress() const { + return GetEndAddress() - 1; + } +}; + +class MemoryBlock final { + friend class MemoryBlockManager; + +private: + VAddr addr{}; + std::size_t num_pages{}; + MemoryState state{MemoryState::None}; + u16 ipc_lock_count{}; + u16 device_use_count{}; + MemoryPermission perm{MemoryPermission::None}; + MemoryPermission original_perm{MemoryPermission::None}; + MemoryAttribute attribute{MemoryAttribute::None}; + +public: + static constexpr int Compare(const MemoryBlock& lhs, const MemoryBlock& rhs) { + if (lhs.GetAddress() < rhs.GetAddress()) { + return -1; + } else if (lhs.GetAddress() <= rhs.GetLastAddress()) { + return 0; + } else { + return 1; + } + } + +public: + constexpr MemoryBlock() = default; + constexpr MemoryBlock(VAddr addr, std::size_t num_pages, MemoryState state, + MemoryPermission perm, MemoryAttribute attribute) + : addr{addr}, num_pages(num_pages), state{state}, perm{perm}, attribute{attribute} {} + + constexpr VAddr GetAddress() const { + return addr; + } + + constexpr std::size_t GetNumPages() const { + return num_pages; + } + + constexpr std::size_t GetSize() const { + return GetNumPages() * PageSize; + } + + constexpr VAddr GetEndAddress() const { + return GetAddress() + GetSize(); + } + + constexpr VAddr GetLastAddress() const { + return GetEndAddress() - 1; + } + + constexpr MemoryInfo GetMemoryInfo() const { + return { + GetAddress(), GetSize(), state, perm, + attribute, original_perm, ipc_lock_count, device_use_count, + }; + } + + void ShareToDevice(MemoryPermission /*new_perm*/) { + ASSERT((attribute & MemoryAttribute::DeviceShared) == MemoryAttribute::DeviceShared || + device_use_count == 0); + attribute |= MemoryAttribute::DeviceShared; + const u16 new_use_count{++device_use_count}; + ASSERT(new_use_count > 0); + } + + void UnshareToDevice(MemoryPermission /*new_perm*/) { + ASSERT((attribute & MemoryAttribute::DeviceShared) == MemoryAttribute::DeviceShared); + const u16 prev_use_count{device_use_count--}; + ASSERT(prev_use_count > 0); + if (prev_use_count == 1) { + attribute &= ~MemoryAttribute::DeviceShared; + } + } + +private: + constexpr bool HasProperties(MemoryState s, MemoryPermission p, MemoryAttribute a) const { + constexpr MemoryAttribute AttributeIgnoreMask{MemoryAttribute::DontCareMask | + MemoryAttribute::IpcLocked | + MemoryAttribute::DeviceShared}; + return state == s && perm == p && + (attribute | AttributeIgnoreMask) == (a | AttributeIgnoreMask); + } + + constexpr bool HasSameProperties(const MemoryBlock& rhs) const { + return state == rhs.state && perm == rhs.perm && original_perm == rhs.original_perm && + attribute == rhs.attribute && ipc_lock_count == rhs.ipc_lock_count && + device_use_count == rhs.device_use_count; + } + + constexpr bool Contains(VAddr start) const { + return GetAddress() <= start && start <= GetEndAddress(); + } + + constexpr void Add(std::size_t count) { + ASSERT(count > 0); + ASSERT(GetAddress() + count * PageSize - 1 < GetEndAddress() + count * PageSize - 1); + + num_pages += count; + } + + constexpr void Update(MemoryState new_state, MemoryPermission new_perm, + MemoryAttribute new_attribute) { + ASSERT(original_perm == MemoryPermission::None); + ASSERT((attribute & MemoryAttribute::IpcLocked) == MemoryAttribute::None); + + state = new_state; + perm = new_perm; + + attribute = static_cast<MemoryAttribute>( + new_attribute | + (attribute & (MemoryAttribute::IpcLocked | MemoryAttribute::DeviceShared))); + } + + constexpr MemoryBlock Split(VAddr split_addr) { + ASSERT(GetAddress() < split_addr); + ASSERT(Contains(split_addr)); + ASSERT(Common::IsAligned(split_addr, PageSize)); + + MemoryBlock block; + block.addr = addr; + block.num_pages = (split_addr - GetAddress()) / PageSize; + block.state = state; + block.ipc_lock_count = ipc_lock_count; + block.device_use_count = device_use_count; + block.perm = perm; + block.original_perm = original_perm; + block.attribute = attribute; + + addr = split_addr; + num_pages -= block.num_pages; + + return block; + } +}; +static_assert(std::is_trivially_destructible<MemoryBlock>::value); + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_block_manager.cpp b/src/core/hle/kernel/memory/memory_block_manager.cpp new file mode 100644 index 000000000..0732fa5a1 --- /dev/null +++ b/src/core/hle/kernel/memory/memory_block_manager.cpp @@ -0,0 +1,223 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/hle/kernel/memory/memory_block_manager.h" +#include "core/hle/kernel/memory/memory_types.h" + +namespace Kernel::Memory { + +MemoryBlockManager::MemoryBlockManager(VAddr start_addr, VAddr end_addr) + : start_addr{start_addr}, end_addr{end_addr} { + const u64 num_pages{(end_addr - start_addr) / PageSize}; + memory_block_tree.emplace_back(start_addr, num_pages, MemoryState::Free, MemoryPermission::None, + MemoryAttribute::None); +} + +MemoryBlockManager::iterator MemoryBlockManager::FindIterator(VAddr addr) { + auto node{memory_block_tree.begin()}; + while (node != end()) { + const VAddr end_addr{node->GetNumPages() * PageSize + node->GetAddress()}; + if (node->GetAddress() <= addr && end_addr - 1 >= addr) { + return node; + } + node = std::next(node); + } + return end(); +} + +VAddr MemoryBlockManager::FindFreeArea(VAddr region_start, std::size_t region_num_pages, + std::size_t num_pages, std::size_t align, std::size_t offset, + std::size_t guard_pages) { + if (num_pages == 0) { + return {}; + } + + const VAddr region_end{region_start + region_num_pages * PageSize}; + const VAddr region_last{region_end - 1}; + for (auto it{FindIterator(region_start)}; it != memory_block_tree.cend(); it++) { + const auto info{it->GetMemoryInfo()}; + if (region_last < info.GetAddress()) { + break; + } + + if (info.state != MemoryState::Free) { + continue; + } + + VAddr area{(info.GetAddress() <= region_start) ? region_start : info.GetAddress()}; + area += guard_pages * PageSize; + + const VAddr offset_area{Common::AlignDown(area, align) + offset}; + area = (area <= offset_area) ? offset_area : offset_area + align; + + const VAddr area_end{area + num_pages * PageSize + guard_pages * PageSize}; + const VAddr area_last{area_end - 1}; + + if (info.GetAddress() <= area && area < area_last && area_last <= region_last && + area_last <= info.GetLastAddress()) { + return area; + } + } + + return {}; +} + +void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState prev_state, + MemoryPermission prev_perm, MemoryAttribute prev_attribute, + MemoryState state, MemoryPermission perm, + MemoryAttribute attribute) { + const VAddr end_addr{addr + num_pages * PageSize}; + iterator node{memory_block_tree.begin()}; + + prev_attribute |= MemoryAttribute::IpcAndDeviceMapped; + + while (node != memory_block_tree.end()) { + MemoryBlock* block{&(*node)}; + iterator next_node{std::next(node)}; + const VAddr cur_addr{block->GetAddress()}; + const VAddr cur_end_addr{block->GetNumPages() * PageSize + cur_addr}; + + if (addr < cur_end_addr && cur_addr < end_addr) { + if (!block->HasProperties(prev_state, prev_perm, prev_attribute)) { + node = next_node; + continue; + } + + iterator new_node{node}; + if (addr > cur_addr) { + memory_block_tree.insert(node, block->Split(addr)); + } + + if (end_addr < cur_end_addr) { + new_node = memory_block_tree.insert(node, block->Split(end_addr)); + } + + new_node->Update(state, perm, attribute); + + MergeAdjacent(new_node, next_node); + } + + if (cur_end_addr - 1 >= end_addr - 1) { + break; + } + + node = next_node; + } +} + +void MemoryBlockManager::Update(VAddr addr, std::size_t num_pages, MemoryState state, + MemoryPermission perm, MemoryAttribute attribute) { + const VAddr end_addr{addr + num_pages * PageSize}; + iterator node{memory_block_tree.begin()}; + + while (node != memory_block_tree.end()) { + MemoryBlock* block{&(*node)}; + iterator next_node{std::next(node)}; + const VAddr cur_addr{block->GetAddress()}; + const VAddr cur_end_addr{block->GetNumPages() * PageSize + cur_addr}; + + if (addr < cur_end_addr && cur_addr < end_addr) { + iterator new_node{node}; + + if (addr > cur_addr) { + memory_block_tree.insert(node, block->Split(addr)); + } + + if (end_addr < cur_end_addr) { + new_node = memory_block_tree.insert(node, block->Split(end_addr)); + } + + new_node->Update(state, perm, attribute); + + MergeAdjacent(new_node, next_node); + } + + if (cur_end_addr - 1 >= end_addr - 1) { + break; + } + + node = next_node; + } +} + +void MemoryBlockManager::UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func, + MemoryPermission perm) { + const VAddr end_addr{addr + num_pages * PageSize}; + iterator node{memory_block_tree.begin()}; + + while (node != memory_block_tree.end()) { + MemoryBlock* block{&(*node)}; + iterator next_node{std::next(node)}; + const VAddr cur_addr{block->GetAddress()}; + const VAddr cur_end_addr{block->GetNumPages() * PageSize + cur_addr}; + + if (addr < cur_end_addr && cur_addr < end_addr) { + iterator new_node{node}; + + if (addr > cur_addr) { + memory_block_tree.insert(node, block->Split(addr)); + } + + if (end_addr < cur_end_addr) { + new_node = memory_block_tree.insert(node, block->Split(end_addr)); + } + + lock_func(new_node, perm); + + MergeAdjacent(new_node, next_node); + } + + if (cur_end_addr - 1 >= end_addr - 1) { + break; + } + + node = next_node; + } +} + +void MemoryBlockManager::IterateForRange(VAddr start, VAddr end, IterateFunc&& func) { + const_iterator it{FindIterator(start)}; + MemoryInfo info{}; + do { + info = it->GetMemoryInfo(); + func(info); + it = std::next(it); + } while (info.addr + info.size - 1 < end - 1 && it != cend()); +} + +void MemoryBlockManager::MergeAdjacent(iterator it, iterator& next_it) { + MemoryBlock* block{&(*it)}; + + auto EraseIt = [&](const iterator it_to_erase) { + if (next_it == it_to_erase) { + next_it = std::next(next_it); + } + memory_block_tree.erase(it_to_erase); + }; + + if (it != memory_block_tree.begin()) { + MemoryBlock* prev{&(*std::prev(it))}; + + if (block->HasSameProperties(*prev)) { + const iterator prev_it{std::prev(it)}; + + prev->Add(block->GetNumPages()); + EraseIt(it); + + it = prev_it; + block = prev; + } + } + + if (it != cend()) { + const MemoryBlock* const next{&(*std::next(it))}; + + if (block->HasSameProperties(*next)) { + block->Add(next->GetNumPages()); + EraseIt(std::next(it)); + } + } +} + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_block_manager.h b/src/core/hle/kernel/memory/memory_block_manager.h new file mode 100644 index 000000000..6e1d41075 --- /dev/null +++ b/src/core/hle/kernel/memory/memory_block_manager.h @@ -0,0 +1,66 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <functional> +#include <list> + +#include "common/common_types.h" +#include "core/hle/kernel/memory/memory_block.h" + +namespace Kernel::Memory { + +class MemoryBlockManager final { +public: + using MemoryBlockTree = std::list<MemoryBlock>; + using iterator = MemoryBlockTree::iterator; + using const_iterator = MemoryBlockTree::const_iterator; + +public: + MemoryBlockManager(VAddr start_addr, VAddr end_addr); + + iterator end() { + return memory_block_tree.end(); + } + const_iterator end() const { + return memory_block_tree.end(); + } + const_iterator cend() const { + return memory_block_tree.cend(); + } + + iterator FindIterator(VAddr addr); + + VAddr FindFreeArea(VAddr region_start, std::size_t region_num_pages, std::size_t num_pages, + std::size_t align, std::size_t offset, std::size_t guard_pages); + + void Update(VAddr addr, std::size_t num_pages, MemoryState prev_state, + MemoryPermission prev_perm, MemoryAttribute prev_attribute, MemoryState state, + MemoryPermission perm, MemoryAttribute attribute); + + void Update(VAddr addr, std::size_t num_pages, MemoryState state, + MemoryPermission perm = MemoryPermission::None, + MemoryAttribute attribute = MemoryAttribute::None); + + using LockFunc = std::function<void(iterator, MemoryPermission)>; + void UpdateLock(VAddr addr, std::size_t num_pages, LockFunc&& lock_func, MemoryPermission perm); + + using IterateFunc = std::function<void(const MemoryInfo&)>; + void IterateForRange(VAddr start, VAddr end, IterateFunc&& func); + + MemoryBlock& FindBlock(VAddr addr) { + return *FindIterator(addr); + } + +private: + void MergeAdjacent(iterator it, iterator& next_it); + + const VAddr start_addr; + const VAddr end_addr; + + MemoryBlockTree memory_block_tree; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_layout.h b/src/core/hle/kernel/memory/memory_layout.h new file mode 100644 index 000000000..9b3d6267a --- /dev/null +++ b/src/core/hle/kernel/memory/memory_layout.h @@ -0,0 +1,71 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace Kernel::Memory { + +class MemoryRegion final { + friend class MemoryLayout; + +public: + constexpr PAddr StartAddress() const { + return start_address; + } + + constexpr PAddr EndAddress() const { + return end_address; + } + +private: + constexpr MemoryRegion() = default; + constexpr MemoryRegion(PAddr start_address, PAddr end_address) + : start_address{start_address}, end_address{end_address} {} + + const PAddr start_address{}; + const PAddr end_address{}; +}; + +class MemoryLayout final { +public: + constexpr const MemoryRegion& Application() const { + return application; + } + + constexpr const MemoryRegion& Applet() const { + return applet; + } + + constexpr const MemoryRegion& System() const { + return system; + } + + static constexpr MemoryLayout GetDefaultLayout() { + constexpr std::size_t application_size{0xcd500000}; + constexpr std::size_t applet_size{0x1fb00000}; + constexpr PAddr application_start_address{Core::DramMemoryMap::End - application_size}; + constexpr PAddr application_end_address{Core::DramMemoryMap::End}; + constexpr PAddr applet_start_address{application_start_address - applet_size}; + constexpr PAddr applet_end_address{applet_start_address + applet_size}; + constexpr PAddr system_start_address{Core::DramMemoryMap::SlabHeapEnd}; + constexpr PAddr system_end_address{applet_start_address}; + return {application_start_address, application_end_address, applet_start_address, + applet_end_address, system_start_address, system_end_address}; + } + +private: + constexpr MemoryLayout(PAddr application_start_address, std::size_t application_size, + PAddr applet_start_address, std::size_t applet_size, + PAddr system_start_address, std::size_t system_size) + : application{application_start_address, application_size}, + applet{applet_start_address, applet_size}, system{system_start_address, system_size} {} + + const MemoryRegion application; + const MemoryRegion applet; + const MemoryRegion system; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_manager.cpp b/src/core/hle/kernel/memory/memory_manager.cpp new file mode 100644 index 000000000..acf13585c --- /dev/null +++ b/src/core/hle/kernel/memory/memory_manager.cpp @@ -0,0 +1,175 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/common_types.h" +#include "common/scope_exit.h" +#include "core/hle/kernel/errors.h" +#include "core/hle/kernel/memory/memory_manager.h" +#include "core/hle/kernel/memory/page_linked_list.h" + +namespace Kernel::Memory { + +std::size_t MemoryManager::Impl::Initialize(Pool new_pool, u64 start_address, u64 end_address) { + const auto size{end_address - start_address}; + + // Calculate metadata sizes + const auto ref_count_size{(size / PageSize) * sizeof(u16)}; + const auto optimize_map_size{(Common::AlignUp((size / PageSize), 64) / 64) * sizeof(u64)}; + const auto manager_size{Common::AlignUp(optimize_map_size + ref_count_size, PageSize)}; + const auto page_heap_size{PageHeap::CalculateMetadataOverheadSize(size)}; + const auto total_metadata_size{manager_size + page_heap_size}; + ASSERT(manager_size <= total_metadata_size); + ASSERT(Common::IsAligned(total_metadata_size, PageSize)); + + // Setup region + pool = new_pool; + + // Initialize the manager's KPageHeap + heap.Initialize(start_address, size, page_heap_size); + + // Free the memory to the heap + heap.Free(start_address, size / PageSize); + + // Update the heap's used size + heap.UpdateUsedSize(); + + return total_metadata_size; +} + +void MemoryManager::InitializeManager(Pool pool, u64 start_address, u64 end_address) { + ASSERT(pool < Pool::Count); + managers[static_cast<std::size_t>(pool)].Initialize(pool, start_address, end_address); +} + +VAddr MemoryManager::AllocateContinuous(std::size_t num_pages, std::size_t align_pages, Pool pool, + Direction dir) { + // Early return if we're allocating no pages + if (num_pages == 0) { + return {}; + } + + // Lock the pool that we're allocating from + const auto pool_index{static_cast<std::size_t>(pool)}; + std::lock_guard lock{pool_locks[pool_index]}; + + // Choose a heap based on our page size request + const s32 heap_index{PageHeap::GetAlignedBlockIndex(num_pages, align_pages)}; + + // Loop, trying to iterate from each block + // TODO (bunnei): Support multiple managers + Impl& chosen_manager{managers[pool_index]}; + VAddr allocated_block{chosen_manager.AllocateBlock(heap_index)}; + + // If we failed to allocate, quit now + if (!allocated_block) { + return {}; + } + + // If we allocated more than we need, free some + const auto allocated_pages{PageHeap::GetBlockNumPages(heap_index)}; + if (allocated_pages > num_pages) { + chosen_manager.Free(allocated_block + num_pages * PageSize, allocated_pages - num_pages); + } + + return allocated_block; +} + +ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pages, Pool pool, + Direction dir) { + ASSERT(page_list.GetNumPages() == 0); + + // Early return if we're allocating no pages + if (num_pages == 0) { + return RESULT_SUCCESS; + } + + // Lock the pool that we're allocating from + const auto pool_index{static_cast<std::size_t>(pool)}; + std::lock_guard lock{pool_locks[pool_index]}; + + // Choose a heap based on our page size request + const s32 heap_index{PageHeap::GetBlockIndex(num_pages)}; + if (heap_index < 0) { + return ERR_OUT_OF_MEMORY; + } + + // TODO (bunnei): Support multiple managers + Impl& chosen_manager{managers[pool_index]}; + + // Ensure that we don't leave anything un-freed + auto group_guard = detail::ScopeExit([&] { + for (const auto& it : page_list.Nodes()) { + const auto min_num_pages{std::min<size_t>( + it.GetNumPages(), (chosen_manager.GetEndAddress() - it.GetAddress()) / PageSize)}; + chosen_manager.Free(it.GetAddress(), min_num_pages); + } + }); + + // Keep allocating until we've allocated all our pages + for (s32 index{heap_index}; index >= 0 && num_pages > 0; index--) { + const auto pages_per_alloc{PageHeap::GetBlockNumPages(index)}; + + while (num_pages >= pages_per_alloc) { + // Allocate a block + VAddr allocated_block{chosen_manager.AllocateBlock(index)}; + if (!allocated_block) { + break; + } + + // Safely add it to our group + { + auto block_guard = detail::ScopeExit( + [&] { chosen_manager.Free(allocated_block, pages_per_alloc); }); + + if (const ResultCode result{page_list.AddBlock(allocated_block, pages_per_alloc)}; + result.IsError()) { + return result; + } + + block_guard.Cancel(); + } + + num_pages -= pages_per_alloc; + } + } + + // Only succeed if we allocated as many pages as we wanted + if (num_pages) { + return ERR_OUT_OF_MEMORY; + } + + // We succeeded! + group_guard.Cancel(); + return RESULT_SUCCESS; +} + +ResultCode MemoryManager::Free(PageLinkedList& page_list, std::size_t num_pages, Pool pool, + Direction dir) { + // Early return if we're freeing no pages + if (!num_pages) { + return RESULT_SUCCESS; + } + + // Lock the pool that we're freeing from + const auto pool_index{static_cast<std::size_t>(pool)}; + std::lock_guard lock{pool_locks[pool_index]}; + + // TODO (bunnei): Support multiple managers + Impl& chosen_manager{managers[pool_index]}; + + // Free all of the pages + for (const auto& it : page_list.Nodes()) { + const auto min_num_pages{std::min<size_t>( + it.GetNumPages(), (chosen_manager.GetEndAddress() - it.GetAddress()) / PageSize)}; + chosen_manager.Free(it.GetAddress(), min_num_pages); + } + + return RESULT_SUCCESS; +} + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_manager.h b/src/core/hle/kernel/memory/memory_manager.h new file mode 100644 index 000000000..3cf444857 --- /dev/null +++ b/src/core/hle/kernel/memory/memory_manager.h @@ -0,0 +1,96 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <mutex> + +#include "common/common_types.h" +#include "core/hle/kernel/memory/page_heap.h" +#include "core/hle/result.h" + +namespace Kernel::Memory { + +class PageLinkedList; + +class MemoryManager final : NonCopyable { +public: + enum class Pool : u32 { + Application = 0, + Applet = 1, + System = 2, + SystemNonSecure = 3, + + Count, + + Shift = 4, + Mask = (0xF << Shift), + }; + + enum class Direction : u32 { + FromFront = 0, + FromBack = 1, + + Shift = 0, + Mask = (0xF << Shift), + }; + + MemoryManager() = default; + + constexpr std::size_t GetSize(Pool pool) const { + return managers[static_cast<std::size_t>(pool)].GetSize(); + } + + void InitializeManager(Pool pool, u64 start_address, u64 end_address); + VAddr AllocateContinuous(std::size_t num_pages, std::size_t align_pages, Pool pool, + Direction dir = Direction::FromFront); + ResultCode Allocate(PageLinkedList& page_list, std::size_t num_pages, Pool pool, + Direction dir = Direction::FromFront); + ResultCode Free(PageLinkedList& page_list, std::size_t num_pages, Pool pool, + Direction dir = Direction::FromFront); + + static constexpr std::size_t MaxManagerCount = 10; + +private: + class Impl final : NonCopyable { + private: + using RefCount = u16; + + private: + PageHeap heap; + Pool pool{}; + + public: + Impl() = default; + + std::size_t Initialize(Pool new_pool, u64 start_address, u64 end_address); + + VAddr AllocateBlock(s32 index) { + return heap.AllocateBlock(index); + } + + void Free(VAddr addr, std::size_t num_pages) { + heap.Free(addr, num_pages); + } + + constexpr std::size_t GetSize() const { + return heap.GetSize(); + } + + constexpr VAddr GetAddress() const { + return heap.GetAddress(); + } + + constexpr VAddr GetEndAddress() const { + return heap.GetEndAddress(); + } + }; + +private: + std::array<std::mutex, static_cast<std::size_t>(Pool::Count)> pool_locks; + std::array<Impl, MaxManagerCount> managers; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_types.h b/src/core/hle/kernel/memory/memory_types.h new file mode 100644 index 000000000..a75bf77c0 --- /dev/null +++ b/src/core/hle/kernel/memory/memory_types.h @@ -0,0 +1,18 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> + +#include "common/common_types.h" + +namespace Kernel::Memory { + +constexpr std::size_t PageBits{12}; +constexpr std::size_t PageSize{1 << PageBits}; + +using Page = std::array<u8, PageSize>; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/page_heap.cpp b/src/core/hle/kernel/memory/page_heap.cpp new file mode 100644 index 000000000..0ab1f7205 --- /dev/null +++ b/src/core/hle/kernel/memory/page_heap.cpp @@ -0,0 +1,119 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#include "core/core.h" +#include "core/hle/kernel/memory/page_heap.h" +#include "core/memory.h" + +namespace Kernel::Memory { + +void PageHeap::Initialize(VAddr address, std::size_t size, std::size_t metadata_size) { + // Check our assumptions + ASSERT(Common::IsAligned((address), PageSize)); + ASSERT(Common::IsAligned(size, PageSize)); + + // Set our members + heap_address = address; + heap_size = size; + + // Setup bitmaps + metadata.resize(metadata_size / sizeof(u64)); + u64* cur_bitmap_storage{metadata.data()}; + for (std::size_t i = 0; i < MemoryBlockPageShifts.size(); i++) { + const std::size_t cur_block_shift{MemoryBlockPageShifts[i]}; + const std::size_t next_block_shift{ + (i != MemoryBlockPageShifts.size() - 1) ? MemoryBlockPageShifts[i + 1] : 0}; + cur_bitmap_storage = blocks[i].Initialize(heap_address, heap_size, cur_block_shift, + next_block_shift, cur_bitmap_storage); + } +} + +VAddr PageHeap::AllocateBlock(s32 index) { + const std::size_t needed_size{blocks[index].GetSize()}; + + for (s32 i{index}; i < static_cast<s32>(MemoryBlockPageShifts.size()); i++) { + if (const VAddr addr{blocks[i].PopBlock()}; addr) { + if (const std::size_t allocated_size{blocks[i].GetSize()}; + allocated_size > needed_size) { + Free(addr + needed_size, (allocated_size - needed_size) / PageSize); + } + return addr; + } + } + + return 0; +} + +void PageHeap::FreeBlock(VAddr block, s32 index) { + do { + block = blocks[index++].PushBlock(block); + } while (block != 0); +} + +void PageHeap::Free(VAddr addr, std::size_t num_pages) { + // Freeing no pages is a no-op + if (num_pages == 0) { + return; + } + + // Find the largest block size that we can free, and free as many as possible + s32 big_index{static_cast<s32>(MemoryBlockPageShifts.size()) - 1}; + const VAddr start{addr}; + const VAddr end{(num_pages * PageSize) + addr}; + VAddr before_start{start}; + VAddr before_end{start}; + VAddr after_start{end}; + VAddr after_end{end}; + while (big_index >= 0) { + const std::size_t block_size{blocks[big_index].GetSize()}; + const VAddr big_start{Common::AlignUp((start), block_size)}; + const VAddr big_end{Common::AlignDown((end), block_size)}; + if (big_start < big_end) { + // Free as many big blocks as we can + for (auto block{big_start}; block < big_end; block += block_size) { + FreeBlock(block, big_index); + } + before_end = big_start; + after_start = big_end; + break; + } + big_index--; + } + ASSERT(big_index >= 0); + + // Free space before the big blocks + for (s32 i{big_index - 1}; i >= 0; i--) { + const std::size_t block_size{blocks[i].GetSize()}; + while (before_start + block_size <= before_end) { + before_end -= block_size; + FreeBlock(before_end, i); + } + } + + // Free space after the big blocks + for (s32 i{big_index - 1}; i >= 0; i--) { + const std::size_t block_size{blocks[i].GetSize()}; + while (after_start + block_size <= after_end) { + FreeBlock(after_start, i); + after_start += block_size; + } + } +} + +std::size_t PageHeap::CalculateMetadataOverheadSize(std::size_t region_size) { + std::size_t overhead_size = 0; + for (std::size_t i = 0; i < MemoryBlockPageShifts.size(); i++) { + const std::size_t cur_block_shift{MemoryBlockPageShifts[i]}; + const std::size_t next_block_shift{ + (i != MemoryBlockPageShifts.size() - 1) ? MemoryBlockPageShifts[i + 1] : 0}; + overhead_size += PageHeap::Block::CalculateMetadataOverheadSize( + region_size, cur_block_shift, next_block_shift); + } + return Common::AlignUp(overhead_size, PageSize); +} + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/page_heap.h b/src/core/hle/kernel/memory/page_heap.h new file mode 100644 index 000000000..22b0de860 --- /dev/null +++ b/src/core/hle/kernel/memory/page_heap.h @@ -0,0 +1,370 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include <array> +#include <vector> + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/bit_util.h" +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "core/hle/kernel/memory/memory_types.h" + +namespace Kernel::Memory { + +class PageHeap final : NonCopyable { +public: + static constexpr s32 GetAlignedBlockIndex(std::size_t num_pages, std::size_t align_pages) { + const auto target_pages{std::max(num_pages, align_pages)}; + for (std::size_t i = 0; i < NumMemoryBlockPageShifts; i++) { + if (target_pages <= + (static_cast<std::size_t>(1) << MemoryBlockPageShifts[i]) / PageSize) { + return static_cast<s32>(i); + } + } + return -1; + } + + static constexpr s32 GetBlockIndex(std::size_t num_pages) { + for (s32 i{static_cast<s32>(NumMemoryBlockPageShifts) - 1}; i >= 0; i--) { + if (num_pages >= (static_cast<std::size_t>(1) << MemoryBlockPageShifts[i]) / PageSize) { + return i; + } + } + return -1; + } + + static constexpr std::size_t GetBlockSize(std::size_t index) { + return static_cast<std::size_t>(1) << MemoryBlockPageShifts[index]; + } + + static constexpr std::size_t GetBlockNumPages(std::size_t index) { + return GetBlockSize(index) / PageSize; + } + +private: + static constexpr std::size_t NumMemoryBlockPageShifts{7}; + static constexpr std::array<std::size_t, NumMemoryBlockPageShifts> MemoryBlockPageShifts{ + 0xC, 0x10, 0x15, 0x16, 0x19, 0x1D, 0x1E, + }; + + class Block final : NonCopyable { + private: + class Bitmap final : NonCopyable { + public: + static constexpr std::size_t MaxDepth{4}; + + private: + std::array<u64*, MaxDepth> bit_storages{}; + std::size_t num_bits{}; + std::size_t used_depths{}; + + public: + constexpr Bitmap() = default; + + constexpr std::size_t GetNumBits() const { + return num_bits; + } + constexpr s32 GetHighestDepthIndex() const { + return static_cast<s32>(used_depths) - 1; + } + + constexpr u64* Initialize(u64* storage, std::size_t size) { + //* Initially, everything is un-set + num_bits = 0; + + // Calculate the needed bitmap depth + used_depths = static_cast<std::size_t>(GetRequiredDepth(size)); + ASSERT(used_depths <= MaxDepth); + + // Set the bitmap pointers + for (s32 depth{GetHighestDepthIndex()}; depth >= 0; depth--) { + bit_storages[depth] = storage; + size = Common::AlignUp(size, 64) / 64; + storage += size; + } + + return storage; + } + + s64 FindFreeBlock() const { + uintptr_t offset{}; + s32 depth{}; + + do { + const u64 v{bit_storages[depth][offset]}; + if (v == 0) { + // Non-zero depth indicates that a previous level had a free block + ASSERT(depth == 0); + return -1; + } + offset = offset * 64 + Common::CountTrailingZeroes64(v); + ++depth; + } while (depth < static_cast<s32>(used_depths)); + + return static_cast<s64>(offset); + } + + constexpr void SetBit(std::size_t offset) { + SetBit(GetHighestDepthIndex(), offset); + num_bits++; + } + + constexpr void ClearBit(std::size_t offset) { + ClearBit(GetHighestDepthIndex(), offset); + num_bits--; + } + + constexpr bool ClearRange(std::size_t offset, std::size_t count) { + const s32 depth{GetHighestDepthIndex()}; + const auto bit_ind{offset / 64}; + u64* bits{bit_storages[depth]}; + if (count < 64) { + const auto shift{offset % 64}; + ASSERT(shift + count <= 64); + // Check that all the bits are set + const u64 mask{((1ULL << count) - 1) << shift}; + u64 v{bits[bit_ind]}; + if ((v & mask) != mask) { + return false; + } + + // Clear the bits + v &= ~mask; + bits[bit_ind] = v; + if (v == 0) { + ClearBit(depth - 1, bit_ind); + } + } else { + ASSERT(offset % 64 == 0); + ASSERT(count % 64 == 0); + // Check that all the bits are set + std::size_t remaining{count}; + std::size_t i = 0; + do { + if (bits[bit_ind + i++] != ~u64(0)) { + return false; + } + remaining -= 64; + } while (remaining > 0); + + // Clear the bits + remaining = count; + i = 0; + do { + bits[bit_ind + i] = 0; + ClearBit(depth - 1, bit_ind + i); + i++; + remaining -= 64; + } while (remaining > 0); + } + + num_bits -= count; + return true; + } + + private: + constexpr void SetBit(s32 depth, std::size_t offset) { + while (depth >= 0) { + const auto ind{offset / 64}; + const auto which{offset % 64}; + const u64 mask{1ULL << which}; + + u64* bit{std::addressof(bit_storages[depth][ind])}; + const u64 v{*bit}; + ASSERT((v & mask) == 0); + *bit = v | mask; + if (v) { + break; + } + offset = ind; + depth--; + } + } + + constexpr void ClearBit(s32 depth, std::size_t offset) { + while (depth >= 0) { + const auto ind{offset / 64}; + const auto which{offset % 64}; + const u64 mask{1ULL << which}; + + u64* bit{std::addressof(bit_storages[depth][ind])}; + u64 v{*bit}; + ASSERT((v & mask) != 0); + v &= ~mask; + *bit = v; + if (v) { + break; + } + offset = ind; + depth--; + } + } + + private: + static constexpr s32 GetRequiredDepth(std::size_t region_size) { + s32 depth = 0; + while (true) { + region_size /= 64; + depth++; + if (region_size == 0) { + return depth; + } + } + } + + public: + static constexpr std::size_t CalculateMetadataOverheadSize(std::size_t region_size) { + std::size_t overhead_bits = 0; + for (s32 depth{GetRequiredDepth(region_size) - 1}; depth >= 0; depth--) { + region_size = Common::AlignUp(region_size, 64) / 64; + overhead_bits += region_size; + } + return overhead_bits * sizeof(u64); + } + }; + + private: + Bitmap bitmap; + VAddr heap_address{}; + uintptr_t end_offset{}; + std::size_t block_shift{}; + std::size_t next_block_shift{}; + + public: + constexpr Block() = default; + + constexpr std::size_t GetShift() const { + return block_shift; + } + constexpr std::size_t GetNextShift() const { + return next_block_shift; + } + constexpr std::size_t GetSize() const { + return static_cast<std::size_t>(1) << GetShift(); + } + constexpr std::size_t GetNumPages() const { + return GetSize() / PageSize; + } + constexpr std::size_t GetNumFreeBlocks() const { + return bitmap.GetNumBits(); + } + constexpr std::size_t GetNumFreePages() const { + return GetNumFreeBlocks() * GetNumPages(); + } + + constexpr u64* Initialize(VAddr addr, std::size_t size, std::size_t bs, std::size_t nbs, + u64* bit_storage) { + // Set shifts + block_shift = bs; + next_block_shift = nbs; + + // Align up the address + VAddr end{addr + size}; + const auto align{(next_block_shift != 0) ? (1ULL << next_block_shift) + : (1ULL << block_shift)}; + addr = Common::AlignDown((addr), align); + end = Common::AlignUp((end), align); + + heap_address = addr; + end_offset = (end - addr) / (1ULL << block_shift); + return bitmap.Initialize(bit_storage, end_offset); + } + + constexpr VAddr PushBlock(VAddr address) { + // Set the bit for the free block + std::size_t offset{(address - heap_address) >> GetShift()}; + bitmap.SetBit(offset); + + // If we have a next shift, try to clear the blocks below and return the address + if (GetNextShift()) { + const auto diff{1ULL << (GetNextShift() - GetShift())}; + offset = Common::AlignDown(offset, diff); + if (bitmap.ClearRange(offset, diff)) { + return heap_address + (offset << GetShift()); + } + } + + // We couldn't coalesce, or we're already as big as possible + return 0; + } + + VAddr PopBlock() { + // Find a free block + const s64 soffset{bitmap.FindFreeBlock()}; + if (soffset < 0) { + return 0; + } + const auto offset{static_cast<std::size_t>(soffset)}; + + // Update our tracking and return it + bitmap.ClearBit(offset); + return heap_address + (offset << GetShift()); + } + + public: + static constexpr std::size_t CalculateMetadataOverheadSize(std::size_t region_size, + std::size_t cur_block_shift, + std::size_t next_block_shift) { + const auto cur_block_size{(1ULL << cur_block_shift)}; + const auto next_block_size{(1ULL << next_block_shift)}; + const auto align{(next_block_shift != 0) ? next_block_size : cur_block_size}; + return Bitmap::CalculateMetadataOverheadSize( + (align * 2 + Common::AlignUp(region_size, align)) / cur_block_size); + } + }; + +public: + PageHeap() = default; + + constexpr VAddr GetAddress() const { + return heap_address; + } + constexpr std::size_t GetSize() const { + return heap_size; + } + constexpr VAddr GetEndAddress() const { + return GetAddress() + GetSize(); + } + constexpr std::size_t GetPageOffset(VAddr block) const { + return (block - GetAddress()) / PageSize; + } + + void Initialize(VAddr heap_address, std::size_t heap_size, std::size_t metadata_size); + VAddr AllocateBlock(s32 index); + void Free(VAddr addr, std::size_t num_pages); + + void UpdateUsedSize() { + used_size = heap_size - (GetNumFreePages() * PageSize); + } + + static std::size_t CalculateMetadataOverheadSize(std::size_t region_size); + +private: + constexpr std::size_t GetNumFreePages() const { + std::size_t num_free{}; + + for (const auto& block : blocks) { + num_free += block.GetNumFreePages(); + } + + return num_free; + } + + void FreeBlock(VAddr block, s32 index); + + VAddr heap_address{}; + std::size_t heap_size{}; + std::size_t used_size{}; + std::array<Block, NumMemoryBlockPageShifts> blocks{}; + std::vector<u64> metadata; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/page_linked_list.h b/src/core/hle/kernel/memory/page_linked_list.h new file mode 100644 index 000000000..45dc13eaf --- /dev/null +++ b/src/core/hle/kernel/memory/page_linked_list.h @@ -0,0 +1,92 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <list> + +#include "common/assert.h" +#include "common/common_types.h" +#include "core/hle/kernel/memory/memory_types.h" +#include "core/hle/result.h" + +namespace Kernel::Memory { + +class PageLinkedList final { +public: + class Node final { + public: + constexpr Node(u64 addr, std::size_t num_pages) : addr{addr}, num_pages{num_pages} {} + + constexpr u64 GetAddress() const { + return addr; + } + + constexpr std::size_t GetNumPages() const { + return num_pages; + } + + private: + u64 addr{}; + std::size_t num_pages{}; + }; + +public: + PageLinkedList() = default; + PageLinkedList(u64 address, u64 num_pages) { + ASSERT(AddBlock(address, num_pages).IsSuccess()); + } + + constexpr std::list<Node>& Nodes() { + return nodes; + } + + constexpr const std::list<Node>& Nodes() const { + return nodes; + } + + std::size_t GetNumPages() const { + std::size_t num_pages = 0; + for (const Node& node : nodes) { + num_pages += node.GetNumPages(); + } + return num_pages; + } + + bool IsEqual(PageLinkedList& other) const { + auto this_node = nodes.begin(); + auto other_node = other.nodes.begin(); + while (this_node != nodes.end() && other_node != other.nodes.end()) { + if (this_node->GetAddress() != other_node->GetAddress() || + this_node->GetNumPages() != other_node->GetNumPages()) { + return false; + } + this_node = std::next(this_node); + other_node = std::next(other_node); + } + + return this_node == nodes.end() && other_node == other.nodes.end(); + } + + ResultCode AddBlock(u64 address, u64 num_pages) { + if (!num_pages) { + return RESULT_SUCCESS; + } + if (!nodes.empty()) { + const auto node = nodes.back(); + if (node.GetAddress() + node.GetNumPages() * PageSize == address) { + address = node.GetAddress(); + num_pages += node.GetNumPages(); + nodes.pop_back(); + } + } + nodes.push_back({address, num_pages}); + return RESULT_SUCCESS; + } + +private: + std::list<Node> nodes; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/page_table.cpp b/src/core/hle/kernel/memory/page_table.cpp new file mode 100644 index 000000000..a3fadb533 --- /dev/null +++ b/src/core/hle/kernel/memory/page_table.cpp @@ -0,0 +1,1174 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/scope_exit.h" +#include "core/core.h" +#include "core/hle/kernel/errors.h" +#include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/address_space_info.h" +#include "core/hle/kernel/memory/memory_block.h" +#include "core/hle/kernel/memory/memory_block_manager.h" +#include "core/hle/kernel/memory/page_linked_list.h" +#include "core/hle/kernel/memory/page_table.h" +#include "core/hle/kernel/memory/system_control.h" +#include "core/hle/kernel/process.h" +#include "core/hle/kernel/resource_limit.h" +#include "core/memory.h" + +namespace Kernel::Memory { + +namespace { + +constexpr std::size_t GetAddressSpaceWidthFromType(FileSys::ProgramAddressSpaceType as_type) { + switch (as_type) { + case FileSys::ProgramAddressSpaceType::Is32Bit: + case FileSys::ProgramAddressSpaceType::Is32BitNoMap: + return 32; + case FileSys::ProgramAddressSpaceType::Is36Bit: + return 36; + case FileSys::ProgramAddressSpaceType::Is39Bit: + return 39; + default: + UNREACHABLE(); + return {}; + } +} + +constexpr u64 GetAddressInRange(const MemoryInfo& info, VAddr addr) { + if (info.GetAddress() < addr) { + return addr; + } + return info.GetAddress(); +} + +constexpr std::size_t GetSizeInRange(const MemoryInfo& info, VAddr start, VAddr end) { + std::size_t size{info.GetSize()}; + if (info.GetAddress() < start) { + size -= start - info.GetAddress(); + } + if (info.GetEndAddress() > end) { + size -= info.GetEndAddress() - end; + } + return size; +} + +} // namespace + +PageTable::PageTable(Core::System& system) : system{system} {} + +ResultCode PageTable::InitializeForProcess(FileSys::ProgramAddressSpaceType as_type, + bool enable_aslr, VAddr code_addr, std::size_t code_size, + Memory::MemoryManager::Pool pool) { + + const auto GetSpaceStart = [this](AddressSpaceInfo::Type type) { + return AddressSpaceInfo::GetAddressSpaceStart(address_space_width, type); + }; + const auto GetSpaceSize = [this](AddressSpaceInfo::Type type) { + return AddressSpaceInfo::GetAddressSpaceSize(address_space_width, type); + }; + + // Set our width and heap/alias sizes + address_space_width = GetAddressSpaceWidthFromType(as_type); + const VAddr start = 0; + const VAddr end{1ULL << address_space_width}; + std::size_t alias_region_size{GetSpaceSize(AddressSpaceInfo::Type::Alias)}; + std::size_t heap_region_size{GetSpaceSize(AddressSpaceInfo::Type::Heap)}; + + ASSERT(start <= code_addr); + ASSERT(code_addr < code_addr + code_size); + ASSERT(code_addr + code_size - 1 <= end - 1); + + // Adjust heap/alias size if we don't have an alias region + if (as_type == FileSys::ProgramAddressSpaceType::Is32BitNoMap) { + heap_region_size += alias_region_size; + alias_region_size = 0; + } + + // Set code regions and determine remaining + constexpr std::size_t RegionAlignment{2 * 1024 * 1024}; + VAddr process_code_start{}; + VAddr process_code_end{}; + std::size_t stack_region_size{}; + std::size_t kernel_map_region_size{}; + + if (address_space_width == 39) { + alias_region_size = GetSpaceSize(AddressSpaceInfo::Type::Alias); + heap_region_size = GetSpaceSize(AddressSpaceInfo::Type::Heap); + stack_region_size = GetSpaceSize(AddressSpaceInfo::Type::Stack); + kernel_map_region_size = GetSpaceSize(AddressSpaceInfo::Type::Is32Bit); + code_region_start = GetSpaceStart(AddressSpaceInfo::Type::Large64Bit); + code_region_end = code_region_start + GetSpaceSize(AddressSpaceInfo::Type::Large64Bit); + alias_code_region_start = code_region_start; + alias_code_region_end = code_region_end; + process_code_start = Common::AlignDown(code_addr, RegionAlignment); + process_code_end = Common::AlignUp(code_addr + code_size, RegionAlignment); + } else { + stack_region_size = 0; + kernel_map_region_size = 0; + code_region_start = GetSpaceStart(AddressSpaceInfo::Type::Is32Bit); + code_region_end = code_region_start + GetSpaceSize(AddressSpaceInfo::Type::Is32Bit); + stack_region_start = code_region_start; + alias_code_region_start = code_region_start; + alias_code_region_end = GetSpaceStart(AddressSpaceInfo::Type::Small64Bit) + + GetSpaceSize(AddressSpaceInfo::Type::Small64Bit); + stack_region_end = code_region_end; + kernel_map_region_start = code_region_start; + kernel_map_region_end = code_region_end; + process_code_start = code_region_start; + process_code_end = code_region_end; + } + + // Set other basic fields + is_aslr_enabled = enable_aslr; + address_space_start = start; + address_space_end = end; + is_kernel = false; + + // Determine the region we can place our undetermineds in + VAddr alloc_start{}; + std::size_t alloc_size{}; + if ((process_code_start - code_region_start) >= (end - process_code_end)) { + alloc_start = code_region_start; + alloc_size = process_code_start - code_region_start; + } else { + alloc_start = process_code_end; + alloc_size = end - process_code_end; + } + const std::size_t needed_size{ + (alias_region_size + heap_region_size + stack_region_size + kernel_map_region_size)}; + if (alloc_size < needed_size) { + UNREACHABLE(); + return ERR_OUT_OF_MEMORY; + } + + const std::size_t remaining_size{alloc_size - needed_size}; + + // Determine random placements for each region + std::size_t alias_rnd{}, heap_rnd{}, stack_rnd{}, kmap_rnd{}; + if (enable_aslr) { + alias_rnd = SystemControl::GenerateRandomRange(0, remaining_size / RegionAlignment) * + RegionAlignment; + heap_rnd = SystemControl::GenerateRandomRange(0, remaining_size / RegionAlignment) * + RegionAlignment; + stack_rnd = SystemControl::GenerateRandomRange(0, remaining_size / RegionAlignment) * + RegionAlignment; + kmap_rnd = SystemControl::GenerateRandomRange(0, remaining_size / RegionAlignment) * + RegionAlignment; + } + + // Setup heap and alias regions + alias_region_start = alloc_start + alias_rnd; + alias_region_end = alias_region_start + alias_region_size; + heap_region_start = alloc_start + heap_rnd; + heap_region_end = heap_region_start + heap_region_size; + + if (alias_rnd <= heap_rnd) { + heap_region_start += alias_region_size; + heap_region_end += alias_region_size; + } else { + alias_region_start += heap_region_size; + alias_region_end += heap_region_size; + } + + // Setup stack region + if (stack_region_size) { + stack_region_start = alloc_start + stack_rnd; + stack_region_end = stack_region_start + stack_region_size; + + if (alias_rnd < stack_rnd) { + stack_region_start += alias_region_size; + stack_region_end += alias_region_size; + } else { + alias_region_start += stack_region_size; + alias_region_end += stack_region_size; + } + + if (heap_rnd < stack_rnd) { + stack_region_start += heap_region_size; + stack_region_end += heap_region_size; + } else { + heap_region_start += stack_region_size; + heap_region_end += stack_region_size; + } + } + + // Setup kernel map region + if (kernel_map_region_size) { + kernel_map_region_start = alloc_start + kmap_rnd; + kernel_map_region_end = kernel_map_region_start + kernel_map_region_size; + + if (alias_rnd < kmap_rnd) { + kernel_map_region_start += alias_region_size; + kernel_map_region_end += alias_region_size; + } else { + alias_region_start += kernel_map_region_size; + alias_region_end += kernel_map_region_size; + } + + if (heap_rnd < kmap_rnd) { + kernel_map_region_start += heap_region_size; + kernel_map_region_end += heap_region_size; + } else { + heap_region_start += kernel_map_region_size; + heap_region_end += kernel_map_region_size; + } + + if (stack_region_size) { + if (stack_rnd < kmap_rnd) { + kernel_map_region_start += stack_region_size; + kernel_map_region_end += stack_region_size; + } else { + stack_region_start += kernel_map_region_size; + stack_region_end += kernel_map_region_size; + } + } + } + + // Set heap members + current_heap_end = heap_region_start; + max_heap_size = 0; + max_physical_memory_size = 0; + + // Ensure that we regions inside our address space + auto IsInAddressSpace = [&](VAddr addr) { + return address_space_start <= addr && addr <= address_space_end; + }; + ASSERT(IsInAddressSpace(alias_region_start)); + ASSERT(IsInAddressSpace(alias_region_end)); + ASSERT(IsInAddressSpace(heap_region_start)); + ASSERT(IsInAddressSpace(heap_region_end)); + ASSERT(IsInAddressSpace(stack_region_start)); + ASSERT(IsInAddressSpace(stack_region_end)); + ASSERT(IsInAddressSpace(kernel_map_region_start)); + ASSERT(IsInAddressSpace(kernel_map_region_end)); + + // Ensure that we selected regions that don't overlap + const VAddr alias_start{alias_region_start}; + const VAddr alias_last{alias_region_end - 1}; + const VAddr heap_start{heap_region_start}; + const VAddr heap_last{heap_region_end - 1}; + const VAddr stack_start{stack_region_start}; + const VAddr stack_last{stack_region_end - 1}; + const VAddr kmap_start{kernel_map_region_start}; + const VAddr kmap_last{kernel_map_region_end - 1}; + ASSERT(alias_last < heap_start || heap_last < alias_start); + ASSERT(alias_last < stack_start || stack_last < alias_start); + ASSERT(alias_last < kmap_start || kmap_last < alias_start); + ASSERT(heap_last < stack_start || stack_last < heap_start); + ASSERT(heap_last < kmap_start || kmap_last < heap_start); + + current_heap_addr = heap_region_start; + heap_capacity = 0; + physical_memory_usage = 0; + memory_pool = pool; + + page_table_impl.Resize(address_space_width, PageBits, true); + + return InitializeMemoryLayout(start, end); +} + +ResultCode PageTable::MapProcessCode(VAddr addr, std::size_t num_pages, MemoryState state, + MemoryPermission perm) { + std::lock_guard lock{page_table_lock}; + + const u64 size{num_pages * PageSize}; + + if (!CanContain(addr, size, state)) { + return ERR_INVALID_ADDRESS_STATE; + } + + if (IsRegionMapped(addr, size)) { + return ERR_INVALID_ADDRESS_STATE; + } + + PageLinkedList page_linked_list; + CASCADE_CODE( + system.Kernel().MemoryManager().Allocate(page_linked_list, num_pages, memory_pool)); + CASCADE_CODE(Operate(addr, num_pages, page_linked_list, OperationType::MapGroup)); + + block_manager->Update(addr, num_pages, state, perm); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::MapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + const std::size_t num_pages{size / PageSize}; + + MemoryState state{}; + MemoryPermission perm{}; + CASCADE_CODE(CheckMemoryState(&state, &perm, nullptr, src_addr, size, MemoryState::All, + MemoryState::Normal, MemoryPermission::Mask, + MemoryPermission::ReadAndWrite, MemoryAttribute::Mask, + MemoryAttribute::None, MemoryAttribute::IpcAndDeviceMapped)); + + if (IsRegionMapped(dst_addr, size)) { + return ERR_INVALID_ADDRESS_STATE; + } + + PageLinkedList page_linked_list; + AddRegionToPages(src_addr, num_pages, page_linked_list); + + { + auto block_guard = detail::ScopeExit( + [&] { Operate(src_addr, num_pages, perm, OperationType::ChangePermissions); }); + + CASCADE_CODE( + Operate(src_addr, num_pages, MemoryPermission::None, OperationType::ChangePermissions)); + CASCADE_CODE(MapPages(dst_addr, page_linked_list, MemoryPermission::None)); + + block_guard.Cancel(); + } + + block_manager->Update(src_addr, num_pages, state, MemoryPermission::None, + MemoryAttribute::Locked); + block_manager->Update(dst_addr, num_pages, MemoryState::AliasCode); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::UnmapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + if (!size) { + return RESULT_SUCCESS; + } + + const std::size_t num_pages{size / PageSize}; + + CASCADE_CODE(CheckMemoryState(nullptr, nullptr, nullptr, src_addr, size, MemoryState::All, + MemoryState::Normal, MemoryPermission::None, + MemoryPermission::None, MemoryAttribute::Mask, + MemoryAttribute::Locked, MemoryAttribute::IpcAndDeviceMapped)); + + MemoryState state{}; + CASCADE_CODE(CheckMemoryState( + &state, nullptr, nullptr, dst_addr, PageSize, MemoryState::FlagCanCodeAlias, + MemoryState::FlagCanCodeAlias, MemoryPermission::None, MemoryPermission::None, + MemoryAttribute::Mask, MemoryAttribute::None, MemoryAttribute::IpcAndDeviceMapped)); + CASCADE_CODE(CheckMemoryState(dst_addr, size, MemoryState::All, state, MemoryPermission::None, + MemoryPermission::None, MemoryAttribute::Mask, + MemoryAttribute::None)); + CASCADE_CODE(Operate(dst_addr, num_pages, MemoryPermission::None, OperationType::Unmap)); + + block_manager->Update(dst_addr, num_pages, MemoryState::Free); + block_manager->Update(src_addr, num_pages, MemoryState::Normal, MemoryPermission::ReadAndWrite); + + return RESULT_SUCCESS; +} + +void PageTable::MapPhysicalMemory(PageLinkedList& page_linked_list, VAddr start, VAddr end) { + auto node{page_linked_list.Nodes().begin()}; + PAddr map_addr{node->GetAddress()}; + std::size_t src_num_pages{node->GetNumPages()}; + + block_manager->IterateForRange(start, end, [&](const MemoryInfo& info) { + if (info.state != MemoryState::Free) { + return; + } + + std::size_t dst_num_pages{GetSizeInRange(info, start, end) / PageSize}; + VAddr dst_addr{GetAddressInRange(info, start)}; + + while (dst_num_pages) { + if (!src_num_pages) { + node = std::next(node); + map_addr = node->GetAddress(); + src_num_pages = node->GetNumPages(); + } + + const std::size_t num_pages{std::min(src_num_pages, dst_num_pages)}; + Operate(dst_addr, num_pages, MemoryPermission::ReadAndWrite, OperationType::Map, + map_addr); + + dst_addr += num_pages * PageSize; + map_addr += num_pages * PageSize; + src_num_pages -= num_pages; + dst_num_pages -= num_pages; + } + }); +} + +ResultCode PageTable::MapPhysicalMemory(VAddr addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + std::size_t mapped_size{}; + const VAddr end_addr{addr + size}; + + block_manager->IterateForRange(addr, end_addr, [&](const MemoryInfo& info) { + if (info.state != MemoryState::Free) { + mapped_size += GetSizeInRange(info, addr, end_addr); + } + }); + + if (mapped_size == size) { + return RESULT_SUCCESS; + } + + auto process{system.Kernel().CurrentProcess()}; + const std::size_t remaining_size{size - mapped_size}; + const std::size_t remaining_pages{remaining_size / PageSize}; + + if (process->GetResourceLimit() && + !process->GetResourceLimit()->Reserve(ResourceType::PhysicalMemory, remaining_size)) { + return ERR_RESOURCE_LIMIT_EXCEEDED; + } + + PageLinkedList page_linked_list; + { + auto block_guard = detail::ScopeExit([&] { + system.Kernel().MemoryManager().Free(page_linked_list, remaining_pages, memory_pool); + process->GetResourceLimit()->Release(ResourceType::PhysicalMemory, remaining_size); + }); + + CASCADE_CODE(system.Kernel().MemoryManager().Allocate(page_linked_list, remaining_pages, + memory_pool)); + + block_guard.Cancel(); + } + + MapPhysicalMemory(page_linked_list, addr, end_addr); + + physical_memory_usage += remaining_size; + + const std::size_t num_pages{size / PageSize}; + block_manager->Update(addr, num_pages, MemoryState::Free, MemoryPermission::None, + MemoryAttribute::None, MemoryState::Normal, + MemoryPermission::ReadAndWrite, MemoryAttribute::None); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::UnmapPhysicalMemory(VAddr addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + const VAddr end_addr{addr + size}; + ResultCode result{RESULT_SUCCESS}; + std::size_t mapped_size{}; + + // Verify that the region can be unmapped + block_manager->IterateForRange(addr, end_addr, [&](const MemoryInfo& info) { + if (info.state == MemoryState::Normal) { + if (info.attribute != MemoryAttribute::None) { + result = ERR_INVALID_ADDRESS_STATE; + return; + } + mapped_size += GetSizeInRange(info, addr, end_addr); + } else if (info.state != MemoryState::Free) { + result = ERR_INVALID_ADDRESS_STATE; + } + }); + + if (result.IsError()) { + return result; + } + + if (!mapped_size) { + return RESULT_SUCCESS; + } + + CASCADE_CODE(UnmapMemory(addr, size)); + + auto process{system.Kernel().CurrentProcess()}; + process->GetResourceLimit()->Release(ResourceType::PhysicalMemory, mapped_size); + physical_memory_usage -= mapped_size; + + return RESULT_SUCCESS; +} + +ResultCode PageTable::UnmapMemory(VAddr addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + const VAddr end_addr{addr + size}; + ResultCode result{RESULT_SUCCESS}; + PageLinkedList page_linked_list; + + // Unmap each region within the range + block_manager->IterateForRange(addr, end_addr, [&](const MemoryInfo& info) { + if (info.state == MemoryState::Normal) { + const std::size_t block_size{GetSizeInRange(info, addr, end_addr)}; + const std::size_t block_num_pages{block_size / PageSize}; + const VAddr block_addr{GetAddressInRange(info, addr)}; + + AddRegionToPages(block_addr, block_size / PageSize, page_linked_list); + + if (result = Operate(block_addr, block_num_pages, MemoryPermission::None, + OperationType::Unmap); + result.IsError()) { + return; + } + } + }); + + if (result.IsError()) { + return result; + } + + const std::size_t num_pages{size / PageSize}; + system.Kernel().MemoryManager().Free(page_linked_list, num_pages, memory_pool); + + block_manager->Update(addr, num_pages, MemoryState::Free); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::Map(VAddr dst_addr, VAddr src_addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + MemoryState src_state{}; + CASCADE_CODE(CheckMemoryState( + &src_state, nullptr, nullptr, src_addr, size, MemoryState::FlagCanAlias, + MemoryState::FlagCanAlias, MemoryPermission::Mask, MemoryPermission::ReadAndWrite, + MemoryAttribute::Mask, MemoryAttribute::None, MemoryAttribute::IpcAndDeviceMapped)); + + if (IsRegionMapped(dst_addr, size)) { + return ERR_INVALID_ADDRESS_STATE; + } + + PageLinkedList page_linked_list; + const std::size_t num_pages{size / PageSize}; + + AddRegionToPages(src_addr, num_pages, page_linked_list); + + { + auto block_guard = detail::ScopeExit([&] { + Operate(src_addr, num_pages, MemoryPermission::ReadAndWrite, + OperationType::ChangePermissions); + }); + + CASCADE_CODE( + Operate(src_addr, num_pages, MemoryPermission::None, OperationType::ChangePermissions)); + CASCADE_CODE(MapPages(dst_addr, page_linked_list, MemoryPermission::ReadAndWrite)); + + block_guard.Cancel(); + } + + block_manager->Update(src_addr, num_pages, src_state, MemoryPermission::None, + MemoryAttribute::Locked); + block_manager->Update(dst_addr, num_pages, MemoryState::Stack, MemoryPermission::ReadAndWrite); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::Unmap(VAddr dst_addr, VAddr src_addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + MemoryState src_state{}; + CASCADE_CODE(CheckMemoryState( + &src_state, nullptr, nullptr, src_addr, size, MemoryState::FlagCanAlias, + MemoryState::FlagCanAlias, MemoryPermission::Mask, MemoryPermission::None, + MemoryAttribute::Mask, MemoryAttribute::Locked, MemoryAttribute::IpcAndDeviceMapped)); + + MemoryPermission dst_perm{}; + CASCADE_CODE(CheckMemoryState(nullptr, &dst_perm, nullptr, dst_addr, size, MemoryState::All, + MemoryState::Stack, MemoryPermission::None, + MemoryPermission::None, MemoryAttribute::Mask, + MemoryAttribute::None, MemoryAttribute::IpcAndDeviceMapped)); + + PageLinkedList src_pages; + PageLinkedList dst_pages; + const std::size_t num_pages{size / PageSize}; + + AddRegionToPages(src_addr, num_pages, src_pages); + AddRegionToPages(dst_addr, num_pages, dst_pages); + + if (!dst_pages.IsEqual(src_pages)) { + return ERR_INVALID_MEMORY_RANGE; + } + + { + auto block_guard = detail::ScopeExit([&] { MapPages(dst_addr, dst_pages, dst_perm); }); + + CASCADE_CODE(Operate(dst_addr, num_pages, MemoryPermission::None, OperationType::Unmap)); + CASCADE_CODE(Operate(src_addr, num_pages, MemoryPermission::ReadAndWrite, + OperationType::ChangePermissions)); + + block_guard.Cancel(); + } + + block_manager->Update(src_addr, num_pages, src_state, MemoryPermission::ReadAndWrite); + block_manager->Update(dst_addr, num_pages, MemoryState::Free); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::MapPages(VAddr addr, const PageLinkedList& page_linked_list, + MemoryPermission perm) { + VAddr cur_addr{addr}; + + for (const auto& node : page_linked_list.Nodes()) { + if (const auto result{ + Operate(cur_addr, node.GetNumPages(), perm, OperationType::Map, node.GetAddress())}; + result.IsError()) { + const std::size_t num_pages{(addr - cur_addr) / PageSize}; + + ASSERT( + Operate(addr, num_pages, MemoryPermission::None, OperationType::Unmap).IsSuccess()); + + return result; + } + + cur_addr += node.GetNumPages() * PageSize; + } + + return RESULT_SUCCESS; +} + +ResultCode PageTable::MapPages(VAddr addr, PageLinkedList& page_linked_list, MemoryState state, + MemoryPermission perm) { + std::lock_guard lock{page_table_lock}; + + const std::size_t num_pages{page_linked_list.GetNumPages()}; + const std::size_t size{num_pages * PageSize}; + + if (!CanContain(addr, size, state)) { + return ERR_INVALID_ADDRESS_STATE; + } + + if (IsRegionMapped(addr, num_pages * PageSize)) { + return ERR_INVALID_ADDRESS_STATE; + } + + CASCADE_CODE(MapPages(addr, page_linked_list, perm)); + + block_manager->Update(addr, num_pages, state, perm); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::SetCodeMemoryPermission(VAddr addr, std::size_t size, MemoryPermission perm) { + + std::lock_guard lock{page_table_lock}; + + MemoryState prev_state{}; + MemoryPermission prev_perm{}; + + CASCADE_CODE(CheckMemoryState( + &prev_state, &prev_perm, nullptr, addr, size, MemoryState::FlagCode, MemoryState::FlagCode, + MemoryPermission::None, MemoryPermission::None, MemoryAttribute::Mask, + MemoryAttribute::None, MemoryAttribute::IpcAndDeviceMapped)); + + MemoryState state{prev_state}; + + // Ensure state is mutable if permission allows write + if ((perm & MemoryPermission::Write) != MemoryPermission::None) { + if (prev_state == MemoryState::Code) { + state = MemoryState::CodeData; + } else if (prev_state == MemoryState::AliasCode) { + state = MemoryState::AliasCodeData; + } else { + UNREACHABLE(); + } + } + + // Return early if there is nothing to change + if (state == prev_state && perm == prev_perm) { + return RESULT_SUCCESS; + } + + const std::size_t num_pages{size / PageSize}; + const OperationType operation{(perm & MemoryPermission::Execute) != MemoryPermission::None + ? OperationType::ChangePermissionsAndRefresh + : OperationType::ChangePermissions}; + + CASCADE_CODE(Operate(addr, num_pages, perm, operation)); + + block_manager->Update(addr, num_pages, state, perm); + + return RESULT_SUCCESS; +} + +MemoryInfo PageTable::QueryInfoImpl(VAddr addr) { + std::lock_guard lock{page_table_lock}; + + return block_manager->FindBlock(addr).GetMemoryInfo(); +} + +MemoryInfo PageTable::QueryInfo(VAddr addr) { + if (!Contains(addr, 1)) { + return {address_space_end, 0 - address_space_end, MemoryState::Inaccessible, + MemoryPermission::None, MemoryAttribute::None, MemoryPermission::None}; + } + + return QueryInfoImpl(addr); +} + +ResultCode PageTable::ReserveTransferMemory(VAddr addr, std::size_t size, MemoryPermission perm) { + std::lock_guard lock{page_table_lock}; + + MemoryState state{}; + MemoryAttribute attribute{}; + + CASCADE_CODE(CheckMemoryState(&state, nullptr, &attribute, addr, size, + MemoryState::FlagCanTransfer | MemoryState::FlagReferenceCounted, + MemoryState::FlagCanTransfer | MemoryState::FlagReferenceCounted, + MemoryPermission::Mask, MemoryPermission::ReadAndWrite, + MemoryAttribute::Mask, MemoryAttribute::None, + MemoryAttribute::IpcAndDeviceMapped)); + + block_manager->Update(addr, size / PageSize, state, perm, attribute | MemoryAttribute::Locked); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::ResetTransferMemory(VAddr addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + MemoryState state{}; + + CASCADE_CODE(CheckMemoryState(&state, nullptr, nullptr, addr, size, + MemoryState::FlagCanTransfer | MemoryState::FlagReferenceCounted, + MemoryState::FlagCanTransfer | MemoryState::FlagReferenceCounted, + MemoryPermission::None, MemoryPermission::None, + MemoryAttribute::Mask, MemoryAttribute::Locked, + MemoryAttribute::IpcAndDeviceMapped)); + + block_manager->Update(addr, size / PageSize, state, MemoryPermission::ReadAndWrite); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::SetMemoryAttribute(VAddr addr, std::size_t size, MemoryAttribute mask, + MemoryAttribute value) { + std::lock_guard lock{page_table_lock}; + + MemoryState state{}; + MemoryPermission perm{}; + MemoryAttribute attribute{}; + + CASCADE_CODE(CheckMemoryState(&state, &perm, &attribute, addr, size, + MemoryState::FlagCanChangeAttribute, + MemoryState::FlagCanChangeAttribute, MemoryPermission::None, + MemoryPermission::None, MemoryAttribute::LockedAndIpcLocked, + MemoryAttribute::None, MemoryAttribute::DeviceSharedAndUncached)); + + attribute = attribute & ~mask; + attribute = attribute | (mask & value); + + block_manager->Update(addr, size / PageSize, state, perm, attribute); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::SetHeapCapacity(std::size_t new_heap_capacity) { + std::lock_guard lock{page_table_lock}; + heap_capacity = new_heap_capacity; + return RESULT_SUCCESS; +} + +ResultVal<VAddr> PageTable::SetHeapSize(std::size_t size) { + + if (size > heap_region_end - heap_region_start) { + return ERR_OUT_OF_MEMORY; + } + + const u64 previous_heap_size{GetHeapSize()}; + + UNIMPLEMENTED_IF_MSG(previous_heap_size > size, "Heap shrink is unimplemented"); + + // Increase the heap size + { + std::lock_guard lock{page_table_lock}; + + const u64 delta{size - previous_heap_size}; + + auto process{system.Kernel().CurrentProcess()}; + if (process->GetResourceLimit() && delta != 0 && + !process->GetResourceLimit()->Reserve(ResourceType::PhysicalMemory, delta)) { + return ERR_RESOURCE_LIMIT_EXCEEDED; + } + + PageLinkedList page_linked_list; + const std::size_t num_pages{delta / PageSize}; + + CASCADE_CODE( + system.Kernel().MemoryManager().Allocate(page_linked_list, num_pages, memory_pool)); + + if (IsRegionMapped(current_heap_addr, delta)) { + return ERR_INVALID_ADDRESS_STATE; + } + + CASCADE_CODE( + Operate(current_heap_addr, num_pages, page_linked_list, OperationType::MapGroup)); + + block_manager->Update(current_heap_addr, num_pages, MemoryState::Normal, + MemoryPermission::ReadAndWrite); + + current_heap_addr = heap_region_start + size; + } + + return MakeResult<VAddr>(heap_region_start); +} + +ResultVal<VAddr> PageTable::AllocateAndMapMemory(std::size_t needed_num_pages, std::size_t align, + bool is_map_only, VAddr region_start, + std::size_t region_num_pages, MemoryState state, + MemoryPermission perm, PAddr map_addr) { + std::lock_guard lock{page_table_lock}; + + if (!CanContain(region_start, region_num_pages * PageSize, state)) { + return ERR_INVALID_ADDRESS_STATE; + } + + if (region_num_pages <= needed_num_pages) { + return ERR_OUT_OF_MEMORY; + } + + const VAddr addr{ + AllocateVirtualMemory(region_start, region_num_pages, needed_num_pages, align)}; + if (!addr) { + return ERR_OUT_OF_MEMORY; + } + + if (is_map_only) { + CASCADE_CODE(Operate(addr, needed_num_pages, perm, OperationType::Map, map_addr)); + } else { + PageLinkedList page_group; + CASCADE_CODE( + system.Kernel().MemoryManager().Allocate(page_group, needed_num_pages, memory_pool)); + CASCADE_CODE(Operate(addr, needed_num_pages, page_group, OperationType::MapGroup)); + } + + block_manager->Update(addr, needed_num_pages, state, perm); + + return MakeResult<VAddr>(addr); +} + +ResultCode PageTable::LockForDeviceAddressSpace(VAddr addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + MemoryPermission perm{}; + if (const ResultCode result{CheckMemoryState( + nullptr, &perm, nullptr, addr, size, MemoryState::FlagCanChangeAttribute, + MemoryState::FlagCanChangeAttribute, MemoryPermission::None, MemoryPermission::None, + MemoryAttribute::LockedAndIpcLocked, MemoryAttribute::None, + MemoryAttribute::DeviceSharedAndUncached)}; + result.IsError()) { + return result; + } + + block_manager->UpdateLock( + addr, size / PageSize, + [](MemoryBlockManager::iterator block, MemoryPermission perm) { + block->ShareToDevice(perm); + }, + perm); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::UnlockForDeviceAddressSpace(VAddr addr, std::size_t size) { + std::lock_guard lock{page_table_lock}; + + MemoryPermission perm{}; + if (const ResultCode result{CheckMemoryState( + nullptr, &perm, nullptr, addr, size, MemoryState::FlagCanChangeAttribute, + MemoryState::FlagCanChangeAttribute, MemoryPermission::None, MemoryPermission::None, + MemoryAttribute::LockedAndIpcLocked, MemoryAttribute::None, + MemoryAttribute::DeviceSharedAndUncached)}; + result.IsError()) { + return result; + } + + block_manager->UpdateLock( + addr, size / PageSize, + [](MemoryBlockManager::iterator block, MemoryPermission perm) { + block->UnshareToDevice(perm); + }, + perm); + + return RESULT_SUCCESS; +} + +ResultCode PageTable::InitializeMemoryLayout(VAddr start, VAddr end) { + block_manager = std::make_unique<MemoryBlockManager>(start, end); + + return RESULT_SUCCESS; +} + +bool PageTable::IsRegionMapped(VAddr address, u64 size) { + return CheckMemoryState(address, size, MemoryState::All, MemoryState::Free, + MemoryPermission::Mask, MemoryPermission::None, MemoryAttribute::Mask, + MemoryAttribute::None, MemoryAttribute::IpcAndDeviceMapped) + .IsError(); +} + +bool PageTable::IsRegionContiguous(VAddr addr, u64 size) const { + auto start_ptr = system.Memory().GetPointer(addr); + for (u64 offset{}; offset < size; offset += PageSize) { + if (start_ptr != system.Memory().GetPointer(addr + offset)) { + return false; + } + start_ptr += PageSize; + } + return true; +} + +void PageTable::AddRegionToPages(VAddr start, std::size_t num_pages, + PageLinkedList& page_linked_list) { + VAddr addr{start}; + while (addr < start + (num_pages * PageSize)) { + const PAddr paddr{GetPhysicalAddr(addr)}; + if (!paddr) { + UNREACHABLE(); + } + page_linked_list.AddBlock(paddr, 1); + addr += PageSize; + } +} + +VAddr PageTable::AllocateVirtualMemory(VAddr start, std::size_t region_num_pages, + u64 needed_num_pages, std::size_t align) { + if (is_aslr_enabled) { + UNIMPLEMENTED(); + } + return block_manager->FindFreeArea(start, region_num_pages, needed_num_pages, align, 0, + IsKernel() ? 1 : 4); +} + +ResultCode PageTable::Operate(VAddr addr, std::size_t num_pages, const PageLinkedList& page_group, + OperationType operation) { + std::lock_guard lock{page_table_lock}; + + ASSERT(Common::IsAligned(addr, PageSize)); + ASSERT(num_pages > 0); + ASSERT(num_pages == page_group.GetNumPages()); + + for (const auto& node : page_group.Nodes()) { + const std::size_t size{node.GetNumPages() * PageSize}; + + switch (operation) { + case OperationType::MapGroup: + system.Memory().MapMemoryRegion(page_table_impl, addr, size, node.GetAddress()); + break; + default: + UNREACHABLE(); + } + + addr += size; + } + + return RESULT_SUCCESS; +} + +ResultCode PageTable::Operate(VAddr addr, std::size_t num_pages, MemoryPermission perm, + OperationType operation, PAddr map_addr) { + std::lock_guard lock{page_table_lock}; + + ASSERT(num_pages > 0); + ASSERT(Common::IsAligned(addr, PageSize)); + ASSERT(ContainsPages(addr, num_pages)); + + switch (operation) { + case OperationType::Unmap: + system.Memory().UnmapRegion(page_table_impl, addr, num_pages * PageSize); + break; + case OperationType::Map: { + ASSERT(map_addr); + ASSERT(Common::IsAligned(map_addr, PageSize)); + system.Memory().MapMemoryRegion(page_table_impl, addr, num_pages * PageSize, map_addr); + break; + } + case OperationType::ChangePermissions: + case OperationType::ChangePermissionsAndRefresh: + break; + default: + UNREACHABLE(); + } + return RESULT_SUCCESS; +} + +constexpr VAddr PageTable::GetRegionAddress(MemoryState state) const { + switch (state) { + case MemoryState::Free: + case MemoryState::Kernel: + return address_space_start; + case MemoryState::Normal: + return heap_region_start; + case MemoryState::Ipc: + case MemoryState::NonSecureIpc: + case MemoryState::NonDeviceIpc: + return alias_region_start; + case MemoryState::Stack: + return stack_region_start; + case MemoryState::Io: + case MemoryState::Static: + case MemoryState::ThreadLocal: + return kernel_map_region_start; + case MemoryState::Shared: + case MemoryState::AliasCode: + case MemoryState::AliasCodeData: + case MemoryState::Transfered: + case MemoryState::SharedTransfered: + case MemoryState::SharedCode: + case MemoryState::GeneratedCode: + case MemoryState::CodeOut: + return alias_code_region_start; + case MemoryState::Code: + case MemoryState::CodeData: + return code_region_start; + default: + UNREACHABLE(); + return {}; + } +} + +constexpr std::size_t PageTable::GetRegionSize(MemoryState state) const { + switch (state) { + case MemoryState::Free: + case MemoryState::Kernel: + return address_space_end - address_space_start; + case MemoryState::Normal: + return heap_region_end - heap_region_start; + case MemoryState::Ipc: + case MemoryState::NonSecureIpc: + case MemoryState::NonDeviceIpc: + return alias_region_end - alias_region_start; + case MemoryState::Stack: + return stack_region_end - stack_region_start; + case MemoryState::Io: + case MemoryState::Static: + case MemoryState::ThreadLocal: + return kernel_map_region_end - kernel_map_region_start; + case MemoryState::Shared: + case MemoryState::AliasCode: + case MemoryState::AliasCodeData: + case MemoryState::Transfered: + case MemoryState::SharedTransfered: + case MemoryState::SharedCode: + case MemoryState::GeneratedCode: + case MemoryState::CodeOut: + return alias_code_region_end - alias_code_region_start; + case MemoryState::Code: + case MemoryState::CodeData: + return code_region_end - code_region_start; + default: + UNREACHABLE(); + return {}; + } +} + +constexpr bool PageTable::CanContain(VAddr addr, std::size_t size, MemoryState state) const { + const VAddr end{addr + size}; + const VAddr last{end - 1}; + const VAddr region_start{GetRegionAddress(state)}; + const std::size_t region_size{GetRegionSize(state)}; + const bool is_in_region{region_start <= addr && addr < end && + last <= region_start + region_size - 1}; + const bool is_in_heap{!(end <= heap_region_start || heap_region_end <= addr)}; + const bool is_in_alias{!(end <= alias_region_start || alias_region_end <= addr)}; + + switch (state) { + case MemoryState::Free: + case MemoryState::Kernel: + return is_in_region; + case MemoryState::Io: + case MemoryState::Static: + case MemoryState::Code: + case MemoryState::CodeData: + case MemoryState::Shared: + case MemoryState::AliasCode: + case MemoryState::AliasCodeData: + case MemoryState::Stack: + case MemoryState::ThreadLocal: + case MemoryState::Transfered: + case MemoryState::SharedTransfered: + case MemoryState::SharedCode: + case MemoryState::GeneratedCode: + case MemoryState::CodeOut: + return is_in_region && !is_in_heap && !is_in_alias; + case MemoryState::Normal: + ASSERT(is_in_heap); + return is_in_region && !is_in_alias; + case MemoryState::Ipc: + case MemoryState::NonSecureIpc: + case MemoryState::NonDeviceIpc: + ASSERT(is_in_alias); + return is_in_region && !is_in_heap; + default: + return false; + } +} + +constexpr ResultCode PageTable::CheckMemoryState(const MemoryInfo& info, MemoryState state_mask, + MemoryState state, MemoryPermission perm_mask, + MemoryPermission perm, MemoryAttribute attr_mask, + MemoryAttribute attr) const { + // Validate the states match expectation + if ((info.state & state_mask) != state) { + return ERR_INVALID_ADDRESS_STATE; + } + if ((info.perm & perm_mask) != perm) { + return ERR_INVALID_ADDRESS_STATE; + } + if ((info.attribute & attr_mask) != attr) { + return ERR_INVALID_ADDRESS_STATE; + } + + return RESULT_SUCCESS; +} + +ResultCode PageTable::CheckMemoryState(MemoryState* out_state, MemoryPermission* out_perm, + MemoryAttribute* out_attr, VAddr addr, std::size_t size, + MemoryState state_mask, MemoryState state, + MemoryPermission perm_mask, MemoryPermission perm, + MemoryAttribute attr_mask, MemoryAttribute attr, + MemoryAttribute ignore_attr) { + std::lock_guard lock{page_table_lock}; + + // Get information about the first block + const VAddr last_addr{addr + size - 1}; + MemoryBlockManager::const_iterator it{block_manager->FindIterator(addr)}; + MemoryInfo info{it->GetMemoryInfo()}; + + // Validate all blocks in the range have correct state + const MemoryState first_state{info.state}; + const MemoryPermission first_perm{info.perm}; + const MemoryAttribute first_attr{info.attribute}; + + while (true) { + // Validate the current block + if (!(info.state == first_state)) { + return ERR_INVALID_ADDRESS_STATE; + } + if (!(info.perm == first_perm)) { + return ERR_INVALID_ADDRESS_STATE; + } + if (!((info.attribute | static_cast<MemoryAttribute>(ignore_attr)) == + (first_attr | static_cast<MemoryAttribute>(ignore_attr)))) { + return ERR_INVALID_ADDRESS_STATE; + } + + // Validate against the provided masks + CASCADE_CODE(CheckMemoryState(info, state_mask, state, perm_mask, perm, attr_mask, attr)); + + // Break once we're done + if (last_addr <= info.GetLastAddress()) { + break; + } + + // Advance our iterator + it++; + ASSERT(it != block_manager->cend()); + info = it->GetMemoryInfo(); + } + + // Write output state + if (out_state) { + *out_state = first_state; + } + if (out_perm) { + *out_perm = first_perm; + } + if (out_attr) { + *out_attr = first_attr & static_cast<MemoryAttribute>(~ignore_attr); + } + + return RESULT_SUCCESS; +} + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/page_table.h b/src/core/hle/kernel/memory/page_table.h new file mode 100644 index 000000000..ce0d38849 --- /dev/null +++ b/src/core/hle/kernel/memory/page_table.h @@ -0,0 +1,277 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <mutex> + +#include "common/common_types.h" +#include "common/page_table.h" +#include "core/file_sys/program_metadata.h" +#include "core/hle/kernel/memory/memory_block.h" +#include "core/hle/kernel/memory/memory_manager.h" +#include "core/hle/result.h" + +namespace Core { +class System; +} + +namespace Kernel::Memory { + +class MemoryBlockManager; + +class PageTable final : NonCopyable { +public: + explicit PageTable(Core::System& system); + + ResultCode InitializeForProcess(FileSys::ProgramAddressSpaceType as_type, bool enable_aslr, + VAddr code_addr, std::size_t code_size, + Memory::MemoryManager::Pool pool); + ResultCode MapProcessCode(VAddr addr, std::size_t pages_count, MemoryState state, + MemoryPermission perm); + ResultCode MapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size); + ResultCode UnmapProcessCodeMemory(VAddr dst_addr, VAddr src_addr, std::size_t size); + ResultCode MapPhysicalMemory(VAddr addr, std::size_t size); + ResultCode UnmapPhysicalMemory(VAddr addr, std::size_t size); + ResultCode UnmapMemory(VAddr addr, std::size_t size); + ResultCode Map(VAddr dst_addr, VAddr src_addr, std::size_t size); + ResultCode Unmap(VAddr dst_addr, VAddr src_addr, std::size_t size); + ResultCode MapPages(VAddr addr, PageLinkedList& page_linked_list, MemoryState state, + MemoryPermission perm); + ResultCode SetCodeMemoryPermission(VAddr addr, std::size_t size, MemoryPermission perm); + MemoryInfo QueryInfo(VAddr addr); + ResultCode ReserveTransferMemory(VAddr addr, std::size_t size, MemoryPermission perm); + ResultCode ResetTransferMemory(VAddr addr, std::size_t size); + ResultCode SetMemoryAttribute(VAddr addr, std::size_t size, MemoryAttribute mask, + MemoryAttribute value); + ResultCode SetHeapCapacity(std::size_t new_heap_capacity); + ResultVal<VAddr> SetHeapSize(std::size_t size); + ResultVal<VAddr> AllocateAndMapMemory(std::size_t needed_num_pages, std::size_t align, + bool is_map_only, VAddr region_start, + std::size_t region_num_pages, MemoryState state, + MemoryPermission perm, PAddr map_addr = 0); + ResultCode LockForDeviceAddressSpace(VAddr addr, std::size_t size); + ResultCode UnlockForDeviceAddressSpace(VAddr addr, std::size_t size); + + Common::PageTable& PageTableImpl() { + return page_table_impl; + } + + const Common::PageTable& PageTableImpl() const { + return page_table_impl; + } + +private: + enum class OperationType : u32 { + Map, + MapGroup, + Unmap, + ChangePermissions, + ChangePermissionsAndRefresh, + }; + + static constexpr MemoryAttribute DefaultMemoryIgnoreAttr = + MemoryAttribute::DontCareMask | MemoryAttribute::IpcLocked | MemoryAttribute::DeviceShared; + + ResultCode InitializeMemoryLayout(VAddr start, VAddr end); + ResultCode MapPages(VAddr addr, const PageLinkedList& page_linked_list, MemoryPermission perm); + void MapPhysicalMemory(PageLinkedList& page_linked_list, VAddr start, VAddr end); + bool IsRegionMapped(VAddr address, u64 size); + bool IsRegionContiguous(VAddr addr, u64 size) const; + void AddRegionToPages(VAddr start, std::size_t num_pages, PageLinkedList& page_linked_list); + MemoryInfo QueryInfoImpl(VAddr addr); + VAddr AllocateVirtualMemory(VAddr start, std::size_t region_num_pages, u64 needed_num_pages, + std::size_t align); + ResultCode Operate(VAddr addr, std::size_t num_pages, const PageLinkedList& page_group, + OperationType operation); + ResultCode Operate(VAddr addr, std::size_t num_pages, MemoryPermission perm, + OperationType operation, PAddr map_addr = 0); + constexpr VAddr GetRegionAddress(MemoryState state) const; + constexpr std::size_t GetRegionSize(MemoryState state) const; + constexpr bool CanContain(VAddr addr, std::size_t size, MemoryState state) const; + + constexpr ResultCode CheckMemoryState(const MemoryInfo& info, MemoryState state_mask, + MemoryState state, MemoryPermission perm_mask, + MemoryPermission perm, MemoryAttribute attr_mask, + MemoryAttribute attr) const; + ResultCode CheckMemoryState(MemoryState* out_state, MemoryPermission* out_perm, + MemoryAttribute* out_attr, VAddr addr, std::size_t size, + MemoryState state_mask, MemoryState state, + MemoryPermission perm_mask, MemoryPermission perm, + MemoryAttribute attr_mask, MemoryAttribute attr, + MemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr); + ResultCode CheckMemoryState(VAddr addr, std::size_t size, MemoryState state_mask, + MemoryState state, MemoryPermission perm_mask, + MemoryPermission perm, MemoryAttribute attr_mask, + MemoryAttribute attr, + MemoryAttribute ignore_attr = DefaultMemoryIgnoreAttr) { + return CheckMemoryState(nullptr, nullptr, nullptr, addr, size, state_mask, state, perm_mask, + perm, attr_mask, attr, ignore_attr); + } + + std::recursive_mutex page_table_lock; + std::unique_ptr<MemoryBlockManager> block_manager; + +public: + constexpr VAddr GetAddressSpaceStart() const { + return address_space_start; + } + constexpr VAddr GetAddressSpaceEnd() const { + return address_space_end; + } + constexpr std::size_t GetAddressSpaceSize() const { + return address_space_end - address_space_start; + } + constexpr VAddr GetHeapRegionStart() const { + return heap_region_start; + } + constexpr VAddr GetHeapRegionEnd() const { + return heap_region_end; + } + constexpr std::size_t GetHeapRegionSize() const { + return heap_region_end - heap_region_start; + } + constexpr VAddr GetAliasRegionStart() const { + return alias_region_start; + } + constexpr VAddr GetAliasRegionEnd() const { + return alias_region_end; + } + constexpr std::size_t GetAliasRegionSize() const { + return alias_region_end - alias_region_start; + } + constexpr VAddr GetStackRegionStart() const { + return stack_region_start; + } + constexpr VAddr GetStackRegionEnd() const { + return stack_region_end; + } + constexpr std::size_t GetStackRegionSize() const { + return stack_region_end - stack_region_start; + } + constexpr VAddr GetKernelMapRegionStart() const { + return kernel_map_region_start; + } + constexpr VAddr GetKernelMapRegionEnd() const { + return kernel_map_region_end; + } + constexpr VAddr GetCodeRegionStart() const { + return code_region_start; + } + constexpr VAddr GetCodeRegionEnd() const { + return code_region_end; + } + constexpr VAddr GetAliasCodeRegionStart() const { + return alias_code_region_start; + } + constexpr VAddr GetAliasCodeRegionSize() const { + return alias_code_region_end - alias_code_region_start; + } + constexpr std::size_t GetAddressSpaceWidth() const { + return address_space_width; + } + constexpr std::size_t GetHeapSize() { + return current_heap_addr - heap_region_start; + } + constexpr std::size_t GetTotalHeapSize() { + return GetHeapSize() + physical_memory_usage; + } + constexpr bool IsInsideAddressSpace(VAddr address, std::size_t size) const { + return address_space_start <= address && address + size - 1 <= address_space_end - 1; + } + constexpr bool IsOutsideAliasRegion(VAddr address, std::size_t size) const { + return alias_region_start > address || address + size - 1 > alias_region_end - 1; + } + constexpr bool IsOutsideStackRegion(VAddr address, std::size_t size) const { + return stack_region_start > address || address + size - 1 > stack_region_end - 1; + } + constexpr bool IsInvalidRegion(VAddr address, std::size_t size) const { + return address + size - 1 > GetAliasCodeRegionStart() + GetAliasCodeRegionSize() - 1; + } + constexpr bool IsInsideHeapRegion(VAddr address, std::size_t size) const { + return address + size > heap_region_start && heap_region_end > address; + } + constexpr bool IsInsideAliasRegion(VAddr address, std::size_t size) const { + return address + size > alias_region_start && alias_region_end > address; + } + constexpr bool IsOutsideASLRRegion(VAddr address, std::size_t size) const { + if (IsInvalidRegion(address, size)) { + return true; + } + if (IsInsideHeapRegion(address, size)) { + return true; + } + if (IsInsideAliasRegion(address, size)) { + return true; + } + return {}; + } + constexpr bool IsInsideASLRRegion(VAddr address, std::size_t size) const { + return !IsOutsideASLRRegion(address, size); + } + constexpr PAddr GetPhysicalAddr(VAddr addr) { + return page_table_impl.backing_addr[addr >> Memory::PageBits] + addr; + } + +private: + constexpr bool Contains(VAddr addr) const { + return address_space_start <= addr && addr <= address_space_end - 1; + } + constexpr bool Contains(VAddr addr, std::size_t size) const { + return address_space_start <= addr && addr < addr + size && + addr + size - 1 <= address_space_end - 1; + } + constexpr bool IsKernel() const { + return is_kernel; + } + constexpr bool IsAslrEnabled() const { + return is_aslr_enabled; + } + + constexpr std::size_t GetNumGuardPages() const { + return IsKernel() ? 1 : 4; + } + + constexpr bool ContainsPages(VAddr addr, std::size_t num_pages) const { + return (address_space_start <= addr) && + (num_pages <= (address_space_end - address_space_start) / PageSize) && + (addr + num_pages * PageSize - 1 <= address_space_end - 1); + } + +private: + VAddr address_space_start{}; + VAddr address_space_end{}; + VAddr heap_region_start{}; + VAddr heap_region_end{}; + VAddr current_heap_end{}; + VAddr alias_region_start{}; + VAddr alias_region_end{}; + VAddr stack_region_start{}; + VAddr stack_region_end{}; + VAddr kernel_map_region_start{}; + VAddr kernel_map_region_end{}; + VAddr code_region_start{}; + VAddr code_region_end{}; + VAddr alias_code_region_start{}; + VAddr alias_code_region_end{}; + VAddr current_heap_addr{}; + + std::size_t heap_capacity{}; + std::size_t physical_memory_usage{}; + std::size_t max_heap_size{}; + std::size_t max_physical_memory_size{}; + std::size_t address_space_width{}; + + bool is_kernel{}; + bool is_aslr_enabled{}; + + MemoryManager::Pool memory_pool{MemoryManager::Pool::Application}; + + Common::PageTable page_table_impl; + + Core::System& system; +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/slab_heap.h b/src/core/hle/kernel/memory/slab_heap.h new file mode 100644 index 000000000..465eaddb3 --- /dev/null +++ b/src/core/hle/kernel/memory/slab_heap.h @@ -0,0 +1,163 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include <atomic> + +#include "common/assert.h" +#include "common/common_types.h" + +namespace Kernel::Memory { + +namespace impl { + +class SlabHeapImpl final : NonCopyable { +public: + struct Node { + Node* next{}; + }; + + constexpr SlabHeapImpl() = default; + + void Initialize(std::size_t size) { + ASSERT(head == nullptr); + obj_size = size; + } + + constexpr std::size_t GetObjectSize() const { + return obj_size; + } + + Node* GetHead() const { + return head; + } + + void* Allocate() { + Node* ret = head.load(); + + do { + if (ret == nullptr) { + break; + } + } while (!head.compare_exchange_weak(ret, ret->next)); + + return ret; + } + + void Free(void* obj) { + Node* node = static_cast<Node*>(obj); + + Node* cur_head = head.load(); + do { + node->next = cur_head; + } while (!head.compare_exchange_weak(cur_head, node)); + } + +private: + std::atomic<Node*> head{}; + std::size_t obj_size{}; +}; + +} // namespace impl + +class SlabHeapBase : NonCopyable { +public: + constexpr SlabHeapBase() = default; + + constexpr bool Contains(uintptr_t addr) const { + return start <= addr && addr < end; + } + + constexpr std::size_t GetSlabHeapSize() const { + return (end - start) / GetObjectSize(); + } + + constexpr std::size_t GetObjectSize() const { + return impl.GetObjectSize(); + } + + constexpr uintptr_t GetSlabHeapAddress() const { + return start; + } + + std::size_t GetObjectIndexImpl(const void* obj) const { + return (reinterpret_cast<uintptr_t>(obj) - start) / GetObjectSize(); + } + + std::size_t GetPeakIndex() const { + return GetObjectIndexImpl(reinterpret_cast<const void*>(peak)); + } + + void* AllocateImpl() { + return impl.Allocate(); + } + + void FreeImpl(void* obj) { + // Don't allow freeing an object that wasn't allocated from this heap + ASSERT(Contains(reinterpret_cast<uintptr_t>(obj))); + impl.Free(obj); + } + + void InitializeImpl(std::size_t obj_size, void* memory, std::size_t memory_size) { + // Ensure we don't initialize a slab using null memory + ASSERT(memory != nullptr); + + // Initialize the base allocator + impl.Initialize(obj_size); + + // Set our tracking variables + const std::size_t num_obj = (memory_size / obj_size); + start = reinterpret_cast<uintptr_t>(memory); + end = start + num_obj * obj_size; + peak = start; + + // Free the objects + u8* cur = reinterpret_cast<u8*>(end); + + for (std::size_t i{}; i < num_obj; i++) { + cur -= obj_size; + impl.Free(cur); + } + } + +private: + using Impl = impl::SlabHeapImpl; + + Impl impl; + uintptr_t peak{}; + uintptr_t start{}; + uintptr_t end{}; +}; + +template <typename T> +class SlabHeap final : public SlabHeapBase { +public: + constexpr SlabHeap() : SlabHeapBase() {} + + void Initialize(void* memory, std::size_t memory_size) { + InitializeImpl(sizeof(T), memory, memory_size); + } + + T* Allocate() { + T* obj = static_cast<T*>(AllocateImpl()); + if (obj != nullptr) { + new (obj) T(); + } + return obj; + } + + void Free(T* obj) { + FreeImpl(obj); + } + + constexpr std::size_t GetObjectIndex(const T* obj) const { + return GetObjectIndexImpl(obj); + } +}; + +} // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/system_control.cpp b/src/core/hle/kernel/memory/system_control.cpp new file mode 100644 index 000000000..11d204bc2 --- /dev/null +++ b/src/core/hle/kernel/memory/system_control.cpp @@ -0,0 +1,40 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <random> + +#include "core/hle/kernel/memory/system_control.h" + +namespace Kernel::Memory::SystemControl { +namespace { +template <typename F> +u64 GenerateUniformRange(u64 min, u64 max, F f) { + // Handle the case where the difference is too large to represent. + if (max == std::numeric_limits<u64>::max() && min == std::numeric_limits<u64>::min()) { + return f(); + } + + // Iterate until we get a value in range. + const u64 range_size = ((max + 1) - min); + const u64 effective_max = (std::numeric_limits<u64>::max() / range_size) * range_size; + while (true) { + if (const u64 rnd = f(); rnd < effective_max) { + return min + (rnd % range_size); + } + } +} + +u64 GenerateRandomU64ForInit() { + static std::random_device device; + static std::mt19937 gen(device()); + static std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max()); + return distribution(gen); +} +} // Anonymous namespace + +u64 GenerateRandomRange(u64 min, u64 max) { + return GenerateUniformRange(min, max, GenerateRandomU64ForInit); +} + +} // namespace Kernel::Memory::SystemControl diff --git a/src/core/hle/kernel/memory/system_control.h b/src/core/hle/kernel/memory/system_control.h new file mode 100644 index 000000000..19cab8cbc --- /dev/null +++ b/src/core/hle/kernel/memory/system_control.h @@ -0,0 +1,13 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace Kernel::Memory::SystemControl { + +u64 GenerateRandomRange(u64 min, u64 max); + +} // namespace Kernel::Memory::SystemControl diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp index eff4e45b0..8f6c944d1 100644 --- a/src/core/hle/kernel/mutex.cpp +++ b/src/core/hle/kernel/mutex.cpp @@ -7,6 +7,7 @@ #include <vector> #include "common/assert.h" +#include "common/logging/log.h" #include "core/core.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" @@ -33,8 +34,6 @@ static std::pair<std::shared_ptr<Thread>, u32> GetHighestPriorityMutexWaitingThr if (thread->GetMutexWaitAddress() != mutex_addr) continue; - ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex); - ++num_waiters; if (highest_priority_thread == nullptr || thread->GetPriority() < highest_priority_thread->GetPriority()) { @@ -48,6 +47,7 @@ static std::pair<std::shared_ptr<Thread>, u32> GetHighestPriorityMutexWaitingThr /// Update the mutex owner field of all threads waiting on the mutex to point to the new owner. static void TransferMutexOwnership(VAddr mutex_addr, std::shared_ptr<Thread> current_thread, std::shared_ptr<Thread> new_owner) { + current_thread->RemoveMutexWaiter(new_owner); const auto threads = current_thread->GetMutexWaitingThreads(); for (const auto& thread : threads) { if (thread->GetMutexWaitAddress() != mutex_addr) @@ -67,85 +67,104 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle, Handle requesting_thread_handle) { // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { + LOG_ERROR(Kernel, "Address is not 4-byte aligned! address={:016X}", address); return ERR_INVALID_ADDRESS; } - const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); + auto& kernel = system.Kernel(); std::shared_ptr<Thread> current_thread = - SharedFrom(system.CurrentScheduler().GetCurrentThread()); - std::shared_ptr<Thread> holding_thread = handle_table.Get<Thread>(holding_thread_handle); - std::shared_ptr<Thread> requesting_thread = handle_table.Get<Thread>(requesting_thread_handle); + SharedFrom(kernel.CurrentScheduler().GetCurrentThread()); + { + SchedulerLock lock(kernel); + // The mutex address must be 4-byte aligned + if ((address % sizeof(u32)) != 0) { + return ERR_INVALID_ADDRESS; + } - // TODO(Subv): It is currently unknown if it is possible to lock a mutex in behalf of another - // thread. - ASSERT(requesting_thread == current_thread); + const auto& handle_table = kernel.CurrentProcess()->GetHandleTable(); + std::shared_ptr<Thread> holding_thread = handle_table.Get<Thread>(holding_thread_handle); + std::shared_ptr<Thread> requesting_thread = + handle_table.Get<Thread>(requesting_thread_handle); - const u32 addr_value = system.Memory().Read32(address); + // TODO(Subv): It is currently unknown if it is possible to lock a mutex in behalf of + // another thread. + ASSERT(requesting_thread == current_thread); - // If the mutex isn't being held, just return success. - if (addr_value != (holding_thread_handle | Mutex::MutexHasWaitersFlag)) { - return RESULT_SUCCESS; - } + current_thread->SetSynchronizationResults(nullptr, RESULT_SUCCESS); - if (holding_thread == nullptr) { - return ERR_INVALID_HANDLE; - } + const u32 addr_value = system.Memory().Read32(address); + + // If the mutex isn't being held, just return success. + if (addr_value != (holding_thread_handle | Mutex::MutexHasWaitersFlag)) { + return RESULT_SUCCESS; + } - // Wait until the mutex is released - current_thread->SetMutexWaitAddress(address); - current_thread->SetWaitHandle(requesting_thread_handle); + if (holding_thread == nullptr) { + return ERR_INVALID_HANDLE; + } - current_thread->SetStatus(ThreadStatus::WaitMutex); - current_thread->InvalidateWakeupCallback(); + // Wait until the mutex is released + current_thread->SetMutexWaitAddress(address); + current_thread->SetWaitHandle(requesting_thread_handle); - // Update the lock holder thread's priority to prevent priority inversion. - holding_thread->AddMutexWaiter(current_thread); + current_thread->SetStatus(ThreadStatus::WaitMutex); - system.PrepareReschedule(); + // Update the lock holder thread's priority to prevent priority inversion. + holding_thread->AddMutexWaiter(current_thread); + } - return RESULT_SUCCESS; + { + SchedulerLock lock(kernel); + auto* owner = current_thread->GetLockOwner(); + if (owner != nullptr) { + owner->RemoveMutexWaiter(current_thread); + } + } + return current_thread->GetSignalingResult(); } -ResultCode Mutex::Release(VAddr address) { +std::pair<ResultCode, std::shared_ptr<Thread>> Mutex::Unlock(std::shared_ptr<Thread> owner, + VAddr address) { // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { - return ERR_INVALID_ADDRESS; + LOG_ERROR(Kernel, "Address is not 4-byte aligned! address={:016X}", address); + return {ERR_INVALID_ADDRESS, nullptr}; } - std::shared_ptr<Thread> current_thread = - SharedFrom(system.CurrentScheduler().GetCurrentThread()); - auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(current_thread, address); - - // There are no more threads waiting for the mutex, release it completely. - if (thread == nullptr) { + auto [new_owner, num_waiters] = GetHighestPriorityMutexWaitingThread(owner, address); + if (new_owner == nullptr) { system.Memory().Write32(address, 0); - return RESULT_SUCCESS; + return {RESULT_SUCCESS, nullptr}; } - // Transfer the ownership of the mutex from the previous owner to the new one. - TransferMutexOwnership(address, current_thread, thread); - - u32 mutex_value = thread->GetWaitHandle(); - + TransferMutexOwnership(address, owner, new_owner); + u32 mutex_value = new_owner->GetWaitHandle(); if (num_waiters >= 2) { // Notify the guest that there are still some threads waiting for the mutex mutex_value |= Mutex::MutexHasWaitersFlag; } + new_owner->SetSynchronizationResults(nullptr, RESULT_SUCCESS); + new_owner->SetLockOwner(nullptr); + new_owner->ResumeFromWait(); - // Grant the mutex to the next waiting thread and resume it. system.Memory().Write32(address, mutex_value); + return {RESULT_SUCCESS, new_owner}; +} - ASSERT(thread->GetStatus() == ThreadStatus::WaitMutex); - thread->ResumeFromWait(); +ResultCode Mutex::Release(VAddr address) { + auto& kernel = system.Kernel(); + SchedulerLock lock(kernel); - thread->SetLockOwner(nullptr); - thread->SetCondVarWaitAddress(0); - thread->SetMutexWaitAddress(0); - thread->SetWaitHandle(0); - thread->SetWaitSynchronizationResult(RESULT_SUCCESS); + std::shared_ptr<Thread> current_thread = + SharedFrom(kernel.CurrentScheduler().GetCurrentThread()); - system.PrepareReschedule(); + auto [result, new_owner] = Unlock(current_thread, address); - return RESULT_SUCCESS; + if (result != RESULT_SUCCESS && new_owner != nullptr) { + new_owner->SetSynchronizationResults(nullptr, result); + } + + return result; } + } // namespace Kernel diff --git a/src/core/hle/kernel/mutex.h b/src/core/hle/kernel/mutex.h index b904de2e8..3b81dc3df 100644 --- a/src/core/hle/kernel/mutex.h +++ b/src/core/hle/kernel/mutex.h @@ -28,6 +28,10 @@ public: ResultCode TryAcquire(VAddr address, Handle holding_thread_handle, Handle requesting_thread_handle); + /// Unlocks a mutex for owner at address + std::pair<ResultCode, std::shared_ptr<Thread>> Unlock(std::shared_ptr<Thread> owner, + VAddr address); + /// Releases the mutex at the specified address. ResultCode Release(VAddr address); diff --git a/src/core/hle/kernel/physical_core.cpp b/src/core/hle/kernel/physical_core.cpp index aa2787467..6e04d025f 100644 --- a/src/core/hle/kernel/physical_core.cpp +++ b/src/core/hle/kernel/physical_core.cpp @@ -2,63 +2,43 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "common/logging/log.h" -#include "core/arm/arm_interface.h" -#ifdef ARCHITECTURE_x86_64 -#include "core/arm/dynarmic/arm_dynarmic_32.h" -#include "core/arm/dynarmic/arm_dynarmic_64.h" -#endif -#include "core/arm/exclusive_monitor.h" -#include "core/arm/unicorn/arm_unicorn.h" +#include "common/spin_lock.h" +#include "core/arm/cpu_interrupt_handler.h" #include "core/core.h" #include "core/hle/kernel/physical_core.h" #include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/thread.h" namespace Kernel { -PhysicalCore::PhysicalCore(Core::System& system, std::size_t id, - Core::ExclusiveMonitor& exclusive_monitor) - : core_index{id} { -#ifdef ARCHITECTURE_x86_64 - arm_interface_32 = - std::make_unique<Core::ARM_Dynarmic_32>(system, exclusive_monitor, core_index); - arm_interface_64 = - std::make_unique<Core::ARM_Dynarmic_64>(system, exclusive_monitor, core_index); - -#else - arm_interface = std::make_shared<Core::ARM_Unicorn>(system); - LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available"); -#endif - - scheduler = std::make_unique<Kernel::Scheduler>(system, core_index); -} +PhysicalCore::PhysicalCore(Core::System& system, std::size_t id, Kernel::Scheduler& scheduler, + Core::CPUInterruptHandler& interrupt_handler) + : interrupt_handler{interrupt_handler}, + core_index{id}, scheduler{scheduler}, guard{std::make_unique<Common::SpinLock>()} {} PhysicalCore::~PhysicalCore() = default; -void PhysicalCore::Run() { - arm_interface->Run(); - arm_interface->ClearExclusiveState(); +void PhysicalCore::Idle() { + interrupt_handler.AwaitInterrupt(); } -void PhysicalCore::Step() { - arm_interface->Step(); +void PhysicalCore::Shutdown() { + scheduler.Shutdown(); } -void PhysicalCore::Stop() { - arm_interface->PrepareReschedule(); +bool PhysicalCore::IsInterrupted() const { + return interrupt_handler.IsInterrupted(); } -void PhysicalCore::Shutdown() { - scheduler->Shutdown(); +void PhysicalCore::Interrupt() { + guard->lock(); + interrupt_handler.SetInterrupt(true); + guard->unlock(); } -void PhysicalCore::SetIs64Bit(bool is_64_bit) { - if (is_64_bit) { - arm_interface = arm_interface_64.get(); - } else { - arm_interface = arm_interface_32.get(); - } +void PhysicalCore::ClearInterrupt() { + guard->lock(); + interrupt_handler.SetInterrupt(false); + guard->unlock(); } } // namespace Kernel diff --git a/src/core/hle/kernel/physical_core.h b/src/core/hle/kernel/physical_core.h index 3269166be..d7a7a951c 100644 --- a/src/core/hle/kernel/physical_core.h +++ b/src/core/hle/kernel/physical_core.h @@ -7,12 +7,17 @@ #include <cstddef> #include <memory> +namespace Common { +class SpinLock; +} + namespace Kernel { class Scheduler; } // namespace Kernel namespace Core { class ARM_Interface; +class CPUInterruptHandler; class ExclusiveMonitor; class System; } // namespace Core @@ -21,7 +26,8 @@ namespace Kernel { class PhysicalCore { public: - PhysicalCore(Core::System& system, std::size_t id, Core::ExclusiveMonitor& exclusive_monitor); + PhysicalCore(Core::System& system, std::size_t id, Kernel::Scheduler& scheduler, + Core::CPUInterruptHandler& interrupt_handler); ~PhysicalCore(); PhysicalCore(const PhysicalCore&) = delete; @@ -30,23 +36,18 @@ public: PhysicalCore(PhysicalCore&&) = default; PhysicalCore& operator=(PhysicalCore&&) = default; - /// Execute current jit state - void Run(); - /// Execute a single instruction in current jit. - void Step(); - /// Stop JIT execution/exit - void Stop(); + void Idle(); + /// Interrupt this physical core. + void Interrupt(); - // Shutdown this physical core. - void Shutdown(); + /// Clear this core's interrupt + void ClearInterrupt(); - Core::ARM_Interface& ArmInterface() { - return *arm_interface; - } + /// Check if this core is interrupted + bool IsInterrupted() const; - const Core::ARM_Interface& ArmInterface() const { - return *arm_interface; - } + // Shutdown this physical core. + void Shutdown(); bool IsMainCore() const { return core_index == 0; @@ -61,21 +62,18 @@ public: } Kernel::Scheduler& Scheduler() { - return *scheduler; + return scheduler; } const Kernel::Scheduler& Scheduler() const { - return *scheduler; + return scheduler; } - void SetIs64Bit(bool is_64_bit); - private: + Core::CPUInterruptHandler& interrupt_handler; std::size_t core_index; - std::unique_ptr<Core::ARM_Interface> arm_interface_32; - std::unique_ptr<Core::ARM_Interface> arm_interface_64; - std::unique_ptr<Kernel::Scheduler> scheduler; - Core::ARM_Interface* arm_interface{}; + Kernel::Scheduler& scheduler; + std::unique_ptr<Common::SpinLock> guard; }; } // namespace Kernel diff --git a/src/core/hle/kernel/physical_memory.h b/src/core/hle/kernel/physical_memory.h index b689e8e8b..7a0266780 100644 --- a/src/core/hle/kernel/physical_memory.h +++ b/src/core/hle/kernel/physical_memory.h @@ -4,6 +4,8 @@ #pragma once +#include <vector> + #include "common/alignment.h" namespace Kernel { diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp index edc414d69..b17529dee 100644 --- a/src/core/hle/kernel/process.cpp +++ b/src/core/hle/kernel/process.cpp @@ -4,21 +4,26 @@ #include <algorithm> #include <bitset> +#include <ctime> #include <memory> #include <random> #include "common/alignment.h" #include "common/assert.h" #include "common/logging/log.h" #include "core/core.h" +#include "core/device_memory.h" #include "core/file_sys/program_metadata.h" #include "core/hle/kernel/code_set.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/memory_block_manager.h" +#include "core/hle/kernel/memory/page_table.h" +#include "core/hle/kernel/memory/slab_heap.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/resource_limit.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" -#include "core/hle/kernel/vm_manager.h" +#include "core/hle/lock.h" #include "core/memory.h" #include "core/settings.h" @@ -27,26 +32,31 @@ namespace { /** * Sets up the primary application thread * + * @param system The system instance to create the main thread under. * @param owner_process The parent process for the main thread - * @param kernel The kernel instance to create the main thread under. * @param priority The priority to give the main thread */ -void SetupMainThread(Process& owner_process, KernelCore& kernel, u32 priority) { - const auto& vm_manager = owner_process.VMManager(); - const VAddr entry_point = vm_manager.GetCodeRegionBaseAddress(); - const VAddr stack_top = vm_manager.GetTLSIORegionEndAddress(); - auto thread_res = Thread::Create(kernel, "main", entry_point, priority, 0, - owner_process.GetIdealCore(), stack_top, owner_process); +void SetupMainThread(Core::System& system, Process& owner_process, u32 priority, VAddr stack_top) { + const VAddr entry_point = owner_process.PageTable().GetCodeRegionStart(); + ThreadType type = THREADTYPE_USER; + auto thread_res = Thread::Create(system, type, "main", entry_point, priority, 0, + owner_process.GetIdealCore(), stack_top, &owner_process); std::shared_ptr<Thread> thread = std::move(thread_res).Unwrap(); // Register 1 must be a handle to the main thread const Handle thread_handle = owner_process.GetHandleTable().Create(thread).Unwrap(); + thread->GetContext32().cpu_registers[0] = 0; + thread->GetContext64().cpu_registers[0] = 0; thread->GetContext32().cpu_registers[1] = thread_handle; thread->GetContext64().cpu_registers[1] = thread_handle; + auto& kernel = system.Kernel(); // Threads by default are dormant, wake up the main thread so it runs when the scheduler fires - thread->ResumeFromWait(); + { + SchedulerLock lock{kernel}; + thread->SetStatus(ThreadStatus::Ready); + } } } // Anonymous namespace @@ -57,7 +67,8 @@ void SetupMainThread(Process& owner_process, KernelCore& kernel, u32 priority) { // (whichever page happens to have an available slot). class TLSPage { public: - static constexpr std::size_t num_slot_entries = Memory::PAGE_SIZE / Memory::TLS_ENTRY_SIZE; + static constexpr std::size_t num_slot_entries = + Core::Memory::PAGE_SIZE / Core::Memory::TLS_ENTRY_SIZE; explicit TLSPage(VAddr address) : base_address{address} {} @@ -76,7 +87,7 @@ public: } is_slot_used[i] = true; - return base_address + (i * Memory::TLS_ENTRY_SIZE); + return base_address + (i * Core::Memory::TLS_ENTRY_SIZE); } return std::nullopt; @@ -86,15 +97,15 @@ public: // Ensure that all given addresses are consistent with how TLS pages // are intended to be used when releasing slots. ASSERT(IsWithinPage(address)); - ASSERT((address % Memory::TLS_ENTRY_SIZE) == 0); + ASSERT((address % Core::Memory::TLS_ENTRY_SIZE) == 0); - const std::size_t index = (address - base_address) / Memory::TLS_ENTRY_SIZE; + const std::size_t index = (address - base_address) / Core::Memory::TLS_ENTRY_SIZE; is_slot_used[index] = false; } private: bool IsWithinPage(VAddr address) const { - return base_address <= address && address < base_address + Memory::PAGE_SIZE; + return base_address <= address && address < base_address + Core::Memory::PAGE_SIZE; } VAddr base_address; @@ -106,14 +117,14 @@ std::shared_ptr<Process> Process::Create(Core::System& system, std::string name, std::shared_ptr<Process> process = std::make_shared<Process>(system); process->name = std::move(name); - process->resource_limit = kernel.GetSystemResourceLimit(); + process->resource_limit = ResourceLimit::Create(kernel); process->status = ProcessStatus::Created; process->program_id = 0; process->process_id = type == ProcessType::KernelInternal ? kernel.CreateNewKernelProcessID() : kernel.CreateNewUserProcessID(); process->capabilities.InitializeForMetadatalessProcess(); - std::mt19937 rng(Settings::values.rng_seed.value_or(0)); + std::mt19937 rng(Settings::values.rng_seed.GetValue().value_or(std::time(nullptr))); std::uniform_int_distribution<u64> distribution; std::generate(process->random_entropy.begin(), process->random_entropy.end(), [&] { return distribution(rng); }); @@ -127,7 +138,15 @@ std::shared_ptr<ResourceLimit> Process::GetResourceLimit() const { } u64 Process::GetTotalPhysicalMemoryAvailable() const { - return vm_manager.GetTotalPhysicalMemoryAvailable(); + const u64 capacity{resource_limit->GetCurrentResourceValue(ResourceType::PhysicalMemory) + + page_table->GetTotalHeapSize() + GetSystemResourceSize() + image_size + + main_thread_stack_size}; + + if (capacity < memory_usage_capacity) { + return capacity; + } + + return memory_usage_capacity; } u64 Process::GetTotalPhysicalMemoryAvailableWithoutSystemResource() const { @@ -135,8 +154,8 @@ u64 Process::GetTotalPhysicalMemoryAvailableWithoutSystemResource() const { } u64 Process::GetTotalPhysicalMemoryUsed() const { - return vm_manager.GetCurrentHeapSize() + main_thread_stack_size + code_memory_size + - GetSystemResourceUsage(); + return image_size + main_thread_stack_size + page_table->GetTotalHeapSize() + + GetSystemResourceSize(); } u64 Process::GetTotalPhysicalMemoryUsedWithoutSystemResource() const { @@ -170,7 +189,6 @@ void Process::RemoveConditionVariableThread(std::shared_ptr<Thread> thread) { } ++it; } - UNREACHABLE(); } std::vector<std::shared_ptr<Thread>> Process::GetConditionVariableThreads( @@ -195,6 +213,7 @@ void Process::UnregisterThread(const Thread* thread) { } ResultCode Process::ClearSignalState() { + SchedulerLock lock(system.Kernel()); if (status == ProcessStatus::Exited) { LOG_ERROR(Kernel, "called on a terminated process instance."); return ERR_INVALID_STATE; @@ -209,33 +228,82 @@ ResultCode Process::ClearSignalState() { return RESULT_SUCCESS; } -ResultCode Process::LoadFromMetadata(const FileSys::ProgramMetadata& metadata) { +ResultCode Process::LoadFromMetadata(const FileSys::ProgramMetadata& metadata, + std::size_t code_size) { program_id = metadata.GetTitleID(); ideal_core = metadata.GetMainThreadCore(); is_64bit_process = metadata.Is64BitProgram(); system_resource_size = metadata.GetSystemResourceSize(); + image_size = code_size; + + // Initialize proces address space + if (const ResultCode result{ + page_table->InitializeForProcess(metadata.GetAddressSpaceType(), false, 0x8000000, + code_size, Memory::MemoryManager::Pool::Application)}; + result.IsError()) { + return result; + } - vm_manager.Reset(metadata.GetAddressSpaceType()); + // Map process code region + if (const ResultCode result{page_table->MapProcessCode( + page_table->GetCodeRegionStart(), code_size / Memory::PageSize, + Memory::MemoryState::Code, Memory::MemoryPermission::None)}; + result.IsError()) { + return result; + } + + // Initialize process capabilities + const auto& caps{metadata.GetKernelCapabilities()}; + if (const ResultCode result{ + capabilities.InitializeForUserProcess(caps.data(), caps.size(), *page_table)}; + result.IsError()) { + return result; + } - const auto& caps = metadata.GetKernelCapabilities(); - const auto capability_init_result = - capabilities.InitializeForUserProcess(caps.data(), caps.size(), vm_manager); - if (capability_init_result.IsError()) { - return capability_init_result; + // Set memory usage capacity + switch (metadata.GetAddressSpaceType()) { + case FileSys::ProgramAddressSpaceType::Is32Bit: + case FileSys::ProgramAddressSpaceType::Is36Bit: + case FileSys::ProgramAddressSpaceType::Is39Bit: + memory_usage_capacity = page_table->GetHeapRegionEnd() - page_table->GetHeapRegionStart(); + break; + + case FileSys::ProgramAddressSpaceType::Is32BitNoMap: + memory_usage_capacity = page_table->GetHeapRegionEnd() - page_table->GetHeapRegionStart() + + page_table->GetAliasRegionEnd() - page_table->GetAliasRegionStart(); + break; + + default: + UNREACHABLE(); } + // Set initial resource limits + resource_limit->SetLimitValue( + ResourceType::PhysicalMemory, + kernel.MemoryManager().GetSize(Memory::MemoryManager::Pool::Application)); + resource_limit->SetLimitValue(ResourceType::Threads, 608); + resource_limit->SetLimitValue(ResourceType::Events, 700); + resource_limit->SetLimitValue(ResourceType::TransferMemory, 128); + resource_limit->SetLimitValue(ResourceType::Sessions, 894); + ASSERT(resource_limit->Reserve(ResourceType::PhysicalMemory, code_size)); + + // Create TLS region + tls_region_address = CreateTLSRegion(); + return handle_table.SetSize(capabilities.GetHandleTableSize()); } void Process::Run(s32 main_thread_priority, u64 stack_size) { AllocateMainThreadStack(stack_size); - tls_region_address = CreateTLSRegion(); - vm_manager.LogLayout(); + const std::size_t heap_capacity{memory_usage_capacity - main_thread_stack_size - image_size}; + ASSERT(!page_table->SetHeapCapacity(heap_capacity).IsError()); ChangeStatus(ProcessStatus::Running); - SetupMainThread(*this, kernel, main_thread_priority); + SetupMainThread(system, *this, main_thread_priority, main_thread_stack_top); + resource_limit->Reserve(ResourceType::Threads, 1); + resource_limit->Reserve(ResourceType::PhysicalMemory, main_thread_stack_size); } void Process::PrepareForTermination() { @@ -279,32 +347,39 @@ static auto FindTLSPageWithAvailableSlots(std::vector<TLSPage>& tls_pages) { } VAddr Process::CreateTLSRegion() { - auto tls_page_iter = FindTLSPageWithAvailableSlots(tls_pages); + SchedulerLock lock(system.Kernel()); + if (auto tls_page_iter{FindTLSPageWithAvailableSlots(tls_pages)}; + tls_page_iter != tls_pages.cend()) { + return *tls_page_iter->ReserveSlot(); + } - if (tls_page_iter == tls_pages.cend()) { - const auto region_address = - vm_manager.FindFreeRegion(vm_manager.GetTLSIORegionBaseAddress(), - vm_manager.GetTLSIORegionEndAddress(), Memory::PAGE_SIZE); - ASSERT(region_address.Succeeded()); + Memory::Page* const tls_page_ptr{kernel.GetUserSlabHeapPages().Allocate()}; + ASSERT(tls_page_ptr); - const auto map_result = vm_manager.MapMemoryBlock( - *region_address, std::make_shared<PhysicalMemory>(Memory::PAGE_SIZE), 0, - Memory::PAGE_SIZE, MemoryState::ThreadLocal); - ASSERT(map_result.Succeeded()); + const VAddr start{page_table->GetKernelMapRegionStart()}; + const VAddr size{page_table->GetKernelMapRegionEnd() - start}; + const PAddr tls_map_addr{system.DeviceMemory().GetPhysicalAddr(tls_page_ptr)}; + const VAddr tls_page_addr{ + page_table + ->AllocateAndMapMemory(1, Memory::PageSize, true, start, size / Memory::PageSize, + Memory::MemoryState::ThreadLocal, + Memory::MemoryPermission::ReadAndWrite, tls_map_addr) + .ValueOr(0)}; - tls_pages.emplace_back(*region_address); + ASSERT(tls_page_addr); - const auto reserve_result = tls_pages.back().ReserveSlot(); - ASSERT(reserve_result.has_value()); + std::memset(tls_page_ptr, 0, Memory::PageSize); + tls_pages.emplace_back(tls_page_addr); - return *reserve_result; - } + const auto reserve_result{tls_pages.back().ReserveSlot()}; + ASSERT(reserve_result.has_value()); - return *tls_page_iter->ReserveSlot(); + return *reserve_result; } void Process::FreeTLSRegion(VAddr tls_address) { - const VAddr aligned_address = Common::AlignDown(tls_address, Memory::PAGE_SIZE); + SchedulerLock lock(system.Kernel()); + const VAddr aligned_address = Common::AlignDown(tls_address, Core::Memory::PAGE_SIZE); auto iter = std::find_if(tls_pages.begin(), tls_pages.end(), [aligned_address](const auto& page) { return page.GetBaseAddress() == aligned_address; @@ -317,29 +392,24 @@ void Process::FreeTLSRegion(VAddr tls_address) { iter->ReleaseSlot(tls_address); } -void Process::LoadModule(CodeSet module_, VAddr base_addr) { - code_memory_size += module_.memory.size(); - - const auto memory = std::make_shared<PhysicalMemory>(std::move(module_.memory)); - - const auto MapSegment = [&](const CodeSet::Segment& segment, VMAPermission permissions, - MemoryState memory_state) { - const auto vma = vm_manager - .MapMemoryBlock(segment.addr + base_addr, memory, segment.offset, - segment.size, memory_state) - .Unwrap(); - vm_manager.Reprotect(vma, permissions); +void Process::LoadModule(CodeSet code_set, VAddr base_addr) { + std::lock_guard lock{HLE::g_hle_lock}; + const auto ReprotectSegment = [&](const CodeSet::Segment& segment, + Memory::MemoryPermission permission) { + page_table->SetCodeMemoryPermission(segment.addr + base_addr, segment.size, permission); }; - // Map CodeSet segments - MapSegment(module_.CodeSegment(), VMAPermission::ReadExecute, MemoryState::Code); - MapSegment(module_.RODataSegment(), VMAPermission::Read, MemoryState::CodeData); - MapSegment(module_.DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeData); + system.Memory().WriteBlock(*this, base_addr, code_set.memory.data(), code_set.memory.size()); + + ReprotectSegment(code_set.CodeSegment(), Memory::MemoryPermission::ReadAndExecute); + ReprotectSegment(code_set.RODataSegment(), Memory::MemoryPermission::Read); + ReprotectSegment(code_set.DataSegment(), Memory::MemoryPermission::ReadAndWrite); } Process::Process(Core::System& system) - : SynchronizationObject{system.Kernel()}, vm_manager{system}, - address_arbiter{system}, mutex{system}, system{system} {} + : SynchronizationObject{system.Kernel()}, page_table{std::make_unique<Memory::PageTable>( + system)}, + handle_table{system.Kernel()}, address_arbiter{system}, mutex{system}, system{system} {} Process::~Process() = default; @@ -361,16 +431,24 @@ void Process::ChangeStatus(ProcessStatus new_status) { Signal(); } -void Process::AllocateMainThreadStack(u64 stack_size) { +ResultCode Process::AllocateMainThreadStack(std::size_t stack_size) { + ASSERT(stack_size); + // The kernel always ensures that the given stack size is page aligned. - main_thread_stack_size = Common::AlignUp(stack_size, Memory::PAGE_SIZE); - - // Allocate and map the main thread stack - const VAddr mapping_address = vm_manager.GetTLSIORegionEndAddress() - main_thread_stack_size; - vm_manager - .MapMemoryBlock(mapping_address, std::make_shared<PhysicalMemory>(main_thread_stack_size), - 0, main_thread_stack_size, MemoryState::Stack) - .Unwrap(); + main_thread_stack_size = Common::AlignUp(stack_size, Memory::PageSize); + + const VAddr start{page_table->GetStackRegionStart()}; + const std::size_t size{page_table->GetStackRegionEnd() - start}; + + CASCADE_RESULT(main_thread_stack_top, + page_table->AllocateAndMapMemory( + main_thread_stack_size / Memory::PageSize, Memory::PageSize, false, start, + size / Memory::PageSize, Memory::MemoryState::Stack, + Memory::MemoryPermission::ReadAndWrite)); + + main_thread_stack_top += main_thread_stack_size; + + return RESULT_SUCCESS; } } // namespace Kernel diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h index 4887132a7..f45cb5674 100644 --- a/src/core/hle/kernel/process.h +++ b/src/core/hle/kernel/process.h @@ -16,7 +16,6 @@ #include "core/hle/kernel/mutex.h" #include "core/hle/kernel/process_capability.h" #include "core/hle/kernel/synchronization_object.h" -#include "core/hle/kernel/vm_manager.h" #include "core/hle/result.h" namespace Core { @@ -36,6 +35,10 @@ class TLSPage; struct CodeSet; +namespace Memory { +class PageTable; +} + enum class MemoryRegion : u16 { APPLICATION = 1, SYSTEM = 2, @@ -100,14 +103,14 @@ public: return HANDLE_TYPE; } - /// Gets a reference to the process' memory manager. - Kernel::VMManager& VMManager() { - return vm_manager; + /// Gets a reference to the process' page table. + Memory::PageTable& PageTable() { + return *page_table; } - /// Gets a const reference to the process' memory manager. - const Kernel::VMManager& VMManager() const { - return vm_manager; + /// Gets const a reference to the process' page table. + const Memory::PageTable& PageTable() const { + return *page_table; } /// Gets a reference to the process' handle table. @@ -273,7 +276,7 @@ public: * @returns RESULT_SUCCESS if all relevant metadata was able to be * loaded and parsed. Otherwise, an error code is returned. */ - ResultCode LoadFromMetadata(const FileSys::ProgramMetadata& metadata); + ResultCode LoadFromMetadata(const FileSys::ProgramMetadata& metadata, std::size_t code_size); /** * Starts the main application thread for this process. @@ -289,7 +292,7 @@ public: */ void PrepareForTermination(); - void LoadModule(CodeSet module_, VAddr base_addr); + void LoadModule(CodeSet code_set, VAddr base_addr); /////////////////////////////////////////////////////////////////////////////////////////////// // Thread-local storage management @@ -313,16 +316,10 @@ private: void ChangeStatus(ProcessStatus new_status); /// Allocates the main thread stack for the process, given the stack size in bytes. - void AllocateMainThreadStack(u64 stack_size); - - /// Memory manager for this process. - Kernel::VMManager vm_manager; + ResultCode AllocateMainThreadStack(std::size_t stack_size); - /// Size of the main thread's stack in bytes. - u64 main_thread_stack_size = 0; - - /// Size of the loaded code memory in bytes. - u64 code_memory_size = 0; + /// Memory manager for this process + std::unique_ptr<Memory::PageTable> page_table; /// Current status of the process ProcessStatus status{}; @@ -385,11 +382,23 @@ private: /// List of threads waiting for a condition variable std::unordered_map<VAddr, std::list<std::shared_ptr<Thread>>> cond_var_threads; - /// System context - Core::System& system; + /// Address of the top of the main thread's stack + VAddr main_thread_stack_top{}; + + /// Size of the main thread's stack + std::size_t main_thread_stack_size{}; + + /// Memory usage capacity for the process + std::size_t memory_usage_capacity{}; + + /// Process total image size + std::size_t image_size{}; /// Name of this process std::string name; + + /// System context + Core::System& system; }; } // namespace Kernel diff --git a/src/core/hle/kernel/process_capability.cpp b/src/core/hle/kernel/process_capability.cpp index 583e35b79..63880f13d 100644 --- a/src/core/hle/kernel/process_capability.cpp +++ b/src/core/hle/kernel/process_capability.cpp @@ -3,10 +3,11 @@ // Refer to the license.txt file included. #include "common/bit_util.h" +#include "common/logging/log.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process_capability.h" -#include "core/hle/kernel/vm_manager.h" namespace Kernel { namespace { @@ -66,7 +67,7 @@ u32 GetFlagBitOffset(CapabilityType type) { ResultCode ProcessCapabilities::InitializeForKernelProcess(const u32* capabilities, std::size_t num_capabilities, - VMManager& vm_manager) { + Memory::PageTable& page_table) { Clear(); // Allow all cores and priorities. @@ -74,15 +75,15 @@ ResultCode ProcessCapabilities::InitializeForKernelProcess(const u32* capabiliti priority_mask = 0xFFFFFFFFFFFFFFFF; kernel_version = PackedKernelVersion; - return ParseCapabilities(capabilities, num_capabilities, vm_manager); + return ParseCapabilities(capabilities, num_capabilities, page_table); } ResultCode ProcessCapabilities::InitializeForUserProcess(const u32* capabilities, std::size_t num_capabilities, - VMManager& vm_manager) { + Memory::PageTable& page_table) { Clear(); - return ParseCapabilities(capabilities, num_capabilities, vm_manager); + return ParseCapabilities(capabilities, num_capabilities, page_table); } void ProcessCapabilities::InitializeForMetadatalessProcess() { @@ -105,7 +106,7 @@ void ProcessCapabilities::InitializeForMetadatalessProcess() { ResultCode ProcessCapabilities::ParseCapabilities(const u32* capabilities, std::size_t num_capabilities, - VMManager& vm_manager) { + Memory::PageTable& page_table) { u32 set_flags = 0; u32 set_svc_bits = 0; @@ -119,22 +120,30 @@ ResultCode ProcessCapabilities::ParseCapabilities(const u32* capabilities, // The MapPhysical type uses two descriptor flags for its parameters. // If there's only one, then there's a problem. if (i >= num_capabilities) { + LOG_ERROR(Kernel, "Invalid combination! i={}", i); return ERR_INVALID_COMBINATION; } const auto size_flags = capabilities[i]; if (GetCapabilityType(size_flags) != CapabilityType::MapPhysical) { + LOG_ERROR(Kernel, "Invalid capability type! size_flags={}", size_flags); return ERR_INVALID_COMBINATION; } - const auto result = HandleMapPhysicalFlags(descriptor, size_flags, vm_manager); + const auto result = HandleMapPhysicalFlags(descriptor, size_flags, page_table); if (result.IsError()) { + LOG_ERROR(Kernel, "Failed to map physical flags! descriptor={}, size_flags={}", + descriptor, size_flags); return result; } } else { const auto result = - ParseSingleFlagCapability(set_flags, set_svc_bits, descriptor, vm_manager); + ParseSingleFlagCapability(set_flags, set_svc_bits, descriptor, page_table); if (result.IsError()) { + LOG_ERROR( + Kernel, + "Failed to parse capability flag! set_flags={}, set_svc_bits={}, descriptor={}", + set_flags, set_svc_bits, descriptor); return result; } } @@ -144,7 +153,7 @@ ResultCode ProcessCapabilities::ParseCapabilities(const u32* capabilities, } ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& set_svc_bits, - u32 flag, VMManager& vm_manager) { + u32 flag, Memory::PageTable& page_table) { const auto type = GetCapabilityType(flag); if (type == CapabilityType::Unset) { @@ -162,6 +171,9 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s const u32 flag_length = GetFlagBitOffset(type); const u32 set_flag = 1U << flag_length; if ((set_flag & set_flags & InitializeOnceMask) != 0) { + LOG_ERROR(Kernel, + "Attempted to initialize flags that may only be initialized once. set_flags={}", + set_flags); return ERR_INVALID_COMBINATION; } set_flags |= set_flag; @@ -172,7 +184,7 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s case CapabilityType::Syscall: return HandleSyscallFlags(set_svc_bits, flag); case CapabilityType::MapIO: - return HandleMapIOFlags(flag, vm_manager); + return HandleMapIOFlags(flag, page_table); case CapabilityType::Interrupt: return HandleInterruptFlags(flag); case CapabilityType::ProgramType: @@ -187,6 +199,7 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s break; } + LOG_ERROR(Kernel, "Invalid capability type! type={}", static_cast<u32>(type)); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } @@ -208,23 +221,31 @@ void ProcessCapabilities::Clear() { ResultCode ProcessCapabilities::HandlePriorityCoreNumFlags(u32 flags) { if (priority_mask != 0 || core_mask != 0) { + LOG_ERROR(Kernel, "Core or priority mask are not zero! priority_mask={}, core_mask={}", + priority_mask, core_mask); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } const u32 core_num_min = (flags >> 16) & 0xFF; const u32 core_num_max = (flags >> 24) & 0xFF; if (core_num_min > core_num_max) { + LOG_ERROR(Kernel, "Core min is greater than core max! core_num_min={}, core_num_max={}", + core_num_min, core_num_max); return ERR_INVALID_COMBINATION; } const u32 priority_min = (flags >> 10) & 0x3F; const u32 priority_max = (flags >> 4) & 0x3F; if (priority_min > priority_max) { + LOG_ERROR(Kernel, + "Priority min is greater than priority max! priority_min={}, priority_max={}", + core_num_min, priority_max); return ERR_INVALID_COMBINATION; } // The switch only has 4 usable cores. if (core_num_max >= 4) { + LOG_ERROR(Kernel, "Invalid max cores specified! core_num_max={}", core_num_max); return ERR_INVALID_PROCESSOR_ID; } @@ -259,6 +280,7 @@ ResultCode ProcessCapabilities::HandleSyscallFlags(u32& set_svc_bits, u32 flags) } if (svc_number >= svc_capabilities.size()) { + LOG_ERROR(Kernel, "Process svc capability is out of range! svc_number={}", svc_number); return ERR_OUT_OF_RANGE; } @@ -269,12 +291,12 @@ ResultCode ProcessCapabilities::HandleSyscallFlags(u32& set_svc_bits, u32 flags) } ResultCode ProcessCapabilities::HandleMapPhysicalFlags(u32 flags, u32 size_flags, - VMManager& vm_manager) { + Memory::PageTable& page_table) { // TODO(Lioncache): Implement once the memory manager can handle this. return RESULT_SUCCESS; } -ResultCode ProcessCapabilities::HandleMapIOFlags(u32 flags, VMManager& vm_manager) { +ResultCode ProcessCapabilities::HandleMapIOFlags(u32 flags, Memory::PageTable& page_table) { // TODO(Lioncache): Implement once the memory manager can handle this. return RESULT_SUCCESS; } @@ -295,6 +317,8 @@ ResultCode ProcessCapabilities::HandleInterruptFlags(u32 flags) { // emulate that, it's sufficient to mark every interrupt as defined. if (interrupt >= interrupt_capabilities.size()) { + LOG_ERROR(Kernel, "Process interrupt capability is out of range! svc_number={}", + interrupt); return ERR_OUT_OF_RANGE; } @@ -307,6 +331,7 @@ ResultCode ProcessCapabilities::HandleInterruptFlags(u32 flags) { ResultCode ProcessCapabilities::HandleProgramTypeFlags(u32 flags) { const u32 reserved = flags >> 17; if (reserved != 0) { + LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved); return ERR_RESERVED_VALUE; } @@ -324,6 +349,9 @@ ResultCode ProcessCapabilities::HandleKernelVersionFlags(u32 flags) { const u32 major_version = kernel_version >> 19; if (major_version != 0 || flags < 0x80000) { + LOG_ERROR(Kernel, + "Kernel version is non zero or flags are too small! major_version={}, flags={}", + major_version, flags); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } @@ -334,6 +362,7 @@ ResultCode ProcessCapabilities::HandleKernelVersionFlags(u32 flags) { ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) { const u32 reserved = flags >> 26; if (reserved != 0) { + LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved); return ERR_RESERVED_VALUE; } @@ -344,6 +373,7 @@ ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) { ResultCode ProcessCapabilities::HandleDebugFlags(u32 flags) { const u32 reserved = flags >> 19; if (reserved != 0) { + LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved); return ERR_RESERVED_VALUE; } diff --git a/src/core/hle/kernel/process_capability.h b/src/core/hle/kernel/process_capability.h index 5cdd80747..ea9d12c16 100644 --- a/src/core/hle/kernel/process_capability.h +++ b/src/core/hle/kernel/process_capability.h @@ -12,7 +12,9 @@ union ResultCode; namespace Kernel { -class VMManager; +namespace Memory { +class PageTable; +} /// The possible types of programs that may be indicated /// by the program type capability descriptor. @@ -81,27 +83,27 @@ public: /// /// @param capabilities The capabilities to parse /// @param num_capabilities The number of capabilities to parse. - /// @param vm_manager The memory manager to use for handling any mapping-related + /// @param page_table The memory manager to use for handling any mapping-related /// operations (such as mapping IO memory, etc). /// /// @returns RESULT_SUCCESS if this capabilities instance was able to be initialized, /// otherwise, an error code upon failure. /// ResultCode InitializeForKernelProcess(const u32* capabilities, std::size_t num_capabilities, - VMManager& vm_manager); + Memory::PageTable& page_table); /// Initializes this process capabilities instance for a userland process. /// /// @param capabilities The capabilities to parse. /// @param num_capabilities The total number of capabilities to parse. - /// @param vm_manager The memory manager to use for handling any mapping-related + /// @param page_table The memory manager to use for handling any mapping-related /// operations (such as mapping IO memory, etc). /// /// @returns RESULT_SUCCESS if this capabilities instance was able to be initialized, /// otherwise, an error code upon failure. /// ResultCode InitializeForUserProcess(const u32* capabilities, std::size_t num_capabilities, - VMManager& vm_manager); + Memory::PageTable& page_table); /// Initializes this process capabilities instance for a process that does not /// have any metadata to parse. @@ -181,13 +183,13 @@ private: /// /// @param capabilities The sequence of capability descriptors to parse. /// @param num_capabilities The number of descriptors within the given sequence. - /// @param vm_manager The memory manager that will perform any memory + /// @param page_table The memory manager that will perform any memory /// mapping if necessary. /// /// @return RESULT_SUCCESS if no errors occur, otherwise an error code. /// ResultCode ParseCapabilities(const u32* capabilities, std::size_t num_capabilities, - VMManager& vm_manager); + Memory::PageTable& page_table); /// Attempts to parse a capability descriptor that is only represented by a /// single flag set. @@ -196,13 +198,13 @@ private: /// flags being initialized more than once when they shouldn't be. /// @param set_svc_bits Running set of bits representing the allowed supervisor calls mask. /// @param flag The flag to attempt to parse. - /// @param vm_manager The memory manager that will perform any memory + /// @param page_table The memory manager that will perform any memory /// mapping if necessary. /// /// @return RESULT_SUCCESS if no errors occurred, otherwise an error code. /// ResultCode ParseSingleFlagCapability(u32& set_flags, u32& set_svc_bits, u32 flag, - VMManager& vm_manager); + Memory::PageTable& page_table); /// Clears the internal state of this process capability instance. Necessary, /// to have a sane starting point due to us allowing running executables without @@ -226,10 +228,10 @@ private: ResultCode HandleSyscallFlags(u32& set_svc_bits, u32 flags); /// Handles flags related to mapping physical memory pages. - ResultCode HandleMapPhysicalFlags(u32 flags, u32 size_flags, VMManager& vm_manager); + ResultCode HandleMapPhysicalFlags(u32 flags, u32 size_flags, Memory::PageTable& page_table); /// Handles flags related to mapping IO pages. - ResultCode HandleMapIOFlags(u32 flags, VMManager& vm_manager); + ResultCode HandleMapIOFlags(u32 flags, Memory::PageTable& page_table); /// Handles flags related to the interrupt capability flags. ResultCode HandleInterruptFlags(u32 flags); diff --git a/src/core/hle/kernel/readable_event.cpp b/src/core/hle/kernel/readable_event.cpp index 9d3d3a81b..6e286419e 100644 --- a/src/core/hle/kernel/readable_event.cpp +++ b/src/core/hle/kernel/readable_event.cpp @@ -4,9 +4,12 @@ #include <algorithm> #include "common/assert.h" +#include "common/logging/log.h" #include "core/hle/kernel/errors.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/readable_event.h" +#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" namespace Kernel { @@ -23,10 +26,12 @@ void ReadableEvent::Acquire(Thread* thread) { } void ReadableEvent::Signal() { - if (!is_signaled) { - is_signaled = true; - SynchronizationObject::Signal(); - }; + if (is_signaled) { + return; + } + + is_signaled = true; + SynchronizationObject::Signal(); } void ReadableEvent::Clear() { @@ -34,7 +39,10 @@ void ReadableEvent::Clear() { } ResultCode ReadableEvent::Reset() { + SchedulerLock lock(kernel); if (!is_signaled) { + LOG_TRACE(Kernel, "Handle is not signaled! object_id={}, object_type={}, object_name={}", + GetObjectId(), GetTypeName(), GetName()); return ERR_INVALID_STATE; } diff --git a/src/core/hle/kernel/resource_limit.cpp b/src/core/hle/kernel/resource_limit.cpp index b53423462..212e442f4 100644 --- a/src/core/hle/kernel/resource_limit.cpp +++ b/src/core/hle/kernel/resource_limit.cpp @@ -16,26 +16,58 @@ constexpr std::size_t ResourceTypeToIndex(ResourceType type) { ResourceLimit::ResourceLimit(KernelCore& kernel) : Object{kernel} {} ResourceLimit::~ResourceLimit() = default; +bool ResourceLimit::Reserve(ResourceType resource, s64 amount) { + return Reserve(resource, amount, 10000000000); +} + +bool ResourceLimit::Reserve(ResourceType resource, s64 amount, u64 timeout) { + const std::size_t index{ResourceTypeToIndex(resource)}; + + s64 new_value = current[index] + amount; + if (new_value > limit[index] && available[index] + amount <= limit[index]) { + // TODO(bunnei): This is wrong for multicore, we should wait the calling thread for timeout + new_value = current[index] + amount; + } + + if (new_value <= limit[index]) { + current[index] = new_value; + return true; + } + return false; +} + +void ResourceLimit::Release(ResourceType resource, u64 amount) { + Release(resource, amount, amount); +} + +void ResourceLimit::Release(ResourceType resource, u64 used_amount, u64 available_amount) { + const std::size_t index{ResourceTypeToIndex(resource)}; + + current[index] -= used_amount; + available[index] -= available_amount; +} + std::shared_ptr<ResourceLimit> ResourceLimit::Create(KernelCore& kernel) { return std::make_shared<ResourceLimit>(kernel); } s64 ResourceLimit::GetCurrentResourceValue(ResourceType resource) const { - return values.at(ResourceTypeToIndex(resource)); + return limit.at(ResourceTypeToIndex(resource)) - current.at(ResourceTypeToIndex(resource)); } s64 ResourceLimit::GetMaxResourceValue(ResourceType resource) const { - return limits.at(ResourceTypeToIndex(resource)); + return limit.at(ResourceTypeToIndex(resource)); } ResultCode ResourceLimit::SetLimitValue(ResourceType resource, s64 value) { - const auto index = ResourceTypeToIndex(resource); - - if (value < values[index]) { + const std::size_t index{ResourceTypeToIndex(resource)}; + if (current[index] <= value) { + limit[index] = value; + return RESULT_SUCCESS; + } else { + LOG_ERROR(Kernel, "Limit value is too large! resource={}, value={}, index={}", + static_cast<u32>(resource), value, index); return ERR_INVALID_STATE; } - - values[index] = value; - return RESULT_SUCCESS; } } // namespace Kernel diff --git a/src/core/hle/kernel/resource_limit.h b/src/core/hle/kernel/resource_limit.h index 53b89e621..936cc4d0f 100644 --- a/src/core/hle/kernel/resource_limit.h +++ b/src/core/hle/kernel/resource_limit.h @@ -51,6 +51,11 @@ public: return HANDLE_TYPE; } + bool Reserve(ResourceType resource, s64 amount); + bool Reserve(ResourceType resource, s64 amount, u64 timeout); + void Release(ResourceType resource, u64 amount); + void Release(ResourceType resource, u64 used_amount, u64 available_amount); + /** * Gets the current value for the specified resource. * @param resource Requested resource type @@ -91,10 +96,9 @@ private: using ResourceArray = std::array<s64, static_cast<std::size_t>(ResourceType::ResourceTypeCount)>; - /// Maximum values a resource type may reach. - ResourceArray limits{}; - /// Current resource limit values. - ResourceArray values{}; + ResourceArray limit{}; + ResourceArray current{}; + ResourceArray available{}; }; } // namespace Kernel diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp index 1140c72a3..6b7db5372 100644 --- a/src/core/hle/kernel/scheduler.cpp +++ b/src/core/hle/kernel/scheduler.cpp @@ -6,16 +6,21 @@ // licensed under GPLv2 or later under exception provided by the author. #include <algorithm> +#include <mutex> #include <set> #include <unordered_set> #include <utility> #include "common/assert.h" +#include "common/bit_util.h" +#include "common/fiber.h" #include "common/logging/log.h" #include "core/arm/arm_interface.h" #include "core/core.h" #include "core/core_timing.h" +#include "core/cpu_manager.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/physical_core.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/time_manager.h" @@ -27,103 +32,149 @@ GlobalScheduler::GlobalScheduler(KernelCore& kernel) : kernel{kernel} {} GlobalScheduler::~GlobalScheduler() = default; void GlobalScheduler::AddThread(std::shared_ptr<Thread> thread) { + std::scoped_lock lock{global_list_guard}; thread_list.push_back(std::move(thread)); } void GlobalScheduler::RemoveThread(std::shared_ptr<Thread> thread) { + std::scoped_lock lock{global_list_guard}; thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread), thread_list.end()); } -void GlobalScheduler::UnloadThread(std::size_t core) { - Scheduler& sched = kernel.Scheduler(core); - sched.UnloadThread(); -} - -void GlobalScheduler::SelectThread(std::size_t core) { +u32 GlobalScheduler::SelectThreads() { + ASSERT(is_locked); const auto update_thread = [](Thread* thread, Scheduler& sched) { - if (thread != sched.selected_thread.get()) { + std::scoped_lock lock{sched.guard}; + if (thread != sched.selected_thread_set.get()) { if (thread == nullptr) { ++sched.idle_selection_count; } - sched.selected_thread = SharedFrom(thread); + sched.selected_thread_set = SharedFrom(thread); } - sched.is_context_switch_pending = sched.selected_thread != sched.current_thread; + const bool reschedule_pending = + sched.is_context_switch_pending || (sched.selected_thread_set != sched.current_thread); + sched.is_context_switch_pending = reschedule_pending; std::atomic_thread_fence(std::memory_order_seq_cst); + return reschedule_pending; }; - Scheduler& sched = kernel.Scheduler(core); - Thread* current_thread = nullptr; - // Step 1: Get top thread in schedule queue. - current_thread = scheduled_queue[core].empty() ? nullptr : scheduled_queue[core].front(); - if (current_thread) { - update_thread(current_thread, sched); - return; + if (!is_reselection_pending.load()) { + return 0; } - // Step 2: Try selecting a suggested thread. - Thread* winner = nullptr; - std::set<s32> sug_cores; - for (auto thread : suggested_queue[core]) { - s32 this_core = thread->GetProcessorID(); - Thread* thread_on_core = nullptr; - if (this_core >= 0) { - thread_on_core = scheduled_queue[this_core].front(); - } - if (this_core < 0 || thread != thread_on_core) { - winner = thread; - break; + std::array<Thread*, Core::Hardware::NUM_CPU_CORES> top_threads{}; + + u32 idle_cores{}; + + // Step 1: Get top thread in schedule queue. + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + Thread* top_thread = + scheduled_queue[core].empty() ? nullptr : scheduled_queue[core].front(); + if (top_thread != nullptr) { + // TODO(Blinkhawk): Implement Thread Pinning + } else { + idle_cores |= (1U << core); } - sug_cores.insert(this_core); + top_threads[core] = top_thread; } - // if we got a suggested thread, select it, else do a second pass. - if (winner && winner->GetPriority() > 2) { - if (winner->IsRunning()) { - UnloadThread(static_cast<u32>(winner->GetProcessorID())); + + while (idle_cores != 0) { + u32 core_id = Common::CountTrailingZeroes32(idle_cores); + + if (!suggested_queue[core_id].empty()) { + std::array<s32, Core::Hardware::NUM_CPU_CORES> migration_candidates{}; + std::size_t num_candidates = 0; + auto iter = suggested_queue[core_id].begin(); + Thread* suggested = nullptr; + // Step 2: Try selecting a suggested thread. + while (iter != suggested_queue[core_id].end()) { + suggested = *iter; + iter++; + s32 suggested_core_id = suggested->GetProcessorID(); + Thread* top_thread = + suggested_core_id >= 0 ? top_threads[suggested_core_id] : nullptr; + if (top_thread != suggested) { + if (top_thread != nullptr && + top_thread->GetPriority() < THREADPRIO_MAX_CORE_MIGRATION) { + suggested = nullptr; + break; + // There's a too high thread to do core migration, cancel + } + TransferToCore(suggested->GetPriority(), static_cast<s32>(core_id), suggested); + break; + } + suggested = nullptr; + migration_candidates[num_candidates++] = suggested_core_id; + } + // Step 3: Select a suggested thread from another core + if (suggested == nullptr) { + for (std::size_t i = 0; i < num_candidates; i++) { + s32 candidate_core = migration_candidates[i]; + suggested = top_threads[candidate_core]; + auto it = scheduled_queue[candidate_core].begin(); + it++; + Thread* next = it != scheduled_queue[candidate_core].end() ? *it : nullptr; + if (next != nullptr) { + TransferToCore(suggested->GetPriority(), static_cast<s32>(core_id), + suggested); + top_threads[candidate_core] = next; + break; + } else { + suggested = nullptr; + } + } + } + top_threads[core_id] = suggested; } - TransferToCore(winner->GetPriority(), static_cast<s32>(core), winner); - update_thread(winner, sched); - return; + + idle_cores &= ~(1U << core_id); } - // Step 3: Select a suggested thread from another core - for (auto& src_core : sug_cores) { - auto it = scheduled_queue[src_core].begin(); - it++; - if (it != scheduled_queue[src_core].end()) { - Thread* thread_on_core = scheduled_queue[src_core].front(); - Thread* to_change = *it; - if (thread_on_core->IsRunning() || to_change->IsRunning()) { - UnloadThread(static_cast<u32>(src_core)); - } - TransferToCore(thread_on_core->GetPriority(), static_cast<s32>(core), thread_on_core); - current_thread = thread_on_core; - break; + u32 cores_needing_context_switch{}; + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + Scheduler& sched = kernel.Scheduler(core); + ASSERT(top_threads[core] == nullptr || + static_cast<u32>(top_threads[core]->GetProcessorID()) == core); + if (update_thread(top_threads[core], sched)) { + cores_needing_context_switch |= (1U << core); } } - update_thread(current_thread, sched); + return cores_needing_context_switch; } bool GlobalScheduler::YieldThread(Thread* yielding_thread) { + ASSERT(is_locked); // Note: caller should use critical section, etc. + if (!yielding_thread->IsRunnable()) { + // Normally this case shouldn't happen except for SetThreadActivity. + is_reselection_pending.store(true, std::memory_order_release); + return false; + } const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID()); const u32 priority = yielding_thread->GetPriority(); // Yield the thread - const Thread* const winner = scheduled_queue[core_id].front(priority); - ASSERT_MSG(yielding_thread == winner, "Thread yielding without being in front"); - scheduled_queue[core_id].yield(priority); + Reschedule(priority, core_id, yielding_thread); + const Thread* const winner = scheduled_queue[core_id].front(); + if (kernel.GetCurrentHostThreadID() != core_id) { + is_reselection_pending.store(true, std::memory_order_release); + } return AskForReselectionOrMarkRedundant(yielding_thread, winner); } bool GlobalScheduler::YieldThreadAndBalanceLoad(Thread* yielding_thread) { + ASSERT(is_locked); // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section, // etc. + if (!yielding_thread->IsRunnable()) { + // Normally this case shouldn't happen except for SetThreadActivity. + is_reselection_pending.store(true, std::memory_order_release); + return false; + } const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID()); const u32 priority = yielding_thread->GetPriority(); // Yield the thread - ASSERT_MSG(yielding_thread == scheduled_queue[core_id].front(priority), - "Thread yielding without being in front"); - scheduled_queue[core_id].yield(priority); + Reschedule(priority, core_id, yielding_thread); std::array<Thread*, Core::Hardware::NUM_CPU_CORES> current_threads; for (std::size_t i = 0; i < current_threads.size(); i++) { @@ -153,21 +204,28 @@ bool GlobalScheduler::YieldThreadAndBalanceLoad(Thread* yielding_thread) { if (winner != nullptr) { if (winner != yielding_thread) { - if (winner->IsRunning()) { - UnloadThread(static_cast<u32>(winner->GetProcessorID())); - } TransferToCore(winner->GetPriority(), s32(core_id), winner); } } else { winner = next_thread; } + if (kernel.GetCurrentHostThreadID() != core_id) { + is_reselection_pending.store(true, std::memory_order_release); + } + return AskForReselectionOrMarkRedundant(yielding_thread, winner); } bool GlobalScheduler::YieldThreadAndWaitForLoadBalancing(Thread* yielding_thread) { + ASSERT(is_locked); // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section, // etc. + if (!yielding_thread->IsRunnable()) { + // Normally this case shouldn't happen except for SetThreadActivity. + is_reselection_pending.store(true, std::memory_order_release); + return false; + } Thread* winner = nullptr; const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID()); @@ -195,25 +253,31 @@ bool GlobalScheduler::YieldThreadAndWaitForLoadBalancing(Thread* yielding_thread } if (winner != nullptr) { if (winner != yielding_thread) { - if (winner->IsRunning()) { - UnloadThread(static_cast<u32>(winner->GetProcessorID())); - } TransferToCore(winner->GetPriority(), static_cast<s32>(core_id), winner); } } else { winner = yielding_thread; } + } else { + winner = scheduled_queue[core_id].front(); + } + + if (kernel.GetCurrentHostThreadID() != core_id) { + is_reselection_pending.store(true, std::memory_order_release); } return AskForReselectionOrMarkRedundant(yielding_thread, winner); } void GlobalScheduler::PreemptThreads() { + ASSERT(is_locked); for (std::size_t core_id = 0; core_id < Core::Hardware::NUM_CPU_CORES; core_id++) { const u32 priority = preemption_priorities[core_id]; if (scheduled_queue[core_id].size(priority) > 0) { - scheduled_queue[core_id].front(priority)->IncrementYieldCount(); + if (scheduled_queue[core_id].size(priority) > 1) { + scheduled_queue[core_id].front(priority)->IncrementYieldCount(); + } scheduled_queue[core_id].yield(priority); if (scheduled_queue[core_id].size(priority) > 1) { scheduled_queue[core_id].front(priority)->IncrementYieldCount(); @@ -247,9 +311,6 @@ void GlobalScheduler::PreemptThreads() { } if (winner != nullptr) { - if (winner->IsRunning()) { - UnloadThread(static_cast<u32>(winner->GetProcessorID())); - } TransferToCore(winner->GetPriority(), s32(core_id), winner); current_thread = winner->GetPriority() <= current_thread->GetPriority() ? winner : current_thread; @@ -280,9 +341,6 @@ void GlobalScheduler::PreemptThreads() { } if (winner != nullptr) { - if (winner->IsRunning()) { - UnloadThread(static_cast<u32>(winner->GetProcessorID())); - } TransferToCore(winner->GetPriority(), s32(core_id), winner); current_thread = winner; } @@ -292,34 +350,65 @@ void GlobalScheduler::PreemptThreads() { } } +void GlobalScheduler::EnableInterruptAndSchedule(u32 cores_pending_reschedule, + Core::EmuThreadHandle global_thread) { + u32 current_core = global_thread.host_handle; + bool must_context_switch = global_thread.guest_handle != InvalidHandle && + (current_core < Core::Hardware::NUM_CPU_CORES); + while (cores_pending_reschedule != 0) { + u32 core = Common::CountTrailingZeroes32(cores_pending_reschedule); + ASSERT(core < Core::Hardware::NUM_CPU_CORES); + if (!must_context_switch || core != current_core) { + auto& phys_core = kernel.PhysicalCore(core); + phys_core.Interrupt(); + } else { + must_context_switch = true; + } + cores_pending_reschedule &= ~(1U << core); + } + if (must_context_switch) { + auto& core_scheduler = kernel.CurrentScheduler(); + kernel.ExitSVCProfile(); + core_scheduler.TryDoContextSwitch(); + kernel.EnterSVCProfile(); + } +} + void GlobalScheduler::Suggest(u32 priority, std::size_t core, Thread* thread) { + ASSERT(is_locked); suggested_queue[core].add(thread, priority); } void GlobalScheduler::Unsuggest(u32 priority, std::size_t core, Thread* thread) { + ASSERT(is_locked); suggested_queue[core].remove(thread, priority); } void GlobalScheduler::Schedule(u32 priority, std::size_t core, Thread* thread) { + ASSERT(is_locked); ASSERT_MSG(thread->GetProcessorID() == s32(core), "Thread must be assigned to this core."); scheduled_queue[core].add(thread, priority); } void GlobalScheduler::SchedulePrepend(u32 priority, std::size_t core, Thread* thread) { + ASSERT(is_locked); ASSERT_MSG(thread->GetProcessorID() == s32(core), "Thread must be assigned to this core."); scheduled_queue[core].add(thread, priority, false); } void GlobalScheduler::Reschedule(u32 priority, std::size_t core, Thread* thread) { + ASSERT(is_locked); scheduled_queue[core].remove(thread, priority); scheduled_queue[core].add(thread, priority); } void GlobalScheduler::Unschedule(u32 priority, std::size_t core, Thread* thread) { + ASSERT(is_locked); scheduled_queue[core].remove(thread, priority); } void GlobalScheduler::TransferToCore(u32 priority, s32 destination_core, Thread* thread) { + ASSERT(is_locked); const bool schedulable = thread->GetPriority() < THREADPRIO_COUNT; const s32 source_core = thread->GetProcessorID(); if (source_core == destination_core || !schedulable) { @@ -349,6 +438,108 @@ bool GlobalScheduler::AskForReselectionOrMarkRedundant(Thread* current_thread, } } +void GlobalScheduler::AdjustSchedulingOnStatus(Thread* thread, u32 old_flags) { + if (old_flags == thread->scheduling_state) { + return; + } + ASSERT(is_locked); + + if (old_flags == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // In this case the thread was running, now it's pausing/exitting + if (thread->processor_id >= 0) { + Unschedule(thread->current_priority, static_cast<u32>(thread->processor_id), thread); + } + + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + if (core != static_cast<u32>(thread->processor_id) && + ((thread->affinity_mask >> core) & 1) != 0) { + Unsuggest(thread->current_priority, core, thread); + } + } + } else if (thread->scheduling_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // The thread is now set to running from being stopped + if (thread->processor_id >= 0) { + Schedule(thread->current_priority, static_cast<u32>(thread->processor_id), thread); + } + + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + if (core != static_cast<u32>(thread->processor_id) && + ((thread->affinity_mask >> core) & 1) != 0) { + Suggest(thread->current_priority, core, thread); + } + } + } + + SetReselectionPending(); +} + +void GlobalScheduler::AdjustSchedulingOnPriority(Thread* thread, u32 old_priority) { + if (thread->scheduling_state != static_cast<u32>(ThreadSchedStatus::Runnable)) { + return; + } + ASSERT(is_locked); + if (thread->processor_id >= 0) { + Unschedule(old_priority, static_cast<u32>(thread->processor_id), thread); + } + + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + if (core != static_cast<u32>(thread->processor_id) && + ((thread->affinity_mask >> core) & 1) != 0) { + Unsuggest(old_priority, core, thread); + } + } + + if (thread->processor_id >= 0) { + if (thread == kernel.CurrentScheduler().GetCurrentThread()) { + SchedulePrepend(thread->current_priority, static_cast<u32>(thread->processor_id), + thread); + } else { + Schedule(thread->current_priority, static_cast<u32>(thread->processor_id), thread); + } + } + + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + if (core != static_cast<u32>(thread->processor_id) && + ((thread->affinity_mask >> core) & 1) != 0) { + Suggest(thread->current_priority, core, thread); + } + } + thread->IncrementYieldCount(); + SetReselectionPending(); +} + +void GlobalScheduler::AdjustSchedulingOnAffinity(Thread* thread, u64 old_affinity_mask, + s32 old_core) { + if (thread->scheduling_state != static_cast<u32>(ThreadSchedStatus::Runnable) || + thread->current_priority >= THREADPRIO_COUNT) { + return; + } + ASSERT(is_locked); + + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + if (((old_affinity_mask >> core) & 1) != 0) { + if (core == static_cast<u32>(old_core)) { + Unschedule(thread->current_priority, core, thread); + } else { + Unsuggest(thread->current_priority, core, thread); + } + } + } + + for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { + if (((thread->affinity_mask >> core) & 1) != 0) { + if (core == static_cast<u32>(thread->processor_id)) { + Schedule(thread->current_priority, core, thread); + } else { + Suggest(thread->current_priority, core, thread); + } + } + } + + thread->IncrementYieldCount(); + SetReselectionPending(); +} + void GlobalScheduler::Shutdown() { for (std::size_t core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { scheduled_queue[core].clear(); @@ -359,10 +550,12 @@ void GlobalScheduler::Shutdown() { void GlobalScheduler::Lock() { Core::EmuThreadHandle current_thread = kernel.GetCurrentEmuThreadID(); + ASSERT(!current_thread.IsInvalid()); if (current_thread == current_owner) { ++scope_lock; } else { inner_lock.lock(); + is_locked = true; current_owner = current_thread; ASSERT(current_owner != Core::EmuThreadHandle::InvalidHandle()); scope_lock = 1; @@ -374,17 +567,18 @@ void GlobalScheduler::Unlock() { ASSERT(scope_lock > 0); return; } - for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { - SelectThread(i); - } + u32 cores_pending_reschedule = SelectThreads(); + Core::EmuThreadHandle leaving_thread = current_owner; current_owner = Core::EmuThreadHandle::InvalidHandle(); scope_lock = 1; + is_locked = false; inner_lock.unlock(); - // TODO(Blinkhawk): Setup the interrupts and change context on current core. + EnableInterruptAndSchedule(cores_pending_reschedule, leaving_thread); } -Scheduler::Scheduler(Core::System& system, std::size_t core_id) - : system{system}, core_id{core_id} {} +Scheduler::Scheduler(Core::System& system, std::size_t core_id) : system(system), core_id(core_id) { + switch_fiber = std::make_shared<Common::Fiber>(std::function<void(void*)>(OnSwitch), this); +} Scheduler::~Scheduler() = default; @@ -393,56 +587,122 @@ bool Scheduler::HaveReadyThreads() const { } Thread* Scheduler::GetCurrentThread() const { - return current_thread.get(); + if (current_thread) { + return current_thread.get(); + } + return idle_thread.get(); } Thread* Scheduler::GetSelectedThread() const { return selected_thread.get(); } -void Scheduler::SelectThreads() { - system.GlobalScheduler().SelectThread(core_id); -} - u64 Scheduler::GetLastContextSwitchTicks() const { return last_context_switch_time; } void Scheduler::TryDoContextSwitch() { + auto& phys_core = system.Kernel().CurrentPhysicalCore(); + if (phys_core.IsInterrupted()) { + phys_core.ClearInterrupt(); + } + guard.lock(); if (is_context_switch_pending) { SwitchContext(); + } else { + guard.unlock(); } } -void Scheduler::UnloadThread() { - Thread* const previous_thread = GetCurrentThread(); - Process* const previous_process = system.Kernel().CurrentProcess(); +void Scheduler::OnThreadStart() { + SwitchContextStep2(); +} - UpdateLastContextSwitchTime(previous_thread, previous_process); +void Scheduler::Unload() { + Thread* thread = current_thread.get(); + if (thread) { + thread->SetContinuousOnSVC(false); + thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); + thread->SetIsRunning(false); + if (!thread->IsHLEThread() && !thread->HasExited()) { + Core::ARM_Interface& cpu_core = thread->ArmInterface(); + cpu_core.SaveContext(thread->GetContext32()); + cpu_core.SaveContext(thread->GetContext64()); + // Save the TPIDR_EL0 system register in case it was modified. + thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0()); + cpu_core.ClearExclusiveState(); + } + thread->context_guard.unlock(); + } +} - // Save context for previous thread - if (previous_thread) { - system.ArmInterface(core_id).SaveContext(previous_thread->GetContext32()); - system.ArmInterface(core_id).SaveContext(previous_thread->GetContext64()); - // Save the TPIDR_EL0 system register in case it was modified. - previous_thread->SetTPIDR_EL0(system.ArmInterface(core_id).GetTPIDR_EL0()); +void Scheduler::Reload() { + Thread* thread = current_thread.get(); + if (thread) { + ASSERT_MSG(thread->GetSchedulingStatus() == ThreadSchedStatus::Runnable, + "Thread must be runnable."); + + // Cancel any outstanding wakeup events for this thread + thread->SetIsRunning(true); + thread->SetWasRunning(false); + thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); - if (previous_thread->GetStatus() == ThreadStatus::Running) { - // This is only the case when a reschedule is triggered without the current thread - // yielding execution (i.e. an event triggered, system core time-sliced, etc) - previous_thread->SetStatus(ThreadStatus::Ready); + auto* const thread_owner_process = thread->GetOwnerProcess(); + if (thread_owner_process != nullptr) { + system.Kernel().MakeCurrentProcess(thread_owner_process); + } + if (!thread->IsHLEThread()) { + Core::ARM_Interface& cpu_core = thread->ArmInterface(); + cpu_core.LoadContext(thread->GetContext32()); + cpu_core.LoadContext(thread->GetContext64()); + cpu_core.SetTlsAddress(thread->GetTLSAddress()); + cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0()); + cpu_core.ChangeProcessorID(this->core_id); + cpu_core.ClearExclusiveState(); } - previous_thread->SetIsRunning(false); } - current_thread = nullptr; +} + +void Scheduler::SwitchContextStep2() { + // Load context of new thread + if (selected_thread) { + ASSERT_MSG(selected_thread->GetSchedulingStatus() == ThreadSchedStatus::Runnable, + "Thread must be runnable."); + + // Cancel any outstanding wakeup events for this thread + selected_thread->SetIsRunning(true); + selected_thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); + selected_thread->SetWasRunning(false); + + auto* const thread_owner_process = current_thread->GetOwnerProcess(); + if (thread_owner_process != nullptr) { + system.Kernel().MakeCurrentProcess(thread_owner_process); + } + if (!selected_thread->IsHLEThread()) { + Core::ARM_Interface& cpu_core = selected_thread->ArmInterface(); + cpu_core.LoadContext(selected_thread->GetContext32()); + cpu_core.LoadContext(selected_thread->GetContext64()); + cpu_core.SetTlsAddress(selected_thread->GetTLSAddress()); + cpu_core.SetTPIDR_EL0(selected_thread->GetTPIDR_EL0()); + cpu_core.ChangeProcessorID(this->core_id); + cpu_core.ClearExclusiveState(); + } + } + + TryDoContextSwitch(); } void Scheduler::SwitchContext() { - Thread* const previous_thread = GetCurrentThread(); - Thread* const new_thread = GetSelectedThread(); + current_thread_prev = current_thread; + selected_thread = selected_thread_set; + Thread* previous_thread = current_thread_prev.get(); + Thread* new_thread = selected_thread.get(); + current_thread = selected_thread; is_context_switch_pending = false; + if (new_thread == previous_thread) { + guard.unlock(); return; } @@ -452,51 +712,80 @@ void Scheduler::SwitchContext() { // Save context for previous thread if (previous_thread) { - system.ArmInterface(core_id).SaveContext(previous_thread->GetContext32()); - system.ArmInterface(core_id).SaveContext(previous_thread->GetContext64()); - // Save the TPIDR_EL0 system register in case it was modified. - previous_thread->SetTPIDR_EL0(system.ArmInterface(core_id).GetTPIDR_EL0()); - - if (previous_thread->GetStatus() == ThreadStatus::Running) { - // This is only the case when a reschedule is triggered without the current thread - // yielding execution (i.e. an event triggered, system core time-sliced, etc) - previous_thread->SetStatus(ThreadStatus::Ready); + if (new_thread != nullptr && new_thread->IsSuspendThread()) { + previous_thread->SetWasRunning(true); } + previous_thread->SetContinuousOnSVC(false); + previous_thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); previous_thread->SetIsRunning(false); + if (!previous_thread->IsHLEThread() && !previous_thread->HasExited()) { + Core::ARM_Interface& cpu_core = previous_thread->ArmInterface(); + cpu_core.SaveContext(previous_thread->GetContext32()); + cpu_core.SaveContext(previous_thread->GetContext64()); + // Save the TPIDR_EL0 system register in case it was modified. + previous_thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0()); + cpu_core.ClearExclusiveState(); + } + previous_thread->context_guard.unlock(); } - // Load context of new thread - if (new_thread) { - ASSERT_MSG(new_thread->GetProcessorID() == s32(this->core_id), - "Thread must be assigned to this core."); - ASSERT_MSG(new_thread->GetStatus() == ThreadStatus::Ready, - "Thread must be ready to become running."); + std::shared_ptr<Common::Fiber>* old_context; + if (previous_thread != nullptr) { + old_context = &previous_thread->GetHostContext(); + } else { + old_context = &idle_thread->GetHostContext(); + } + guard.unlock(); - // Cancel any outstanding wakeup events for this thread - new_thread->CancelWakeupTimer(); - current_thread = SharedFrom(new_thread); - new_thread->SetStatus(ThreadStatus::Running); - new_thread->SetIsRunning(true); + Common::Fiber::YieldTo(*old_context, switch_fiber); + /// When a thread wakes up, the scheduler may have changed to other in another core. + auto& next_scheduler = system.Kernel().CurrentScheduler(); + next_scheduler.SwitchContextStep2(); +} - auto* const thread_owner_process = current_thread->GetOwnerProcess(); - if (previous_process != thread_owner_process) { - system.Kernel().MakeCurrentProcess(thread_owner_process); - } +void Scheduler::OnSwitch(void* this_scheduler) { + Scheduler* sched = static_cast<Scheduler*>(this_scheduler); + sched->SwitchToCurrent(); +} - system.ArmInterface(core_id).LoadContext(new_thread->GetContext32()); - system.ArmInterface(core_id).LoadContext(new_thread->GetContext64()); - system.ArmInterface(core_id).SetTlsAddress(new_thread->GetTLSAddress()); - system.ArmInterface(core_id).SetTPIDR_EL0(new_thread->GetTPIDR_EL0()); - } else { - current_thread = nullptr; - // Note: We do not reset the current process and current page table when idling because - // technically we haven't changed processes, our threads are just paused. +void Scheduler::SwitchToCurrent() { + while (true) { + { + std::scoped_lock lock{guard}; + selected_thread = selected_thread_set; + current_thread = selected_thread; + is_context_switch_pending = false; + } + const auto is_switch_pending = [this] { + std::scoped_lock lock{guard}; + return is_context_switch_pending; + }; + do { + if (current_thread != nullptr && !current_thread->IsHLEThread()) { + current_thread->context_guard.lock(); + if (!current_thread->IsRunnable()) { + current_thread->context_guard.unlock(); + break; + } + if (static_cast<u32>(current_thread->GetProcessorID()) != core_id) { + current_thread->context_guard.unlock(); + break; + } + } + std::shared_ptr<Common::Fiber>* next_context; + if (current_thread != nullptr) { + next_context = ¤t_thread->GetHostContext(); + } else { + next_context = &idle_thread->GetHostContext(); + } + Common::Fiber::YieldTo(switch_fiber, *next_context); + } while (!is_switch_pending()); } } void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) { const u64 prev_switch_ticks = last_context_switch_time; - const u64 most_recent_switch_ticks = system.CoreTiming().GetTicks(); + const u64 most_recent_switch_ticks = system.CoreTiming().GetCPUTicks(); const u64 update_ticks = most_recent_switch_ticks - prev_switch_ticks; if (thread != nullptr) { @@ -510,6 +799,16 @@ void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) { last_context_switch_time = most_recent_switch_ticks; } +void Scheduler::Initialize() { + std::string name = "Idle Thread Id:" + std::to_string(core_id); + std::function<void(void*)> init_func = Core::CpuManager::GetIdleThreadStartFunc(); + void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater(); + ThreadType type = static_cast<ThreadType>(THREADTYPE_KERNEL | THREADTYPE_HLE | THREADTYPE_IDLE); + auto thread_res = Thread::Create(system, type, name, 0, 64, 0, static_cast<u32>(core_id), 0, + nullptr, std::move(init_func), init_func_parameter); + idle_thread = std::move(thread_res).Unwrap(); +} + void Scheduler::Shutdown() { current_thread = nullptr; selected_thread = nullptr; @@ -538,4 +837,13 @@ SchedulerLockAndSleep::~SchedulerLockAndSleep() { time_manager.ScheduleTimeEvent(event_handle, time_task, nanoseconds); } +void SchedulerLockAndSleep::Release() { + if (sleep_cancelled) { + return; + } + auto& time_manager = kernel.TimeManager(); + time_manager.ScheduleTimeEvent(event_handle, time_task, nanoseconds); + sleep_cancelled = true; +} + } // namespace Kernel diff --git a/src/core/hle/kernel/scheduler.h b/src/core/hle/kernel/scheduler.h index 07df33f9c..b6f04dcea 100644 --- a/src/core/hle/kernel/scheduler.h +++ b/src/core/hle/kernel/scheduler.h @@ -11,9 +11,14 @@ #include "common/common_types.h" #include "common/multi_level_queue.h" +#include "common/spin_lock.h" #include "core/hardware_properties.h" #include "core/hle/kernel/thread.h" +namespace Common { +class Fiber; +} + namespace Core { class ARM_Interface; class System; @@ -41,41 +46,17 @@ public: return thread_list; } - /** - * Add a thread to the suggested queue of a cpu core. Suggested threads may be - * picked if no thread is scheduled to run on the core. - */ - void Suggest(u32 priority, std::size_t core, Thread* thread); - - /** - * Remove a thread to the suggested queue of a cpu core. Suggested threads may be - * picked if no thread is scheduled to run on the core. - */ - void Unsuggest(u32 priority, std::size_t core, Thread* thread); - - /** - * Add a thread to the scheduling queue of a cpu core. The thread is added at the - * back the queue in its priority level. - */ - void Schedule(u32 priority, std::size_t core, Thread* thread); - - /** - * Add a thread to the scheduling queue of a cpu core. The thread is added at the - * front the queue in its priority level. - */ - void SchedulePrepend(u32 priority, std::size_t core, Thread* thread); + /// Notify the scheduler a thread's status has changed. + void AdjustSchedulingOnStatus(Thread* thread, u32 old_flags); - /// Reschedule an already scheduled thread based on a new priority - void Reschedule(u32 priority, std::size_t core, Thread* thread); - - /// Unschedules a thread. - void Unschedule(u32 priority, std::size_t core, Thread* thread); + /// Notify the scheduler a thread's priority has changed. + void AdjustSchedulingOnPriority(Thread* thread, u32 old_priority); - /// Selects a core and forces it to unload its current thread's context - void UnloadThread(std::size_t core); + /// Notify the scheduler a thread's core and/or affinity mask has changed. + void AdjustSchedulingOnAffinity(Thread* thread, u64 old_affinity_mask, s32 old_core); /** - * Takes care of selecting the new scheduled thread in three steps: + * Takes care of selecting the new scheduled threads in three steps: * * 1. First a thread is selected from the top of the priority queue. If no thread * is obtained then we move to step two, else we are done. @@ -85,8 +66,10 @@ public: * * 3. Third is no suggested thread is found, we do a second pass and pick a running * thread in another core and swap it with its current thread. + * + * returns the cores needing scheduling. */ - void SelectThread(std::size_t core); + u32 SelectThreads(); bool HaveReadyThreads(std::size_t core_id) const { return !scheduled_queue[core_id].empty(); @@ -149,6 +132,40 @@ private: /// Unlocks the scheduler, reselects threads, interrupts cores for rescheduling /// and reschedules current core if needed. void Unlock(); + + void EnableInterruptAndSchedule(u32 cores_pending_reschedule, + Core::EmuThreadHandle global_thread); + + /** + * Add a thread to the suggested queue of a cpu core. Suggested threads may be + * picked if no thread is scheduled to run on the core. + */ + void Suggest(u32 priority, std::size_t core, Thread* thread); + + /** + * Remove a thread to the suggested queue of a cpu core. Suggested threads may be + * picked if no thread is scheduled to run on the core. + */ + void Unsuggest(u32 priority, std::size_t core, Thread* thread); + + /** + * Add a thread to the scheduling queue of a cpu core. The thread is added at the + * back the queue in its priority level. + */ + void Schedule(u32 priority, std::size_t core, Thread* thread); + + /** + * Add a thread to the scheduling queue of a cpu core. The thread is added at the + * front the queue in its priority level. + */ + void SchedulePrepend(u32 priority, std::size_t core, Thread* thread); + + /// Reschedule an already scheduled thread based on a new priority + void Reschedule(u32 priority, std::size_t core, Thread* thread); + + /// Unschedules a thread. + void Unschedule(u32 priority, std::size_t core, Thread* thread); + /** * Transfers a thread into an specific core. If the destination_core is -1 * it will be unscheduled from its source code and added into its suggested @@ -170,10 +187,13 @@ private: std::array<u32, Core::Hardware::NUM_CPU_CORES> preemption_priorities = {59, 59, 59, 62}; /// Scheduler lock mechanisms. - std::mutex inner_lock{}; // TODO(Blinkhawk): Replace for a SpinLock + bool is_locked{}; + std::mutex inner_lock; std::atomic<s64> scope_lock{}; Core::EmuThreadHandle current_owner{Core::EmuThreadHandle::InvalidHandle()}; + Common::SpinLock global_list_guard{}; + /// Lists all thread ids that aren't deleted/etc. std::vector<std::shared_ptr<Thread>> thread_list; KernelCore& kernel; @@ -190,11 +210,11 @@ public: /// Reschedules to the next available thread (call after current thread is suspended) void TryDoContextSwitch(); - /// Unloads currently running thread - void UnloadThread(); - - /// Select the threads in top of the scheduling multilist. - void SelectThreads(); + /// The next two are for SingleCore Only. + /// Unload current thread before preempting core. + void Unload(); + /// Reload current thread after core preemption. + void Reload(); /// Gets the current running thread Thread* GetCurrentThread() const; @@ -209,15 +229,30 @@ public: return is_context_switch_pending; } + void Initialize(); + /// Shutdowns the scheduler. void Shutdown(); + void OnThreadStart(); + + std::shared_ptr<Common::Fiber>& ControlContext() { + return switch_fiber; + } + + const std::shared_ptr<Common::Fiber>& ControlContext() const { + return switch_fiber; + } + private: friend class GlobalScheduler; /// Switches the CPU's active thread context to that of the specified thread void SwitchContext(); + /// When a thread wakes up, it must run this through it's new scheduler + void SwitchContextStep2(); + /** * Called on every context switch to update the internal timestamp * This also updates the running time ticks for the given thread and @@ -231,20 +266,30 @@ private: */ void UpdateLastContextSwitchTime(Thread* thread, Process* process); + static void OnSwitch(void* this_scheduler); + void SwitchToCurrent(); + std::shared_ptr<Thread> current_thread = nullptr; std::shared_ptr<Thread> selected_thread = nullptr; + std::shared_ptr<Thread> current_thread_prev = nullptr; + std::shared_ptr<Thread> selected_thread_set = nullptr; + std::shared_ptr<Thread> idle_thread = nullptr; + + std::shared_ptr<Common::Fiber> switch_fiber = nullptr; Core::System& system; u64 last_context_switch_time = 0; u64 idle_selection_count = 0; const std::size_t core_id; + Common::SpinLock guard{}; + bool is_context_switch_pending = false; }; class SchedulerLock { public: - explicit SchedulerLock(KernelCore& kernel); + [[nodiscard]] explicit SchedulerLock(KernelCore& kernel); ~SchedulerLock(); protected: @@ -261,6 +306,8 @@ public: sleep_cancelled = true; } + void Release(); + private: Handle& event_handle; Thread* time_task; diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp index 4604e35c5..8c19f2534 100644 --- a/src/core/hle/kernel/server_session.cpp +++ b/src/core/hle/kernel/server_session.cpp @@ -8,7 +8,6 @@ #include "common/assert.h" #include "common/common_types.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/core_timing.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/client_port.h" @@ -17,6 +16,7 @@ #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" +#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/server_session.h" #include "core/hle/kernel/session.h" #include "core/hle/kernel/thread.h" @@ -32,8 +32,10 @@ ResultVal<std::shared_ptr<ServerSession>> ServerSession::Create(KernelCore& kern std::string name) { std::shared_ptr<ServerSession> session{std::make_shared<ServerSession>(kernel)}; - session->request_event = Core::Timing::CreateEvent( - name, [session](u64 userdata, s64 cycles_late) { session->CompleteSyncRequest(); }); + session->request_event = + Core::Timing::CreateEvent(name, [session](std::uintptr_t, std::chrono::nanoseconds) { + session->CompleteSyncRequest(); + }); session->name = std::move(name); session->parent = std::move(parent); @@ -134,10 +136,11 @@ ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& con return RESULT_SUCCESS; } -ResultCode ServerSession::QueueSyncRequest(std::shared_ptr<Thread> thread, Memory::Memory& memory) { +ResultCode ServerSession::QueueSyncRequest(std::shared_ptr<Thread> thread, + Core::Memory::Memory& memory) { u32* cmd_buf{reinterpret_cast<u32*>(memory.GetPointer(thread->GetTLSAddress()))}; - std::shared_ptr<Kernel::HLERequestContext> context{ - std::make_shared<Kernel::HLERequestContext>(SharedFrom(this), std::move(thread))}; + auto context = + std::make_shared<HLERequestContext>(kernel, memory, SharedFrom(this), std::move(thread)); context->PopulateFromIncomingCommandBuffer(kernel.CurrentProcess()->GetHandleTable(), cmd_buf); request_queue.Push(std::move(context)); @@ -167,9 +170,12 @@ ResultCode ServerSession::CompleteSyncRequest() { } // Some service requests require the thread to block - if (!context.IsThreadWaiting()) { - context.GetThread().ResumeFromWait(); - context.GetThread().SetWaitSynchronizationResult(result); + { + SchedulerLock lock(kernel); + if (!context.IsThreadWaiting()) { + context.GetThread().ResumeFromWait(); + context.GetThread().SetSynchronizationResults(nullptr, result); + } } request_queue.Pop(); @@ -178,9 +184,12 @@ ResultCode ServerSession::CompleteSyncRequest() { } ResultCode ServerSession::HandleSyncRequest(std::shared_ptr<Thread> thread, - Memory::Memory& memory) { - Core::System::GetInstance().CoreTiming().ScheduleEvent(20000, request_event, {}); - return QueueSyncRequest(std::move(thread), memory); + Core::Memory::Memory& memory, + Core::Timing::CoreTiming& core_timing) { + const ResultCode result = QueueSyncRequest(std::move(thread), memory); + const auto delay = std::chrono::nanoseconds{kernel.IsMulticore() ? 0 : 20000}; + core_timing.ScheduleEvent(delay, request_event, {}); + return result; } } // namespace Kernel diff --git a/src/core/hle/kernel/server_session.h b/src/core/hle/kernel/server_session.h index 77e4f6721..d23e9ec68 100644 --- a/src/core/hle/kernel/server_session.h +++ b/src/core/hle/kernel/server_session.h @@ -13,13 +13,14 @@ #include "core/hle/kernel/synchronization_object.h" #include "core/hle/result.h" -namespace Memory { +namespace Core::Memory { class Memory; } namespace Core::Timing { +class CoreTiming; struct EventType; -} +} // namespace Core::Timing namespace Kernel { @@ -87,12 +88,14 @@ public: /** * Handle a sync request from the emulated application. * - * @param thread Thread that initiated the request. - * @param memory Memory context to handle the sync request under. + * @param thread Thread that initiated the request. + * @param memory Memory context to handle the sync request under. + * @param core_timing Core timing context to schedule the request event under. * * @returns ResultCode from the operation. */ - ResultCode HandleSyncRequest(std::shared_ptr<Thread> thread, Memory::Memory& memory); + ResultCode HandleSyncRequest(std::shared_ptr<Thread> thread, Core::Memory::Memory& memory, + Core::Timing::CoreTiming& core_timing); bool ShouldWait(const Thread* thread) const override; @@ -126,7 +129,7 @@ public: private: /// Queues a sync request from the emulated application. - ResultCode QueueSyncRequest(std::shared_ptr<Thread> thread, Memory::Memory& memory); + ResultCode QueueSyncRequest(std::shared_ptr<Thread> thread, Core::Memory::Memory& memory); /// Completes a sync request from the emulated application. ResultCode CompleteSyncRequest(); diff --git a/src/core/hle/kernel/shared_memory.cpp b/src/core/hle/kernel/shared_memory.cpp index afb2e3fc2..0cd467110 100644 --- a/src/core/hle/kernel/shared_memory.cpp +++ b/src/core/hle/kernel/shared_memory.cpp @@ -2,149 +2,56 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <utility> - #include "common/assert.h" -#include "common/logging/log.h" -#include "core/hle/kernel/errors.h" +#include "core/core.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/shared_memory.h" namespace Kernel { -SharedMemory::SharedMemory(KernelCore& kernel) : Object{kernel} {} -SharedMemory::~SharedMemory() = default; - -std::shared_ptr<SharedMemory> SharedMemory::Create(KernelCore& kernel, Process* owner_process, - u64 size, MemoryPermission permissions, - MemoryPermission other_permissions, - VAddr address, MemoryRegion region, - std::string name) { - std::shared_ptr<SharedMemory> shared_memory = std::make_shared<SharedMemory>(kernel); - - shared_memory->owner_process = owner_process; - shared_memory->name = std::move(name); - shared_memory->size = size; - shared_memory->permissions = permissions; - shared_memory->other_permissions = other_permissions; - - if (address == 0) { - shared_memory->backing_block = std::make_shared<Kernel::PhysicalMemory>(size); - shared_memory->backing_block_offset = 0; - - // Refresh the address mappings for the current process. - if (kernel.CurrentProcess() != nullptr) { - kernel.CurrentProcess()->VMManager().RefreshMemoryBlockMappings( - shared_memory->backing_block.get()); - } - } else { - const auto& vm_manager = shared_memory->owner_process->VMManager(); +SharedMemory::SharedMemory(KernelCore& kernel, Core::DeviceMemory& device_memory) + : Object{kernel}, device_memory{device_memory} {} - // The memory is already available and mapped in the owner process. - const auto vma = vm_manager.FindVMA(address); - ASSERT_MSG(vm_manager.IsValidHandle(vma), "Invalid memory address"); - ASSERT_MSG(vma->second.backing_block, "Backing block doesn't exist for address"); - - // The returned VMA might be a bigger one encompassing the desired address. - const auto vma_offset = address - vma->first; - ASSERT_MSG(vma_offset + size <= vma->second.size, - "Shared memory exceeds bounds of mapped block"); - - shared_memory->backing_block = vma->second.backing_block; - shared_memory->backing_block_offset = vma->second.offset + vma_offset; - } - - shared_memory->base_address = address; +SharedMemory::~SharedMemory() = default; - return shared_memory; -} +std::shared_ptr<SharedMemory> SharedMemory::Create( + KernelCore& kernel, Core::DeviceMemory& device_memory, Process* owner_process, + Memory::PageLinkedList&& page_list, Memory::MemoryPermission owner_permission, + Memory::MemoryPermission user_permission, PAddr physical_address, std::size_t size, + std::string name) { -std::shared_ptr<SharedMemory> SharedMemory::CreateForApplet( - KernelCore& kernel, std::shared_ptr<Kernel::PhysicalMemory> heap_block, std::size_t offset, - u64 size, MemoryPermission permissions, MemoryPermission other_permissions, std::string name) { - std::shared_ptr<SharedMemory> shared_memory = std::make_shared<SharedMemory>(kernel); + std::shared_ptr<SharedMemory> shared_memory{ + std::make_shared<SharedMemory>(kernel, device_memory)}; - shared_memory->owner_process = nullptr; - shared_memory->name = std::move(name); + shared_memory->owner_process = owner_process; + shared_memory->page_list = std::move(page_list); + shared_memory->owner_permission = owner_permission; + shared_memory->user_permission = user_permission; + shared_memory->physical_address = physical_address; shared_memory->size = size; - shared_memory->permissions = permissions; - shared_memory->other_permissions = other_permissions; - shared_memory->backing_block = std::move(heap_block); - shared_memory->backing_block_offset = offset; - shared_memory->base_address = - kernel.CurrentProcess()->VMManager().GetHeapRegionBaseAddress() + offset; + shared_memory->name = name; return shared_memory; } -ResultCode SharedMemory::Map(Process& target_process, VAddr address, MemoryPermission permissions, - MemoryPermission other_permissions) { - const MemoryPermission own_other_permissions = - &target_process == owner_process ? this->permissions : this->other_permissions; - - // Automatically allocated memory blocks can only be mapped with other_permissions = DontCare - if (base_address == 0 && other_permissions != MemoryPermission::DontCare) { - return ERR_INVALID_MEMORY_PERMISSIONS; - } - - // Error out if the requested permissions don't match what the creator process allows. - if (static_cast<u32>(permissions) & ~static_cast<u32>(own_other_permissions)) { - LOG_ERROR(Kernel, "cannot map id={}, address=0x{:X} name={}, permissions don't match", - GetObjectId(), address, name); - return ERR_INVALID_MEMORY_PERMISSIONS; - } +ResultCode SharedMemory::Map(Process& target_process, VAddr address, std::size_t size, + Memory::MemoryPermission permissions) { + const u64 page_count{(size + Memory::PageSize - 1) / Memory::PageSize}; - // Error out if the provided permissions are not compatible with what the creator process needs. - if (other_permissions != MemoryPermission::DontCare && - static_cast<u32>(this->permissions) & ~static_cast<u32>(other_permissions)) { - LOG_ERROR(Kernel, "cannot map id={}, address=0x{:X} name={}, permissions don't match", - GetObjectId(), address, name); - return ERR_INVALID_MEMORY_PERMISSIONS; + if (page_list.GetNumPages() != page_count) { + UNIMPLEMENTED_MSG("Page count does not match"); } - VAddr target_address = address; + const Memory::MemoryPermission expected = + &target_process == owner_process ? owner_permission : user_permission; - // Map the memory block into the target process - auto result = target_process.VMManager().MapMemoryBlock( - target_address, backing_block, backing_block_offset, size, MemoryState::Shared); - if (result.Failed()) { - LOG_ERROR( - Kernel, - "cannot map id={}, target_address=0x{:X} name={}, error mapping to virtual memory", - GetObjectId(), target_address, name); - return result.Code(); + if (permissions != expected) { + UNIMPLEMENTED_MSG("Permission does not match"); } - return target_process.VMManager().ReprotectRange(target_address, size, - ConvertPermissions(permissions)); -} - -ResultCode SharedMemory::Unmap(Process& target_process, VAddr address, u64 unmap_size) { - if (unmap_size != size) { - LOG_ERROR(Kernel, - "Invalid size passed to Unmap. Size must be equal to the size of the " - "memory managed. Shared memory size=0x{:016X}, Unmap size=0x{:016X}", - size, unmap_size); - return ERR_INVALID_SIZE; - } - - // TODO(Subv): Verify what happens if the application tries to unmap an address that is not - // mapped to a SharedMemory. - return target_process.VMManager().UnmapRange(address, size); -} - -VMAPermission SharedMemory::ConvertPermissions(MemoryPermission permission) { - u32 masked_permissions = - static_cast<u32>(permission) & static_cast<u32>(MemoryPermission::ReadWriteExecute); - return static_cast<VMAPermission>(masked_permissions); -} - -u8* SharedMemory::GetPointer(std::size_t offset) { - return backing_block->data() + backing_block_offset + offset; -} - -const u8* SharedMemory::GetPointer(std::size_t offset) const { - return backing_block->data() + backing_block_offset + offset; + return target_process.PageTable().MapPages(address, page_list, Memory::MemoryState::Shared, + permissions); } } // namespace Kernel diff --git a/src/core/hle/kernel/shared_memory.h b/src/core/hle/kernel/shared_memory.h index 014951d82..0ef87235c 100644 --- a/src/core/hle/kernel/shared_memory.h +++ b/src/core/hle/kernel/shared_memory.h @@ -8,8 +8,10 @@ #include <string> #include "common/common_types.h" +#include "core/device_memory.h" +#include "core/hle/kernel/memory/memory_block.h" +#include "core/hle/kernel/memory/page_linked_list.h" #include "core/hle/kernel/object.h" -#include "core/hle/kernel/physical_memory.h" #include "core/hle/kernel/process.h" #include "core/hle/result.h" @@ -17,63 +19,21 @@ namespace Kernel { class KernelCore; -/// Permissions for mapped shared memory blocks -enum class MemoryPermission : u32 { - None = 0, - Read = (1u << 0), - Write = (1u << 1), - ReadWrite = (Read | Write), - Execute = (1u << 2), - ReadExecute = (Read | Execute), - WriteExecute = (Write | Execute), - ReadWriteExecute = (Read | Write | Execute), - DontCare = (1u << 28) -}; - class SharedMemory final : public Object { public: - explicit SharedMemory(KernelCore& kernel); + explicit SharedMemory(KernelCore& kernel, Core::DeviceMemory& device_memory); ~SharedMemory() override; - /** - * Creates a shared memory object. - * @param kernel The kernel instance to create a shared memory instance under. - * @param owner_process Process that created this shared memory object. - * @param size Size of the memory block. Must be page-aligned. - * @param permissions Permission restrictions applied to the process which created the block. - * @param other_permissions Permission restrictions applied to other processes mapping the - * block. - * @param address The address from which to map the Shared Memory. - * @param region If the address is 0, the shared memory will be allocated in this region of the - * linear heap. - * @param name Optional object name, used for debugging purposes. - */ - static std::shared_ptr<SharedMemory> Create(KernelCore& kernel, Process* owner_process, - u64 size, MemoryPermission permissions, - MemoryPermission other_permissions, - VAddr address = 0, - MemoryRegion region = MemoryRegion::BASE, - std::string name = "Unknown"); - - /** - * Creates a shared memory object from a block of memory managed by an HLE applet. - * @param kernel The kernel instance to create a shared memory instance under. - * @param heap_block Heap block of the HLE applet. - * @param offset The offset into the heap block that the SharedMemory will map. - * @param size Size of the memory block. Must be page-aligned. - * @param permissions Permission restrictions applied to the process which created the block. - * @param other_permissions Permission restrictions applied to other processes mapping the - * block. - * @param name Optional object name, used for debugging purposes. - */ - static std::shared_ptr<SharedMemory> CreateForApplet( - KernelCore& kernel, std::shared_ptr<Kernel::PhysicalMemory> heap_block, std::size_t offset, - u64 size, MemoryPermission permissions, MemoryPermission other_permissions, - std::string name = "Unknown Applet"); + static std::shared_ptr<SharedMemory> Create( + KernelCore& kernel, Core::DeviceMemory& device_memory, Process* owner_process, + Memory::PageLinkedList&& page_list, Memory::MemoryPermission owner_permission, + Memory::MemoryPermission user_permission, PAddr physical_address, std::size_t size, + std::string name); std::string GetTypeName() const override { return "SharedMemory"; } + std::string GetName() const override { return name; } @@ -83,71 +43,42 @@ public: return HANDLE_TYPE; } - /// Gets the size of the underlying memory block in bytes. - u64 GetSize() const { - return size; - } - - /** - * Converts the specified MemoryPermission into the equivalent VMAPermission. - * @param permission The MemoryPermission to convert. - */ - static VMAPermission ConvertPermissions(MemoryPermission permission); - /** * Maps a shared memory block to an address in the target process' address space - * @param target_process Process on which to map the memory block. + * @param target_process Process on which to map the memory block * @param address Address in system memory to map shared memory block to + * @param size Size of the shared memory block to map * @param permissions Memory block map permissions (specified by SVC field) - * @param other_permissions Memory block map other permissions (specified by SVC field) - */ - ResultCode Map(Process& target_process, VAddr address, MemoryPermission permissions, - MemoryPermission other_permissions); - - /** - * Unmaps a shared memory block from the specified address in system memory - * - * @param target_process Process from which to unmap the memory block. - * @param address Address in system memory where the shared memory block is mapped. - * @param unmap_size The amount of bytes to unmap from this shared memory instance. - * - * @return Result code of the unmap operation - * - * @pre The given size to unmap must be the same size as the amount of memory managed by - * the SharedMemory instance itself, otherwise ERR_INVALID_SIZE will be returned. */ - ResultCode Unmap(Process& target_process, VAddr address, u64 unmap_size); + ResultCode Map(Process& target_process, VAddr address, std::size_t size, + Memory::MemoryPermission permissions); /** * Gets a pointer to the shared memory block * @param offset Offset from the start of the shared memory block to get pointer * @return A pointer to the shared memory block from the specified offset */ - u8* GetPointer(std::size_t offset = 0); + u8* GetPointer(std::size_t offset = 0) { + return device_memory.GetPointer(physical_address + offset); + } /** - * Gets a constant pointer to the shared memory block + * Gets a pointer to the shared memory block * @param offset Offset from the start of the shared memory block to get pointer - * @return A constant pointer to the shared memory block from the specified offset + * @return A pointer to the shared memory block from the specified offset */ - const u8* GetPointer(std::size_t offset = 0) const; + const u8* GetPointer(std::size_t offset = 0) const { + return device_memory.GetPointer(physical_address + offset); + } private: - /// Backing memory for this shared memory block. - std::shared_ptr<PhysicalMemory> backing_block; - /// Offset into the backing block for this shared memory. - std::size_t backing_block_offset = 0; - /// Size of the memory block. Page-aligned. - u64 size = 0; - /// Permission restrictions applied to the process which created the block. - MemoryPermission permissions{}; - /// Permission restrictions applied to other processes mapping the block. - MemoryPermission other_permissions{}; - /// Process that created this shared memory block. - Process* owner_process; - /// Address of shared memory block in the owner process if specified. - VAddr base_address = 0; - /// Name of shared memory object. + Core::DeviceMemory& device_memory; + Process* owner_process{}; + Memory::PageLinkedList page_list; + Memory::MemoryPermission owner_permission{}; + Memory::MemoryPermission user_permission{}; + PAddr physical_address{}; + std::size_t size{}; std::string name; }; diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index 4ffc113c2..e3b770d66 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -10,30 +10,36 @@ #include "common/alignment.h" #include "common/assert.h" +#include "common/fiber.h" #include "common/logging/log.h" #include "common/microprofile.h" #include "common/string_util.h" #include "core/arm/exclusive_monitor.h" #include "core/core.h" -#include "core/core_manager.h" #include "core/core_timing.h" #include "core/core_timing_util.h" +#include "core/cpu_manager.h" #include "core/hle/kernel/address_arbiter.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/memory_block.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/mutex.h" +#include "core/hle/kernel/physical_core.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/resource_limit.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/shared_memory.h" #include "core/hle/kernel/svc.h" +#include "core/hle/kernel/svc_types.h" #include "core/hle/kernel/svc_wrap.h" #include "core/hle/kernel/synchronization.h" #include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" #include "core/hle/kernel/transfer_memory.h" #include "core/hle/kernel/writable_event.h" #include "core/hle/lock.h" @@ -42,7 +48,7 @@ #include "core/memory.h" #include "core/reporter.h" -namespace Kernel { +namespace Kernel::Svc { namespace { // Checks if address + size is greater than the given address @@ -52,14 +58,11 @@ constexpr bool IsValidAddressRange(VAddr address, u64 size) { return address + size > address; } -// 8 GiB -constexpr u64 MAIN_MEMORY_SIZE = 0x200000000; - // Helper function that performs the common sanity checks for svcMapMemory // and svcUnmapMemory. This is doable, as both functions perform their sanitizing // in the same order. -ResultCode MapUnmapMemorySanityChecks(const VMManager& vm_manager, VAddr dst_addr, VAddr src_addr, - u64 size) { +ResultCode MapUnmapMemorySanityChecks(const Memory::PageTable& manager, VAddr dst_addr, + VAddr src_addr, u64 size) { if (!Common::Is4KBAligned(dst_addr)) { LOG_ERROR(Kernel_SVC, "Destination address is not aligned to 4KB, 0x{:016X}", dst_addr); return ERR_INVALID_ADDRESS; @@ -93,36 +96,33 @@ ResultCode MapUnmapMemorySanityChecks(const VMManager& vm_manager, VAddr dst_add return ERR_INVALID_ADDRESS_STATE; } - if (!vm_manager.IsWithinAddressSpace(src_addr, size)) { + if (!manager.IsInsideAddressSpace(src_addr, size)) { LOG_ERROR(Kernel_SVC, "Source is not within the address space, addr=0x{:016X}, size=0x{:016X}", src_addr, size); return ERR_INVALID_ADDRESS_STATE; } - if (!vm_manager.IsWithinStackRegion(dst_addr, size)) { + if (manager.IsOutsideStackRegion(dst_addr, size)) { LOG_ERROR(Kernel_SVC, "Destination is not within the stack region, addr=0x{:016X}, size=0x{:016X}", dst_addr, size); return ERR_INVALID_MEMORY_RANGE; } - const VAddr dst_end_address = dst_addr + size; - if (dst_end_address > vm_manager.GetHeapRegionBaseAddress() && - vm_manager.GetHeapRegionEndAddress() > dst_addr) { + if (manager.IsInsideHeapRegion(dst_addr, size)) { LOG_ERROR(Kernel_SVC, "Destination does not fit within the heap region, addr=0x{:016X}, " - "size=0x{:016X}, end_addr=0x{:016X}", - dst_addr, size, dst_end_address); + "size=0x{:016X}", + dst_addr, size); return ERR_INVALID_MEMORY_RANGE; } - if (dst_end_address > vm_manager.GetMapRegionBaseAddress() && - vm_manager.GetMapRegionEndAddress() > dst_addr) { + if (manager.IsInsideAliasRegion(dst_addr, size)) { LOG_ERROR(Kernel_SVC, "Destination does not fit within the map region, addr=0x{:016X}, " - "size=0x{:016X}, end_addr=0x{:016X}", - dst_addr, size, dst_end_address); + "size=0x{:016X}", + dst_addr, size); return ERR_INVALID_MEMORY_RANGE; } @@ -136,6 +136,7 @@ enum class ResourceLimitValueType { ResultVal<s64> RetrieveResourceLimitValue(Core::System& system, Handle resource_limit, u32 resource_type, ResourceLimitValueType value_type) { + std::lock_guard lock{HLE::g_hle_lock}; const auto type = static_cast<ResourceType>(resource_type); if (!IsValidResourceType(type)) { LOG_ERROR(Kernel_SVC, "Invalid resource limit type: '{}'", resource_type); @@ -163,6 +164,7 @@ ResultVal<s64> RetrieveResourceLimitValue(Core::System& system, Handle resource_ /// Set the process heap to a given Size. It can both extend and shrink the heap. static ResultCode SetHeapSize(Core::System& system, VAddr* heap_addr, u64 heap_size) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_TRACE(Kernel_SVC, "called, heap_size=0x{:X}", heap_size); // Size must be a multiple of 0x200000 (2MB) and be equal to or less than 8GB. @@ -177,13 +179,10 @@ static ResultCode SetHeapSize(Core::System& system, VAddr* heap_addr, u64 heap_s return ERR_INVALID_SIZE; } - auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); - const auto alloc_result = vm_manager.SetHeapSize(heap_size); - if (alloc_result.Failed()) { - return alloc_result.Code(); - } + auto& page_table{system.Kernel().CurrentProcess()->PageTable()}; + + CASCADE_RESULT(*heap_addr, page_table.SetHeapSize(heap_size)); - *heap_addr = *alloc_result; return RESULT_SUCCESS; } @@ -194,65 +193,9 @@ static ResultCode SetHeapSize32(Core::System& system, u32* heap_addr, u32 heap_s return result; } -static ResultCode SetMemoryPermission(Core::System& system, VAddr addr, u64 size, u32 prot) { - LOG_TRACE(Kernel_SVC, "called, addr=0x{:X}, size=0x{:X}, prot=0x{:X}", addr, size, prot); - - if (!Common::Is4KBAligned(addr)) { - LOG_ERROR(Kernel_SVC, "Address is not aligned to 4KB, addr=0x{:016X}", addr); - return ERR_INVALID_ADDRESS; - } - - if (size == 0) { - LOG_ERROR(Kernel_SVC, "Size is 0"); - return ERR_INVALID_SIZE; - } - - if (!Common::Is4KBAligned(size)) { - LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, size=0x{:016X}", size); - return ERR_INVALID_SIZE; - } - - if (!IsValidAddressRange(addr, size)) { - LOG_ERROR(Kernel_SVC, "Region is not a valid address range, addr=0x{:016X}, size=0x{:016X}", - addr, size); - return ERR_INVALID_ADDRESS_STATE; - } - - const auto permission = static_cast<MemoryPermission>(prot); - if (permission != MemoryPermission::None && permission != MemoryPermission::Read && - permission != MemoryPermission::ReadWrite) { - LOG_ERROR(Kernel_SVC, "Invalid memory permission specified, Got memory permission=0x{:08X}", - static_cast<u32>(permission)); - return ERR_INVALID_MEMORY_PERMISSIONS; - } - - auto* const current_process = system.Kernel().CurrentProcess(); - auto& vm_manager = current_process->VMManager(); - - if (!vm_manager.IsWithinAddressSpace(addr, size)) { - LOG_ERROR(Kernel_SVC, - "Source is not within the address space, addr=0x{:016X}, size=0x{:016X}", addr, - size); - return ERR_INVALID_ADDRESS_STATE; - } - - const VMManager::VMAHandle iter = vm_manager.FindVMA(addr); - if (!vm_manager.IsValidHandle(iter)) { - LOG_ERROR(Kernel_SVC, "Unable to find VMA for address=0x{:016X}", addr); - return ERR_INVALID_ADDRESS_STATE; - } - - LOG_WARNING(Kernel_SVC, "Uniformity check on protected memory is not implemented."); - // TODO: Performs a uniformity check to make sure only protected memory is changed (it doesn't - // make sense to allow changing permissions on kernel memory itself, etc). - - const auto converted_permissions = SharedMemory::ConvertPermissions(permission); - - return vm_manager.ReprotectRange(addr, size, converted_permissions); -} - static ResultCode SetMemoryAttribute(Core::System& system, VAddr address, u64 size, u32 mask, u32 attribute) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_DEBUG(Kernel_SVC, "called, address=0x{:016X}, size=0x{:X}, mask=0x{:08X}, attribute=0x{:08X}", address, size, mask, attribute); @@ -274,72 +217,73 @@ static ResultCode SetMemoryAttribute(Core::System& system, VAddr address, u64 si return ERR_INVALID_ADDRESS_STATE; } - const auto mem_attribute = static_cast<MemoryAttribute>(attribute); - const auto mem_mask = static_cast<MemoryAttribute>(mask); - const auto attribute_with_mask = mem_attribute | mem_mask; - - if (attribute_with_mask != mem_mask) { + const auto attributes{static_cast<Memory::MemoryAttribute>(mask | attribute)}; + if (attributes != static_cast<Memory::MemoryAttribute>(mask) || + (attributes | Memory::MemoryAttribute::Uncached) != Memory::MemoryAttribute::Uncached) { LOG_ERROR(Kernel_SVC, "Memory attribute doesn't match the given mask (Attribute: 0x{:X}, Mask: {:X}", attribute, mask); return ERR_INVALID_COMBINATION; } - if ((attribute_with_mask | MemoryAttribute::Uncached) != MemoryAttribute::Uncached) { - LOG_ERROR(Kernel_SVC, "Specified attribute isn't equal to MemoryAttributeUncached (8)."); - return ERR_INVALID_COMBINATION; - } + auto& page_table{system.Kernel().CurrentProcess()->PageTable()}; - auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); - if (!vm_manager.IsWithinAddressSpace(address, size)) { - LOG_ERROR(Kernel_SVC, - "Given address (0x{:016X}) is outside the bounds of the address space.", address); - return ERR_INVALID_ADDRESS_STATE; - } + return page_table.SetMemoryAttribute(address, size, static_cast<Memory::MemoryAttribute>(mask), + static_cast<Memory::MemoryAttribute>(attribute)); +} - return vm_manager.SetMemoryAttribute(address, size, mem_mask, mem_attribute); +static ResultCode SetMemoryAttribute32(Core::System& system, u32 address, u32 size, u32 mask, + u32 attribute) { + return SetMemoryAttribute(system, static_cast<VAddr>(address), static_cast<std::size_t>(size), + mask, attribute); } /// Maps a memory range into a different range. static ResultCode MapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr, u64 size) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr, src_addr, size); - auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); - const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size); + auto& page_table{system.Kernel().CurrentProcess()->PageTable()}; - if (result.IsError()) { + if (const ResultCode result{MapUnmapMemorySanityChecks(page_table, dst_addr, src_addr, size)}; + result.IsError()) { return result; } - return vm_manager.MirrorMemory(dst_addr, src_addr, size, MemoryState::Stack); + return page_table.Map(dst_addr, src_addr, size); +} + +static ResultCode MapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) { + return MapMemory(system, static_cast<VAddr>(dst_addr), static_cast<VAddr>(src_addr), + static_cast<std::size_t>(size)); } /// Unmaps a region that was previously mapped with svcMapMemory static ResultCode UnmapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr, u64 size) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr, src_addr, size); - auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); - const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size); + auto& page_table{system.Kernel().CurrentProcess()->PageTable()}; - if (result.IsError()) { + if (const ResultCode result{MapUnmapMemorySanityChecks(page_table, dst_addr, src_addr, size)}; + result.IsError()) { return result; } - const auto unmap_res = vm_manager.UnmapRange(dst_addr, size); - - // Reprotect the source mapping on success - if (unmap_res.IsSuccess()) { - ASSERT(vm_manager.ReprotectRange(src_addr, size, VMAPermission::ReadWrite).IsSuccess()); - } + return page_table.Unmap(dst_addr, src_addr, size); +} - return unmap_res; +static ResultCode UnmapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) { + return UnmapMemory(system, static_cast<VAddr>(dst_addr), static_cast<VAddr>(src_addr), + static_cast<std::size_t>(size)); } /// Connect to an OS service given the port name, returns the handle to the port to out static ResultCode ConnectToNamedPort(Core::System& system, Handle* out_handle, VAddr port_name_address) { + std::lock_guard lock{HLE::g_hle_lock}; auto& memory = system.Memory(); if (!memory.IsValidVirtualAddress(port_name_address)) { @@ -367,6 +311,8 @@ static ResultCode ConnectToNamedPort(Core::System& system, Handle* out_handle, return ERR_NOT_FOUND; } + ASSERT(kernel.CurrentProcess()->GetResourceLimit()->Reserve(ResourceType::Sessions, 1)); + auto client_port = it->second; std::shared_ptr<ClientSession> client_session; @@ -396,11 +342,30 @@ static ResultCode SendSyncRequest(Core::System& system, Handle handle) { LOG_TRACE(Kernel_SVC, "called handle=0x{:08X}({})", handle, session->GetName()); auto thread = system.CurrentScheduler().GetCurrentThread(); - thread->InvalidateWakeupCallback(); - thread->SetStatus(ThreadStatus::WaitIPC); - system.PrepareReschedule(thread->GetProcessorID()); + { + SchedulerLock lock(system.Kernel()); + thread->InvalidateHLECallback(); + thread->SetStatus(ThreadStatus::WaitIPC); + session->SendSyncRequest(SharedFrom(thread), system.Memory(), system.CoreTiming()); + } + + if (thread->HasHLECallback()) { + Handle event_handle = thread->GetHLETimeEvent(); + if (event_handle != InvalidHandle) { + auto& time_manager = system.Kernel().TimeManager(); + time_manager.UnscheduleTimeEvent(event_handle); + } + + { + SchedulerLock lock(system.Kernel()); + auto* sync_object = thread->GetHLESyncObject(); + sync_object->RemoveWaitingThread(SharedFrom(thread)); + } + + thread->InvokeHLECallback(SharedFrom(thread)); + } - return session->SendSyncRequest(SharedFrom(thread), system.Memory()); + return thread->GetSignalingResult(); } static ResultCode SendSyncRequest32(Core::System& system, Handle handle) { @@ -462,6 +427,15 @@ static ResultCode GetProcessId(Core::System& system, u64* process_id, Handle han return ERR_INVALID_HANDLE; } +static ResultCode GetProcessId32(Core::System& system, u32* process_id_low, u32* process_id_high, + Handle handle) { + u64 process_id{}; + const auto result = GetProcessId(system, &process_id, handle); + *process_id_low = static_cast<u32>(process_id); + *process_id_high = static_cast<u32>(process_id >> 32); + return result; +} + /// Wait for the given handles to synchronize, timeout after the specified nanoseconds static ResultCode WaitSynchronization(Core::System& system, Handle* index, VAddr handles_address, u64 handle_count, s64 nano_seconds) { @@ -484,9 +458,7 @@ static ResultCode WaitSynchronization(Core::System& system, Handle* index, VAddr return ERR_OUT_OF_RANGE; } - auto* const thread = system.CurrentScheduler().GetCurrentThread(); auto& kernel = system.Kernel(); - using ObjectPtr = Thread::ThreadSynchronizationObjects::value_type; Thread::ThreadSynchronizationObjects objects(handle_count); const auto& handle_table = kernel.CurrentProcess()->GetHandleTable(); @@ -526,10 +498,13 @@ static ResultCode CancelSynchronization(Core::System& system, Handle thread_hand } thread->CancelWait(); - system.PrepareReschedule(thread->GetProcessorID()); return RESULT_SUCCESS; } +static ResultCode CancelSynchronization32(Core::System& system, Handle thread_handle) { + return CancelSynchronization(system, thread_handle); +} + /// Attempts to locks a mutex, creating it if it does not already exist static ResultCode ArbitrateLock(Core::System& system, Handle holding_thread_handle, VAddr mutex_addr, Handle requesting_thread_handle) { @@ -538,7 +513,7 @@ static ResultCode ArbitrateLock(Core::System& system, Handle holding_thread_hand "requesting_current_thread_handle=0x{:08X}", holding_thread_handle, mutex_addr, requesting_thread_handle); - if (Memory::IsKernelVirtualAddress(mutex_addr)) { + if (Core::Memory::IsKernelVirtualAddress(mutex_addr)) { LOG_ERROR(Kernel_SVC, "Mutex Address is a kernel virtual address, mutex_addr={:016X}", mutex_addr); return ERR_INVALID_ADDRESS_STATE; @@ -554,11 +529,17 @@ static ResultCode ArbitrateLock(Core::System& system, Handle holding_thread_hand requesting_thread_handle); } +static ResultCode ArbitrateLock32(Core::System& system, Handle holding_thread_handle, + u32 mutex_addr, Handle requesting_thread_handle) { + return ArbitrateLock(system, holding_thread_handle, static_cast<VAddr>(mutex_addr), + requesting_thread_handle); +} + /// Unlock a mutex static ResultCode ArbitrateUnlock(Core::System& system, VAddr mutex_addr) { LOG_TRACE(Kernel_SVC, "called mutex_addr=0x{:X}", mutex_addr); - if (Memory::IsKernelVirtualAddress(mutex_addr)) { + if (Core::Memory::IsKernelVirtualAddress(mutex_addr)) { LOG_ERROR(Kernel_SVC, "Mutex Address is a kernel virtual address, mutex_addr={:016X}", mutex_addr); return ERR_INVALID_ADDRESS_STATE; @@ -573,6 +554,10 @@ static ResultCode ArbitrateUnlock(Core::System& system, VAddr mutex_addr) { return current_process->GetMutex().Release(mutex_addr); } +static ResultCode ArbitrateUnlock32(Core::System& system, u32 mutex_addr) { + return ArbitrateUnlock(system, static_cast<VAddr>(mutex_addr)); +} + enum class BreakType : u32 { Panic = 0, AssertionFailed = 1, @@ -673,6 +658,7 @@ static void Break(Core::System& system, u32 reason, u64 info1, u64 info2) { info2, has_dumped_buffer ? std::make_optional(debug_buffer) : std::nullopt); if (!break_reason.signal_debugger) { + SchedulerLock lock(system.Kernel()); LOG_CRITICAL( Debug_Emulated, "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}", @@ -683,18 +669,19 @@ static void Break(Core::System& system, u32 reason, u64 info1, u64 info2) { auto* const current_thread = system.CurrentScheduler().GetCurrentThread(); const auto thread_processor_id = current_thread->GetProcessorID(); system.ArmInterface(static_cast<std::size_t>(thread_processor_id)).LogBacktrace(); - ASSERT(false); - - system.Kernel().CurrentProcess()->PrepareForTermination(); // Kill the current thread + system.Kernel().ExceptionalExit(); current_thread->Stop(); - system.PrepareReschedule(); } } +static void Break32(Core::System& system, u32 reason, u32 info1, u32 info2) { + Break(system, reason, static_cast<u64>(info1), static_cast<u64>(info2)); +} + /// Used to output a message on a debug hardware unit - does nothing on a retail unit -static void OutputDebugString([[maybe_unused]] Core::System& system, VAddr address, u64 len) { +static void OutputDebugString(Core::System& system, VAddr address, u64 len) { if (len == 0) { return; } @@ -707,6 +694,7 @@ static void OutputDebugString([[maybe_unused]] Core::System& system, VAddr addre /// Gets system/memory information for the current process static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 handle, u64 info_sub_id) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_TRACE(Kernel_SVC, "called info_id=0x{:X}, info_sub_id=0x{:X}, handle=0x{:08X}", info_id, info_sub_id, handle); @@ -765,6 +753,8 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha case GetInfoType::TotalPhysicalMemoryAvailableWithoutSystemResource: case GetInfoType::TotalPhysicalMemoryUsedWithoutSystemResource: { if (info_sub_id != 0) { + LOG_ERROR(Kernel_SVC, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id, + info_sub_id); return ERR_INVALID_ENUM_VALUE; } @@ -772,6 +762,8 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha system.Kernel().CurrentProcess()->GetHandleTable(); const auto process = current_process_handle_table.Get<Process>(static_cast<Handle>(handle)); if (!process) { + LOG_ERROR(Kernel_SVC, "Process is not valid! info_id={}, info_sub_id={}, handle={:08X}", + info_id, info_sub_id, handle); return ERR_INVALID_HANDLE; } @@ -785,35 +777,35 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha return RESULT_SUCCESS; case GetInfoType::MapRegionBaseAddr: - *result = process->VMManager().GetMapRegionBaseAddress(); + *result = process->PageTable().GetAliasRegionStart(); return RESULT_SUCCESS; case GetInfoType::MapRegionSize: - *result = process->VMManager().GetMapRegionSize(); + *result = process->PageTable().GetAliasRegionSize(); return RESULT_SUCCESS; case GetInfoType::HeapRegionBaseAddr: - *result = process->VMManager().GetHeapRegionBaseAddress(); + *result = process->PageTable().GetHeapRegionStart(); return RESULT_SUCCESS; case GetInfoType::HeapRegionSize: - *result = process->VMManager().GetHeapRegionSize(); + *result = process->PageTable().GetHeapRegionSize(); return RESULT_SUCCESS; case GetInfoType::ASLRRegionBaseAddr: - *result = process->VMManager().GetASLRRegionBaseAddress(); + *result = process->PageTable().GetAliasCodeRegionStart(); return RESULT_SUCCESS; case GetInfoType::ASLRRegionSize: - *result = process->VMManager().GetASLRRegionSize(); + *result = process->PageTable().GetAliasCodeRegionSize(); return RESULT_SUCCESS; case GetInfoType::StackRegionBaseAddr: - *result = process->VMManager().GetStackRegionBaseAddress(); + *result = process->PageTable().GetStackRegionStart(); return RESULT_SUCCESS; case GetInfoType::StackRegionSize: - *result = process->VMManager().GetStackRegionSize(); + *result = process->PageTable().GetStackRegionSize(); return RESULT_SUCCESS; case GetInfoType::TotalPhysicalMemoryAvailable: @@ -853,7 +845,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha break; } - LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id); + LOG_ERROR(Kernel_SVC, "Unimplemented svcGetInfo id=0x{:016X}", info_id); return ERR_INVALID_ENUM_VALUE; } @@ -863,10 +855,13 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha case GetInfoType::RegisterResourceLimit: { if (handle != 0) { + LOG_ERROR(Kernel, "Handle is non zero! handle={:08X}", handle); return ERR_INVALID_HANDLE; } if (info_sub_id != 0) { + LOG_ERROR(Kernel, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id, + info_sub_id); return ERR_INVALID_COMBINATION; } @@ -936,9 +931,9 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha if (same_thread && info_sub_id == 0xFFFFFFFFFFFFFFFF) { const u64 thread_ticks = current_thread->GetTotalCPUTimeTicks(); - out_ticks = thread_ticks + (core_timing.GetTicks() - prev_ctx_ticks); + out_ticks = thread_ticks + (core_timing.GetCPUTicks() - prev_ctx_ticks); } else if (same_thread && info_sub_id == system.CurrentCoreIndex()) { - out_ticks = core_timing.GetTicks() - prev_ctx_ticks; + out_ticks = core_timing.GetCPUTicks() - prev_ctx_ticks; } *result = out_ticks; @@ -946,7 +941,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha } default: - LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id); + LOG_ERROR(Kernel_SVC, "Unimplemented svcGetInfo id=0x{:016X}", info_id); return ERR_INVALID_ENUM_VALUE; } } @@ -965,6 +960,7 @@ static ResultCode GetInfo32(Core::System& system, u32* result_low, u32* result_h /// Maps memory at a desired address static ResultCode MapPhysicalMemory(Core::System& system, VAddr addr, u64 size) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_DEBUG(Kernel_SVC, "called, addr=0x{:016X}, size=0x{:X}", addr, size); if (!Common::Is4KBAligned(addr)) { @@ -987,24 +983,38 @@ static ResultCode MapPhysicalMemory(Core::System& system, VAddr addr, u64 size) return ERR_INVALID_MEMORY_RANGE; } - Process* const current_process = system.Kernel().CurrentProcess(); - auto& vm_manager = current_process->VMManager(); + Process* const current_process{system.Kernel().CurrentProcess()}; + auto& page_table{current_process->PageTable()}; if (current_process->GetSystemResourceSize() == 0) { LOG_ERROR(Kernel_SVC, "System Resource Size is zero"); return ERR_INVALID_STATE; } - if (!vm_manager.IsWithinMapRegion(addr, size)) { - LOG_ERROR(Kernel_SVC, "Range not within map region"); + if (!page_table.IsInsideAddressSpace(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Address is not within the address space, addr=0x{:016X}, size=0x{:016X}", addr, + size); + return ERR_INVALID_MEMORY_RANGE; + } + + if (page_table.IsOutsideAliasRegion(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Address is not within the alias region, addr=0x{:016X}, size=0x{:016X}", addr, + size); return ERR_INVALID_MEMORY_RANGE; } - return vm_manager.MapPhysicalMemory(addr, size); + return page_table.MapPhysicalMemory(addr, size); +} + +static ResultCode MapPhysicalMemory32(Core::System& system, u32 addr, u32 size) { + return MapPhysicalMemory(system, static_cast<VAddr>(addr), static_cast<std::size_t>(size)); } /// Unmaps memory previously mapped via MapPhysicalMemory static ResultCode UnmapPhysicalMemory(Core::System& system, VAddr addr, u64 size) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_DEBUG(Kernel_SVC, "called, addr=0x{:016X}, size=0x{:X}", addr, size); if (!Common::Is4KBAligned(addr)) { @@ -1027,20 +1037,33 @@ static ResultCode UnmapPhysicalMemory(Core::System& system, VAddr addr, u64 size return ERR_INVALID_MEMORY_RANGE; } - Process* const current_process = system.Kernel().CurrentProcess(); - auto& vm_manager = current_process->VMManager(); + Process* const current_process{system.Kernel().CurrentProcess()}; + auto& page_table{current_process->PageTable()}; if (current_process->GetSystemResourceSize() == 0) { LOG_ERROR(Kernel_SVC, "System Resource Size is zero"); return ERR_INVALID_STATE; } - if (!vm_manager.IsWithinMapRegion(addr, size)) { - LOG_ERROR(Kernel_SVC, "Range not within map region"); + if (!page_table.IsInsideAddressSpace(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Address is not within the address space, addr=0x{:016X}, size=0x{:016X}", addr, + size); + return ERR_INVALID_MEMORY_RANGE; + } + + if (page_table.IsOutsideAliasRegion(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Address is not within the alias region, addr=0x{:016X}, size=0x{:016X}", addr, + size); return ERR_INVALID_MEMORY_RANGE; } - return vm_manager.UnmapPhysicalMemory(addr, size); + return page_table.UnmapPhysicalMemory(addr, size); +} + +static ResultCode UnmapPhysicalMemory32(Core::System& system, u32 addr, u32 size) { + return UnmapPhysicalMemory(system, static_cast<VAddr>(addr), static_cast<std::size_t>(size)); } /// Sets the thread activity @@ -1072,10 +1095,11 @@ static ResultCode SetThreadActivity(Core::System& system, Handle handle, u32 act return ERR_BUSY; } - thread->SetActivity(static_cast<ThreadActivity>(activity)); + return thread->SetActivity(static_cast<ThreadActivity>(activity)); +} - system.PrepareReschedule(thread->GetProcessorID()); - return RESULT_SUCCESS; +static ResultCode SetThreadActivity32(Core::System& system, Handle handle, u32 activity) { + return SetThreadActivity(system, handle, activity); } /// Gets the thread context @@ -1119,6 +1143,10 @@ static ResultCode GetThreadContext(Core::System& system, VAddr thread_context, H return RESULT_SUCCESS; } +static ResultCode GetThreadContext32(Core::System& system, u32 thread_context, Handle handle) { + return GetThreadContext(system, static_cast<VAddr>(thread_context), handle); +} + /// Gets the priority for the specified thread static ResultCode GetThreadPriority(Core::System& system, u32* priority, Handle handle) { LOG_TRACE(Kernel_SVC, "called"); @@ -1126,6 +1154,7 @@ static ResultCode GetThreadPriority(Core::System& system, u32* priority, Handle const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const std::shared_ptr<Thread> thread = handle_table.Get<Thread>(handle); if (!thread) { + *priority = 0; LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle); return ERR_INVALID_HANDLE; } @@ -1160,18 +1189,26 @@ static ResultCode SetThreadPriority(Core::System& system, Handle handle, u32 pri thread->SetPriority(priority); - system.PrepareReschedule(thread->GetProcessorID()); return RESULT_SUCCESS; } +static ResultCode SetThreadPriority32(Core::System& system, Handle handle, u32 priority) { + return SetThreadPriority(system, handle, priority); +} + /// Get which CPU core is executing the current thread static u32 GetCurrentProcessorNumber(Core::System& system) { LOG_TRACE(Kernel_SVC, "called"); - return system.CurrentScheduler().GetCurrentThread()->GetProcessorID(); + return static_cast<u32>(system.CurrentPhysicalCore().CoreIndex()); +} + +static u32 GetCurrentProcessorNumber32(Core::System& system) { + return GetCurrentProcessorNumber(system); } static ResultCode MapSharedMemory(Core::System& system, Handle shared_memory_handle, VAddr addr, u64 size, u32 permissions) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_TRACE(Kernel_SVC, "called, shared_memory_handle=0x{:X}, addr=0x{:X}, size=0x{:X}, permissions=0x{:08X}", shared_memory_handle, addr, size, permissions); @@ -1197,79 +1234,61 @@ static ResultCode MapSharedMemory(Core::System& system, Handle shared_memory_han return ERR_INVALID_ADDRESS_STATE; } - const auto permissions_type = static_cast<MemoryPermission>(permissions); - if (permissions_type != MemoryPermission::Read && - permissions_type != MemoryPermission::ReadWrite) { + const auto permission_type = static_cast<Memory::MemoryPermission>(permissions); + if ((permission_type | Memory::MemoryPermission::Write) != + Memory::MemoryPermission::ReadAndWrite) { LOG_ERROR(Kernel_SVC, "Expected Read or ReadWrite permission but got permissions=0x{:08X}", permissions); return ERR_INVALID_MEMORY_PERMISSIONS; } - auto* const current_process = system.Kernel().CurrentProcess(); - auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle); - if (!shared_memory) { - LOG_ERROR(Kernel_SVC, "Shared memory does not exist, shared_memory_handle=0x{:08X}", - shared_memory_handle); - return ERR_INVALID_HANDLE; - } + auto* const current_process{system.Kernel().CurrentProcess()}; + auto& page_table{current_process->PageTable()}; - const auto& vm_manager = current_process->VMManager(); - if (!vm_manager.IsWithinASLRRegion(addr, size)) { - LOG_ERROR(Kernel_SVC, "Region is not within the ASLR region. addr=0x{:016X}, size={:016X}", + if (page_table.IsInvalidRegion(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Addr does not fit within the valid region, addr=0x{:016X}, " + "size=0x{:016X}", addr, size); return ERR_INVALID_MEMORY_RANGE; } - return shared_memory->Map(*current_process, addr, permissions_type, MemoryPermission::DontCare); -} - -static ResultCode UnmapSharedMemory(Core::System& system, Handle shared_memory_handle, VAddr addr, - u64 size) { - LOG_WARNING(Kernel_SVC, "called, shared_memory_handle=0x{:08X}, addr=0x{:X}, size=0x{:X}", - shared_memory_handle, addr, size); - - if (!Common::Is4KBAligned(addr)) { - LOG_ERROR(Kernel_SVC, "Address is not aligned to 4KB, addr=0x{:016X}", addr); - return ERR_INVALID_ADDRESS; - } - - if (size == 0) { - LOG_ERROR(Kernel_SVC, "Size is 0"); - return ERR_INVALID_SIZE; - } - - if (!Common::Is4KBAligned(size)) { - LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, size=0x{:016X}", size); - return ERR_INVALID_SIZE; + if (page_table.IsInsideHeapRegion(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Addr does not fit within the heap region, addr=0x{:016X}, " + "size=0x{:016X}", + addr, size); + return ERR_INVALID_MEMORY_RANGE; } - if (!IsValidAddressRange(addr, size)) { - LOG_ERROR(Kernel_SVC, "Region is not a valid address range, addr=0x{:016X}, size=0x{:016X}", + if (page_table.IsInsideAliasRegion(addr, size)) { + LOG_ERROR(Kernel_SVC, + "Address does not fit within the map region, addr=0x{:016X}, " + "size=0x{:016X}", addr, size); - return ERR_INVALID_ADDRESS_STATE; + return ERR_INVALID_MEMORY_RANGE; } - auto* const current_process = system.Kernel().CurrentProcess(); - auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle); + auto shared_memory{current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle)}; if (!shared_memory) { LOG_ERROR(Kernel_SVC, "Shared memory does not exist, shared_memory_handle=0x{:08X}", shared_memory_handle); return ERR_INVALID_HANDLE; } - const auto& vm_manager = current_process->VMManager(); - if (!vm_manager.IsWithinASLRRegion(addr, size)) { - LOG_ERROR(Kernel_SVC, "Region is not within the ASLR region. addr=0x{:016X}, size={:016X}", - addr, size); - return ERR_INVALID_MEMORY_RANGE; - } + return shared_memory->Map(*current_process, addr, size, permission_type); +} - return shared_memory->Unmap(*current_process, addr, size); +static ResultCode MapSharedMemory32(Core::System& system, Handle shared_memory_handle, u32 addr, + u32 size, u32 permissions) { + return MapSharedMemory(system, shared_memory_handle, static_cast<VAddr>(addr), + static_cast<std::size_t>(size), permissions); } static ResultCode QueryProcessMemory(Core::System& system, VAddr memory_info_address, VAddr page_info_address, Handle process_handle, VAddr address) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_TRACE(Kernel_SVC, "called process=0x{:08X} address={:X}", process_handle, address); const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); std::shared_ptr<Process> process = handle_table.Get<Process>(process_handle); @@ -1279,18 +1298,17 @@ static ResultCode QueryProcessMemory(Core::System& system, VAddr memory_info_add return ERR_INVALID_HANDLE; } - auto& memory = system.Memory(); - const auto& vm_manager = process->VMManager(); - const MemoryInfo memory_info = vm_manager.QueryMemory(address); - - memory.Write64(memory_info_address, memory_info.base_address); - memory.Write64(memory_info_address + 8, memory_info.size); - memory.Write32(memory_info_address + 16, memory_info.state); - memory.Write32(memory_info_address + 20, memory_info.attributes); - memory.Write32(memory_info_address + 24, memory_info.permission); - memory.Write32(memory_info_address + 32, memory_info.ipc_ref_count); - memory.Write32(memory_info_address + 28, memory_info.device_ref_count); - memory.Write32(memory_info_address + 36, 0); + auto& memory{system.Memory()}; + const auto memory_info{process->PageTable().QueryInfo(address).GetSvcMemoryInfo()}; + + memory.Write64(memory_info_address + 0x00, memory_info.addr); + memory.Write64(memory_info_address + 0x08, memory_info.size); + memory.Write32(memory_info_address + 0x10, static_cast<u32>(memory_info.state) & 0xff); + memory.Write32(memory_info_address + 0x14, static_cast<u32>(memory_info.attr)); + memory.Write32(memory_info_address + 0x18, static_cast<u32>(memory_info.perm)); + memory.Write32(memory_info_address + 0x1c, memory_info.ipc_refcount); + memory.Write32(memory_info_address + 0x20, memory_info.device_refcount); + memory.Write32(memory_info_address + 0x24, 0); // Page info appears to be currently unused by the kernel and is always set to zero. memory.Write32(page_info_address, 0); @@ -1362,8 +1380,8 @@ static ResultCode MapProcessCodeMemory(Core::System& system, Handle process_hand return ERR_INVALID_HANDLE; } - auto& vm_manager = process->VMManager(); - if (!vm_manager.IsWithinAddressSpace(src_address, size)) { + auto& page_table = process->PageTable(); + if (!page_table.IsInsideAddressSpace(src_address, size)) { LOG_ERROR(Kernel_SVC, "Source address range is not within the address space (src_address=0x{:016X}, " "size=0x{:016X}).", @@ -1371,7 +1389,7 @@ static ResultCode MapProcessCodeMemory(Core::System& system, Handle process_hand return ERR_INVALID_ADDRESS_STATE; } - if (!vm_manager.IsWithinASLRRegion(dst_address, size)) { + if (!page_table.IsInsideASLRRegion(dst_address, size)) { LOG_ERROR(Kernel_SVC, "Destination address range is not within the ASLR region (dst_address=0x{:016X}, " "size=0x{:016X}).", @@ -1379,7 +1397,7 @@ static ResultCode MapProcessCodeMemory(Core::System& system, Handle process_hand return ERR_INVALID_MEMORY_RANGE; } - return vm_manager.MapCodeMemory(dst_address, src_address, size); + return page_table.MapProcessCodeMemory(dst_address, src_address, size); } static ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_handle, @@ -1430,8 +1448,8 @@ static ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_ha return ERR_INVALID_HANDLE; } - auto& vm_manager = process->VMManager(); - if (!vm_manager.IsWithinAddressSpace(src_address, size)) { + auto& page_table = process->PageTable(); + if (!page_table.IsInsideAddressSpace(src_address, size)) { LOG_ERROR(Kernel_SVC, "Source address range is not within the address space (src_address=0x{:016X}, " "size=0x{:016X}).", @@ -1439,7 +1457,7 @@ static ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_ha return ERR_INVALID_ADDRESS_STATE; } - if (!vm_manager.IsWithinASLRRegion(dst_address, size)) { + if (!page_table.IsInsideASLRRegion(dst_address, size)) { LOG_ERROR(Kernel_SVC, "Destination address range is not within the ASLR region (dst_address=0x{:016X}, " "size=0x{:016X}).", @@ -1447,12 +1465,13 @@ static ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_ha return ERR_INVALID_MEMORY_RANGE; } - return vm_manager.UnmapCodeMemory(dst_address, src_address, size); + return page_table.UnmapProcessCodeMemory(dst_address, src_address, size); } /// Exits the current process static void ExitProcess(Core::System& system) { auto* current_process = system.Kernel().CurrentProcess(); + UNIMPLEMENTED(); LOG_INFO(Kernel_SVC, "Process {} exiting", current_process->GetProcessID()); ASSERT_MSG(current_process->GetStatus() == ProcessStatus::Running, @@ -1462,8 +1481,10 @@ static void ExitProcess(Core::System& system) { // Kill the current thread system.CurrentScheduler().GetCurrentThread()->Stop(); +} - system.PrepareReschedule(); +static void ExitProcess32(Core::System& system) { + ExitProcess(system); } /// Creates a new thread @@ -1506,9 +1527,13 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e } auto& kernel = system.Kernel(); + + ASSERT(kernel.CurrentProcess()->GetResourceLimit()->Reserve(ResourceType::Threads, 1)); + + ThreadType type = THREADTYPE_USER; CASCADE_RESULT(std::shared_ptr<Thread> thread, - Thread::Create(kernel, "", entry_point, priority, arg, processor_id, stack_top, - *current_process)); + Thread::Create(system, type, "", entry_point, priority, arg, processor_id, + stack_top, current_process)); const auto new_thread_handle = current_process->GetHandleTable().Create(thread); if (new_thread_handle.Failed()) { @@ -1522,11 +1547,15 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e thread->SetName( fmt::format("thread[entry_point={:X}, handle={:X}]", entry_point, *new_thread_handle)); - system.PrepareReschedule(thread->GetProcessorID()); - return RESULT_SUCCESS; } +static ResultCode CreateThread32(Core::System& system, Handle* out_handle, u32 priority, + u32 entry_point, u32 arg, u32 stack_top, s32 processor_id) { + return CreateThread(system, out_handle, static_cast<VAddr>(entry_point), static_cast<u64>(arg), + static_cast<VAddr>(stack_top), priority, processor_id); +} + /// Starts the thread for the provided handle static ResultCode StartThread(Core::System& system, Handle thread_handle) { LOG_DEBUG(Kernel_SVC, "called thread=0x{:08X}", thread_handle); @@ -1541,13 +1570,11 @@ static ResultCode StartThread(Core::System& system, Handle thread_handle) { ASSERT(thread->GetStatus() == ThreadStatus::Dormant); - thread->ResumeFromWait(); - - if (thread->GetStatus() == ThreadStatus::Ready) { - system.PrepareReschedule(thread->GetProcessorID()); - } + return thread->Start(); +} - return RESULT_SUCCESS; +static ResultCode StartThread32(Core::System& system, Handle thread_handle) { + return StartThread(system, thread_handle); } /// Called when a thread exits @@ -1555,9 +1582,12 @@ static void ExitThread(Core::System& system) { LOG_DEBUG(Kernel_SVC, "called, pc=0x{:08X}", system.CurrentArmInterface().GetPC()); auto* const current_thread = system.CurrentScheduler().GetCurrentThread(); - current_thread->Stop(); system.GlobalScheduler().RemoveThread(SharedFrom(current_thread)); - system.PrepareReschedule(); + current_thread->Stop(); +} + +static void ExitThread32(Core::System& system) { + ExitThread(system); } /// Sleep the current thread @@ -1576,15 +1606,21 @@ static void SleepThread(Core::System& system, s64 nanoseconds) { if (nanoseconds <= 0) { switch (static_cast<SleepType>(nanoseconds)) { - case SleepType::YieldWithoutLoadBalancing: - is_redundant = current_thread->YieldSimple(); + case SleepType::YieldWithoutLoadBalancing: { + auto pair = current_thread->YieldSimple(); + is_redundant = pair.second; break; - case SleepType::YieldWithLoadBalancing: - is_redundant = current_thread->YieldAndBalanceLoad(); + } + case SleepType::YieldWithLoadBalancing: { + auto pair = current_thread->YieldAndBalanceLoad(); + is_redundant = pair.second; break; - case SleepType::YieldAndWaitForLoadBalancing: - is_redundant = current_thread->YieldAndWaitForLoadBalancing(); + } + case SleepType::YieldAndWaitForLoadBalancing: { + auto pair = current_thread->YieldAndWaitForLoadBalancing(); + is_redundant = pair.second; break; + } default: UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds); } @@ -1592,13 +1628,18 @@ static void SleepThread(Core::System& system, s64 nanoseconds) { current_thread->Sleep(nanoseconds); } - if (is_redundant) { - // If it's redundant, the core is pretty much idle. Some games keep idling - // a core while it's doing nothing, we advance timing to avoid costly continuous - // calls. - system.CoreTiming().AddTicks(2000); + if (is_redundant && !system.Kernel().IsMulticore()) { + system.Kernel().ExitSVCProfile(); + system.CoreTiming().AddTicks(1000U); + system.GetCpuManager().PreemptSingleCore(); + system.Kernel().EnterSVCProfile(); } - system.PrepareReschedule(current_thread->GetProcessorID()); +} + +static void SleepThread32(Core::System& system, u32 nanoseconds_low, u32 nanoseconds_high) { + const s64 nanoseconds = static_cast<s64>(static_cast<u64>(nanoseconds_low) | + (static_cast<u64>(nanoseconds_high) << 32)); + SleepThread(system, nanoseconds); } /// Wait process wide key atomic @@ -1610,7 +1651,7 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add "called mutex_addr={:X}, condition_variable_addr={:X}, thread_handle=0x{:08X}, timeout={}", mutex_addr, condition_variable_addr, thread_handle, nano_seconds); - if (Memory::IsKernelVirtualAddress(mutex_addr)) { + if (Core::Memory::IsKernelVirtualAddress(mutex_addr)) { LOG_ERROR( Kernel_SVC, "Given mutex address must not be within the kernel address space. address=0x{:016X}", @@ -1625,31 +1666,69 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add } ASSERT(condition_variable_addr == Common::AlignDown(condition_variable_addr, 4)); - + auto& kernel = system.Kernel(); + Handle event_handle; + Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); auto* const current_process = system.Kernel().CurrentProcess(); - const auto& handle_table = current_process->GetHandleTable(); - std::shared_ptr<Thread> thread = handle_table.Get<Thread>(thread_handle); - ASSERT(thread); + { + SchedulerLockAndSleep lock(kernel, event_handle, current_thread, nano_seconds); + const auto& handle_table = current_process->GetHandleTable(); + std::shared_ptr<Thread> thread = handle_table.Get<Thread>(thread_handle); + ASSERT(thread); + + current_thread->SetSynchronizationResults(nullptr, RESULT_TIMEOUT); + + if (thread->IsPendingTermination()) { + lock.CancelSleep(); + return ERR_THREAD_TERMINATING; + } + + const auto release_result = current_process->GetMutex().Release(mutex_addr); + if (release_result.IsError()) { + lock.CancelSleep(); + return release_result; + } - const auto release_result = current_process->GetMutex().Release(mutex_addr); - if (release_result.IsError()) { - return release_result; + if (nano_seconds == 0) { + lock.CancelSleep(); + return RESULT_TIMEOUT; + } + + current_thread->SetCondVarWaitAddress(condition_variable_addr); + current_thread->SetMutexWaitAddress(mutex_addr); + current_thread->SetWaitHandle(thread_handle); + current_thread->SetStatus(ThreadStatus::WaitCondVar); + current_process->InsertConditionVariableThread(SharedFrom(current_thread)); } - Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); - current_thread->SetCondVarWaitAddress(condition_variable_addr); - current_thread->SetMutexWaitAddress(mutex_addr); - current_thread->SetWaitHandle(thread_handle); - current_thread->SetStatus(ThreadStatus::WaitCondVar); - current_thread->InvalidateWakeupCallback(); - current_process->InsertConditionVariableThread(SharedFrom(current_thread)); + if (event_handle != InvalidHandle) { + auto& time_manager = kernel.TimeManager(); + time_manager.UnscheduleTimeEvent(event_handle); + } + + { + SchedulerLock lock(kernel); - current_thread->WakeAfterDelay(nano_seconds); + auto* owner = current_thread->GetLockOwner(); + if (owner != nullptr) { + owner->RemoveMutexWaiter(SharedFrom(current_thread)); + } + current_process->RemoveConditionVariableThread(SharedFrom(current_thread)); + } // Note: Deliberately don't attempt to inherit the lock owner's priority. - system.PrepareReschedule(current_thread->GetProcessorID()); - return RESULT_SUCCESS; + return current_thread->GetSignalingResult(); +} + +static ResultCode WaitProcessWideKeyAtomic32(Core::System& system, u32 mutex_addr, + u32 condition_variable_addr, Handle thread_handle, + u32 nanoseconds_low, u32 nanoseconds_high) { + const s64 nanoseconds = + static_cast<s64>(nanoseconds_low | (static_cast<u64>(nanoseconds_high) << 32)); + return WaitProcessWideKeyAtomic(system, static_cast<VAddr>(mutex_addr), + static_cast<VAddr>(condition_variable_addr), thread_handle, + nanoseconds); } /// Signal process wide key @@ -1660,16 +1739,18 @@ static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_ ASSERT(condition_variable_addr == Common::AlignDown(condition_variable_addr, 4)); // Retrieve a list of all threads that are waiting for this condition variable. - auto* const current_process = system.Kernel().CurrentProcess(); + auto& kernel = system.Kernel(); + SchedulerLock lock(kernel); + auto* const current_process = kernel.CurrentProcess(); std::vector<std::shared_ptr<Thread>> waiting_threads = current_process->GetConditionVariableThreads(condition_variable_addr); // Only process up to 'target' threads, unless 'target' is less equal 0, in which case process // them all. std::size_t last = waiting_threads.size(); - if (target > 0) + if (target > 0) { last = std::min(waiting_threads.size(), static_cast<std::size_t>(target)); - + } for (std::size_t index = 0; index < last; ++index) { auto& thread = waiting_threads[index]; @@ -1677,21 +1758,17 @@ static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_ // liberate Cond Var Thread. current_process->RemoveConditionVariableThread(thread); - thread->SetCondVarWaitAddress(0); const std::size_t current_core = system.CurrentCoreIndex(); auto& monitor = system.Monitor(); - auto& memory = system.Memory(); // Atomically read the value of the mutex. u32 mutex_val = 0; u32 update_val = 0; const VAddr mutex_address = thread->GetMutexWaitAddress(); do { - monitor.SetExclusive(current_core, mutex_address); - // If the mutex is not yet acquired, acquire it. - mutex_val = memory.Read32(mutex_address); + mutex_val = monitor.ExclusiveRead32(current_core, mutex_address); if (mutex_val != 0) { update_val = mutex_val | Mutex::MutexHasWaitersFlag; @@ -1699,33 +1776,28 @@ static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_ update_val = thread->GetWaitHandle(); } } while (!monitor.ExclusiveWrite32(current_core, mutex_address, update_val)); + monitor.ClearExclusive(); if (mutex_val == 0) { // We were able to acquire the mutex, resume this thread. - ASSERT(thread->GetStatus() == ThreadStatus::WaitCondVar); - thread->ResumeFromWait(); - auto* const lock_owner = thread->GetLockOwner(); if (lock_owner != nullptr) { lock_owner->RemoveMutexWaiter(thread); } thread->SetLockOwner(nullptr); - thread->SetMutexWaitAddress(0); - thread->SetWaitHandle(0); - thread->SetWaitSynchronizationResult(RESULT_SUCCESS); - system.PrepareReschedule(thread->GetProcessorID()); + thread->SetSynchronizationResults(nullptr, RESULT_SUCCESS); + thread->ResumeFromWait(); } else { // The mutex is already owned by some other thread, make this thread wait on it. const Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask); const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); auto owner = handle_table.Get<Thread>(owner_handle); ASSERT(owner); - ASSERT(thread->GetStatus() == ThreadStatus::WaitCondVar); - thread->InvalidateWakeupCallback(); - thread->SetStatus(ThreadStatus::WaitMutex); + if (thread->GetStatus() == ThreadStatus::WaitCondVar) { + thread->SetStatus(ThreadStatus::WaitMutex); + } owner->AddMutexWaiter(thread); - system.PrepareReschedule(thread->GetProcessorID()); } } } @@ -1741,7 +1813,7 @@ static ResultCode WaitForAddress(Core::System& system, VAddr address, u32 type, type, value, timeout); // If the passed address is a kernel virtual address, return invalid memory state. - if (Memory::IsKernelVirtualAddress(address)) { + if (Core::Memory::IsKernelVirtualAddress(address)) { LOG_ERROR(Kernel_SVC, "Address is a kernel virtual address, address={:016X}", address); return ERR_INVALID_ADDRESS_STATE; } @@ -1756,12 +1828,15 @@ static ResultCode WaitForAddress(Core::System& system, VAddr address, u32 type, auto& address_arbiter = system.Kernel().CurrentProcess()->GetAddressArbiter(); const ResultCode result = address_arbiter.WaitForAddress(address, arbitration_type, value, timeout); - if (result == RESULT_SUCCESS) { - system.PrepareReschedule(); - } return result; } +static ResultCode WaitForAddress32(Core::System& system, u32 address, u32 type, s32 value, + u32 timeout_low, u32 timeout_high) { + s64 timeout = static_cast<s64>(timeout_low | (static_cast<u64>(timeout_high) << 32)); + return WaitForAddress(system, static_cast<VAddr>(address), type, value, timeout); +} + // Signals to an address (via Address Arbiter) static ResultCode SignalToAddress(Core::System& system, VAddr address, u32 type, s32 value, s32 num_to_wake) { @@ -1769,7 +1844,7 @@ static ResultCode SignalToAddress(Core::System& system, VAddr address, u32 type, address, type, value, num_to_wake); // If the passed address is a kernel virtual address, return invalid memory state. - if (Memory::IsKernelVirtualAddress(address)) { + if (Core::Memory::IsKernelVirtualAddress(address)) { LOG_ERROR(Kernel_SVC, "Address is a kernel virtual address, address={:016X}", address); return ERR_INVALID_ADDRESS_STATE; } @@ -1785,6 +1860,11 @@ static ResultCode SignalToAddress(Core::System& system, VAddr address, u32 type, return address_arbiter.SignalToAddress(address, signal_type, value, num_to_wake); } +static ResultCode SignalToAddress32(Core::System& system, u32 address, u32 type, s32 value, + s32 num_to_wake) { + return SignalToAddress(system, static_cast<VAddr>(address), type, value, num_to_wake); +} + static void KernelDebug([[maybe_unused]] Core::System& system, [[maybe_unused]] u32 kernel_debug_type, [[maybe_unused]] u64 param1, [[maybe_unused]] u64 param2, [[maybe_unused]] u64 param3) { @@ -1803,14 +1883,21 @@ static u64 GetSystemTick(Core::System& system) { auto& core_timing = system.CoreTiming(); // Returns the value of cntpct_el0 (https://switchbrew.org/wiki/SVC#svcGetSystemTick) - const u64 result{Core::Timing::CpuCyclesToClockCycles(system.CoreTiming().GetTicks())}; + const u64 result{system.CoreTiming().GetClockTicks()}; - // Advance time to defeat dumb games that busy-wait for the frame to end. - core_timing.AddTicks(400); + if (!system.Kernel().IsMulticore()) { + core_timing.AddTicks(400U); + } return result; } +static void GetSystemTick32(Core::System& system, u32* time_low, u32* time_high) { + u64 time = GetSystemTick(system); + *time_low = static_cast<u32>(time); + *time_high = static_cast<u32>(time >> 32); +} + /// Close a handle static ResultCode CloseHandle(Core::System& system, Handle handle) { LOG_TRACE(Kernel_SVC, "Closing handle 0x{:08X}", handle); @@ -1843,9 +1930,14 @@ static ResultCode ResetSignal(Core::System& system, Handle handle) { return ERR_INVALID_HANDLE; } +static ResultCode ResetSignal32(Core::System& system, Handle handle) { + return ResetSignal(system, handle); +} + /// Creates a TransferMemory object static ResultCode CreateTransferMemory(Core::System& system, Handle* handle, VAddr addr, u64 size, u32 permissions) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_DEBUG(Kernel_SVC, "called addr=0x{:X}, size=0x{:X}, perms=0x{:08X}", addr, size, permissions); @@ -1865,9 +1957,9 @@ static ResultCode CreateTransferMemory(Core::System& system, Handle* handle, VAd return ERR_INVALID_ADDRESS_STATE; } - const auto perms = static_cast<MemoryPermission>(permissions); - if (perms != MemoryPermission::None && perms != MemoryPermission::Read && - perms != MemoryPermission::ReadWrite) { + const auto perms{static_cast<Memory::MemoryPermission>(permissions)}; + if (perms > Memory::MemoryPermission::ReadAndWrite || + perms == Memory::MemoryPermission::Write) { LOG_ERROR(Kernel_SVC, "Invalid memory permissions for transfer memory! (perms={:08X})", permissions); return ERR_INVALID_MEMORY_PERMISSIONS; @@ -1890,109 +1982,10 @@ static ResultCode CreateTransferMemory(Core::System& system, Handle* handle, VAd return RESULT_SUCCESS; } -static ResultCode MapTransferMemory(Core::System& system, Handle handle, VAddr address, u64 size, - u32 permission_raw) { - LOG_DEBUG(Kernel_SVC, - "called. handle=0x{:08X}, address=0x{:016X}, size=0x{:016X}, permissions=0x{:08X}", - handle, address, size, permission_raw); - - if (!Common::Is4KBAligned(address)) { - LOG_ERROR(Kernel_SVC, "Transfer memory addresses must be 4KB aligned (size=0x{:016X}).", - address); - return ERR_INVALID_ADDRESS; - } - - if (size == 0 || !Common::Is4KBAligned(size)) { - LOG_ERROR(Kernel_SVC, - "Transfer memory sizes must be 4KB aligned and not be zero (size=0x{:016X}).", - size); - return ERR_INVALID_SIZE; - } - - if (!IsValidAddressRange(address, size)) { - LOG_ERROR(Kernel_SVC, - "Given address and size overflows the 64-bit range (address=0x{:016X}, " - "size=0x{:016X}).", - address, size); - return ERR_INVALID_ADDRESS_STATE; - } - - const auto permissions = static_cast<MemoryPermission>(permission_raw); - if (permissions != MemoryPermission::None && permissions != MemoryPermission::Read && - permissions != MemoryPermission::ReadWrite) { - LOG_ERROR(Kernel_SVC, "Invalid transfer memory permissions given (permissions=0x{:08X}).", - permission_raw); - return ERR_INVALID_STATE; - } - - const auto& kernel = system.Kernel(); - const auto* const current_process = kernel.CurrentProcess(); - const auto& handle_table = current_process->GetHandleTable(); - - auto transfer_memory = handle_table.Get<TransferMemory>(handle); - if (!transfer_memory) { - LOG_ERROR(Kernel_SVC, "Nonexistent transfer memory handle given (handle=0x{:08X}).", - handle); - return ERR_INVALID_HANDLE; - } - - if (!current_process->VMManager().IsWithinASLRRegion(address, size)) { - LOG_ERROR(Kernel_SVC, - "Given address and size don't fully fit within the ASLR region " - "(address=0x{:016X}, size=0x{:016X}).", - address, size); - return ERR_INVALID_MEMORY_RANGE; - } - - return transfer_memory->MapMemory(address, size, permissions); -} - -static ResultCode UnmapTransferMemory(Core::System& system, Handle handle, VAddr address, - u64 size) { - LOG_DEBUG(Kernel_SVC, "called. handle=0x{:08X}, address=0x{:016X}, size=0x{:016X}", handle, - address, size); - - if (!Common::Is4KBAligned(address)) { - LOG_ERROR(Kernel_SVC, "Transfer memory addresses must be 4KB aligned (size=0x{:016X}).", - address); - return ERR_INVALID_ADDRESS; - } - - if (size == 0 || !Common::Is4KBAligned(size)) { - LOG_ERROR(Kernel_SVC, - "Transfer memory sizes must be 4KB aligned and not be zero (size=0x{:016X}).", - size); - return ERR_INVALID_SIZE; - } - - if (!IsValidAddressRange(address, size)) { - LOG_ERROR(Kernel_SVC, - "Given address and size overflows the 64-bit range (address=0x{:016X}, " - "size=0x{:016X}).", - address, size); - return ERR_INVALID_ADDRESS_STATE; - } - - const auto& kernel = system.Kernel(); - const auto* const current_process = kernel.CurrentProcess(); - const auto& handle_table = current_process->GetHandleTable(); - - auto transfer_memory = handle_table.Get<TransferMemory>(handle); - if (!transfer_memory) { - LOG_ERROR(Kernel_SVC, "Nonexistent transfer memory handle given (handle=0x{:08X}).", - handle); - return ERR_INVALID_HANDLE; - } - - if (!current_process->VMManager().IsWithinASLRRegion(address, size)) { - LOG_ERROR(Kernel_SVC, - "Given address and size don't fully fit within the ASLR region " - "(address=0x{:016X}, size=0x{:016X}).", - address, size); - return ERR_INVALID_MEMORY_RANGE; - } - - return transfer_memory->UnmapMemory(address, size); +static ResultCode CreateTransferMemory32(Core::System& system, Handle* handle, u32 addr, u32 size, + u32 permissions) { + return CreateTransferMemory(system, handle, static_cast<VAddr>(addr), + static_cast<std::size_t>(size), permissions); } static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, u32* core, @@ -2004,6 +1997,8 @@ static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}", thread_handle); + *core = 0; + *mask = 0; return ERR_INVALID_HANDLE; } @@ -2013,6 +2008,15 @@ static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, return RESULT_SUCCESS; } +static ResultCode GetThreadCoreMask32(Core::System& system, Handle thread_handle, u32* core, + u32* mask_low, u32* mask_high) { + u64 mask{}; + const auto result = GetThreadCoreMask(system, thread_handle, core, &mask); + *mask_high = static_cast<u32>(mask >> 32); + *mask_low = static_cast<u32>(mask); + return result; +} + static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, u32 core, u64 affinity_mask) { LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, core=0x{:X}, affinity_mask=0x{:016X}", @@ -2044,7 +2048,7 @@ static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, return ERR_INVALID_COMBINATION; } - if (core < Core::NUM_CPU_CORES) { + if (core < Core::Hardware::NUM_CPU_CORES) { if ((affinity_mask & (1ULL << core)) == 0) { LOG_ERROR(Kernel_SVC, "Core is not enabled for the current mask, core={}, mask={:016X}", core, @@ -2066,57 +2070,14 @@ static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, return ERR_INVALID_HANDLE; } - system.PrepareReschedule(thread->GetProcessorID()); - thread->ChangeCore(core, affinity_mask); - system.PrepareReschedule(thread->GetProcessorID()); - - return RESULT_SUCCESS; + return thread->SetCoreAndAffinityMask(core, affinity_mask); } -static ResultCode CreateSharedMemory(Core::System& system, Handle* handle, u64 size, - u32 local_permissions, u32 remote_permissions) { - LOG_TRACE(Kernel_SVC, "called, size=0x{:X}, localPerms=0x{:08X}, remotePerms=0x{:08X}", size, - local_permissions, remote_permissions); - if (size == 0) { - LOG_ERROR(Kernel_SVC, "Size is 0"); - return ERR_INVALID_SIZE; - } - if (!Common::Is4KBAligned(size)) { - LOG_ERROR(Kernel_SVC, "Size is not aligned to 4KB, 0x{:016X}", size); - return ERR_INVALID_SIZE; - } - - if (size >= MAIN_MEMORY_SIZE) { - LOG_ERROR(Kernel_SVC, "Size is not less than 8GB, 0x{:016X}", size); - return ERR_INVALID_SIZE; - } - - const auto local_perms = static_cast<MemoryPermission>(local_permissions); - if (local_perms != MemoryPermission::Read && local_perms != MemoryPermission::ReadWrite) { - LOG_ERROR(Kernel_SVC, - "Invalid local memory permissions, expected Read or ReadWrite but got " - "local_permissions={}", - static_cast<u32>(local_permissions)); - return ERR_INVALID_MEMORY_PERMISSIONS; - } - - const auto remote_perms = static_cast<MemoryPermission>(remote_permissions); - if (remote_perms != MemoryPermission::Read && remote_perms != MemoryPermission::ReadWrite && - remote_perms != MemoryPermission::DontCare) { - LOG_ERROR(Kernel_SVC, - "Invalid remote memory permissions, expected Read, ReadWrite or DontCare but got " - "remote_permissions={}", - static_cast<u32>(remote_permissions)); - return ERR_INVALID_MEMORY_PERMISSIONS; - } - - auto& kernel = system.Kernel(); - auto process = kernel.CurrentProcess(); - auto& handle_table = process->GetHandleTable(); - auto shared_mem_handle = SharedMemory::Create(kernel, process, size, local_perms, remote_perms); - - CASCADE_RESULT(*handle, handle_table.Create(shared_mem_handle)); - return RESULT_SUCCESS; +static ResultCode SetThreadCoreMask32(Core::System& system, Handle thread_handle, u32 core, + u32 affinity_mask_low, u32 affinity_mask_high) { + const u64 affinity_mask = + static_cast<u64>(affinity_mask_low) | (static_cast<u64>(affinity_mask_high) << 32); + return SetThreadCoreMask(system, thread_handle, core, affinity_mask); } static ResultCode CreateEvent(Core::System& system, Handle* write_handle, Handle* read_handle) { @@ -2147,6 +2108,10 @@ static ResultCode CreateEvent(Core::System& system, Handle* write_handle, Handle return RESULT_SUCCESS; } +static ResultCode CreateEvent32(Core::System& system, Handle* write_handle, Handle* read_handle) { + return CreateEvent(system, write_handle, read_handle); +} + static ResultCode ClearEvent(Core::System& system, Handle handle) { LOG_TRACE(Kernel_SVC, "called, event=0x{:08X}", handle); @@ -2168,6 +2133,10 @@ static ResultCode ClearEvent(Core::System& system, Handle handle) { return ERR_INVALID_HANDLE; } +static ResultCode ClearEvent32(Core::System& system, Handle handle) { + return ClearEvent(system, handle); +} + static ResultCode SignalEvent(Core::System& system, Handle handle) { LOG_DEBUG(Kernel_SVC, "called. Handle=0x{:08X}", handle); @@ -2180,10 +2149,13 @@ static ResultCode SignalEvent(Core::System& system, Handle handle) { } writable_event->Signal(); - system.PrepareReschedule(); return RESULT_SUCCESS; } +static ResultCode SignalEvent32(Core::System& system, Handle handle) { + return SignalEvent(system, handle); +} + static ResultCode GetProcessInfo(Core::System& system, u64* out, Handle process_handle, u32 type) { LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, type=0x{:X}", process_handle, type); @@ -2211,6 +2183,7 @@ static ResultCode GetProcessInfo(Core::System& system, u64* out, Handle process_ } static ResultCode CreateResourceLimit(Core::System& system, Handle* out_handle) { + std::lock_guard lock{HLE::g_hle_lock}; LOG_DEBUG(Kernel_SVC, "called"); auto& kernel = system.Kernel(); @@ -2305,11 +2278,10 @@ static ResultCode GetProcessList(Core::System& system, u32* out_num_processes, } const auto& kernel = system.Kernel(); - const auto& vm_manager = kernel.CurrentProcess()->VMManager(); const auto total_copy_size = out_process_ids_size * sizeof(u64); - if (out_process_ids_size > 0 && - !vm_manager.IsWithinAddressSpace(out_process_ids, total_copy_size)) { + if (out_process_ids_size > 0 && !kernel.CurrentProcess()->PageTable().IsInsideAddressSpace( + out_process_ids, total_copy_size)) { LOG_ERROR(Kernel_SVC, "Address range outside address space. begin=0x{:016X}, end=0x{:016X}", out_process_ids, out_process_ids + total_copy_size); return ERR_INVALID_ADDRESS_STATE; @@ -2345,11 +2317,10 @@ static ResultCode GetThreadList(Core::System& system, u32* out_num_threads, VAdd } const auto* const current_process = system.Kernel().CurrentProcess(); - const auto& vm_manager = current_process->VMManager(); const auto total_copy_size = out_thread_ids_size * sizeof(u64); if (out_thread_ids_size > 0 && - !vm_manager.IsWithinAddressSpace(out_thread_ids, total_copy_size)) { + !current_process->PageTable().IsInsideAddressSpace(out_thread_ids, total_copy_size)) { LOG_ERROR(Kernel_SVC, "Address range outside address space. begin=0x{:016X}, end=0x{:016X}", out_thread_ids, out_thread_ids + total_copy_size); return ERR_INVALID_ADDRESS_STATE; @@ -2370,6 +2341,15 @@ static ResultCode GetThreadList(Core::System& system, u32* out_num_threads, VAdd return RESULT_SUCCESS; } +static ResultCode FlushProcessDataCache32(Core::System& system, Handle handle, u32 address, + u32 size) { + // Note(Blinkhawk): For emulation purposes of the data cache this is mostly a nope + // as all emulation is done in the same cache level in host architecture, thus data cache + // does not need flushing. + LOG_DEBUG(Kernel_SVC, "called"); + return RESULT_SUCCESS; +} + namespace { struct FunctionDef { using Func = void(Core::System&); @@ -2384,57 +2364,57 @@ static const FunctionDef SVC_Table_32[] = { {0x00, nullptr, "Unknown"}, {0x01, SvcWrap32<SetHeapSize32>, "SetHeapSize32"}, {0x02, nullptr, "Unknown"}, - {0x03, nullptr, "SetMemoryAttribute32"}, - {0x04, nullptr, "MapMemory32"}, - {0x05, nullptr, "UnmapMemory32"}, + {0x03, SvcWrap32<SetMemoryAttribute32>, "SetMemoryAttribute32"}, + {0x04, SvcWrap32<MapMemory32>, "MapMemory32"}, + {0x05, SvcWrap32<UnmapMemory32>, "UnmapMemory32"}, {0x06, SvcWrap32<QueryMemory32>, "QueryMemory32"}, - {0x07, nullptr, "ExitProcess32"}, - {0x08, nullptr, "CreateThread32"}, - {0x09, nullptr, "StartThread32"}, - {0x0a, nullptr, "ExitThread32"}, - {0x0b, nullptr, "SleepThread32"}, + {0x07, SvcWrap32<ExitProcess32>, "ExitProcess32"}, + {0x08, SvcWrap32<CreateThread32>, "CreateThread32"}, + {0x09, SvcWrap32<StartThread32>, "StartThread32"}, + {0x0a, SvcWrap32<ExitThread32>, "ExitThread32"}, + {0x0b, SvcWrap32<SleepThread32>, "SleepThread32"}, {0x0c, SvcWrap32<GetThreadPriority32>, "GetThreadPriority32"}, - {0x0d, nullptr, "SetThreadPriority32"}, - {0x0e, nullptr, "GetThreadCoreMask32"}, - {0x0f, nullptr, "SetThreadCoreMask32"}, - {0x10, nullptr, "GetCurrentProcessorNumber32"}, - {0x11, nullptr, "SignalEvent32"}, - {0x12, nullptr, "ClearEvent32"}, - {0x13, nullptr, "MapSharedMemory32"}, + {0x0d, SvcWrap32<SetThreadPriority32>, "SetThreadPriority32"}, + {0x0e, SvcWrap32<GetThreadCoreMask32>, "GetThreadCoreMask32"}, + {0x0f, SvcWrap32<SetThreadCoreMask32>, "SetThreadCoreMask32"}, + {0x10, SvcWrap32<GetCurrentProcessorNumber32>, "GetCurrentProcessorNumber32"}, + {0x11, SvcWrap32<SignalEvent32>, "SignalEvent32"}, + {0x12, SvcWrap32<ClearEvent32>, "ClearEvent32"}, + {0x13, SvcWrap32<MapSharedMemory32>, "MapSharedMemory32"}, {0x14, nullptr, "UnmapSharedMemory32"}, - {0x15, nullptr, "CreateTransferMemory32"}, + {0x15, SvcWrap32<CreateTransferMemory32>, "CreateTransferMemory32"}, {0x16, SvcWrap32<CloseHandle32>, "CloseHandle32"}, - {0x17, nullptr, "ResetSignal32"}, + {0x17, SvcWrap32<ResetSignal32>, "ResetSignal32"}, {0x18, SvcWrap32<WaitSynchronization32>, "WaitSynchronization32"}, - {0x19, nullptr, "CancelSynchronization32"}, - {0x1a, nullptr, "ArbitrateLock32"}, - {0x1b, nullptr, "ArbitrateUnlock32"}, - {0x1c, nullptr, "WaitProcessWideKeyAtomic32"}, + {0x19, SvcWrap32<CancelSynchronization32>, "CancelSynchronization32"}, + {0x1a, SvcWrap32<ArbitrateLock32>, "ArbitrateLock32"}, + {0x1b, SvcWrap32<ArbitrateUnlock32>, "ArbitrateUnlock32"}, + {0x1c, SvcWrap32<WaitProcessWideKeyAtomic32>, "WaitProcessWideKeyAtomic32"}, {0x1d, SvcWrap32<SignalProcessWideKey32>, "SignalProcessWideKey32"}, - {0x1e, nullptr, "GetSystemTick32"}, + {0x1e, SvcWrap32<GetSystemTick32>, "GetSystemTick32"}, {0x1f, SvcWrap32<ConnectToNamedPort32>, "ConnectToNamedPort32"}, {0x20, nullptr, "Unknown"}, {0x21, SvcWrap32<SendSyncRequest32>, "SendSyncRequest32"}, {0x22, nullptr, "SendSyncRequestWithUserBuffer32"}, {0x23, nullptr, "Unknown"}, - {0x24, nullptr, "GetProcessId32"}, + {0x24, SvcWrap32<GetProcessId32>, "GetProcessId32"}, {0x25, SvcWrap32<GetThreadId32>, "GetThreadId32"}, - {0x26, nullptr, "Break32"}, + {0x26, SvcWrap32<Break32>, "Break32"}, {0x27, nullptr, "OutputDebugString32"}, {0x28, nullptr, "Unknown"}, {0x29, SvcWrap32<GetInfo32>, "GetInfo32"}, {0x2a, nullptr, "Unknown"}, {0x2b, nullptr, "Unknown"}, - {0x2c, nullptr, "MapPhysicalMemory32"}, - {0x2d, nullptr, "UnmapPhysicalMemory32"}, + {0x2c, SvcWrap32<MapPhysicalMemory32>, "MapPhysicalMemory32"}, + {0x2d, SvcWrap32<UnmapPhysicalMemory32>, "UnmapPhysicalMemory32"}, {0x2e, nullptr, "Unknown"}, {0x2f, nullptr, "Unknown"}, {0x30, nullptr, "Unknown"}, {0x31, nullptr, "Unknown"}, - {0x32, nullptr, "SetThreadActivity32"}, - {0x33, nullptr, "GetThreadContext32"}, - {0x34, nullptr, "WaitForAddress32"}, - {0x35, nullptr, "SignalToAddress32"}, + {0x32, SvcWrap32<SetThreadActivity32>, "SetThreadActivity32"}, + {0x33, SvcWrap32<GetThreadContext32>, "GetThreadContext32"}, + {0x34, SvcWrap32<WaitForAddress32>, "WaitForAddress32"}, + {0x35, SvcWrap32<SignalToAddress32>, "SignalToAddress32"}, {0x36, nullptr, "Unknown"}, {0x37, nullptr, "Unknown"}, {0x38, nullptr, "Unknown"}, @@ -2450,7 +2430,7 @@ static const FunctionDef SVC_Table_32[] = { {0x42, nullptr, "Unknown"}, {0x43, nullptr, "ReplyAndReceive32"}, {0x44, nullptr, "Unknown"}, - {0x45, nullptr, "CreateEvent32"}, + {0x45, SvcWrap32<CreateEvent32>, "CreateEvent32"}, {0x46, nullptr, "Unknown"}, {0x47, nullptr, "Unknown"}, {0x48, nullptr, "Unknown"}, @@ -2476,7 +2456,7 @@ static const FunctionDef SVC_Table_32[] = { {0x5c, nullptr, "Unknown"}, {0x5d, nullptr, "Unknown"}, {0x5e, nullptr, "Unknown"}, - {0x5F, nullptr, "FlushProcessDataCache32"}, + {0x5F, SvcWrap32<FlushProcessDataCache32>, "FlushProcessDataCache32"}, {0x60, nullptr, "Unknown"}, {0x61, nullptr, "Unknown"}, {0x62, nullptr, "Unknown"}, @@ -2510,7 +2490,7 @@ static const FunctionDef SVC_Table_32[] = { static const FunctionDef SVC_Table_64[] = { {0x00, nullptr, "Unknown"}, {0x01, SvcWrap64<SetHeapSize>, "SetHeapSize"}, - {0x02, SvcWrap64<SetMemoryPermission>, "SetMemoryPermission"}, + {0x02, nullptr, "SetMemoryPermission"}, {0x03, SvcWrap64<SetMemoryAttribute>, "SetMemoryAttribute"}, {0x04, SvcWrap64<MapMemory>, "MapMemory"}, {0x05, SvcWrap64<UnmapMemory>, "UnmapMemory"}, @@ -2528,7 +2508,7 @@ static const FunctionDef SVC_Table_64[] = { {0x11, SvcWrap64<SignalEvent>, "SignalEvent"}, {0x12, SvcWrap64<ClearEvent>, "ClearEvent"}, {0x13, SvcWrap64<MapSharedMemory>, "MapSharedMemory"}, - {0x14, SvcWrap64<UnmapSharedMemory>, "UnmapSharedMemory"}, + {0x14, nullptr, "UnmapSharedMemory"}, {0x15, SvcWrap64<CreateTransferMemory>, "CreateTransferMemory"}, {0x16, SvcWrap64<CloseHandle>, "CloseHandle"}, {0x17, SvcWrap64<ResetSignal>, "ResetSignal"}, @@ -2588,9 +2568,9 @@ static const FunctionDef SVC_Table_64[] = { {0x4D, nullptr, "SleepSystem"}, {0x4E, nullptr, "ReadWriteRegister"}, {0x4F, nullptr, "SetProcessActivity"}, - {0x50, SvcWrap64<CreateSharedMemory>, "CreateSharedMemory"}, - {0x51, SvcWrap64<MapTransferMemory>, "MapTransferMemory"}, - {0x52, SvcWrap64<UnmapTransferMemory>, "UnmapTransferMemory"}, + {0x50, nullptr, "CreateSharedMemory"}, + {0x51, nullptr, "MapTransferMemory"}, + {0x52, nullptr, "UnmapTransferMemory"}, {0x53, nullptr, "CreateInterruptEvent"}, {0x54, nullptr, "QueryPhysicalAddress"}, {0x55, nullptr, "QueryIoMapping"}, @@ -2654,13 +2634,10 @@ static const FunctionDef* GetSVCInfo64(u32 func_num) { return &SVC_Table_64[func_num]; } -MICROPROFILE_DEFINE(Kernel_SVC, "Kernel", "SVC", MP_RGB(70, 200, 70)); - -void CallSVC(Core::System& system, u32 immediate) { - MICROPROFILE_SCOPE(Kernel_SVC); - - // Lock the global kernel mutex when we enter the kernel HLE. - std::lock_guard lock{HLE::g_hle_lock}; +void Call(Core::System& system, u32 immediate) { + system.ExitDynarmicProfile(); + auto& kernel = system.Kernel(); + kernel.EnterSVCProfile(); const FunctionDef* info = system.CurrentProcess()->Is64BitProcess() ? GetSVCInfo64(immediate) : GetSVCInfo32(immediate); @@ -2673,6 +2650,9 @@ void CallSVC(Core::System& system, u32 immediate) { } else { LOG_CRITICAL(Kernel_SVC, "Unknown SVC function 0x{:X}", immediate); } + + kernel.ExitSVCProfile(); + system.EnterDynarmicProfile(); } -} // namespace Kernel +} // namespace Kernel::Svc diff --git a/src/core/hle/kernel/svc.h b/src/core/hle/kernel/svc.h index c5539ac1c..46e64277e 100644 --- a/src/core/hle/kernel/svc.h +++ b/src/core/hle/kernel/svc.h @@ -10,8 +10,8 @@ namespace Core { class System; } -namespace Kernel { +namespace Kernel::Svc { -void CallSVC(Core::System& system, u32 immediate); +void Call(Core::System& system, u32 immediate); -} // namespace Kernel +} // namespace Kernel::Svc diff --git a/src/core/hle/kernel/svc_types.h b/src/core/hle/kernel/svc_types.h new file mode 100644 index 000000000..986724beb --- /dev/null +++ b/src/core/hle/kernel/svc_types.h @@ -0,0 +1,68 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Kernel::Svc { + +enum class MemoryState : u32 { + Free = 0x00, + Io = 0x01, + Static = 0x02, + Code = 0x03, + CodeData = 0x04, + Normal = 0x05, + Shared = 0x06, + Alias = 0x07, + AliasCode = 0x08, + AliasCodeData = 0x09, + Ipc = 0x0A, + Stack = 0x0B, + ThreadLocal = 0x0C, + Transfered = 0x0D, + SharedTransfered = 0x0E, + SharedCode = 0x0F, + Inaccessible = 0x10, + NonSecureIpc = 0x11, + NonDeviceIpc = 0x12, + Kernel = 0x13, + GeneratedCode = 0x14, + CodeOut = 0x15, +}; +DECLARE_ENUM_FLAG_OPERATORS(MemoryState); + +enum class MemoryAttribute : u32 { + Locked = (1 << 0), + IpcLocked = (1 << 1), + DeviceShared = (1 << 2), + Uncached = (1 << 3), +}; +DECLARE_ENUM_FLAG_OPERATORS(MemoryAttribute); + +enum class MemoryPermission : u32 { + None = (0 << 0), + Read = (1 << 0), + Write = (1 << 1), + Execute = (1 << 2), + ReadWrite = Read | Write, + ReadExecute = Read | Execute, + DontCare = (1 << 28), +}; +DECLARE_ENUM_FLAG_OPERATORS(MemoryPermission); + +struct MemoryInfo { + u64 addr{}; + u64 size{}; + MemoryState state{}; + MemoryAttribute attr{}; + MemoryPermission perm{}; + u32 ipc_refcount{}; + u32 device_refcount{}; + u32 padding{}; +}; + +} // namespace Kernel::Svc diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h index 7d735e3fa..0b6dd9df0 100644 --- a/src/core/hle/kernel/svc_wrap.h +++ b/src/core/hle/kernel/svc_wrap.h @@ -350,13 +350,50 @@ void SvcWrap64(Core::System& system) { func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), Param(system, 2)); } -// Used by QueryMemory32 +// Used by QueryMemory32, ArbitrateLock32 template <ResultCode func(Core::System&, u32, u32, u32)> void SvcWrap32(Core::System& system) { FuncReturn32(system, func(system, Param32(system, 0), Param32(system, 1), Param32(system, 2)).raw); } +// Used by Break32 +template <void func(Core::System&, u32, u32, u32)> +void SvcWrap32(Core::System& system) { + func(system, Param32(system, 0), Param32(system, 1), Param32(system, 2)); +} + +// Used by ExitProcess32, ExitThread32 +template <void func(Core::System&)> +void SvcWrap32(Core::System& system) { + func(system); +} + +// Used by GetCurrentProcessorNumber32 +template <u32 func(Core::System&)> +void SvcWrap32(Core::System& system) { + FuncReturn32(system, func(system)); +} + +// Used by SleepThread32 +template <void func(Core::System&, u32, u32)> +void SvcWrap32(Core::System& system) { + func(system, Param32(system, 0), Param32(system, 1)); +} + +// Used by CreateThread32 +template <ResultCode func(Core::System&, Handle*, u32, u32, u32, u32, s32)> +void SvcWrap32(Core::System& system) { + Handle param_1 = 0; + + const u32 retval = func(system, ¶m_1, Param32(system, 0), Param32(system, 1), + Param32(system, 2), Param32(system, 3), Param32(system, 4)) + .raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); +} + // Used by GetInfo32 template <ResultCode func(Core::System&, u32*, u32*, u32, u32, u32, u32)> void SvcWrap32(Core::System& system) { @@ -393,18 +430,114 @@ void SvcWrap32(Core::System& system) { FuncReturn(system, retval); } +// Used by GetSystemTick32 +template <void func(Core::System&, u32*, u32*)> +void SvcWrap32(Core::System& system) { + u32 param_1 = 0; + u32 param_2 = 0; + + func(system, ¶m_1, ¶m_2); + system.CurrentArmInterface().SetReg(0, param_1); + system.CurrentArmInterface().SetReg(1, param_2); +} + +// Used by CreateEvent32 +template <ResultCode func(Core::System&, Handle*, Handle*)> +void SvcWrap32(Core::System& system) { + Handle param_1 = 0; + Handle param_2 = 0; + + const u32 retval = func(system, ¶m_1, ¶m_2).raw; + system.CurrentArmInterface().SetReg(1, param_1); + system.CurrentArmInterface().SetReg(2, param_2); + FuncReturn(system, retval); +} + +// Used by GetThreadId32 +template <ResultCode func(Core::System&, Handle, u32*, u32*, u32*)> +void SvcWrap32(Core::System& system) { + u32 param_1 = 0; + u32 param_2 = 0; + u32 param_3 = 0; + + const u32 retval = func(system, Param32(system, 2), ¶m_1, ¶m_2, ¶m_3).raw; + system.CurrentArmInterface().SetReg(1, param_1); + system.CurrentArmInterface().SetReg(2, param_2); + system.CurrentArmInterface().SetReg(3, param_3); + FuncReturn(system, retval); +} + // Used by SignalProcessWideKey32 template <void func(Core::System&, u32, s32)> void SvcWrap32(Core::System& system) { func(system, static_cast<u32>(Param(system, 0)), static_cast<s32>(Param(system, 1))); } -// Used by SendSyncRequest32 +// Used by SetThreadPriority32 +template <ResultCode func(Core::System&, Handle, u32)> +void SvcWrap32(Core::System& system) { + const u32 retval = + func(system, static_cast<Handle>(Param(system, 0)), static_cast<u32>(Param(system, 1))).raw; + FuncReturn(system, retval); +} + +// Used by SetThreadCoreMask32 +template <ResultCode func(Core::System&, Handle, u32, u32, u32)> +void SvcWrap32(Core::System& system) { + const u32 retval = + func(system, static_cast<Handle>(Param(system, 0)), static_cast<u32>(Param(system, 1)), + static_cast<u32>(Param(system, 2)), static_cast<u32>(Param(system, 3))) + .raw; + FuncReturn(system, retval); +} + +// Used by WaitProcessWideKeyAtomic32 +template <ResultCode func(Core::System&, u32, u32, Handle, u32, u32)> +void SvcWrap32(Core::System& system) { + const u32 retval = + func(system, static_cast<u32>(Param(system, 0)), static_cast<u32>(Param(system, 1)), + static_cast<Handle>(Param(system, 2)), static_cast<u32>(Param(system, 3)), + static_cast<u32>(Param(system, 4))) + .raw; + FuncReturn(system, retval); +} + +// Used by WaitForAddress32 +template <ResultCode func(Core::System&, u32, u32, s32, u32, u32)> +void SvcWrap32(Core::System& system) { + const u32 retval = func(system, static_cast<u32>(Param(system, 0)), + static_cast<u32>(Param(system, 1)), static_cast<s32>(Param(system, 2)), + static_cast<u32>(Param(system, 3)), static_cast<u32>(Param(system, 4))) + .raw; + FuncReturn(system, retval); +} + +// Used by SignalToAddress32 +template <ResultCode func(Core::System&, u32, u32, s32, s32)> +void SvcWrap32(Core::System& system) { + const u32 retval = + func(system, static_cast<u32>(Param(system, 0)), static_cast<u32>(Param(system, 1)), + static_cast<s32>(Param(system, 2)), static_cast<s32>(Param(system, 3))) + .raw; + FuncReturn(system, retval); +} + +// Used by SendSyncRequest32, ArbitrateUnlock32 template <ResultCode func(Core::System&, u32)> void SvcWrap32(Core::System& system) { FuncReturn(system, func(system, static_cast<u32>(Param(system, 0))).raw); } +// Used by CreateTransferMemory32 +template <ResultCode func(Core::System&, Handle*, u32, u32, u32)> +void SvcWrap32(Core::System& system) { + Handle handle = 0; + const u32 retval = + func(system, &handle, Param32(system, 1), Param32(system, 2), Param32(system, 3)).raw; + system.CurrentArmInterface().SetReg(1, handle); + FuncReturn(system, retval); +} + // Used by WaitSynchronization32 template <ResultCode func(Core::System&, u32, u32, s32, u32, Handle*)> void SvcWrap32(Core::System& system) { diff --git a/src/core/hle/kernel/synchronization.cpp b/src/core/hle/kernel/synchronization.cpp index dc37fad1a..8b875d853 100644 --- a/src/core/hle/kernel/synchronization.cpp +++ b/src/core/hle/kernel/synchronization.cpp @@ -10,78 +10,106 @@ #include "core/hle/kernel/synchronization.h" #include "core/hle/kernel/synchronization_object.h" #include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" namespace Kernel { -/// Default thread wakeup callback for WaitSynchronization -static bool DefaultThreadWakeupCallback(ThreadWakeupReason reason, std::shared_ptr<Thread> thread, - std::shared_ptr<SynchronizationObject> object, - std::size_t index) { - ASSERT(thread->GetStatus() == ThreadStatus::WaitSynch); - - if (reason == ThreadWakeupReason::Timeout) { - thread->SetWaitSynchronizationResult(RESULT_TIMEOUT); - return true; - } - - ASSERT(reason == ThreadWakeupReason::Signal); - thread->SetWaitSynchronizationResult(RESULT_SUCCESS); - thread->SetWaitSynchronizationOutput(static_cast<u32>(index)); - return true; -} - Synchronization::Synchronization(Core::System& system) : system{system} {} void Synchronization::SignalObject(SynchronizationObject& obj) const { + auto& kernel = system.Kernel(); + SchedulerLock lock(kernel); if (obj.IsSignaled()) { - obj.WakeupAllWaitingThreads(); + for (auto thread : obj.GetWaitingThreads()) { + if (thread->GetSchedulingStatus() == ThreadSchedStatus::Paused) { + if (thread->GetStatus() != ThreadStatus::WaitHLEEvent) { + ASSERT(thread->GetStatus() == ThreadStatus::WaitSynch); + ASSERT(thread->IsWaitingSync()); + } + thread->SetSynchronizationResults(&obj, RESULT_SUCCESS); + thread->ResumeFromWait(); + } + } + obj.ClearWaitingThreads(); } } std::pair<ResultCode, Handle> Synchronization::WaitFor( std::vector<std::shared_ptr<SynchronizationObject>>& sync_objects, s64 nano_seconds) { + auto& kernel = system.Kernel(); auto* const thread = system.CurrentScheduler().GetCurrentThread(); - // Find the first object that is acquirable in the provided list of objects - const auto itr = std::find_if(sync_objects.begin(), sync_objects.end(), - [thread](const std::shared_ptr<SynchronizationObject>& object) { - return object->IsSignaled(); - }); - - if (itr != sync_objects.end()) { - // We found a ready object, acquire it and set the result value - SynchronizationObject* object = itr->get(); - object->Acquire(thread); - const u32 index = static_cast<s32>(std::distance(sync_objects.begin(), itr)); - return {RESULT_SUCCESS, index}; + Handle event_handle = InvalidHandle; + { + SchedulerLockAndSleep lock(kernel, event_handle, thread, nano_seconds); + const auto itr = + std::find_if(sync_objects.begin(), sync_objects.end(), + [thread](const std::shared_ptr<SynchronizationObject>& object) { + return object->IsSignaled(); + }); + + if (itr != sync_objects.end()) { + // We found a ready object, acquire it and set the result value + SynchronizationObject* object = itr->get(); + object->Acquire(thread); + const u32 index = static_cast<s32>(std::distance(sync_objects.begin(), itr)); + lock.CancelSleep(); + return {RESULT_SUCCESS, index}; + } + + if (nano_seconds == 0) { + lock.CancelSleep(); + return {RESULT_TIMEOUT, InvalidHandle}; + } + + if (thread->IsPendingTermination()) { + lock.CancelSleep(); + return {ERR_THREAD_TERMINATING, InvalidHandle}; + } + + if (thread->IsSyncCancelled()) { + thread->SetSyncCancelled(false); + lock.CancelSleep(); + return {ERR_SYNCHRONIZATION_CANCELED, InvalidHandle}; + } + + for (auto& object : sync_objects) { + object->AddWaitingThread(SharedFrom(thread)); + } + + thread->SetSynchronizationObjects(&sync_objects); + thread->SetSynchronizationResults(nullptr, RESULT_TIMEOUT); + thread->SetStatus(ThreadStatus::WaitSynch); + thread->SetWaitingSync(true); } + thread->SetWaitingSync(false); - // No objects were ready to be acquired, prepare to suspend the thread. - - // If a timeout value of 0 was provided, just return the Timeout error code instead of - // suspending the thread. - if (nano_seconds == 0) { - return {RESULT_TIMEOUT, InvalidHandle}; + if (event_handle != InvalidHandle) { + auto& time_manager = kernel.TimeManager(); + time_manager.UnscheduleTimeEvent(event_handle); } - if (thread->IsSyncCancelled()) { - thread->SetSyncCancelled(false); - return {ERR_SYNCHRONIZATION_CANCELED, InvalidHandle}; + { + SchedulerLock lock(kernel); + ResultCode signaling_result = thread->GetSignalingResult(); + SynchronizationObject* signaling_object = thread->GetSignalingObject(); + thread->SetSynchronizationObjects(nullptr); + auto shared_thread = SharedFrom(thread); + for (auto& obj : sync_objects) { + obj->RemoveWaitingThread(shared_thread); + } + if (signaling_object != nullptr) { + const auto itr = std::find_if( + sync_objects.begin(), sync_objects.end(), + [signaling_object](const std::shared_ptr<SynchronizationObject>& object) { + return object.get() == signaling_object; + }); + ASSERT(itr != sync_objects.end()); + signaling_object->Acquire(thread); + const u32 index = static_cast<s32>(std::distance(sync_objects.begin(), itr)); + return {signaling_result, index}; + } + return {signaling_result, -1}; } - - for (auto& object : sync_objects) { - object->AddWaitingThread(SharedFrom(thread)); - } - - thread->SetSynchronizationObjects(std::move(sync_objects)); - thread->SetStatus(ThreadStatus::WaitSynch); - - // Create an event to wake the thread up after the specified nanosecond delay has passed - thread->WakeAfterDelay(nano_seconds); - thread->SetWakeupCallback(DefaultThreadWakeupCallback); - - system.PrepareReschedule(thread->GetProcessorID()); - - return {RESULT_TIMEOUT, InvalidHandle}; } } // namespace Kernel diff --git a/src/core/hle/kernel/synchronization_object.cpp b/src/core/hle/kernel/synchronization_object.cpp index 43f3eef18..ba4d39157 100644 --- a/src/core/hle/kernel/synchronization_object.cpp +++ b/src/core/hle/kernel/synchronization_object.cpp @@ -38,68 +38,8 @@ void SynchronizationObject::RemoveWaitingThread(std::shared_ptr<Thread> thread) waiting_threads.erase(itr); } -std::shared_ptr<Thread> SynchronizationObject::GetHighestPriorityReadyThread() const { - Thread* candidate = nullptr; - u32 candidate_priority = THREADPRIO_LOWEST + 1; - - for (const auto& thread : waiting_threads) { - const ThreadStatus thread_status = thread->GetStatus(); - - // The list of waiting threads must not contain threads that are not waiting to be awakened. - ASSERT_MSG(thread_status == ThreadStatus::WaitSynch || - thread_status == ThreadStatus::WaitHLEEvent, - "Inconsistent thread statuses in waiting_threads"); - - if (thread->GetPriority() >= candidate_priority) - continue; - - if (ShouldWait(thread.get())) - continue; - - candidate = thread.get(); - candidate_priority = thread->GetPriority(); - } - - return SharedFrom(candidate); -} - -void SynchronizationObject::WakeupWaitingThread(std::shared_ptr<Thread> thread) { - ASSERT(!ShouldWait(thread.get())); - - if (!thread) { - return; - } - - if (thread->IsSleepingOnWait()) { - for (const auto& object : thread->GetSynchronizationObjects()) { - ASSERT(!object->ShouldWait(thread.get())); - object->Acquire(thread.get()); - } - } else { - Acquire(thread.get()); - } - - const std::size_t index = thread->GetSynchronizationObjectIndex(SharedFrom(this)); - - thread->ClearSynchronizationObjects(); - - thread->CancelWakeupTimer(); - - bool resume = true; - if (thread->HasWakeupCallback()) { - resume = thread->InvokeWakeupCallback(ThreadWakeupReason::Signal, thread, SharedFrom(this), - index); - } - if (resume) { - thread->ResumeFromWait(); - kernel.PrepareReschedule(thread->GetProcessorID()); - } -} - -void SynchronizationObject::WakeupAllWaitingThreads() { - while (auto thread = GetHighestPriorityReadyThread()) { - WakeupWaitingThread(thread); - } +void SynchronizationObject::ClearWaitingThreads() { + waiting_threads.clear(); } const std::vector<std::shared_ptr<Thread>>& SynchronizationObject::GetWaitingThreads() const { diff --git a/src/core/hle/kernel/synchronization_object.h b/src/core/hle/kernel/synchronization_object.h index 741c31faf..f89b24204 100644 --- a/src/core/hle/kernel/synchronization_object.h +++ b/src/core/hle/kernel/synchronization_object.h @@ -12,6 +12,7 @@ namespace Kernel { class KernelCore; +class Synchronization; class Thread; /// Class that represents a Kernel object that a thread can be waiting on @@ -49,24 +50,11 @@ public: */ void RemoveWaitingThread(std::shared_ptr<Thread> thread); - /** - * Wake up all threads waiting on this object that can be awoken, in priority order, - * and set the synchronization result and output of the thread. - */ - void WakeupAllWaitingThreads(); - - /** - * Wakes up a single thread waiting on this object. - * @param thread Thread that is waiting on this object to wakeup. - */ - void WakeupWaitingThread(std::shared_ptr<Thread> thread); - - /// Obtains the highest priority thread that is ready to run from this object's waiting list. - std::shared_ptr<Thread> GetHighestPriorityReadyThread() const; - /// Get a const reference to the waiting threads list for debug use const std::vector<std::shared_ptr<Thread>>& GetWaitingThreads() const; + void ClearWaitingThreads(); + protected: bool is_signaled{}; // Tells if this sync object is signalled; diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index 83e956036..da0cb26b6 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -9,12 +9,12 @@ #include "common/assert.h" #include "common/common_types.h" +#include "common/fiber.h" #include "common/logging/log.h" #include "common/thread_queue_list.h" #include "core/arm/arm_interface.h" #include "core/core.h" -#include "core/core_timing.h" -#include "core/core_timing_util.h" +#include "core/cpu_manager.h" #include "core/hardware_properties.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" @@ -23,9 +23,15 @@ #include "core/hle/kernel/process.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" #include "core/hle/result.h" #include "core/memory.h" +#ifdef ARCHITECTURE_x86_64 +#include "core/arm/dynarmic/arm_dynarmic_32.h" +#include "core/arm/dynarmic/arm_dynarmic_64.h" +#endif + namespace Kernel { bool Thread::ShouldWait(const Thread* thread) const { @@ -44,47 +50,28 @@ Thread::Thread(KernelCore& kernel) : SynchronizationObject{kernel} {} Thread::~Thread() = default; void Thread::Stop() { - // Cancel any outstanding wakeup events for this thread - Core::System::GetInstance().CoreTiming().UnscheduleEvent(kernel.ThreadWakeupCallbackEventType(), - global_handle); - kernel.GlobalHandleTable().Close(global_handle); - global_handle = 0; - SetStatus(ThreadStatus::Dead); - Signal(); - - // Clean up any dangling references in objects that this thread was waiting for - for (auto& wait_object : wait_objects) { - wait_object->RemoveWaitingThread(SharedFrom(this)); - } - wait_objects.clear(); - - owner_process->UnregisterThread(this); - - // Mark the TLS slot in the thread's page as free. - owner_process->FreeTLSRegion(tls_address); -} + { + SchedulerLock lock(kernel); + SetStatus(ThreadStatus::Dead); + Signal(); + kernel.GlobalHandleTable().Close(global_handle); -void Thread::WakeAfterDelay(s64 nanoseconds) { - // Don't schedule a wakeup if the thread wants to wait forever - if (nanoseconds == -1) - return; - - // This function might be called from any thread so we have to be cautious and use the - // thread-safe version of ScheduleEvent. - const s64 cycles = Core::Timing::nsToCycles(std::chrono::nanoseconds{nanoseconds}); - Core::System::GetInstance().CoreTiming().ScheduleEvent( - cycles, kernel.ThreadWakeupCallbackEventType(), global_handle); -} + if (owner_process) { + owner_process->UnregisterThread(this); -void Thread::CancelWakeupTimer() { - Core::System::GetInstance().CoreTiming().UnscheduleEvent(kernel.ThreadWakeupCallbackEventType(), - global_handle); + // Mark the TLS slot in the thread's page as free. + owner_process->FreeTLSRegion(tls_address); + } + arm_interface.reset(); + has_exited = true; + } + global_handle = 0; } void Thread::ResumeFromWait() { - ASSERT_MSG(wait_objects.empty(), "Thread is waking up while waiting for objects"); - + SchedulerLock lock(kernel); switch (status) { + case ThreadStatus::Paused: case ThreadStatus::WaitSynch: case ThreadStatus::WaitHLEEvent: case ThreadStatus::WaitSleep: @@ -92,12 +79,13 @@ void Thread::ResumeFromWait() { case ThreadStatus::WaitMutex: case ThreadStatus::WaitCondVar: case ThreadStatus::WaitArb: + case ThreadStatus::Dormant: break; case ThreadStatus::Ready: // The thread's wakeup callback must have already been cleared when the thread was first // awoken. - ASSERT(wakeup_callback == nullptr); + ASSERT(hle_callback == nullptr); // If the thread is waiting on multiple wait objects, it might be awoken more than once // before actually resuming. We can ignore subsequent wakeups if the thread status has // already been set to ThreadStatus::Ready. @@ -113,24 +101,31 @@ void Thread::ResumeFromWait() { return; } - wakeup_callback = nullptr; + SetStatus(ThreadStatus::Ready); +} - if (activity == ThreadActivity::Paused) { - SetStatus(ThreadStatus::Paused); - return; - } +void Thread::OnWakeUp() { + SchedulerLock lock(kernel); + + SetStatus(ThreadStatus::Ready); +} +ResultCode Thread::Start() { + SchedulerLock lock(kernel); SetStatus(ThreadStatus::Ready); + return RESULT_SUCCESS; } void Thread::CancelWait() { - if (GetSchedulingStatus() != ThreadSchedStatus::Paused) { + SchedulerLock lock(kernel); + if (GetSchedulingStatus() != ThreadSchedStatus::Paused || !is_waiting_on_sync) { is_sync_cancelled = true; return; } + // TODO(Blinkhawk): Implement cancel of server session is_sync_cancelled = false; - SetWaitSynchronizationResult(ERR_SYNCHRONIZATION_CANCELED); - ResumeFromWait(); + SetSynchronizationResults(nullptr, ERR_SYNCHRONIZATION_CANCELED); + SetStatus(ThreadStatus::Ready); } static void ResetThreadContext32(Core::ARM_Interface::ThreadContext32& context, u32 stack_top, @@ -148,16 +143,32 @@ static void ResetThreadContext64(Core::ARM_Interface::ThreadContext64& context, context.pc = entry_point; context.sp = stack_top; // TODO(merry): Perform a hardware test to determine the below value. - // AHP = 0, DN = 1, FTZ = 1, RMode = Round towards zero - context.fpcr = 0x03C00000; + context.fpcr = 0; +} + +std::shared_ptr<Common::Fiber>& Thread::GetHostContext() { + return host_context; +} + +ResultVal<std::shared_ptr<Thread>> Thread::Create(Core::System& system, ThreadType type_flags, + std::string name, VAddr entry_point, u32 priority, + u64 arg, s32 processor_id, VAddr stack_top, + Process* owner_process) { + std::function<void(void*)> init_func = Core::CpuManager::GetGuestThreadStartFunc(); + void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater(); + return Create(system, type_flags, name, entry_point, priority, arg, processor_id, stack_top, + owner_process, std::move(init_func), init_func_parameter); } -ResultVal<std::shared_ptr<Thread>> Thread::Create(KernelCore& kernel, std::string name, - VAddr entry_point, u32 priority, u64 arg, - s32 processor_id, VAddr stack_top, - Process& owner_process) { +ResultVal<std::shared_ptr<Thread>> Thread::Create(Core::System& system, ThreadType type_flags, + std::string name, VAddr entry_point, u32 priority, + u64 arg, s32 processor_id, VAddr stack_top, + Process* owner_process, + std::function<void(void*)>&& thread_start_func, + void* thread_start_parameter) { + auto& kernel = system.Kernel(); // Check if priority is in ranged. Lowest priority -> highest priority id. - if (priority > THREADPRIO_LOWEST) { + if (priority > THREADPRIO_LOWEST && ((type_flags & THREADTYPE_IDLE) == 0)) { LOG_ERROR(Kernel_SVC, "Invalid thread priority: {}", priority); return ERR_INVALID_THREAD_PRIORITY; } @@ -167,11 +178,12 @@ ResultVal<std::shared_ptr<Thread>> Thread::Create(KernelCore& kernel, std::strin return ERR_INVALID_PROCESSOR_ID; } - auto& system = Core::System::GetInstance(); - if (!system.Memory().IsValidVirtualAddress(owner_process, entry_point)) { - LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:016X}", name, entry_point); - // TODO (bunnei): Find the correct error code to use here - return RESULT_UNKNOWN; + if (owner_process) { + if (!system.Memory().IsValidVirtualAddress(*owner_process, entry_point)) { + LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:016X}", name, entry_point); + // TODO (bunnei): Find the correct error code to use here + return RESULT_UNKNOWN; + } } std::shared_ptr<Thread> thread = std::make_shared<Thread>(kernel); @@ -182,51 +194,72 @@ ResultVal<std::shared_ptr<Thread>> Thread::Create(KernelCore& kernel, std::strin thread->stack_top = stack_top; thread->tpidr_el0 = 0; thread->nominal_priority = thread->current_priority = priority; - thread->last_running_ticks = system.CoreTiming().GetTicks(); + thread->last_running_ticks = 0; thread->processor_id = processor_id; thread->ideal_core = processor_id; thread->affinity_mask = 1ULL << processor_id; - thread->wait_objects.clear(); + thread->wait_objects = nullptr; thread->mutex_wait_address = 0; thread->condvar_wait_address = 0; thread->wait_handle = 0; thread->name = std::move(name); thread->global_handle = kernel.GlobalHandleTable().Create(thread).Unwrap(); - thread->owner_process = &owner_process; - auto& scheduler = kernel.GlobalScheduler(); - scheduler.AddThread(thread); - thread->tls_address = thread->owner_process->CreateTLSRegion(); + thread->owner_process = owner_process; + thread->type = type_flags; + if ((type_flags & THREADTYPE_IDLE) == 0) { + auto& scheduler = kernel.GlobalScheduler(); + scheduler.AddThread(thread); + } + if (owner_process) { + thread->tls_address = thread->owner_process->CreateTLSRegion(); + thread->owner_process->RegisterThread(thread.get()); + } else { + thread->tls_address = 0; + } - thread->owner_process->RegisterThread(thread.get()); + thread->arm_interface.reset(); + if ((type_flags & THREADTYPE_HLE) == 0) { +#ifdef ARCHITECTURE_x86_64 + if (owner_process && !owner_process->Is64BitProcess()) { + thread->arm_interface = std::make_unique<Core::ARM_Dynarmic_32>( + system, kernel.Interrupts(), kernel.IsMulticore(), kernel.GetExclusiveMonitor(), + processor_id); + } else { + thread->arm_interface = std::make_unique<Core::ARM_Dynarmic_64>( + system, kernel.Interrupts(), kernel.IsMulticore(), kernel.GetExclusiveMonitor(), + processor_id); + } +#else +#error Platform not supported yet. +#endif - ResetThreadContext32(thread->context_32, static_cast<u32>(stack_top), - static_cast<u32>(entry_point), static_cast<u32>(arg)); - ResetThreadContext64(thread->context_64, stack_top, entry_point, arg); + ResetThreadContext32(thread->context_32, static_cast<u32>(stack_top), + static_cast<u32>(entry_point), static_cast<u32>(arg)); + ResetThreadContext64(thread->context_64, stack_top, entry_point, arg); + } + thread->host_context = + std::make_shared<Common::Fiber>(std::move(thread_start_func), thread_start_parameter); return MakeResult<std::shared_ptr<Thread>>(std::move(thread)); } void Thread::SetPriority(u32 priority) { + SchedulerLock lock(kernel); ASSERT_MSG(priority <= THREADPRIO_LOWEST && priority >= THREADPRIO_HIGHEST, "Invalid priority value."); nominal_priority = priority; UpdatePriority(); } -void Thread::SetWaitSynchronizationResult(ResultCode result) { - context_32.cpu_registers[0] = result.raw; - context_64.cpu_registers[0] = result.raw; -} - -void Thread::SetWaitSynchronizationOutput(s32 output) { - context_32.cpu_registers[1] = output; - context_64.cpu_registers[1] = output; +void Thread::SetSynchronizationResults(SynchronizationObject* object, ResultCode result) { + signaling_object = object; + signaling_result = result; } s32 Thread::GetSynchronizationObjectIndex(std::shared_ptr<SynchronizationObject> object) const { - ASSERT_MSG(!wait_objects.empty(), "Thread is not waiting for anything"); - const auto match = std::find(wait_objects.rbegin(), wait_objects.rend(), object); - return static_cast<s32>(std::distance(match, wait_objects.rend()) - 1); + ASSERT_MSG(!wait_objects->empty(), "Thread is not waiting for anything"); + const auto match = std::find(wait_objects->rbegin(), wait_objects->rend(), object); + return static_cast<s32>(std::distance(match, wait_objects->rend()) - 1); } VAddr Thread::GetCommandBufferAddress() const { @@ -235,6 +268,14 @@ VAddr Thread::GetCommandBufferAddress() const { return GetTLSAddress() + command_header_offset; } +Core::ARM_Interface& Thread::ArmInterface() { + return *arm_interface; +} + +const Core::ARM_Interface& Thread::ArmInterface() const { + return *arm_interface; +} + void Thread::SetStatus(ThreadStatus new_status) { if (new_status == status) { return; @@ -256,10 +297,6 @@ void Thread::SetStatus(ThreadStatus new_status) { break; } - if (status == ThreadStatus::Running) { - last_running_ticks = Core::System::GetInstance().CoreTiming().GetTicks(); - } - status = new_status; } @@ -340,75 +377,116 @@ void Thread::UpdatePriority() { lock_owner->UpdatePriority(); } -void Thread::ChangeCore(u32 core, u64 mask) { - SetCoreAndAffinityMask(core, mask); -} - bool Thread::AllSynchronizationObjectsReady() const { - return std::none_of(wait_objects.begin(), wait_objects.end(), + return std::none_of(wait_objects->begin(), wait_objects->end(), [this](const std::shared_ptr<SynchronizationObject>& object) { return object->ShouldWait(this); }); } -bool Thread::InvokeWakeupCallback(ThreadWakeupReason reason, std::shared_ptr<Thread> thread, - std::shared_ptr<SynchronizationObject> object, - std::size_t index) { - ASSERT(wakeup_callback); - return wakeup_callback(reason, std::move(thread), std::move(object), index); +bool Thread::InvokeHLECallback(std::shared_ptr<Thread> thread) { + ASSERT(hle_callback); + return hle_callback(std::move(thread)); } -void Thread::SetActivity(ThreadActivity value) { - activity = value; +ResultCode Thread::SetActivity(ThreadActivity value) { + SchedulerLock lock(kernel); + + auto sched_status = GetSchedulingStatus(); + + if (sched_status != ThreadSchedStatus::Runnable && sched_status != ThreadSchedStatus::Paused) { + return ERR_INVALID_STATE; + } + + if (IsPendingTermination()) { + return RESULT_SUCCESS; + } if (value == ThreadActivity::Paused) { - // Set status if not waiting - if (status == ThreadStatus::Ready || status == ThreadStatus::Running) { - SetStatus(ThreadStatus::Paused); - kernel.PrepareReschedule(processor_id); + if ((pausing_state & static_cast<u32>(ThreadSchedFlags::ThreadPauseFlag)) != 0) { + return ERR_INVALID_STATE; + } + AddSchedulingFlag(ThreadSchedFlags::ThreadPauseFlag); + } else { + if ((pausing_state & static_cast<u32>(ThreadSchedFlags::ThreadPauseFlag)) == 0) { + return ERR_INVALID_STATE; } - } else if (status == ThreadStatus::Paused) { - // Ready to reschedule - ResumeFromWait(); + RemoveSchedulingFlag(ThreadSchedFlags::ThreadPauseFlag); } + return RESULT_SUCCESS; } -void Thread::Sleep(s64 nanoseconds) { - // Sleep current thread and check for next thread to schedule - SetStatus(ThreadStatus::WaitSleep); +ResultCode Thread::Sleep(s64 nanoseconds) { + Handle event_handle{}; + { + SchedulerLockAndSleep lock(kernel, event_handle, this, nanoseconds); + SetStatus(ThreadStatus::WaitSleep); + } - // Create an event to wake the thread up after the specified nanosecond delay has passed - WakeAfterDelay(nanoseconds); + if (event_handle != InvalidHandle) { + auto& time_manager = kernel.TimeManager(); + time_manager.UnscheduleTimeEvent(event_handle); + } + return RESULT_SUCCESS; } -bool Thread::YieldSimple() { - auto& scheduler = kernel.GlobalScheduler(); - return scheduler.YieldThread(this); +std::pair<ResultCode, bool> Thread::YieldSimple() { + bool is_redundant = false; + { + SchedulerLock lock(kernel); + is_redundant = kernel.GlobalScheduler().YieldThread(this); + } + return {RESULT_SUCCESS, is_redundant}; +} + +std::pair<ResultCode, bool> Thread::YieldAndBalanceLoad() { + bool is_redundant = false; + { + SchedulerLock lock(kernel); + is_redundant = kernel.GlobalScheduler().YieldThreadAndBalanceLoad(this); + } + return {RESULT_SUCCESS, is_redundant}; +} + +std::pair<ResultCode, bool> Thread::YieldAndWaitForLoadBalancing() { + bool is_redundant = false; + { + SchedulerLock lock(kernel); + is_redundant = kernel.GlobalScheduler().YieldThreadAndWaitForLoadBalancing(this); + } + return {RESULT_SUCCESS, is_redundant}; } -bool Thread::YieldAndBalanceLoad() { - auto& scheduler = kernel.GlobalScheduler(); - return scheduler.YieldThreadAndBalanceLoad(this); +void Thread::AddSchedulingFlag(ThreadSchedFlags flag) { + const u32 old_state = scheduling_state; + pausing_state |= static_cast<u32>(flag); + const u32 base_scheduling = static_cast<u32>(GetSchedulingStatus()); + scheduling_state = base_scheduling | pausing_state; + kernel.GlobalScheduler().AdjustSchedulingOnStatus(this, old_state); } -bool Thread::YieldAndWaitForLoadBalancing() { - auto& scheduler = kernel.GlobalScheduler(); - return scheduler.YieldThreadAndWaitForLoadBalancing(this); +void Thread::RemoveSchedulingFlag(ThreadSchedFlags flag) { + const u32 old_state = scheduling_state; + pausing_state &= ~static_cast<u32>(flag); + const u32 base_scheduling = static_cast<u32>(GetSchedulingStatus()); + scheduling_state = base_scheduling | pausing_state; + kernel.GlobalScheduler().AdjustSchedulingOnStatus(this, old_state); } void Thread::SetSchedulingStatus(ThreadSchedStatus new_status) { - const u32 old_flags = scheduling_state; + const u32 old_state = scheduling_state; scheduling_state = (scheduling_state & static_cast<u32>(ThreadSchedMasks::HighMask)) | static_cast<u32>(new_status); - AdjustSchedulingOnStatus(old_flags); + kernel.GlobalScheduler().AdjustSchedulingOnStatus(this, old_state); } void Thread::SetCurrentPriority(u32 new_priority) { const u32 old_priority = std::exchange(current_priority, new_priority); - AdjustSchedulingOnPriority(old_priority); + kernel.GlobalScheduler().AdjustSchedulingOnPriority(this, old_priority); } ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) { + SchedulerLock lock(kernel); const auto HighestSetCore = [](u64 mask, u32 max_cores) { for (s32 core = static_cast<s32>(max_cores - 1); core >= 0; core--) { if (((mask >> core) & 1) != 0) { @@ -422,6 +500,8 @@ ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) { if (new_core == THREADPROCESSORID_DONT_UPDATE) { new_core = use_override ? ideal_core_override : ideal_core; if ((new_affinity_mask & (1ULL << new_core)) == 0) { + LOG_ERROR(Kernel, "New affinity mask is incorrect! new_core={}, new_affinity_mask={}", + new_core, new_affinity_mask); return ERR_INVALID_COMBINATION; } } @@ -440,118 +520,10 @@ ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) { processor_id = ideal_core; } } - AdjustSchedulingOnAffinity(old_affinity_mask, old_core); + kernel.GlobalScheduler().AdjustSchedulingOnAffinity(this, old_affinity_mask, old_core); } } return RESULT_SUCCESS; } -void Thread::AdjustSchedulingOnStatus(u32 old_flags) { - if (old_flags == scheduling_state) { - return; - } - - auto& scheduler = kernel.GlobalScheduler(); - if (static_cast<ThreadSchedStatus>(old_flags & static_cast<u32>(ThreadSchedMasks::LowMask)) == - ThreadSchedStatus::Runnable) { - // In this case the thread was running, now it's pausing/exitting - if (processor_id >= 0) { - scheduler.Unschedule(current_priority, static_cast<u32>(processor_id), this); - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) { - scheduler.Unsuggest(current_priority, core, this); - } - } - } else if (GetSchedulingStatus() == ThreadSchedStatus::Runnable) { - // The thread is now set to running from being stopped - if (processor_id >= 0) { - scheduler.Schedule(current_priority, static_cast<u32>(processor_id), this); - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) { - scheduler.Suggest(current_priority, core, this); - } - } - } - - scheduler.SetReselectionPending(); -} - -void Thread::AdjustSchedulingOnPriority(u32 old_priority) { - if (GetSchedulingStatus() != ThreadSchedStatus::Runnable) { - return; - } - auto& scheduler = kernel.GlobalScheduler(); - if (processor_id >= 0) { - scheduler.Unschedule(old_priority, static_cast<u32>(processor_id), this); - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) { - scheduler.Unsuggest(old_priority, core, this); - } - } - - // Add thread to the new priority queues. - Thread* current_thread = GetCurrentThread(); - - if (processor_id >= 0) { - if (current_thread == this) { - scheduler.SchedulePrepend(current_priority, static_cast<u32>(processor_id), this); - } else { - scheduler.Schedule(current_priority, static_cast<u32>(processor_id), this); - } - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(processor_id) && ((affinity_mask >> core) & 1) != 0) { - scheduler.Suggest(current_priority, core, this); - } - } - - scheduler.SetReselectionPending(); -} - -void Thread::AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core) { - auto& scheduler = kernel.GlobalScheduler(); - if (GetSchedulingStatus() != ThreadSchedStatus::Runnable || - current_priority >= THREADPRIO_COUNT) { - return; - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (((old_affinity_mask >> core) & 1) != 0) { - if (core == static_cast<u32>(old_core)) { - scheduler.Unschedule(current_priority, core, this); - } else { - scheduler.Unsuggest(current_priority, core, this); - } - } - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (((affinity_mask >> core) & 1) != 0) { - if (core == static_cast<u32>(processor_id)) { - scheduler.Schedule(current_priority, core, this); - } else { - scheduler.Suggest(current_priority, core, this); - } - } - } - - scheduler.SetReselectionPending(); -} - -//////////////////////////////////////////////////////////////////////////////////////////////////// - -/** - * Gets the current thread - */ -Thread* GetCurrentThread() { - return Core::System::GetInstance().CurrentScheduler().GetCurrentThread(); -} - } // namespace Kernel diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h index 23fdef8a4..8daf79fac 100644 --- a/src/core/hle/kernel/thread.h +++ b/src/core/hle/kernel/thread.h @@ -6,26 +6,47 @@ #include <functional> #include <string> +#include <utility> #include <vector> #include "common/common_types.h" +#include "common/spin_lock.h" #include "core/arm/arm_interface.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/synchronization_object.h" #include "core/hle/result.h" +namespace Common { +class Fiber; +} + +namespace Core { +class ARM_Interface; +class System; +} // namespace Core + namespace Kernel { +class GlobalScheduler; class KernelCore; class Process; class Scheduler; enum ThreadPriority : u32 { - THREADPRIO_HIGHEST = 0, ///< Highest thread priority - THREADPRIO_USERLAND_MAX = 24, ///< Highest thread priority for userland apps - THREADPRIO_DEFAULT = 44, ///< Default thread priority for userland apps - THREADPRIO_LOWEST = 63, ///< Lowest thread priority - THREADPRIO_COUNT = 64, ///< Total number of possible thread priorities. + THREADPRIO_HIGHEST = 0, ///< Highest thread priority + THREADPRIO_MAX_CORE_MIGRATION = 2, ///< Highest priority for a core migration + THREADPRIO_USERLAND_MAX = 24, ///< Highest thread priority for userland apps + THREADPRIO_DEFAULT = 44, ///< Default thread priority for userland apps + THREADPRIO_LOWEST = 63, ///< Lowest thread priority + THREADPRIO_COUNT = 64, ///< Total number of possible thread priorities. +}; + +enum ThreadType : u32 { + THREADTYPE_USER = 0x1, + THREADTYPE_KERNEL = 0x2, + THREADTYPE_HLE = 0x4, + THREADTYPE_IDLE = 0x8, + THREADTYPE_SUSPEND = 0x10, }; enum ThreadProcessorId : s32 { @@ -107,26 +128,45 @@ public: using ThreadSynchronizationObjects = std::vector<std::shared_ptr<SynchronizationObject>>; - using WakeupCallback = - std::function<bool(ThreadWakeupReason reason, std::shared_ptr<Thread> thread, - std::shared_ptr<SynchronizationObject> object, std::size_t index)>; + using HLECallback = std::function<bool(std::shared_ptr<Thread> thread)>; + + /** + * Creates and returns a new thread. The new thread is immediately scheduled + * @param system The instance of the whole system + * @param name The friendly name desired for the thread + * @param entry_point The address at which the thread should start execution + * @param priority The thread's priority + * @param arg User data to pass to the thread + * @param processor_id The ID(s) of the processors on which the thread is desired to be run + * @param stack_top The address of the thread's stack top + * @param owner_process The parent process for the thread, if null, it's a kernel thread + * @return A shared pointer to the newly created thread + */ + static ResultVal<std::shared_ptr<Thread>> Create(Core::System& system, ThreadType type_flags, + std::string name, VAddr entry_point, + u32 priority, u64 arg, s32 processor_id, + VAddr stack_top, Process* owner_process); /** * Creates and returns a new thread. The new thread is immediately scheduled - * @param kernel The kernel instance this thread will be created under. + * @param system The instance of the whole system * @param name The friendly name desired for the thread * @param entry_point The address at which the thread should start execution * @param priority The thread's priority * @param arg User data to pass to the thread * @param processor_id The ID(s) of the processors on which the thread is desired to be run * @param stack_top The address of the thread's stack top - * @param owner_process The parent process for the thread + * @param owner_process The parent process for the thread, if null, it's a kernel thread + * @param thread_start_func The function where the host context will start. + * @param thread_start_parameter The parameter which will passed to host context on init * @return A shared pointer to the newly created thread */ - static ResultVal<std::shared_ptr<Thread>> Create(KernelCore& kernel, std::string name, - VAddr entry_point, u32 priority, u64 arg, - s32 processor_id, VAddr stack_top, - Process& owner_process); + static ResultVal<std::shared_ptr<Thread>> Create(Core::System& system, ThreadType type_flags, + std::string name, VAddr entry_point, + u32 priority, u64 arg, s32 processor_id, + VAddr stack_top, Process* owner_process, + std::function<void(void*)>&& thread_start_func, + void* thread_start_parameter); std::string GetName() const override { return name; @@ -181,7 +221,7 @@ public: void UpdatePriority(); /// Changes the core that the thread is running or scheduled to run on. - void ChangeCore(u32 core, u64 mask); + ResultCode SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask); /** * Gets the thread's thread ID @@ -194,6 +234,10 @@ public: /// Resumes a thread from waiting void ResumeFromWait(); + void OnWakeUp(); + + ResultCode Start(); + /// Cancels a waiting operation that this thread may or may not be within. /// /// When the thread is within a waiting state, this will set the thread's @@ -202,26 +246,19 @@ public: /// void CancelWait(); - /** - * Schedules an event to wake up the specified thread after the specified delay - * @param nanoseconds The time this thread will be allowed to sleep for - */ - void WakeAfterDelay(s64 nanoseconds); + void SetSynchronizationResults(SynchronizationObject* object, ResultCode result); - /// Cancel any outstanding wakeup events for this thread - void CancelWakeupTimer(); + Core::ARM_Interface& ArmInterface(); - /** - * Sets the result after the thread awakens (from svcWaitSynchronization) - * @param result Value to set to the returned result - */ - void SetWaitSynchronizationResult(ResultCode result); + const Core::ARM_Interface& ArmInterface() const; - /** - * Sets the output parameter value after the thread awakens (from svcWaitSynchronization) - * @param output Value to set to the output parameter - */ - void SetWaitSynchronizationOutput(s32 output); + SynchronizationObject* GetSignalingObject() const { + return signaling_object; + } + + ResultCode GetSignalingResult() const { + return signaling_result; + } /** * Retrieves the index that this particular object occupies in the list of objects @@ -269,11 +306,6 @@ public: */ VAddr GetCommandBufferAddress() const; - /// Returns whether this thread is waiting on objects from a WaitSynchronization call. - bool IsSleepingOnWait() const { - return status == ThreadStatus::WaitSynch; - } - ThreadContext32& GetContext32() { return context_32; } @@ -290,6 +322,28 @@ public: return context_64; } + bool IsHLEThread() const { + return (type & THREADTYPE_HLE) != 0; + } + + bool IsSuspendThread() const { + return (type & THREADTYPE_SUSPEND) != 0; + } + + bool IsIdleThread() const { + return (type & THREADTYPE_IDLE) != 0; + } + + bool WasRunning() const { + return was_running; + } + + void SetWasRunning(bool value) { + was_running = value; + } + + std::shared_ptr<Common::Fiber>& GetHostContext(); + ThreadStatus GetStatus() const { return status; } @@ -325,18 +379,18 @@ public: } const ThreadSynchronizationObjects& GetSynchronizationObjects() const { - return wait_objects; + return *wait_objects; } - void SetSynchronizationObjects(ThreadSynchronizationObjects objects) { - wait_objects = std::move(objects); + void SetSynchronizationObjects(ThreadSynchronizationObjects* objects) { + wait_objects = objects; } void ClearSynchronizationObjects() { - for (const auto& waiting_object : wait_objects) { + for (const auto& waiting_object : *wait_objects) { waiting_object->RemoveWaitingThread(SharedFrom(this)); } - wait_objects.clear(); + wait_objects->clear(); } /// Determines whether all the objects this thread is waiting on are ready. @@ -386,26 +440,35 @@ public: arb_wait_address = address; } - bool HasWakeupCallback() const { - return wakeup_callback != nullptr; + bool HasHLECallback() const { + return hle_callback != nullptr; } - void SetWakeupCallback(WakeupCallback callback) { - wakeup_callback = std::move(callback); + void SetHLECallback(HLECallback callback) { + hle_callback = std::move(callback); } - void InvalidateWakeupCallback() { - SetWakeupCallback(nullptr); + void SetHLETimeEvent(Handle time_event) { + hle_time_event = time_event; } - /** - * Invokes the thread's wakeup callback. - * - * @pre A valid wakeup callback has been set. Violating this precondition - * will cause an assertion to trigger. - */ - bool InvokeWakeupCallback(ThreadWakeupReason reason, std::shared_ptr<Thread> thread, - std::shared_ptr<SynchronizationObject> object, std::size_t index); + void SetHLESyncObject(SynchronizationObject* object) { + hle_object = object; + } + + Handle GetHLETimeEvent() const { + return hle_time_event; + } + + SynchronizationObject* GetHLESyncObject() const { + return hle_object; + } + + void InvalidateHLECallback() { + SetHLECallback(nullptr); + } + + bool InvokeHLECallback(std::shared_ptr<Thread> thread); u32 GetIdealCore() const { return ideal_core; @@ -415,23 +478,19 @@ public: return affinity_mask; } - ThreadActivity GetActivity() const { - return activity; - } - - void SetActivity(ThreadActivity value); + ResultCode SetActivity(ThreadActivity value); /// Sleeps this thread for the given amount of nanoseconds. - void Sleep(s64 nanoseconds); + ResultCode Sleep(s64 nanoseconds); /// Yields this thread without rebalancing loads. - bool YieldSimple(); + std::pair<ResultCode, bool> YieldSimple(); /// Yields this thread and does a load rebalancing. - bool YieldAndBalanceLoad(); + std::pair<ResultCode, bool> YieldAndBalanceLoad(); /// Yields this thread and if the core is left idle, loads are rebalanced - bool YieldAndWaitForLoadBalancing(); + std::pair<ResultCode, bool> YieldAndWaitForLoadBalancing(); void IncrementYieldCount() { yield_count++; @@ -446,6 +505,10 @@ public: static_cast<u32>(ThreadSchedMasks::LowMask)); } + bool IsRunnable() const { + return scheduling_state == static_cast<u32>(ThreadSchedStatus::Runnable); + } + bool IsRunning() const { return is_running; } @@ -466,17 +529,65 @@ public: return global_handle; } + bool IsWaitingForArbitration() const { + return waiting_for_arbitration; + } + + void WaitForArbitration(bool set) { + waiting_for_arbitration = set; + } + + bool IsWaitingSync() const { + return is_waiting_on_sync; + } + + void SetWaitingSync(bool is_waiting) { + is_waiting_on_sync = is_waiting; + } + + bool IsPendingTermination() const { + return will_be_terminated || GetSchedulingStatus() == ThreadSchedStatus::Exited; + } + + bool IsPaused() const { + return pausing_state != 0; + } + + bool IsContinuousOnSVC() const { + return is_continuous_on_svc; + } + + void SetContinuousOnSVC(bool is_continuous) { + is_continuous_on_svc = is_continuous; + } + + bool IsPhantomMode() const { + return is_phantom_mode; + } + + void SetPhantomMode(bool phantom) { + is_phantom_mode = phantom; + } + + bool HasExited() const { + return has_exited; + } + private: + friend class GlobalScheduler; + friend class Scheduler; + void SetSchedulingStatus(ThreadSchedStatus new_status); - void SetCurrentPriority(u32 new_priority); - ResultCode SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask); + void AddSchedulingFlag(ThreadSchedFlags flag); + void RemoveSchedulingFlag(ThreadSchedFlags flag); - void AdjustSchedulingOnStatus(u32 old_flags); - void AdjustSchedulingOnPriority(u32 old_priority); - void AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core); + void SetCurrentPriority(u32 new_priority); + Common::SpinLock context_guard{}; ThreadContext32 context_32{}; ThreadContext64 context_64{}; + std::unique_ptr<Core::ARM_Interface> arm_interface{}; + std::shared_ptr<Common::Fiber> host_context{}; u64 thread_id = 0; @@ -485,6 +596,8 @@ private: VAddr entry_point = 0; VAddr stack_top = 0; + ThreadType type; + /// Nominal thread priority, as set by the emulated application. /// The nominal priority is the thread priority without priority /// inheritance taken into account. @@ -509,7 +622,10 @@ private: /// Objects that the thread is waiting on, in the same order as they were /// passed to WaitSynchronization. - ThreadSynchronizationObjects wait_objects; + ThreadSynchronizationObjects* wait_objects; + + SynchronizationObject* signaling_object; + ResultCode signaling_result{RESULT_SUCCESS}; /// List of threads that are waiting for a mutex that is held by this thread. MutexWaitingThreads wait_mutex_threads; @@ -526,36 +642,40 @@ private: /// If waiting for an AddressArbiter, this is the address being waited on. VAddr arb_wait_address{0}; + bool waiting_for_arbitration{}; /// Handle used as userdata to reference this object when inserting into the CoreTiming queue. Handle global_handle = 0; - /// Callback that will be invoked when the thread is resumed from a waiting state. If the thread - /// was waiting via WaitSynchronization then the object will be the last object that became - /// available. In case of a timeout, the object will be nullptr. - WakeupCallback wakeup_callback; + /// Callback for HLE Events + HLECallback hle_callback; + Handle hle_time_event; + SynchronizationObject* hle_object; Scheduler* scheduler = nullptr; u32 ideal_core{0xFFFFFFFF}; u64 affinity_mask{0x1}; - ThreadActivity activity = ThreadActivity::Normal; - s32 ideal_core_override = -1; u64 affinity_mask_override = 0x1; u32 affinity_override_count = 0; u32 scheduling_state = 0; + u32 pausing_state = 0; bool is_running = false; + bool is_waiting_on_sync = false; bool is_sync_cancelled = false; + bool is_continuous_on_svc = false; + + bool will_be_terminated = false; + bool is_phantom_mode = false; + bool has_exited = false; + + bool was_running = false; + std::string name; }; -/** - * Gets the current thread - */ -Thread* GetCurrentThread(); - } // namespace Kernel diff --git a/src/core/hle/kernel/time_manager.cpp b/src/core/hle/kernel/time_manager.cpp index 21b290468..95f2446c9 100644 --- a/src/core/hle/kernel/time_manager.cpp +++ b/src/core/hle/kernel/time_manager.cpp @@ -8,30 +8,38 @@ #include "core/core_timing_util.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/kernel/time_manager.h" namespace Kernel { -TimeManager::TimeManager(Core::System& system) : system{system} { +TimeManager::TimeManager(Core::System& system_) : system{system_} { time_manager_event_type = Core::Timing::CreateEvent( - "Kernel::TimeManagerCallback", [this](u64 thread_handle, [[maybe_unused]] s64 cycles_late) { - Handle proper_handle = static_cast<Handle>(thread_handle); - std::shared_ptr<Thread> thread = - this->system.Kernel().RetrieveThreadFromGlobalHandleTable(proper_handle); - thread->ResumeFromWait(); + "Kernel::TimeManagerCallback", + [this](std::uintptr_t thread_handle, std::chrono::nanoseconds) { + const SchedulerLock lock(system.Kernel()); + const auto proper_handle = static_cast<Handle>(thread_handle); + if (cancelled_events[proper_handle]) { + return; + } + auto thread = this->system.Kernel().RetrieveThreadFromGlobalHandleTable(proper_handle); + thread->OnWakeUp(); }); } void TimeManager::ScheduleTimeEvent(Handle& event_handle, Thread* timetask, s64 nanoseconds) { + event_handle = timetask->GetGlobalHandle(); if (nanoseconds > 0) { ASSERT(timetask); - event_handle = timetask->GetGlobalHandle(); - const s64 cycles = Core::Timing::nsToCycles(std::chrono::nanoseconds{nanoseconds}); - system.CoreTiming().ScheduleEvent(cycles, time_manager_event_type, event_handle); + ASSERT(timetask->GetStatus() != ThreadStatus::Ready); + ASSERT(timetask->GetStatus() != ThreadStatus::WaitMutex); + system.CoreTiming().ScheduleEvent(std::chrono::nanoseconds{nanoseconds}, + time_manager_event_type, event_handle); } else { event_handle = InvalidHandle; } + cancelled_events[event_handle] = false; } void TimeManager::UnscheduleTimeEvent(Handle event_handle) { @@ -39,6 +47,12 @@ void TimeManager::UnscheduleTimeEvent(Handle event_handle) { return; } system.CoreTiming().UnscheduleEvent(time_manager_event_type, event_handle); + cancelled_events[event_handle] = true; +} + +void TimeManager::CancelTimeEvent(Thread* time_task) { + Handle event_handle = time_task->GetGlobalHandle(); + UnscheduleTimeEvent(event_handle); } } // namespace Kernel diff --git a/src/core/hle/kernel/time_manager.h b/src/core/hle/kernel/time_manager.h index eaec486d1..307a18765 100644 --- a/src/core/hle/kernel/time_manager.h +++ b/src/core/hle/kernel/time_manager.h @@ -5,6 +5,7 @@ #pragma once #include <memory> +#include <unordered_map> #include "core/hle/kernel/object.h" @@ -35,9 +36,12 @@ public: /// Unschedule an existing time event void UnscheduleTimeEvent(Handle event_handle); + void CancelTimeEvent(Thread* time_task); + private: Core::System& system; std::shared_ptr<Core::Timing::EventType> time_manager_event_type; + std::unordered_map<Handle, bool> cancelled_events; }; } // namespace Kernel diff --git a/src/core/hle/kernel/transfer_memory.cpp b/src/core/hle/kernel/transfer_memory.cpp index f2d3f8b49..765f408c3 100644 --- a/src/core/hle/kernel/transfer_memory.cpp +++ b/src/core/hle/kernel/transfer_memory.cpp @@ -2,17 +2,16 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "core/hle/kernel/errors.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/shared_memory.h" #include "core/hle/kernel/transfer_memory.h" #include "core/hle/result.h" #include "core/memory.h" namespace Kernel { -TransferMemory::TransferMemory(KernelCore& kernel, Memory::Memory& memory) +TransferMemory::TransferMemory(KernelCore& kernel, Core::Memory::Memory& memory) : Object{kernel}, memory{memory} {} TransferMemory::~TransferMemory() { @@ -20,14 +19,15 @@ TransferMemory::~TransferMemory() { Reset(); } -std::shared_ptr<TransferMemory> TransferMemory::Create(KernelCore& kernel, Memory::Memory& memory, - VAddr base_address, u64 size, - MemoryPermission permissions) { +std::shared_ptr<TransferMemory> TransferMemory::Create(KernelCore& kernel, + Core::Memory::Memory& memory, + VAddr base_address, std::size_t size, + Memory::MemoryPermission permissions) { std::shared_ptr<TransferMemory> transfer_memory{ std::make_shared<TransferMemory>(kernel, memory)}; transfer_memory->base_address = base_address; - transfer_memory->memory_size = size; + transfer_memory->size = size; transfer_memory->owner_permissions = permissions; transfer_memory->owner_process = kernel.CurrentProcess(); @@ -38,98 +38,12 @@ const u8* TransferMemory::GetPointer() const { return memory.GetPointer(base_address); } -u64 TransferMemory::GetSize() const { - return memory_size; -} - -ResultCode TransferMemory::MapMemory(VAddr address, u64 size, MemoryPermission permissions) { - if (memory_size != size) { - return ERR_INVALID_SIZE; - } - - if (owner_permissions != permissions) { - return ERR_INVALID_STATE; - } - - if (is_mapped) { - return ERR_INVALID_STATE; - } - - backing_block = std::make_shared<PhysicalMemory>(size); - - const auto map_state = owner_permissions == MemoryPermission::None - ? MemoryState::TransferMemoryIsolated - : MemoryState::TransferMemory; - auto& vm_manager = owner_process->VMManager(); - const auto map_result = vm_manager.MapMemoryBlock(address, backing_block, 0, size, map_state); - if (map_result.Failed()) { - return map_result.Code(); - } - - is_mapped = true; - return RESULT_SUCCESS; -} - ResultCode TransferMemory::Reserve() { - auto& vm_manager{owner_process->VMManager()}; - const auto check_range_result{vm_manager.CheckRangeState( - base_address, memory_size, MemoryState::FlagTransfer | MemoryState::FlagMemoryPoolAllocated, - MemoryState::FlagTransfer | MemoryState::FlagMemoryPoolAllocated, VMAPermission::All, - VMAPermission::ReadWrite, MemoryAttribute::Mask, MemoryAttribute::None, - MemoryAttribute::IpcAndDeviceMapped)}; - - if (check_range_result.Failed()) { - return check_range_result.Code(); - } - - auto [state_, permissions_, attribute] = *check_range_result; - - if (const auto result{vm_manager.ReprotectRange( - base_address, memory_size, SharedMemory::ConvertPermissions(owner_permissions))}; - result.IsError()) { - return result; - } - - return vm_manager.SetMemoryAttribute(base_address, memory_size, MemoryAttribute::Mask, - attribute | MemoryAttribute::Locked); + return owner_process->PageTable().ReserveTransferMemory(base_address, size, owner_permissions); } ResultCode TransferMemory::Reset() { - auto& vm_manager{owner_process->VMManager()}; - if (const auto result{vm_manager.CheckRangeState( - base_address, memory_size, - MemoryState::FlagTransfer | MemoryState::FlagMemoryPoolAllocated, - MemoryState::FlagTransfer | MemoryState::FlagMemoryPoolAllocated, VMAPermission::None, - VMAPermission::None, MemoryAttribute::Mask, MemoryAttribute::Locked, - MemoryAttribute::IpcAndDeviceMapped)}; - result.Failed()) { - return result.Code(); - } - - if (const auto result{ - vm_manager.ReprotectRange(base_address, memory_size, VMAPermission::ReadWrite)}; - result.IsError()) { - return result; - } - - return vm_manager.SetMemoryAttribute(base_address, memory_size, MemoryAttribute::Mask, - MemoryAttribute::None); -} - -ResultCode TransferMemory::UnmapMemory(VAddr address, u64 size) { - if (memory_size != size) { - return ERR_INVALID_SIZE; - } - - auto& vm_manager = owner_process->VMManager(); - const auto result = vm_manager.UnmapRange(address, size); - - if (result.IsError()) { - return result; - } - - is_mapped = false; - return RESULT_SUCCESS; + return owner_process->PageTable().ResetTransferMemory(base_address, size); } } // namespace Kernel diff --git a/src/core/hle/kernel/transfer_memory.h b/src/core/hle/kernel/transfer_memory.h index 6e388536a..05e9f7464 100644 --- a/src/core/hle/kernel/transfer_memory.h +++ b/src/core/hle/kernel/transfer_memory.h @@ -6,12 +6,13 @@ #include <memory> +#include "core/hle/kernel/memory/memory_block.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/physical_memory.h" union ResultCode; -namespace Memory { +namespace Core::Memory { class Memory; } @@ -20,8 +21,6 @@ namespace Kernel { class KernelCore; class Process; -enum class MemoryPermission : u32; - /// Defines the interface for transfer memory objects. /// /// Transfer memory is typically used for the purpose of @@ -30,14 +29,14 @@ enum class MemoryPermission : u32; /// class TransferMemory final : public Object { public: - explicit TransferMemory(KernelCore& kernel, Memory::Memory& memory); + explicit TransferMemory(KernelCore& kernel, Core::Memory::Memory& memory); ~TransferMemory() override; static constexpr HandleType HANDLE_TYPE = HandleType::TransferMemory; - static std::shared_ptr<TransferMemory> Create(KernelCore& kernel, Memory::Memory& memory, - VAddr base_address, u64 size, - MemoryPermission permissions); + static std::shared_ptr<TransferMemory> Create(KernelCore& kernel, Core::Memory::Memory& memory, + VAddr base_address, std::size_t size, + Memory::MemoryPermission permissions); TransferMemory(const TransferMemory&) = delete; TransferMemory& operator=(const TransferMemory&) = delete; @@ -61,29 +60,9 @@ public: const u8* GetPointer() const; /// Gets the size of the memory backing this instance in bytes. - u64 GetSize() const; - - /// Attempts to map transfer memory with the given range and memory permissions. - /// - /// @param address The base address to being mapping memory at. - /// @param size The size of the memory to map, in bytes. - /// @param permissions The memory permissions to check against when mapping memory. - /// - /// @pre The given address, size, and memory permissions must all match - /// the same values that were given when creating the transfer memory - /// instance. - /// - ResultCode MapMemory(VAddr address, u64 size, MemoryPermission permissions); - - /// Unmaps the transfer memory with the given range - /// - /// @param address The base address to begin unmapping memory at. - /// @param size The size of the memory to unmap, in bytes. - /// - /// @pre The given address and size must be the same as the ones used - /// to create the transfer memory instance. - /// - ResultCode UnmapMemory(VAddr address, u64 size); + constexpr std::size_t GetSize() const { + return size; + } /// Reserves the region to be used for the transfer memory, called after the transfer memory is /// created. @@ -94,25 +73,19 @@ public: ResultCode Reset(); private: - /// Memory block backing this instance. - std::shared_ptr<PhysicalMemory> backing_block; - /// The base address for the memory managed by this instance. - VAddr base_address = 0; + VAddr base_address{}; /// Size of the memory, in bytes, that this instance manages. - u64 memory_size = 0; + std::size_t size{}; /// The memory permissions that are applied to this instance. - MemoryPermission owner_permissions{}; + Memory::MemoryPermission owner_permissions{}; /// The process that this transfer memory instance was created under. - Process* owner_process = nullptr; - - /// Whether or not this transfer memory instance has mapped memory. - bool is_mapped = false; + Process* owner_process{}; - Memory::Memory& memory; + Core::Memory::Memory& memory; }; } // namespace Kernel diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp deleted file mode 100644 index 024c22901..000000000 --- a/src/core/hle/kernel/vm_manager.cpp +++ /dev/null @@ -1,1175 +0,0 @@ -// Copyright 2015 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <cstring> -#include <iterator> -#include <utility> -#include "common/alignment.h" -#include "common/assert.h" -#include "common/logging/log.h" -#include "common/memory_hook.h" -#include "core/core.h" -#include "core/file_sys/program_metadata.h" -#include "core/hle/kernel/errors.h" -#include "core/hle/kernel/process.h" -#include "core/hle/kernel/resource_limit.h" -#include "core/hle/kernel/vm_manager.h" -#include "core/memory.h" - -namespace Kernel { -namespace { -const char* GetMemoryStateName(MemoryState state) { - static constexpr const char* names[] = { - "Unmapped", "Io", - "Normal", "Code", - "CodeData", "Heap", - "Shared", "Unknown1", - "ModuleCode", "ModuleCodeData", - "IpcBuffer0", "Stack", - "ThreadLocal", "TransferMemoryIsolated", - "TransferMemory", "ProcessMemory", - "Inaccessible", "IpcBuffer1", - "IpcBuffer3", "KernelStack", - }; - - return names[ToSvcMemoryState(state)]; -} - -// Checks if a given address range lies within a larger address range. -constexpr bool IsInsideAddressRange(VAddr address, u64 size, VAddr address_range_begin, - VAddr address_range_end) { - const VAddr end_address = address + size - 1; - return address_range_begin <= address && end_address <= address_range_end - 1; -} -} // Anonymous namespace - -bool VirtualMemoryArea::CanBeMergedWith(const VirtualMemoryArea& next) const { - ASSERT(base + size == next.base); - if (permissions != next.permissions || state != next.state || attribute != next.attribute || - type != next.type) { - return false; - } - if ((attribute & MemoryAttribute::DeviceMapped) == MemoryAttribute::DeviceMapped) { - // TODO: Can device mapped memory be merged sanely? - // Not merging it may cause inaccuracies versus hardware when memory layout is queried. - return false; - } - if (type == VMAType::AllocatedMemoryBlock) { - return true; - } - if (type == VMAType::BackingMemory && backing_memory + size != next.backing_memory) { - return false; - } - if (type == VMAType::MMIO && paddr + size != next.paddr) { - return false; - } - return true; -} - -VMManager::VMManager(Core::System& system) : system{system} { - // Default to assuming a 39-bit address space. This way we have a sane - // starting point with executables that don't provide metadata. - Reset(FileSys::ProgramAddressSpaceType::Is39Bit); -} - -VMManager::~VMManager() = default; - -void VMManager::Reset(FileSys::ProgramAddressSpaceType type) { - Clear(); - - InitializeMemoryRegionRanges(type); - - page_table.Resize(address_space_width); - - // Initialize the map with a single free region covering the entire managed space. - VirtualMemoryArea initial_vma; - initial_vma.size = address_space_end; - vma_map.emplace(initial_vma.base, initial_vma); - - UpdatePageTableForVMA(initial_vma); -} - -VMManager::VMAHandle VMManager::FindVMA(VAddr target) const { - if (target >= address_space_end) { - return vma_map.end(); - } else { - return std::prev(vma_map.upper_bound(target)); - } -} - -bool VMManager::IsValidHandle(VMAHandle handle) const { - return handle != vma_map.cend(); -} - -ResultVal<VMManager::VMAHandle> VMManager::MapMemoryBlock(VAddr target, - std::shared_ptr<PhysicalMemory> block, - std::size_t offset, u64 size, - MemoryState state, VMAPermission perm) { - ASSERT(block != nullptr); - ASSERT(offset + size <= block->size()); - - // This is the appropriately sized VMA that will turn into our allocation. - CASCADE_RESULT(VMAIter vma_handle, CarveVMA(target, size)); - VirtualMemoryArea& final_vma = vma_handle->second; - ASSERT(final_vma.size == size); - - final_vma.type = VMAType::AllocatedMemoryBlock; - final_vma.permissions = perm; - final_vma.state = state; - final_vma.backing_block = std::move(block); - final_vma.offset = offset; - UpdatePageTableForVMA(final_vma); - - return MakeResult<VMAHandle>(MergeAdjacent(vma_handle)); -} - -ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* memory, u64 size, - MemoryState state) { - ASSERT(memory != nullptr); - - // This is the appropriately sized VMA that will turn into our allocation. - CASCADE_RESULT(VMAIter vma_handle, CarveVMA(target, size)); - VirtualMemoryArea& final_vma = vma_handle->second; - ASSERT(final_vma.size == size); - - final_vma.type = VMAType::BackingMemory; - final_vma.permissions = VMAPermission::ReadWrite; - final_vma.state = state; - final_vma.backing_memory = memory; - UpdatePageTableForVMA(final_vma); - - return MakeResult<VMAHandle>(MergeAdjacent(vma_handle)); -} - -ResultVal<VAddr> VMManager::FindFreeRegion(u64 size) const { - return FindFreeRegion(GetASLRRegionBaseAddress(), GetASLRRegionEndAddress(), size); -} - -ResultVal<VAddr> VMManager::FindFreeRegion(VAddr begin, VAddr end, u64 size) const { - ASSERT(begin < end); - ASSERT(size <= end - begin); - - const VMAHandle vma_handle = - std::find_if(vma_map.begin(), vma_map.end(), [begin, end, size](const auto& vma) { - if (vma.second.type != VMAType::Free) { - return false; - } - const VAddr vma_base = vma.second.base; - const VAddr vma_end = vma_base + vma.second.size; - const VAddr assumed_base = (begin < vma_base) ? vma_base : begin; - const VAddr used_range = assumed_base + size; - - return vma_base <= assumed_base && assumed_base < used_range && used_range < end && - used_range <= vma_end; - }); - - if (vma_handle == vma_map.cend()) { - // TODO(Subv): Find the correct error code here. - return RESULT_UNKNOWN; - } - - const VAddr target = std::max(begin, vma_handle->second.base); - return MakeResult<VAddr>(target); -} - -ResultVal<VMManager::VMAHandle> VMManager::MapMMIO(VAddr target, PAddr paddr, u64 size, - MemoryState state, - Common::MemoryHookPointer mmio_handler) { - // This is the appropriately sized VMA that will turn into our allocation. - CASCADE_RESULT(VMAIter vma_handle, CarveVMA(target, size)); - VirtualMemoryArea& final_vma = vma_handle->second; - ASSERT(final_vma.size == size); - - final_vma.type = VMAType::MMIO; - final_vma.permissions = VMAPermission::ReadWrite; - final_vma.state = state; - final_vma.paddr = paddr; - final_vma.mmio_handler = std::move(mmio_handler); - UpdatePageTableForVMA(final_vma); - - return MakeResult<VMAHandle>(MergeAdjacent(vma_handle)); -} - -VMManager::VMAIter VMManager::Unmap(VMAIter vma_handle) { - VirtualMemoryArea& vma = vma_handle->second; - vma.type = VMAType::Free; - vma.permissions = VMAPermission::None; - vma.state = MemoryState::Unmapped; - vma.attribute = MemoryAttribute::None; - - vma.backing_block = nullptr; - vma.offset = 0; - vma.backing_memory = nullptr; - vma.paddr = 0; - - UpdatePageTableForVMA(vma); - - return MergeAdjacent(vma_handle); -} - -ResultCode VMManager::UnmapRange(VAddr target, u64 size) { - CASCADE_RESULT(VMAIter vma, CarveVMARange(target, size)); - const VAddr target_end = target + size; - - const VMAIter end = vma_map.end(); - // The comparison against the end of the range must be done using addresses since VMAs can be - // merged during this process, causing invalidation of the iterators. - while (vma != end && vma->second.base < target_end) { - vma = std::next(Unmap(vma)); - } - - ASSERT(FindVMA(target)->second.size >= size); - - return RESULT_SUCCESS; -} - -VMManager::VMAHandle VMManager::Reprotect(VMAHandle vma_handle, VMAPermission new_perms) { - VMAIter iter = StripIterConstness(vma_handle); - - VirtualMemoryArea& vma = iter->second; - vma.permissions = new_perms; - UpdatePageTableForVMA(vma); - - return MergeAdjacent(iter); -} - -ResultCode VMManager::ReprotectRange(VAddr target, u64 size, VMAPermission new_perms) { - CASCADE_RESULT(VMAIter vma, CarveVMARange(target, size)); - const VAddr target_end = target + size; - - const VMAIter end = vma_map.end(); - // The comparison against the end of the range must be done using addresses since VMAs can be - // merged during this process, causing invalidation of the iterators. - while (vma != end && vma->second.base < target_end) { - vma = std::next(StripIterConstness(Reprotect(vma, new_perms))); - } - - return RESULT_SUCCESS; -} - -ResultVal<VAddr> VMManager::SetHeapSize(u64 size) { - if (size > GetHeapRegionSize()) { - return ERR_OUT_OF_MEMORY; - } - - // No need to do any additional work if the heap is already the given size. - if (size == GetCurrentHeapSize()) { - return MakeResult(heap_region_base); - } - - if (heap_memory == nullptr) { - // Initialize heap - heap_memory = std::make_shared<PhysicalMemory>(size); - heap_end = heap_region_base + size; - } else { - UnmapRange(heap_region_base, GetCurrentHeapSize()); - } - - // If necessary, expand backing vector to cover new heap extents in - // the case of allocating. Otherwise, shrink the backing memory, - // if a smaller heap has been requested. - heap_memory->resize(size); - heap_memory->shrink_to_fit(); - RefreshMemoryBlockMappings(heap_memory.get()); - - heap_end = heap_region_base + size; - ASSERT(GetCurrentHeapSize() == heap_memory->size()); - - const auto mapping_result = - MapMemoryBlock(heap_region_base, heap_memory, 0, size, MemoryState::Heap); - if (mapping_result.Failed()) { - return mapping_result.Code(); - } - - return MakeResult<VAddr>(heap_region_base); -} - -ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) { - // Check how much memory we've already mapped. - const auto mapped_size_result = SizeOfAllocatedVMAsInRange(target, size); - if (mapped_size_result.Failed()) { - return mapped_size_result.Code(); - } - - // If we've already mapped the desired amount, return early. - const std::size_t mapped_size = *mapped_size_result; - if (mapped_size == size) { - return RESULT_SUCCESS; - } - - // Check that we can map the memory we want. - const auto res_limit = system.CurrentProcess()->GetResourceLimit(); - const u64 physmem_remaining = res_limit->GetMaxResourceValue(ResourceType::PhysicalMemory) - - res_limit->GetCurrentResourceValue(ResourceType::PhysicalMemory); - if (physmem_remaining < (size - mapped_size)) { - return ERR_RESOURCE_LIMIT_EXCEEDED; - } - - // Keep track of the memory regions we unmap. - std::vector<std::pair<u64, u64>> mapped_regions; - ResultCode result = RESULT_SUCCESS; - - // Iterate, trying to map memory. - { - const auto end_addr = target + size; - const auto last_addr = end_addr - 1; - VAddr cur_addr = target; - - auto iter = FindVMA(target); - ASSERT(iter != vma_map.end()); - - while (true) { - const auto& vma = iter->second; - const auto vma_start = vma.base; - const auto vma_end = vma_start + vma.size; - const auto vma_last = vma_end - 1; - - // Map the memory block - const auto map_size = std::min(end_addr - cur_addr, vma_end - cur_addr); - if (vma.state == MemoryState::Unmapped) { - const auto map_res = - MapMemoryBlock(cur_addr, std::make_shared<PhysicalMemory>(map_size), 0, - map_size, MemoryState::Heap, VMAPermission::ReadWrite); - result = map_res.Code(); - if (result.IsError()) { - break; - } - - mapped_regions.emplace_back(cur_addr, map_size); - } - - // Break once we hit the end of the range. - if (last_addr <= vma_last) { - break; - } - - // Advance to the next block. - cur_addr = vma_end; - iter = FindVMA(cur_addr); - ASSERT(iter != vma_map.end()); - } - } - - // If we failed, unmap memory. - if (result.IsError()) { - for (const auto [unmap_address, unmap_size] : mapped_regions) { - ASSERT_MSG(UnmapRange(unmap_address, unmap_size).IsSuccess(), - "Failed to unmap memory range."); - } - - return result; - } - - // Update amount of mapped physical memory. - physical_memory_mapped += size - mapped_size; - - return RESULT_SUCCESS; -} - -ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) { - // Check how much memory is currently mapped. - const auto mapped_size_result = SizeOfUnmappablePhysicalMemoryInRange(target, size); - if (mapped_size_result.Failed()) { - return mapped_size_result.Code(); - } - - // If we've already unmapped all the memory, return early. - const std::size_t mapped_size = *mapped_size_result; - if (mapped_size == 0) { - return RESULT_SUCCESS; - } - - // Keep track of the memory regions we unmap. - std::vector<std::pair<u64, u64>> unmapped_regions; - ResultCode result = RESULT_SUCCESS; - - // Try to unmap regions. - { - const auto end_addr = target + size; - const auto last_addr = end_addr - 1; - VAddr cur_addr = target; - - auto iter = FindVMA(target); - ASSERT(iter != vma_map.end()); - - while (true) { - const auto& vma = iter->second; - const auto vma_start = vma.base; - const auto vma_end = vma_start + vma.size; - const auto vma_last = vma_end - 1; - - // Unmap the memory block - const auto unmap_size = std::min(end_addr - cur_addr, vma_end - cur_addr); - if (vma.state == MemoryState::Heap) { - result = UnmapRange(cur_addr, unmap_size); - if (result.IsError()) { - break; - } - - unmapped_regions.emplace_back(cur_addr, unmap_size); - } - - // Break once we hit the end of the range. - if (last_addr <= vma_last) { - break; - } - - // Advance to the next block. - cur_addr = vma_end; - iter = FindVMA(cur_addr); - ASSERT(iter != vma_map.end()); - } - } - - // If we failed, re-map regions. - // TODO: Preserve memory contents? - if (result.IsError()) { - for (const auto [map_address, map_size] : unmapped_regions) { - const auto remap_res = - MapMemoryBlock(map_address, std::make_shared<PhysicalMemory>(map_size), 0, map_size, - MemoryState::Heap, VMAPermission::None); - ASSERT_MSG(remap_res.Succeeded(), "Failed to remap a memory block."); - } - - return result; - } - - // Update mapped amount - physical_memory_mapped -= mapped_size; - - return RESULT_SUCCESS; -} - -ResultCode VMManager::MapCodeMemory(VAddr dst_address, VAddr src_address, u64 size) { - constexpr auto ignore_attribute = MemoryAttribute::LockedForIPC | MemoryAttribute::DeviceMapped; - const auto src_check_result = CheckRangeState( - src_address, size, MemoryState::All, MemoryState::Heap, VMAPermission::All, - VMAPermission::ReadWrite, MemoryAttribute::Mask, MemoryAttribute::None, ignore_attribute); - - if (src_check_result.Failed()) { - return src_check_result.Code(); - } - - const auto mirror_result = - MirrorMemory(dst_address, src_address, size, MemoryState::ModuleCode); - if (mirror_result.IsError()) { - return mirror_result; - } - - // Ensure we lock the source memory region. - const auto src_vma_result = CarveVMARange(src_address, size); - if (src_vma_result.Failed()) { - return src_vma_result.Code(); - } - auto src_vma_iter = *src_vma_result; - src_vma_iter->second.attribute = MemoryAttribute::Locked; - Reprotect(src_vma_iter, VMAPermission::Read); - - // The destination memory region is fine as is, however we need to make it read-only. - return ReprotectRange(dst_address, size, VMAPermission::Read); -} - -ResultCode VMManager::UnmapCodeMemory(VAddr dst_address, VAddr src_address, u64 size) { - constexpr auto ignore_attribute = MemoryAttribute::LockedForIPC | MemoryAttribute::DeviceMapped; - const auto src_check_result = CheckRangeState( - src_address, size, MemoryState::All, MemoryState::Heap, VMAPermission::None, - VMAPermission::None, MemoryAttribute::Mask, MemoryAttribute::Locked, ignore_attribute); - - if (src_check_result.Failed()) { - return src_check_result.Code(); - } - - // Yes, the kernel only checks the first page of the region. - const auto dst_check_result = - CheckRangeState(dst_address, Memory::PAGE_SIZE, MemoryState::FlagModule, - MemoryState::FlagModule, VMAPermission::None, VMAPermission::None, - MemoryAttribute::Mask, MemoryAttribute::None, ignore_attribute); - - if (dst_check_result.Failed()) { - return dst_check_result.Code(); - } - - const auto dst_memory_state = std::get<MemoryState>(*dst_check_result); - const auto dst_contiguous_check_result = CheckRangeState( - dst_address, size, MemoryState::All, dst_memory_state, VMAPermission::None, - VMAPermission::None, MemoryAttribute::Mask, MemoryAttribute::None, ignore_attribute); - - if (dst_contiguous_check_result.Failed()) { - return dst_contiguous_check_result.Code(); - } - - const auto unmap_result = UnmapRange(dst_address, size); - if (unmap_result.IsError()) { - return unmap_result; - } - - // With the mirrored portion unmapped, restore the original region's traits. - const auto src_vma_result = CarveVMARange(src_address, size); - if (src_vma_result.Failed()) { - return src_vma_result.Code(); - } - auto src_vma_iter = *src_vma_result; - src_vma_iter->second.state = MemoryState::Heap; - src_vma_iter->second.attribute = MemoryAttribute::None; - Reprotect(src_vma_iter, VMAPermission::ReadWrite); - - if (dst_memory_state == MemoryState::ModuleCode) { - system.InvalidateCpuInstructionCaches(); - } - - return unmap_result; -} - -MemoryInfo VMManager::QueryMemory(VAddr address) const { - const auto vma = FindVMA(address); - MemoryInfo memory_info{}; - - if (IsValidHandle(vma)) { - memory_info.base_address = vma->second.base; - memory_info.attributes = ToSvcMemoryAttribute(vma->second.attribute); - memory_info.permission = static_cast<u32>(vma->second.permissions); - memory_info.size = vma->second.size; - memory_info.state = ToSvcMemoryState(vma->second.state); - } else { - memory_info.base_address = address_space_end; - memory_info.permission = static_cast<u32>(VMAPermission::None); - memory_info.size = 0 - address_space_end; - memory_info.state = static_cast<u32>(MemoryState::Inaccessible); - } - - return memory_info; -} - -ResultCode VMManager::SetMemoryAttribute(VAddr address, u64 size, MemoryAttribute mask, - MemoryAttribute attribute) { - constexpr auto ignore_mask = - MemoryAttribute::Uncached | MemoryAttribute::DeviceMapped | MemoryAttribute::Locked; - constexpr auto attribute_mask = ~ignore_mask; - - const auto result = CheckRangeState( - address, size, MemoryState::FlagUncached, MemoryState::FlagUncached, VMAPermission::None, - VMAPermission::None, attribute_mask, MemoryAttribute::None, ignore_mask); - - if (result.Failed()) { - return result.Code(); - } - - const auto [prev_state, prev_permissions, prev_attributes] = *result; - const auto new_attribute = (prev_attributes & ~mask) | (mask & attribute); - - const auto carve_result = CarveVMARange(address, size); - if (carve_result.Failed()) { - return carve_result.Code(); - } - - auto vma_iter = *carve_result; - vma_iter->second.attribute = new_attribute; - - MergeAdjacent(vma_iter); - return RESULT_SUCCESS; -} - -ResultCode VMManager::MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, MemoryState state) { - const auto vma = FindVMA(src_addr); - - ASSERT_MSG(vma != vma_map.end(), "Invalid memory address"); - ASSERT_MSG(vma->second.backing_block, "Backing block doesn't exist for address"); - - // The returned VMA might be a bigger one encompassing the desired address. - const auto vma_offset = src_addr - vma->first; - ASSERT_MSG(vma_offset + size <= vma->second.size, - "Shared memory exceeds bounds of mapped block"); - - const std::shared_ptr<PhysicalMemory>& backing_block = vma->second.backing_block; - const std::size_t backing_block_offset = vma->second.offset + vma_offset; - - CASCADE_RESULT(auto new_vma, - MapMemoryBlock(dst_addr, backing_block, backing_block_offset, size, state)); - // Protect mirror with permissions from old region - Reprotect(new_vma, vma->second.permissions); - // Remove permissions from old region - ReprotectRange(src_addr, size, VMAPermission::None); - - return RESULT_SUCCESS; -} - -void VMManager::RefreshMemoryBlockMappings(const PhysicalMemory* block) { - // If this ever proves to have a noticeable performance impact, allow users of the function to - // specify a specific range of addresses to limit the scan to. - for (const auto& p : vma_map) { - const VirtualMemoryArea& vma = p.second; - if (block == vma.backing_block.get()) { - UpdatePageTableForVMA(vma); - } - } -} - -void VMManager::LogLayout() const { - for (const auto& p : vma_map) { - const VirtualMemoryArea& vma = p.second; - LOG_DEBUG(Kernel, "{:016X} - {:016X} size: {:016X} {}{}{} {}", vma.base, - vma.base + vma.size, vma.size, - (u8)vma.permissions & (u8)VMAPermission::Read ? 'R' : '-', - (u8)vma.permissions & (u8)VMAPermission::Write ? 'W' : '-', - (u8)vma.permissions & (u8)VMAPermission::Execute ? 'X' : '-', - GetMemoryStateName(vma.state)); - } -} - -VMManager::VMAIter VMManager::StripIterConstness(const VMAHandle& iter) { - // This uses a neat C++ trick to convert a const_iterator to a regular iterator, given - // non-const access to its container. - return vma_map.erase(iter, iter); // Erases an empty range of elements -} - -ResultVal<VMManager::VMAIter> VMManager::CarveVMA(VAddr base, u64 size) { - ASSERT_MSG((size & Memory::PAGE_MASK) == 0, "non-page aligned size: 0x{:016X}", size); - ASSERT_MSG((base & Memory::PAGE_MASK) == 0, "non-page aligned base: 0x{:016X}", base); - - VMAIter vma_handle = StripIterConstness(FindVMA(base)); - if (vma_handle == vma_map.end()) { - // Target address is outside the range managed by the kernel - return ERR_INVALID_ADDRESS; - } - - const VirtualMemoryArea& vma = vma_handle->second; - if (vma.type != VMAType::Free) { - // Region is already allocated - return ERR_INVALID_ADDRESS_STATE; - } - - const VAddr start_in_vma = base - vma.base; - const VAddr end_in_vma = start_in_vma + size; - - if (end_in_vma > vma.size) { - // Requested allocation doesn't fit inside VMA - return ERR_INVALID_ADDRESS_STATE; - } - - if (end_in_vma != vma.size) { - // Split VMA at the end of the allocated region - SplitVMA(vma_handle, end_in_vma); - } - if (start_in_vma != 0) { - // Split VMA at the start of the allocated region - vma_handle = SplitVMA(vma_handle, start_in_vma); - } - - return MakeResult<VMAIter>(vma_handle); -} - -ResultVal<VMManager::VMAIter> VMManager::CarveVMARange(VAddr target, u64 size) { - ASSERT_MSG((size & Memory::PAGE_MASK) == 0, "non-page aligned size: 0x{:016X}", size); - ASSERT_MSG((target & Memory::PAGE_MASK) == 0, "non-page aligned base: 0x{:016X}", target); - - const VAddr target_end = target + size; - ASSERT(target_end >= target); - ASSERT(target_end <= address_space_end); - ASSERT(size > 0); - - VMAIter begin_vma = StripIterConstness(FindVMA(target)); - const VMAIter i_end = vma_map.lower_bound(target_end); - if (std::any_of(begin_vma, i_end, - [](const auto& entry) { return entry.second.type == VMAType::Free; })) { - return ERR_INVALID_ADDRESS_STATE; - } - - if (target != begin_vma->second.base) { - begin_vma = SplitVMA(begin_vma, target - begin_vma->second.base); - } - - VMAIter end_vma = StripIterConstness(FindVMA(target_end)); - if (end_vma != vma_map.end() && target_end != end_vma->second.base) { - end_vma = SplitVMA(end_vma, target_end - end_vma->second.base); - } - - return MakeResult<VMAIter>(begin_vma); -} - -VMManager::VMAIter VMManager::SplitVMA(VMAIter vma_handle, u64 offset_in_vma) { - VirtualMemoryArea& old_vma = vma_handle->second; - VirtualMemoryArea new_vma = old_vma; // Make a copy of the VMA - - // For now, don't allow no-op VMA splits (trying to split at a boundary) because it's probably - // a bug. This restriction might be removed later. - ASSERT(offset_in_vma < old_vma.size); - ASSERT(offset_in_vma > 0); - - old_vma.size = offset_in_vma; - new_vma.base += offset_in_vma; - new_vma.size -= offset_in_vma; - - switch (new_vma.type) { - case VMAType::Free: - break; - case VMAType::AllocatedMemoryBlock: - new_vma.offset += offset_in_vma; - break; - case VMAType::BackingMemory: - new_vma.backing_memory += offset_in_vma; - break; - case VMAType::MMIO: - new_vma.paddr += offset_in_vma; - break; - } - - ASSERT(old_vma.CanBeMergedWith(new_vma)); - - return vma_map.emplace_hint(std::next(vma_handle), new_vma.base, new_vma); -} - -VMManager::VMAIter VMManager::MergeAdjacent(VMAIter iter) { - const VMAIter next_vma = std::next(iter); - if (next_vma != vma_map.end() && iter->second.CanBeMergedWith(next_vma->second)) { - MergeAdjacentVMA(iter->second, next_vma->second); - vma_map.erase(next_vma); - } - - if (iter != vma_map.begin()) { - VMAIter prev_vma = std::prev(iter); - if (prev_vma->second.CanBeMergedWith(iter->second)) { - MergeAdjacentVMA(prev_vma->second, iter->second); - vma_map.erase(iter); - iter = prev_vma; - } - } - - return iter; -} - -void VMManager::MergeAdjacentVMA(VirtualMemoryArea& left, const VirtualMemoryArea& right) { - ASSERT(left.CanBeMergedWith(right)); - - // Always merge allocated memory blocks, even when they don't share the same backing block. - if (left.type == VMAType::AllocatedMemoryBlock && - (left.backing_block != right.backing_block || left.offset + left.size != right.offset)) { - - // Check if we can save work. - if (left.offset == 0 && left.size == left.backing_block->size()) { - // Fast case: left is an entire backing block. - left.backing_block->resize(left.size + right.size); - std::memcpy(left.backing_block->data() + left.size, - right.backing_block->data() + right.offset, right.size); - } else { - // Slow case: make a new memory block for left and right. - auto new_memory = std::make_shared<PhysicalMemory>(); - new_memory->resize(left.size + right.size); - std::memcpy(new_memory->data(), left.backing_block->data() + left.offset, left.size); - std::memcpy(new_memory->data() + left.size, right.backing_block->data() + right.offset, - right.size); - - left.backing_block = std::move(new_memory); - left.offset = 0; - } - - // Page table update is needed, because backing memory changed. - left.size += right.size; - UpdatePageTableForVMA(left); - } else { - // Just update the size. - left.size += right.size; - } -} - -void VMManager::UpdatePageTableForVMA(const VirtualMemoryArea& vma) { - auto& memory = system.Memory(); - - switch (vma.type) { - case VMAType::Free: - memory.UnmapRegion(page_table, vma.base, vma.size); - break; - case VMAType::AllocatedMemoryBlock: - memory.MapMemoryRegion(page_table, vma.base, vma.size, *vma.backing_block, vma.offset); - break; - case VMAType::BackingMemory: - memory.MapMemoryRegion(page_table, vma.base, vma.size, vma.backing_memory); - break; - case VMAType::MMIO: - memory.MapIoRegion(page_table, vma.base, vma.size, vma.mmio_handler); - break; - } -} - -void VMManager::InitializeMemoryRegionRanges(FileSys::ProgramAddressSpaceType type) { - u64 map_region_size = 0; - u64 heap_region_size = 0; - u64 stack_region_size = 0; - u64 tls_io_region_size = 0; - - u64 stack_and_tls_io_end = 0; - - switch (type) { - case FileSys::ProgramAddressSpaceType::Is32Bit: - case FileSys::ProgramAddressSpaceType::Is32BitNoMap: - address_space_width = 32; - code_region_base = 0x200000; - code_region_end = code_region_base + 0x3FE00000; - aslr_region_base = 0x200000; - aslr_region_end = aslr_region_base + 0xFFE00000; - if (type == FileSys::ProgramAddressSpaceType::Is32Bit) { - map_region_size = 0x40000000; - heap_region_size = 0x40000000; - } else { - map_region_size = 0; - heap_region_size = 0x80000000; - } - stack_and_tls_io_end = 0x40000000; - break; - case FileSys::ProgramAddressSpaceType::Is36Bit: - address_space_width = 36; - code_region_base = 0x8000000; - code_region_end = code_region_base + 0x78000000; - aslr_region_base = 0x8000000; - aslr_region_end = aslr_region_base + 0xFF8000000; - map_region_size = 0x180000000; - heap_region_size = 0x180000000; - stack_and_tls_io_end = 0x80000000; - break; - case FileSys::ProgramAddressSpaceType::Is39Bit: - address_space_width = 39; - code_region_base = 0x8000000; - code_region_end = code_region_base + 0x80000000; - aslr_region_base = 0x8000000; - aslr_region_end = aslr_region_base + 0x7FF8000000; - map_region_size = 0x1000000000; - heap_region_size = 0x180000000; - stack_region_size = 0x80000000; - tls_io_region_size = 0x1000000000; - break; - default: - UNREACHABLE_MSG("Invalid address space type specified: {}", static_cast<u32>(type)); - return; - } - - const u64 stack_and_tls_io_begin = aslr_region_base; - - address_space_base = 0; - address_space_end = 1ULL << address_space_width; - - map_region_base = code_region_end; - map_region_end = map_region_base + map_region_size; - - heap_region_base = map_region_end; - heap_region_end = heap_region_base + heap_region_size; - heap_end = heap_region_base; - - stack_region_base = heap_region_end; - stack_region_end = stack_region_base + stack_region_size; - - tls_io_region_base = stack_region_end; - tls_io_region_end = tls_io_region_base + tls_io_region_size; - - if (stack_region_size == 0) { - stack_region_base = stack_and_tls_io_begin; - stack_region_end = stack_and_tls_io_end; - } - - if (tls_io_region_size == 0) { - tls_io_region_base = stack_and_tls_io_begin; - tls_io_region_end = stack_and_tls_io_end; - } -} - -void VMManager::Clear() { - ClearVMAMap(); - ClearPageTable(); -} - -void VMManager::ClearVMAMap() { - vma_map.clear(); -} - -void VMManager::ClearPageTable() { - std::fill(page_table.pointers.begin(), page_table.pointers.end(), nullptr); - page_table.special_regions.clear(); - std::fill(page_table.attributes.begin(), page_table.attributes.end(), - Common::PageType::Unmapped); -} - -VMManager::CheckResults VMManager::CheckRangeState(VAddr address, u64 size, MemoryState state_mask, - MemoryState state, VMAPermission permission_mask, - VMAPermission permissions, - MemoryAttribute attribute_mask, - MemoryAttribute attribute, - MemoryAttribute ignore_mask) const { - auto iter = FindVMA(address); - - // If we don't have a valid VMA handle at this point, then it means this is - // being called with an address outside of the address space, which is definitely - // indicative of a bug, as this function only operates on mapped memory regions. - DEBUG_ASSERT(IsValidHandle(iter)); - - const VAddr end_address = address + size - 1; - const MemoryAttribute initial_attributes = iter->second.attribute; - const VMAPermission initial_permissions = iter->second.permissions; - const MemoryState initial_state = iter->second.state; - - while (true) { - // The iterator should be valid throughout the traversal. Hitting the end of - // the mapped VMA regions is unquestionably indicative of a bug. - DEBUG_ASSERT(IsValidHandle(iter)); - - const auto& vma = iter->second; - - if (vma.state != initial_state) { - return ERR_INVALID_ADDRESS_STATE; - } - - if ((vma.state & state_mask) != state) { - return ERR_INVALID_ADDRESS_STATE; - } - - if (vma.permissions != initial_permissions) { - return ERR_INVALID_ADDRESS_STATE; - } - - if ((vma.permissions & permission_mask) != permissions) { - return ERR_INVALID_ADDRESS_STATE; - } - - if ((vma.attribute | ignore_mask) != (initial_attributes | ignore_mask)) { - return ERR_INVALID_ADDRESS_STATE; - } - - if ((vma.attribute & attribute_mask) != attribute) { - return ERR_INVALID_ADDRESS_STATE; - } - - if (end_address <= vma.EndAddress()) { - break; - } - - ++iter; - } - - return MakeResult( - std::make_tuple(initial_state, initial_permissions, initial_attributes & ~ignore_mask)); -} - -ResultVal<std::size_t> VMManager::SizeOfAllocatedVMAsInRange(VAddr address, - std::size_t size) const { - const VAddr end_addr = address + size; - const VAddr last_addr = end_addr - 1; - std::size_t mapped_size = 0; - - VAddr cur_addr = address; - auto iter = FindVMA(cur_addr); - ASSERT(iter != vma_map.end()); - - while (true) { - const auto& vma = iter->second; - const VAddr vma_start = vma.base; - const VAddr vma_end = vma_start + vma.size; - const VAddr vma_last = vma_end - 1; - - // Add size if relevant. - if (vma.state != MemoryState::Unmapped) { - mapped_size += std::min(end_addr - cur_addr, vma_end - cur_addr); - } - - // Break once we hit the end of the range. - if (last_addr <= vma_last) { - break; - } - - // Advance to the next block. - cur_addr = vma_end; - iter = std::next(iter); - ASSERT(iter != vma_map.end()); - } - - return MakeResult(mapped_size); -} - -ResultVal<std::size_t> VMManager::SizeOfUnmappablePhysicalMemoryInRange(VAddr address, - std::size_t size) const { - const VAddr end_addr = address + size; - const VAddr last_addr = end_addr - 1; - std::size_t mapped_size = 0; - - VAddr cur_addr = address; - auto iter = FindVMA(cur_addr); - ASSERT(iter != vma_map.end()); - - while (true) { - const auto& vma = iter->second; - const auto vma_start = vma.base; - const auto vma_end = vma_start + vma.size; - const auto vma_last = vma_end - 1; - const auto state = vma.state; - const auto attr = vma.attribute; - - // Memory within region must be free or mapped heap. - if (!((state == MemoryState::Heap && attr == MemoryAttribute::None) || - (state == MemoryState::Unmapped))) { - return ERR_INVALID_ADDRESS_STATE; - } - - // Add size if relevant. - if (state != MemoryState::Unmapped) { - mapped_size += std::min(end_addr - cur_addr, vma_end - cur_addr); - } - - // Break once we hit the end of the range. - if (last_addr <= vma_last) { - break; - } - - // Advance to the next block. - cur_addr = vma_end; - iter = std::next(iter); - ASSERT(iter != vma_map.end()); - } - - return MakeResult(mapped_size); -} - -u64 VMManager::GetTotalPhysicalMemoryAvailable() const { - LOG_WARNING(Kernel, "(STUBBED) called"); - return 0xF8000000; -} - -VAddr VMManager::GetAddressSpaceBaseAddress() const { - return address_space_base; -} - -VAddr VMManager::GetAddressSpaceEndAddress() const { - return address_space_end; -} - -u64 VMManager::GetAddressSpaceSize() const { - return address_space_end - address_space_base; -} - -u64 VMManager::GetAddressSpaceWidth() const { - return address_space_width; -} - -bool VMManager::IsWithinAddressSpace(VAddr address, u64 size) const { - return IsInsideAddressRange(address, size, GetAddressSpaceBaseAddress(), - GetAddressSpaceEndAddress()); -} - -VAddr VMManager::GetASLRRegionBaseAddress() const { - return aslr_region_base; -} - -VAddr VMManager::GetASLRRegionEndAddress() const { - return aslr_region_end; -} - -u64 VMManager::GetASLRRegionSize() const { - return aslr_region_end - aslr_region_base; -} - -bool VMManager::IsWithinASLRRegion(VAddr begin, u64 size) const { - const VAddr range_end = begin + size; - const VAddr aslr_start = GetASLRRegionBaseAddress(); - const VAddr aslr_end = GetASLRRegionEndAddress(); - - if (aslr_start > begin || begin > range_end || range_end - 1 > aslr_end - 1) { - return false; - } - - if (range_end > heap_region_base && heap_region_end > begin) { - return false; - } - - if (range_end > map_region_base && map_region_end > begin) { - return false; - } - - return true; -} - -VAddr VMManager::GetCodeRegionBaseAddress() const { - return code_region_base; -} - -VAddr VMManager::GetCodeRegionEndAddress() const { - return code_region_end; -} - -u64 VMManager::GetCodeRegionSize() const { - return code_region_end - code_region_base; -} - -bool VMManager::IsWithinCodeRegion(VAddr address, u64 size) const { - return IsInsideAddressRange(address, size, GetCodeRegionBaseAddress(), - GetCodeRegionEndAddress()); -} - -VAddr VMManager::GetHeapRegionBaseAddress() const { - return heap_region_base; -} - -VAddr VMManager::GetHeapRegionEndAddress() const { - return heap_region_end; -} - -u64 VMManager::GetHeapRegionSize() const { - return heap_region_end - heap_region_base; -} - -u64 VMManager::GetCurrentHeapSize() const { - return heap_end - heap_region_base; -} - -bool VMManager::IsWithinHeapRegion(VAddr address, u64 size) const { - return IsInsideAddressRange(address, size, GetHeapRegionBaseAddress(), - GetHeapRegionEndAddress()); -} - -VAddr VMManager::GetMapRegionBaseAddress() const { - return map_region_base; -} - -VAddr VMManager::GetMapRegionEndAddress() const { - return map_region_end; -} - -u64 VMManager::GetMapRegionSize() const { - return map_region_end - map_region_base; -} - -bool VMManager::IsWithinMapRegion(VAddr address, u64 size) const { - return IsInsideAddressRange(address, size, GetMapRegionBaseAddress(), GetMapRegionEndAddress()); -} - -VAddr VMManager::GetStackRegionBaseAddress() const { - return stack_region_base; -} - -VAddr VMManager::GetStackRegionEndAddress() const { - return stack_region_end; -} - -u64 VMManager::GetStackRegionSize() const { - return stack_region_end - stack_region_base; -} - -bool VMManager::IsWithinStackRegion(VAddr address, u64 size) const { - return IsInsideAddressRange(address, size, GetStackRegionBaseAddress(), - GetStackRegionEndAddress()); -} - -VAddr VMManager::GetTLSIORegionBaseAddress() const { - return tls_io_region_base; -} - -VAddr VMManager::GetTLSIORegionEndAddress() const { - return tls_io_region_end; -} - -u64 VMManager::GetTLSIORegionSize() const { - return tls_io_region_end - tls_io_region_base; -} - -bool VMManager::IsWithinTLSIORegion(VAddr address, u64 size) const { - return IsInsideAddressRange(address, size, GetTLSIORegionBaseAddress(), - GetTLSIORegionEndAddress()); -} - -} // namespace Kernel diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h deleted file mode 100644 index 90b4b006a..000000000 --- a/src/core/hle/kernel/vm_manager.h +++ /dev/null @@ -1,796 +0,0 @@ -// Copyright 2015 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <map> -#include <memory> -#include <tuple> -#include <vector> -#include "common/common_types.h" -#include "common/memory_hook.h" -#include "common/page_table.h" -#include "core/hle/kernel/physical_memory.h" -#include "core/hle/result.h" -#include "core/memory.h" - -namespace Core { -class System; -} - -namespace FileSys { -enum class ProgramAddressSpaceType : u8; -} - -namespace Kernel { - -enum class VMAType : u8 { - /// VMA represents an unmapped region of the address space. - Free, - /// VMA is backed by a ref-counted allocate memory block. - AllocatedMemoryBlock, - /// VMA is backed by a raw, unmanaged pointer. - BackingMemory, - /// VMA is mapped to MMIO registers at a fixed PAddr. - MMIO, - // TODO(yuriks): Implement MemoryAlias to support MAP/UNMAP -}; - -/// Permissions for mapped memory blocks -enum class VMAPermission : u8 { - None = 0, - Read = 1, - Write = 2, - Execute = 4, - - ReadWrite = Read | Write, - ReadExecute = Read | Execute, - WriteExecute = Write | Execute, - ReadWriteExecute = Read | Write | Execute, - - // Used as a wildcard when checking permissions across memory ranges - All = 0xFF, -}; - -constexpr VMAPermission operator|(VMAPermission lhs, VMAPermission rhs) { - return static_cast<VMAPermission>(u32(lhs) | u32(rhs)); -} - -constexpr VMAPermission operator&(VMAPermission lhs, VMAPermission rhs) { - return static_cast<VMAPermission>(u32(lhs) & u32(rhs)); -} - -constexpr VMAPermission operator^(VMAPermission lhs, VMAPermission rhs) { - return static_cast<VMAPermission>(u32(lhs) ^ u32(rhs)); -} - -constexpr VMAPermission operator~(VMAPermission permission) { - return static_cast<VMAPermission>(~u32(permission)); -} - -constexpr VMAPermission& operator|=(VMAPermission& lhs, VMAPermission rhs) { - lhs = lhs | rhs; - return lhs; -} - -constexpr VMAPermission& operator&=(VMAPermission& lhs, VMAPermission rhs) { - lhs = lhs & rhs; - return lhs; -} - -constexpr VMAPermission& operator^=(VMAPermission& lhs, VMAPermission rhs) { - lhs = lhs ^ rhs; - return lhs; -} - -/// Attribute flags that can be applied to a VMA -enum class MemoryAttribute : u32 { - Mask = 0xFF, - - /// No particular qualities - None = 0, - /// Memory locked/borrowed for use. e.g. This would be used by transfer memory. - Locked = 1, - /// Memory locked for use by IPC-related internals. - LockedForIPC = 2, - /// Mapped as part of the device address space. - DeviceMapped = 4, - /// Uncached memory - Uncached = 8, - - IpcAndDeviceMapped = LockedForIPC | DeviceMapped, -}; - -constexpr MemoryAttribute operator|(MemoryAttribute lhs, MemoryAttribute rhs) { - return static_cast<MemoryAttribute>(u32(lhs) | u32(rhs)); -} - -constexpr MemoryAttribute operator&(MemoryAttribute lhs, MemoryAttribute rhs) { - return static_cast<MemoryAttribute>(u32(lhs) & u32(rhs)); -} - -constexpr MemoryAttribute operator^(MemoryAttribute lhs, MemoryAttribute rhs) { - return static_cast<MemoryAttribute>(u32(lhs) ^ u32(rhs)); -} - -constexpr MemoryAttribute operator~(MemoryAttribute attribute) { - return static_cast<MemoryAttribute>(~u32(attribute)); -} - -constexpr MemoryAttribute& operator|=(MemoryAttribute& lhs, MemoryAttribute rhs) { - lhs = lhs | rhs; - return lhs; -} - -constexpr MemoryAttribute& operator&=(MemoryAttribute& lhs, MemoryAttribute rhs) { - lhs = lhs & rhs; - return lhs; -} - -constexpr MemoryAttribute& operator^=(MemoryAttribute& lhs, MemoryAttribute rhs) { - lhs = lhs ^ rhs; - return lhs; -} - -constexpr u32 ToSvcMemoryAttribute(MemoryAttribute attribute) { - return static_cast<u32>(attribute & MemoryAttribute::Mask); -} - -// clang-format off -/// Represents memory states and any relevant flags, as used by the kernel. -/// svcQueryMemory interprets these by masking away all but the first eight -/// bits when storing memory state into a MemoryInfo instance. -enum class MemoryState : u32 { - Mask = 0xFF, - FlagProtect = 1U << 8, - FlagDebug = 1U << 9, - FlagIPC0 = 1U << 10, - FlagIPC3 = 1U << 11, - FlagIPC1 = 1U << 12, - FlagMapped = 1U << 13, - FlagCode = 1U << 14, - FlagAlias = 1U << 15, - FlagModule = 1U << 16, - FlagTransfer = 1U << 17, - FlagQueryPhysicalAddressAllowed = 1U << 18, - FlagSharedDevice = 1U << 19, - FlagSharedDeviceAligned = 1U << 20, - FlagIPCBuffer = 1U << 21, - FlagMemoryPoolAllocated = 1U << 22, - FlagMapProcess = 1U << 23, - FlagUncached = 1U << 24, - FlagCodeMemory = 1U << 25, - - // Wildcard used in range checking to indicate all states. - All = 0xFFFFFFFF, - - // Convenience flag sets to reduce repetition - IPCFlags = FlagIPC0 | FlagIPC3 | FlagIPC1, - - CodeFlags = FlagDebug | IPCFlags | FlagMapped | FlagCode | FlagQueryPhysicalAddressAllowed | - FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated, - - DataFlags = FlagProtect | IPCFlags | FlagMapped | FlagAlias | FlagTransfer | - FlagQueryPhysicalAddressAllowed | FlagSharedDevice | FlagSharedDeviceAligned | - FlagMemoryPoolAllocated | FlagIPCBuffer | FlagUncached, - - Unmapped = 0x00, - Io = 0x01 | FlagMapped, - Normal = 0x02 | FlagMapped | FlagQueryPhysicalAddressAllowed, - Code = 0x03 | CodeFlags | FlagMapProcess, - CodeData = 0x04 | DataFlags | FlagMapProcess | FlagCodeMemory, - Heap = 0x05 | DataFlags | FlagCodeMemory, - Shared = 0x06 | FlagMapped | FlagMemoryPoolAllocated, - ModuleCode = 0x08 | CodeFlags | FlagModule | FlagMapProcess, - ModuleCodeData = 0x09 | DataFlags | FlagModule | FlagMapProcess | FlagCodeMemory, - - IpcBuffer0 = 0x0A | FlagMapped | FlagQueryPhysicalAddressAllowed | FlagMemoryPoolAllocated | - IPCFlags | FlagSharedDevice | FlagSharedDeviceAligned, - - Stack = 0x0B | FlagMapped | IPCFlags | FlagQueryPhysicalAddressAllowed | - FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated, - - ThreadLocal = 0x0C | FlagMapped | FlagMemoryPoolAllocated, - - TransferMemoryIsolated = 0x0D | IPCFlags | FlagMapped | FlagQueryPhysicalAddressAllowed | - FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated | - FlagUncached, - - TransferMemory = 0x0E | FlagIPC3 | FlagIPC1 | FlagMapped | FlagQueryPhysicalAddressAllowed | - FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated, - - ProcessMemory = 0x0F | FlagIPC3 | FlagIPC1 | FlagMapped | FlagMemoryPoolAllocated, - - // Used to signify an inaccessible or invalid memory region with memory queries - Inaccessible = 0x10, - - IpcBuffer1 = 0x11 | FlagIPC3 | FlagIPC1 | FlagMapped | FlagQueryPhysicalAddressAllowed | - FlagSharedDevice | FlagSharedDeviceAligned | FlagMemoryPoolAllocated, - - IpcBuffer3 = 0x12 | FlagIPC3 | FlagMapped | FlagQueryPhysicalAddressAllowed | - FlagSharedDeviceAligned | FlagMemoryPoolAllocated, - - KernelStack = 0x13 | FlagMapped, -}; -// clang-format on - -constexpr MemoryState operator|(MemoryState lhs, MemoryState rhs) { - return static_cast<MemoryState>(u32(lhs) | u32(rhs)); -} - -constexpr MemoryState operator&(MemoryState lhs, MemoryState rhs) { - return static_cast<MemoryState>(u32(lhs) & u32(rhs)); -} - -constexpr MemoryState operator^(MemoryState lhs, MemoryState rhs) { - return static_cast<MemoryState>(u32(lhs) ^ u32(rhs)); -} - -constexpr MemoryState operator~(MemoryState lhs) { - return static_cast<MemoryState>(~u32(lhs)); -} - -constexpr MemoryState& operator|=(MemoryState& lhs, MemoryState rhs) { - lhs = lhs | rhs; - return lhs; -} - -constexpr MemoryState& operator&=(MemoryState& lhs, MemoryState rhs) { - lhs = lhs & rhs; - return lhs; -} - -constexpr MemoryState& operator^=(MemoryState& lhs, MemoryState rhs) { - lhs = lhs ^ rhs; - return lhs; -} - -constexpr u32 ToSvcMemoryState(MemoryState state) { - return static_cast<u32>(state & MemoryState::Mask); -} - -struct MemoryInfo { - u64 base_address; - u64 size; - u32 state; - u32 attributes; - u32 permission; - u32 ipc_ref_count; - u32 device_ref_count; -}; -static_assert(sizeof(MemoryInfo) == 0x28, "MemoryInfo has incorrect size."); - -struct PageInfo { - u32 flags; -}; - -/** - * Represents a VMA in an address space. A VMA is a contiguous region of virtual addressing space - * with homogeneous attributes across its extents. In this particular implementation each VMA is - * also backed by a single host memory allocation. - */ -struct VirtualMemoryArea { - /// Gets the starting (base) address of this VMA. - VAddr StartAddress() const { - return base; - } - - /// Gets the ending address of this VMA. - VAddr EndAddress() const { - return base + size - 1; - } - - /// Virtual base address of the region. - VAddr base = 0; - /// Size of the region. - u64 size = 0; - - VMAType type = VMAType::Free; - VMAPermission permissions = VMAPermission::None; - MemoryState state = MemoryState::Unmapped; - MemoryAttribute attribute = MemoryAttribute::None; - - // Settings for type = AllocatedMemoryBlock - /// Memory block backing this VMA. - std::shared_ptr<PhysicalMemory> backing_block = nullptr; - /// Offset into the backing_memory the mapping starts from. - std::size_t offset = 0; - - // Settings for type = BackingMemory - /// Pointer backing this VMA. It will not be destroyed or freed when the VMA is removed. - u8* backing_memory = nullptr; - - // Settings for type = MMIO - /// Physical address of the register area this VMA maps to. - PAddr paddr = 0; - Common::MemoryHookPointer mmio_handler = nullptr; - - /// Tests if this area can be merged to the right with `next`. - bool CanBeMergedWith(const VirtualMemoryArea& next) const; -}; - -/** - * Manages a process' virtual addressing space. This class maintains a list of allocated and free - * regions in the address space, along with their attributes, and allows kernel clients to - * manipulate it, adjusting the page table to match. - * - * This is similar in idea and purpose to the VM manager present in operating system kernels, with - * the main difference being that it doesn't have to support swapping or memory mapping of files. - * The implementation is also simplified by not having to allocate page frames. See these articles - * about the Linux kernel for an explantion of the concept and implementation: - * - http://duartes.org/gustavo/blog/post/how-the-kernel-manages-your-memory/ - * - http://duartes.org/gustavo/blog/post/page-cache-the-affair-between-memory-and-files/ - */ -class VMManager final { - using VMAMap = std::map<VAddr, VirtualMemoryArea>; - -public: - using VMAHandle = VMAMap::const_iterator; - - explicit VMManager(Core::System& system); - ~VMManager(); - - /// Clears the address space map, re-initializing with a single free area. - void Reset(FileSys::ProgramAddressSpaceType type); - - /// Finds the VMA in which the given address is included in, or `vma_map.end()`. - VMAHandle FindVMA(VAddr target) const; - - /// Indicates whether or not the given handle is within the VMA map. - bool IsValidHandle(VMAHandle handle) const; - - // TODO(yuriks): Should these functions actually return the handle? - - /** - * Maps part of a ref-counted block of memory at a given address. - * - * @param target The guest address to start the mapping at. - * @param block The block to be mapped. - * @param offset Offset into `block` to map from. - * @param size Size of the mapping. - * @param state MemoryState tag to attach to the VMA. - */ - ResultVal<VMAHandle> MapMemoryBlock(VAddr target, std::shared_ptr<PhysicalMemory> block, - std::size_t offset, u64 size, MemoryState state, - VMAPermission perm = VMAPermission::ReadWrite); - - /** - * Maps an unmanaged host memory pointer at a given address. - * - * @param target The guest address to start the mapping at. - * @param memory The memory to be mapped. - * @param size Size of the mapping. - * @param state MemoryState tag to attach to the VMA. - */ - ResultVal<VMAHandle> MapBackingMemory(VAddr target, u8* memory, u64 size, MemoryState state); - - /** - * Finds the first free memory region of the given size within - * the user-addressable ASLR memory region. - * - * @param size The size of the desired region in bytes. - * - * @returns If successful, the base address of the free region with - * the given size. - */ - ResultVal<VAddr> FindFreeRegion(u64 size) const; - - /** - * Finds the first free address range that can hold a region of the desired size - * - * @param begin The starting address of the range. - * This is treated as an inclusive beginning address. - * - * @param end The ending address of the range. - * This is treated as an exclusive ending address. - * - * @param size The size of the free region to attempt to locate, - * in bytes. - * - * @returns If successful, the base address of the free region with - * the given size. - * - * @returns If unsuccessful, a result containing an error code. - * - * @pre The starting address must be less than the ending address. - * @pre The size must not exceed the address range itself. - */ - ResultVal<VAddr> FindFreeRegion(VAddr begin, VAddr end, u64 size) const; - - /** - * Maps a memory-mapped IO region at a given address. - * - * @param target The guest address to start the mapping at. - * @param paddr The physical address where the registers are present. - * @param size Size of the mapping. - * @param state MemoryState tag to attach to the VMA. - * @param mmio_handler The handler that will implement read and write for this MMIO region. - */ - ResultVal<VMAHandle> MapMMIO(VAddr target, PAddr paddr, u64 size, MemoryState state, - Common::MemoryHookPointer mmio_handler); - - /// Unmaps a range of addresses, splitting VMAs as necessary. - ResultCode UnmapRange(VAddr target, u64 size); - - /// Changes the permissions of the given VMA. - VMAHandle Reprotect(VMAHandle vma, VMAPermission new_perms); - - /// Changes the permissions of a range of addresses, splitting VMAs as necessary. - ResultCode ReprotectRange(VAddr target, u64 size, VMAPermission new_perms); - - ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, MemoryState state); - - /// Attempts to allocate a heap with the given size. - /// - /// @param size The size of the heap to allocate in bytes. - /// - /// @note If a heap is currently allocated, and this is called - /// with a size that is equal to the size of the current heap, - /// then this function will do nothing and return the current - /// heap's starting address, as there's no need to perform - /// any additional heap allocation work. - /// - /// @note If a heap is currently allocated, and this is called - /// with a size less than the current heap's size, then - /// this function will attempt to shrink the heap. - /// - /// @note If a heap is currently allocated, and this is called - /// with a size larger than the current heap's size, then - /// this function will attempt to extend the size of the heap. - /// - /// @returns A result indicating either success or failure. - /// <p> - /// If successful, this function will return a result - /// containing the starting address to the allocated heap. - /// <p> - /// If unsuccessful, this function will return a result - /// containing an error code. - /// - /// @pre The given size must lie within the allowable heap - /// memory region managed by this VMManager instance. - /// Failure to abide by this will result in ERR_OUT_OF_MEMORY - /// being returned as the result. - /// - ResultVal<VAddr> SetHeapSize(u64 size); - - /// Maps memory at a given address. - /// - /// @param target The virtual address to map memory at. - /// @param size The amount of memory to map. - /// - /// @note The destination address must lie within the Map region. - /// - /// @note This function requires that SystemResourceSize be non-zero, - /// however, this is just because if it were not then the - /// resulting page tables could be exploited on hardware by - /// a malicious program. SystemResource usage does not need - /// to be explicitly checked or updated here. - ResultCode MapPhysicalMemory(VAddr target, u64 size); - - /// Unmaps memory at a given address. - /// - /// @param target The virtual address to unmap memory at. - /// @param size The amount of memory to unmap. - /// - /// @note The destination address must lie within the Map region. - /// - /// @note This function requires that SystemResourceSize be non-zero, - /// however, this is just because if it were not then the - /// resulting page tables could be exploited on hardware by - /// a malicious program. SystemResource usage does not need - /// to be explicitly checked or updated here. - ResultCode UnmapPhysicalMemory(VAddr target, u64 size); - - /// Maps a region of memory as code memory. - /// - /// @param dst_address The base address of the region to create the aliasing memory region. - /// @param src_address The base address of the region to be aliased. - /// @param size The total amount of memory to map in bytes. - /// - /// @pre Both memory regions lie within the actual addressable address space. - /// - /// @post After this function finishes execution, assuming success, then the address range - /// [dst_address, dst_address+size) will alias the memory region, - /// [src_address, src_address+size). - /// <p> - /// What this also entails is as follows: - /// 1. The aliased region gains the Locked memory attribute. - /// 2. The aliased region becomes read-only. - /// 3. The aliasing region becomes read-only. - /// 4. The aliasing region is created with a memory state of MemoryState::CodeModule. - /// - ResultCode MapCodeMemory(VAddr dst_address, VAddr src_address, u64 size); - - /// Unmaps a region of memory designated as code module memory. - /// - /// @param dst_address The base address of the memory region aliasing the source memory region. - /// @param src_address The base address of the memory region being aliased. - /// @param size The size of the memory region to unmap in bytes. - /// - /// @pre Both memory ranges lie within the actual addressable address space. - /// - /// @pre The memory region being unmapped has been previously been mapped - /// by a call to MapCodeMemory. - /// - /// @post After execution of the function, if successful. the aliasing memory region - /// will be unmapped and the aliased region will have various traits about it - /// restored to what they were prior to the original mapping call preceding - /// this function call. - /// <p> - /// What this also entails is as follows: - /// 1. The state of the memory region will now indicate a general heap region. - /// 2. All memory attributes for the memory region are cleared. - /// 3. Memory permissions for the region are restored to user read/write. - /// - ResultCode UnmapCodeMemory(VAddr dst_address, VAddr src_address, u64 size); - - /// Queries the memory manager for information about the given address. - /// - /// @param address The address to query the memory manager about for information. - /// - /// @return A MemoryInfo instance containing information about the given address. - /// - MemoryInfo QueryMemory(VAddr address) const; - - /// Sets an attribute across the given address range. - /// - /// @param address The starting address - /// @param size The size of the range to set the attribute on. - /// @param mask The attribute mask - /// @param attribute The attribute to set across the given address range - /// - /// @returns RESULT_SUCCESS if successful - /// @returns ERR_INVALID_ADDRESS_STATE if the attribute could not be set. - /// - ResultCode SetMemoryAttribute(VAddr address, u64 size, MemoryAttribute mask, - MemoryAttribute attribute); - - /** - * Scans all VMAs and updates the page table range of any that use the given vector as backing - * memory. This should be called after any operation that causes reallocation of the vector. - */ - void RefreshMemoryBlockMappings(const PhysicalMemory* block); - - /// Dumps the address space layout to the log, for debugging - void LogLayout() const; - - /// Gets the total memory usage, used by svcGetInfo - u64 GetTotalPhysicalMemoryAvailable() const; - - /// Gets the address space base address - VAddr GetAddressSpaceBaseAddress() const; - - /// Gets the address space end address - VAddr GetAddressSpaceEndAddress() const; - - /// Gets the total address space address size in bytes - u64 GetAddressSpaceSize() const; - - /// Gets the address space width in bits. - u64 GetAddressSpaceWidth() const; - - /// Determines whether or not the given address range lies within the address space. - bool IsWithinAddressSpace(VAddr address, u64 size) const; - - /// Gets the base address of the ASLR region. - VAddr GetASLRRegionBaseAddress() const; - - /// Gets the end address of the ASLR region. - VAddr GetASLRRegionEndAddress() const; - - /// Gets the size of the ASLR region - u64 GetASLRRegionSize() const; - - /// Determines whether or not the specified address range is within the ASLR region. - bool IsWithinASLRRegion(VAddr address, u64 size) const; - - /// Gets the base address of the code region. - VAddr GetCodeRegionBaseAddress() const; - - /// Gets the end address of the code region. - VAddr GetCodeRegionEndAddress() const; - - /// Gets the total size of the code region in bytes. - u64 GetCodeRegionSize() const; - - /// Determines whether or not the specified range is within the code region. - bool IsWithinCodeRegion(VAddr address, u64 size) const; - - /// Gets the base address of the heap region. - VAddr GetHeapRegionBaseAddress() const; - - /// Gets the end address of the heap region; - VAddr GetHeapRegionEndAddress() const; - - /// Gets the total size of the heap region in bytes. - u64 GetHeapRegionSize() const; - - /// Gets the total size of the current heap in bytes. - /// - /// @note This is the current allocated heap size, not the size - /// of the region it's allowed to exist within. - /// - u64 GetCurrentHeapSize() const; - - /// Determines whether or not the specified range is within the heap region. - bool IsWithinHeapRegion(VAddr address, u64 size) const; - - /// Gets the base address of the map region. - VAddr GetMapRegionBaseAddress() const; - - /// Gets the end address of the map region. - VAddr GetMapRegionEndAddress() const; - - /// Gets the total size of the map region in bytes. - u64 GetMapRegionSize() const; - - /// Determines whether or not the specified range is within the map region. - bool IsWithinMapRegion(VAddr address, u64 size) const; - - /// Gets the base address of the stack region. - VAddr GetStackRegionBaseAddress() const; - - /// Gets the end address of the stack region. - VAddr GetStackRegionEndAddress() const; - - /// Gets the total size of the stack region in bytes. - u64 GetStackRegionSize() const; - - /// Determines whether or not the given address range is within the stack region - bool IsWithinStackRegion(VAddr address, u64 size) const; - - /// Gets the base address of the TLS IO region. - VAddr GetTLSIORegionBaseAddress() const; - - /// Gets the end address of the TLS IO region. - VAddr GetTLSIORegionEndAddress() const; - - /// Gets the total size of the TLS IO region in bytes. - u64 GetTLSIORegionSize() const; - - /// Determines if the given address range is within the TLS IO region. - bool IsWithinTLSIORegion(VAddr address, u64 size) const; - - /// Each VMManager has its own page table, which is set as the main one when the owning process - /// is scheduled. - Common::PageTable page_table{Memory::PAGE_BITS}; - - using CheckResults = ResultVal<std::tuple<MemoryState, VMAPermission, MemoryAttribute>>; - - /// Checks if an address range adheres to the specified states provided. - /// - /// @param address The starting address of the address range. - /// @param size The size of the address range. - /// @param state_mask The memory state mask. - /// @param state The state to compare the individual VMA states against, - /// which is done in the form of: (vma.state & state_mask) != state. - /// @param permission_mask The memory permissions mask. - /// @param permissions The permission to compare the individual VMA permissions against, - /// which is done in the form of: - /// (vma.permission & permission_mask) != permission. - /// @param attribute_mask The memory attribute mask. - /// @param attribute The memory attributes to compare the individual VMA attributes - /// against, which is done in the form of: - /// (vma.attributes & attribute_mask) != attribute. - /// @param ignore_mask The memory attributes to ignore during the check. - /// - /// @returns If successful, returns a tuple containing the memory attributes - /// (with ignored bits specified by ignore_mask unset), memory permissions, and - /// memory state across the memory range. - /// @returns If not successful, returns ERR_INVALID_ADDRESS_STATE. - /// - CheckResults CheckRangeState(VAddr address, u64 size, MemoryState state_mask, MemoryState state, - VMAPermission permission_mask, VMAPermission permissions, - MemoryAttribute attribute_mask, MemoryAttribute attribute, - MemoryAttribute ignore_mask) const; - -private: - using VMAIter = VMAMap::iterator; - - /// Converts a VMAHandle to a mutable VMAIter. - VMAIter StripIterConstness(const VMAHandle& iter); - - /// Unmaps the given VMA. - VMAIter Unmap(VMAIter vma); - - /** - * Carves a VMA of a specific size at the specified address by splitting Free VMAs while doing - * the appropriate error checking. - */ - ResultVal<VMAIter> CarveVMA(VAddr base, u64 size); - - /** - * Splits the edges of the given range of non-Free VMAs so that there is a VMA split at each - * end of the range. - */ - ResultVal<VMAIter> CarveVMARange(VAddr base, u64 size); - - /** - * Splits a VMA in two, at the specified offset. - * @returns the right side of the split, with the original iterator becoming the left side. - */ - VMAIter SplitVMA(VMAIter vma, u64 offset_in_vma); - - /** - * Checks for and merges the specified VMA with adjacent ones if possible. - * @returns the merged VMA or the original if no merging was possible. - */ - VMAIter MergeAdjacent(VMAIter vma); - - /** - * Merges two adjacent VMAs. - */ - void MergeAdjacentVMA(VirtualMemoryArea& left, const VirtualMemoryArea& right); - - /// Updates the pages corresponding to this VMA so they match the VMA's attributes. - void UpdatePageTableForVMA(const VirtualMemoryArea& vma); - - /// Initializes memory region ranges to adhere to a given address space type. - void InitializeMemoryRegionRanges(FileSys::ProgramAddressSpaceType type); - - /// Clears the underlying map and page table. - void Clear(); - - /// Clears out the VMA map, unmapping any previously mapped ranges. - void ClearVMAMap(); - - /// Clears out the page table - void ClearPageTable(); - - /// Gets the amount of memory currently mapped (state != Unmapped) in a range. - ResultVal<std::size_t> SizeOfAllocatedVMAsInRange(VAddr address, std::size_t size) const; - - /// Gets the amount of memory unmappable by UnmapPhysicalMemory in a range. - ResultVal<std::size_t> SizeOfUnmappablePhysicalMemoryInRange(VAddr address, - std::size_t size) const; - - /** - * A map covering the entirety of the managed address space, keyed by the `base` field of each - * VMA. It must always be modified by splitting or merging VMAs, so that the invariant - * `elem.base + elem.size == next.base` is preserved, and mergeable regions must always be - * merged when possible so that no two similar and adjacent regions exist that have not been - * merged. - */ - VMAMap vma_map; - - u32 address_space_width = 0; - VAddr address_space_base = 0; - VAddr address_space_end = 0; - - VAddr aslr_region_base = 0; - VAddr aslr_region_end = 0; - - VAddr code_region_base = 0; - VAddr code_region_end = 0; - - VAddr heap_region_base = 0; - VAddr heap_region_end = 0; - - VAddr map_region_base = 0; - VAddr map_region_end = 0; - - VAddr stack_region_base = 0; - VAddr stack_region_end = 0; - - VAddr tls_io_region_base = 0; - VAddr tls_io_region_end = 0; - - // Memory used to back the allocations in the regular heap. A single vector is used to cover - // the entire virtual address space extents that bound the allocations, including any holes. - // This makes deallocation and reallocation of holes fast and keeps process memory contiguous - // in the emulator address space, allowing Memory::GetPointer to be reasonably safe. - std::shared_ptr<PhysicalMemory> heap_memory; - - // The end of the currently allocated heap. This is not an inclusive - // end of the range. This is essentially 'base_address + current_size'. - VAddr heap_end = 0; - - // The current amount of memory mapped via MapPhysicalMemory. - // This is used here (and in Nintendo's kernel) only for debugging, and does not impact - // any behavior. - u64 physical_memory_mapped = 0; - - Core::System& system; -}; -} // namespace Kernel diff --git a/src/core/hle/result.h b/src/core/hle/result.h index 450f61fea..b6bdbd988 100644 --- a/src/core/hle/result.h +++ b/src/core/hle/result.h @@ -342,8 +342,9 @@ ResultVal<std::remove_reference_t<Arg>> MakeResult(Arg&& arg) { */ #define CASCADE_RESULT(target, source) \ auto CONCAT2(check_result_L, __LINE__) = source; \ - if (CONCAT2(check_result_L, __LINE__).Failed()) \ + if (CONCAT2(check_result_L, __LINE__).Failed()) { \ return CONCAT2(check_result_L, __LINE__).Code(); \ + } \ target = std::move(*CONCAT2(check_result_L, __LINE__)) /** @@ -351,6 +352,9 @@ ResultVal<std::remove_reference_t<Arg>> MakeResult(Arg&& arg) { * non-success, or discarded otherwise. */ #define CASCADE_CODE(source) \ - auto CONCAT2(check_result_L, __LINE__) = source; \ - if (CONCAT2(check_result_L, __LINE__).IsError()) \ - return CONCAT2(check_result_L, __LINE__); + do { \ + auto CONCAT2(check_result_L, __LINE__) = source; \ + if (CONCAT2(check_result_L, __LINE__).IsError()) { \ + return CONCAT2(check_result_L, __LINE__); \ + } \ + } while (false) diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp index cfac8ca9a..c2c11dbcb 100644 --- a/src/core/hle/service/acc/acc.cpp +++ b/src/core/hle/service/acc/acc.cpp @@ -11,6 +11,7 @@ #include "common/string_util.h" #include "common/swap.h" #include "core/constants.h" +#include "core/core.h" #include "core/core_timing.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" @@ -35,7 +36,7 @@ constexpr ResultCode ERR_INVALID_BUFFER_SIZE{ErrorModule::Account, 30}; constexpr ResultCode ERR_FAILED_SAVE_DATA{ErrorModule::Account, 100}; static std::string GetImagePath(Common::UUID uuid) { - return FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + + return Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg"; } @@ -44,6 +45,218 @@ static constexpr u32 SanitizeJPEGSize(std::size_t size) { return static_cast<u32>(std::min(size, max_jpeg_image_size)); } +class IManagerForSystemService final : public ServiceFramework<IManagerForSystemService> { +public: + explicit IManagerForSystemService(Common::UUID user_id) + : ServiceFramework("IManagerForSystemService") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "CheckAvailability"}, + {1, nullptr, "GetAccountId"}, + {2, nullptr, "EnsureIdTokenCacheAsync"}, + {3, nullptr, "LoadIdTokenCache"}, + {100, nullptr, "SetSystemProgramIdentification"}, + {101, nullptr, "RefreshNotificationTokenAsync"}, // 7.0.0+ + {110, nullptr, "GetServiceEntryRequirementCache"}, // 4.0.0+ + {111, nullptr, "InvalidateServiceEntryRequirementCache"}, // 4.0.0+ + {112, nullptr, "InvalidateTokenCache"}, // 4.0.0 - 6.2.0 + {113, nullptr, "GetServiceEntryRequirementCacheForOnlinePlay"}, // 6.1.0+ + {120, nullptr, "GetNintendoAccountId"}, + {121, nullptr, "CalculateNintendoAccountAuthenticationFingerprint"}, // 9.0.0+ + {130, nullptr, "GetNintendoAccountUserResourceCache"}, + {131, nullptr, "RefreshNintendoAccountUserResourceCacheAsync"}, + {132, nullptr, "RefreshNintendoAccountUserResourceCacheAsyncIfSecondsElapsed"}, + {133, nullptr, "GetNintendoAccountVerificationUrlCache"}, // 9.0.0+ + {134, nullptr, "RefreshNintendoAccountVerificationUrlCache"}, // 9.0.0+ + {135, nullptr, "RefreshNintendoAccountVerificationUrlCacheAsyncIfSecondsElapsed"}, // 9.0.0+ + {140, nullptr, "GetNetworkServiceLicenseCache"}, // 5.0.0+ + {141, nullptr, "RefreshNetworkServiceLicenseCacheAsync"}, // 5.0.0+ + {142, nullptr, "RefreshNetworkServiceLicenseCacheAsyncIfSecondsElapsed"}, // 5.0.0+ + {150, nullptr, "CreateAuthorizationRequest"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +// 3.0.0+ +class IFloatingRegistrationRequest final : public ServiceFramework<IFloatingRegistrationRequest> { +public: + explicit IFloatingRegistrationRequest(Common::UUID user_id) + : ServiceFramework("IFloatingRegistrationRequest") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetSessionId"}, + {12, nullptr, "GetAccountId"}, + {13, nullptr, "GetLinkedNintendoAccountId"}, + {14, nullptr, "GetNickname"}, + {15, nullptr, "GetProfileImage"}, + {21, nullptr, "LoadIdTokenCache"}, + {100, nullptr, "RegisterUser"}, // [1.0.0-3.0.2] RegisterAsync + {101, nullptr, "RegisterUserWithUid"}, // [1.0.0-3.0.2] RegisterWithUidAsync + {102, nullptr, "RegisterNetworkServiceAccountAsync"}, // 4.0.0+ + {103, nullptr, "RegisterNetworkServiceAccountWithUidAsync"}, // 4.0.0+ + {110, nullptr, "SetSystemProgramIdentification"}, + {111, nullptr, "EnsureIdTokenCacheAsync"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IAdministrator final : public ServiceFramework<IAdministrator> { +public: + explicit IAdministrator(Common::UUID user_id) : ServiceFramework("IAdministrator") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "CheckAvailability"}, + {1, nullptr, "GetAccountId"}, + {2, nullptr, "EnsureIdTokenCacheAsync"}, + {3, nullptr, "LoadIdTokenCache"}, + {100, nullptr, "SetSystemProgramIdentification"}, + {101, nullptr, "RefreshNotificationTokenAsync"}, // 7.0.0+ + {110, nullptr, "GetServiceEntryRequirementCache"}, // 4.0.0+ + {111, nullptr, "InvalidateServiceEntryRequirementCache"}, // 4.0.0+ + {112, nullptr, "InvalidateTokenCache"}, // 4.0.0 - 6.2.0 + {113, nullptr, "GetServiceEntryRequirementCacheForOnlinePlay"}, // 6.1.0+ + {120, nullptr, "GetNintendoAccountId"}, + {121, nullptr, "CalculateNintendoAccountAuthenticationFingerprint"}, // 9.0.0+ + {130, nullptr, "GetNintendoAccountUserResourceCache"}, + {131, nullptr, "RefreshNintendoAccountUserResourceCacheAsync"}, + {132, nullptr, "RefreshNintendoAccountUserResourceCacheAsyncIfSecondsElapsed"}, + {133, nullptr, "GetNintendoAccountVerificationUrlCache"}, // 9.0.0+ + {134, nullptr, "RefreshNintendoAccountVerificationUrlCacheAsync"}, // 9.0.0+ + {135, nullptr, "RefreshNintendoAccountVerificationUrlCacheAsyncIfSecondsElapsed"}, // 9.0.0+ + {140, nullptr, "GetNetworkServiceLicenseCache"}, // 5.0.0+ + {141, nullptr, "RefreshNetworkServiceLicenseCacheAsync"}, // 5.0.0+ + {142, nullptr, "RefreshNetworkServiceLicenseCacheAsyncIfSecondsElapsed"}, // 5.0.0+ + {150, nullptr, "CreateAuthorizationRequest"}, + {200, nullptr, "IsRegistered"}, + {201, nullptr, "RegisterAsync"}, + {202, nullptr, "UnregisterAsync"}, + {203, nullptr, "DeleteRegistrationInfoLocally"}, + {220, nullptr, "SynchronizeProfileAsync"}, + {221, nullptr, "UploadProfileAsync"}, + {222, nullptr, "SynchronizaProfileAsyncIfSecondsElapsed"}, + {250, nullptr, "IsLinkedWithNintendoAccount"}, + {251, nullptr, "CreateProcedureToLinkWithNintendoAccount"}, + {252, nullptr, "ResumeProcedureToLinkWithNintendoAccount"}, + {255, nullptr, "CreateProcedureToUpdateLinkageStateOfNintendoAccount"}, + {256, nullptr, "ResumeProcedureToUpdateLinkageStateOfNintendoAccount"}, + {260, nullptr, "CreateProcedureToLinkNnidWithNintendoAccount"}, // 3.0.0+ + {261, nullptr, "ResumeProcedureToLinkNnidWithNintendoAccount"}, // 3.0.0+ + {280, nullptr, "ProxyProcedureToAcquireApplicationAuthorizationForNintendoAccount"}, + {290, nullptr, "GetRequestForNintendoAccountUserResourceView"}, // 8.0.0+ + {300, nullptr, "TryRecoverNintendoAccountUserStateAsync"}, // 6.0.0+ + {400, nullptr, "IsServiceEntryRequirementCacheRefreshRequiredForOnlinePlay"}, // 6.1.0+ + {401, nullptr, "RefreshServiceEntryRequirementCacheForOnlinePlayAsync"}, // 6.1.0+ + {900, nullptr, "GetAuthenticationInfoForWin"}, // 9.0.0+ + {901, nullptr, "ImportAsyncForWin"}, // 9.0.0+ + {997, nullptr, "DebugUnlinkNintendoAccountAsync"}, + {998, nullptr, "DebugSetAvailabilityErrorDetail"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IAuthorizationRequest final : public ServiceFramework<IAuthorizationRequest> { +public: + explicit IAuthorizationRequest(Common::UUID user_id) + : ServiceFramework("IAuthorizationRequest") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetSessionId"}, + {10, nullptr, "InvokeWithoutInteractionAsync"}, + {19, nullptr, "IsAuthorized"}, + {20, nullptr, "GetAuthorizationCode"}, + {21, nullptr, "GetIdToken"}, + {22, nullptr, "GetState"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IOAuthProcedure final : public ServiceFramework<IOAuthProcedure> { +public: + explicit IOAuthProcedure(Common::UUID user_id) : ServiceFramework("IOAuthProcedure") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "PrepareAsync"}, + {1, nullptr, "GetRequest"}, + {2, nullptr, "ApplyResponse"}, + {3, nullptr, "ApplyResponseAsync"}, + {10, nullptr, "Suspend"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +// 3.0.0+ +class IOAuthProcedureForExternalNsa final : public ServiceFramework<IOAuthProcedureForExternalNsa> { +public: + explicit IOAuthProcedureForExternalNsa(Common::UUID user_id) + : ServiceFramework("IOAuthProcedureForExternalNsa") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "PrepareAsync"}, + {1, nullptr, "GetRequest"}, + {2, nullptr, "ApplyResponse"}, + {3, nullptr, "ApplyResponseAsync"}, + {10, nullptr, "Suspend"}, + {100, nullptr, "GetAccountId"}, + {101, nullptr, "GetLinkedNintendoAccountId"}, + {102, nullptr, "GetNickname"}, + {103, nullptr, "GetProfileImage"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IOAuthProcedureForNintendoAccountLinkage final + : public ServiceFramework<IOAuthProcedureForNintendoAccountLinkage> { +public: + explicit IOAuthProcedureForNintendoAccountLinkage(Common::UUID user_id) + : ServiceFramework("IOAuthProcedureForNintendoAccountLinkage") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "PrepareAsync"}, + {1, nullptr, "GetRequest"}, + {2, nullptr, "ApplyResponse"}, + {3, nullptr, "ApplyResponseAsync"}, + {10, nullptr, "Suspend"}, + {100, nullptr, "GetRequestWithTheme"}, + {101, nullptr, "IsNetworkServiceAccountReplaced"}, + {199, nullptr, "GetUrlForIntroductionOfExtraMembership"}, // 2.0.0 - 5.1.0 + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class INotifier final : public ServiceFramework<INotifier> { +public: + explicit INotifier(Common::UUID user_id) : ServiceFramework("INotifier") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetSystemEvent"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + class IProfileCommon : public ServiceFramework<IProfileCommon> { public: explicit IProfileCommon(const char* name, bool editor_commands, Common::UUID user_id, @@ -74,9 +287,7 @@ protected: ProfileBase profile_base{}; ProfileData data{}; if (profile_manager.GetProfileBaseAndData(user_id, profile_base, data)) { - std::array<u8, sizeof(ProfileData)> raw_data; - std::memcpy(raw_data.data(), &data, sizeof(ProfileData)); - ctx.WriteBuffer(raw_data); + ctx.WriteBuffer(data); IPC::ResponseBuilder rb{ctx, 16}; rb.Push(RESULT_SUCCESS); rb.PushRaw(profile_base); @@ -108,7 +319,7 @@ protected: IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - const FileUtil::IOFile image(GetImagePath(user_id), "rb"); + const Common::FS::IOFile image(GetImagePath(user_id), "rb"); if (!image.IsOpen()) { LOG_WARNING(Service_ACC, "Failed to load user provided image! Falling back to built-in backup..."); @@ -121,7 +332,7 @@ protected: std::vector<u8> buffer(size); image.ReadBytes(buffer.data(), buffer.size()); - ctx.WriteBuffer(buffer.data(), buffer.size()); + ctx.WriteBuffer(buffer); rb.Push<u32>(size); } @@ -130,7 +341,7 @@ protected: IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - const FileUtil::IOFile image(GetImagePath(user_id), "rb"); + const Common::FS::IOFile image(GetImagePath(user_id), "rb"); if (!image.IsOpen()) { LOG_WARNING(Service_ACC, @@ -195,7 +406,7 @@ protected: ProfileData data; std::memcpy(&data, user_data.data(), sizeof(ProfileData)); - FileUtil::IOFile image(GetImagePath(user_id), "wb"); + Common::FS::IOFile image(GetImagePath(user_id), "wb"); if (!image.IsOpen() || !image.Resize(image_data.size()) || image.WriteBytes(image_data.data(), image_data.size()) != image_data.size() || @@ -226,9 +437,58 @@ public: : IProfileCommon("IProfileEditor", true, user_id, profile_manager) {} }; +class IAsyncContext final : public ServiceFramework<IAsyncContext> { +public: + explicit IAsyncContext(Common::UUID user_id) : ServiceFramework("IAsyncContext") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetSystemEvent"}, + {1, nullptr, "Cancel"}, + {2, nullptr, "HasDone"}, + {3, nullptr, "GetResult"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class ISessionObject final : public ServiceFramework<ISessionObject> { +public: + explicit ISessionObject(Common::UUID user_id) : ServiceFramework("ISessionObject") { + // clang-format off + static const FunctionInfo functions[] = { + {999, nullptr, "Dummy"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IGuestLoginRequest final : public ServiceFramework<IGuestLoginRequest> { +public: + explicit IGuestLoginRequest(Common::UUID) : ServiceFramework("IGuestLoginRequest") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetSessionId"}, + {11, nullptr, "Unknown"}, // 1.0.0 - 2.3.0 (the name is blank on Switchbrew) + {12, nullptr, "GetAccountId"}, + {13, nullptr, "GetLinkedNintendoAccountId"}, + {14, nullptr, "GetNickname"}, + {15, nullptr, "GetProfileImage"}, + {21, nullptr, "LoadIdTokenCache"}, // 3.0.0+ + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + class IManagerForApplication final : public ServiceFramework<IManagerForApplication> { public: - IManagerForApplication() : ServiceFramework("IManagerForApplication") { + explicit IManagerForApplication(Common::UUID user_id) + : ServiceFramework("IManagerForApplication"), user_id(user_id) { // clang-format off static const FunctionInfo functions[] = { {0, &IManagerForApplication::CheckAvailability, "CheckAvailability"}, @@ -237,7 +497,7 @@ public: {3, nullptr, "LoadIdTokenCache"}, {130, nullptr, "GetNintendoAccountUserResourceCacheForApplication"}, {150, nullptr, "CreateAuthorizationRequest"}, - {160, nullptr, "StoreOpenContext"}, + {160, &IManagerForApplication::StoreOpenContext, "StoreOpenContext"}, {170, nullptr, "LoadNetworkServiceLicenseKindAsync"}, }; // clang-format on @@ -254,11 +514,100 @@ private: } void GetAccountId(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_ACC, "(STUBBED) called"); - // Should return a nintendo account ID + LOG_DEBUG(Service_ACC, "called"); + IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<u64>(1); + rb.PushRaw<u64>(user_id.GetNintendoID()); + } + + void StoreOpenContext(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_ACC, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + Common::UUID user_id; +}; + +// 6.0.0+ +class IAsyncNetworkServiceLicenseKindContext final + : public ServiceFramework<IAsyncNetworkServiceLicenseKindContext> { +public: + explicit IAsyncNetworkServiceLicenseKindContext(Common::UUID user_id) + : ServiceFramework("IAsyncNetworkServiceLicenseKindContext") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetSystemEvent"}, + {1, nullptr, "Cancel"}, + {2, nullptr, "HasDone"}, + {3, nullptr, "GetResult"}, + {4, nullptr, "GetNetworkServiceLicenseKind"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +// 8.0.0+ +class IOAuthProcedureForUserRegistration final + : public ServiceFramework<IOAuthProcedureForUserRegistration> { +public: + explicit IOAuthProcedureForUserRegistration(Common::UUID user_id) + : ServiceFramework("IOAuthProcedureForUserRegistration") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "PrepareAsync"}, + {1, nullptr, "GetRequest"}, + {2, nullptr, "ApplyResponse"}, + {3, nullptr, "ApplyResponseAsync"}, + {10, nullptr, "Suspend"}, + {100, nullptr, "GetAccountId"}, + {101, nullptr, "GetLinkedNintendoAccountId"}, + {102, nullptr, "GetNickname"}, + {103, nullptr, "GetProfileImage"}, + {110, nullptr, "RegisterUserAsync"}, + {111, nullptr, "GetUid"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class DAUTH_O final : public ServiceFramework<DAUTH_O> { +public: + explicit DAUTH_O(Common::UUID) : ServiceFramework("dauth:o") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "EnsureAuthenticationTokenCacheAsync"}, // [5.0.0-5.1.0] GeneratePostData + {1, nullptr, "LoadAuthenticationTokenCache"}, // 6.0.0+ + {2, nullptr, "InvalidateAuthenticationTokenCache"}, // 6.0.0+ + {10, nullptr, "EnsureEdgeTokenCacheAsync"}, // 6.0.0+ + {11, nullptr, "LoadEdgeTokenCache"}, // 6.0.0+ + {12, nullptr, "InvalidateEdgeTokenCache"}, // 6.0.0+ + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +// 6.0.0+ +class IAsyncResult final : public ServiceFramework<IAsyncResult> { +public: + explicit IAsyncResult(Common::UUID user_id) : ServiceFramework("IAsyncResult") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetResult"}, + {1, nullptr, "Cancel"}, + {2, nullptr, "IsAvailable"}, + {3, nullptr, "GetSystemEvent"}, + }; + // clang-format on + + RegisterHandlers(functions); } }; @@ -319,46 +668,37 @@ void Module::Interface::IsUserRegistrationRequestPermitted(Kernel::HLERequestCon void Module::Interface::InitializeApplicationInfo(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto pid = rp.Pop<u64>(); - LOG_DEBUG(Service_ACC, "called, process_id={}", pid); + LOG_DEBUG(Service_ACC, "called"); IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(InitializeApplicationInfoBase(pid)); + rb.Push(InitializeApplicationInfoBase()); } void Module::Interface::InitializeApplicationInfoRestricted(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto pid = rp.Pop<u64>(); - LOG_WARNING(Service_ACC, "(Partial implementation) called, process_id={}", pid); + LOG_WARNING(Service_ACC, "(Partial implementation) called"); // TODO(ogniK): We require checking if the user actually owns the title and what not. As of // currently, we assume the user owns the title. InitializeApplicationInfoBase SHOULD be called // first then we do extra checks if the game is a digital copy. IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(InitializeApplicationInfoBase(pid)); + rb.Push(InitializeApplicationInfoBase()); } -ResultCode Module::Interface::InitializeApplicationInfoBase(u64 process_id) { +ResultCode Module::Interface::InitializeApplicationInfoBase() { if (application_info) { LOG_ERROR(Service_ACC, "Application already initialized"); return ERR_ACCOUNTINFO_ALREADY_INITIALIZED; } - const auto& list = system.Kernel().GetProcessList(); - const auto iter = std::find_if(list.begin(), list.end(), [&process_id](const auto& process) { - return process->GetProcessID() == process_id; - }); - - if (iter == list.end()) { - LOG_ERROR(Service_ACC, "Failed to find process ID"); - application_info.application_type = ApplicationType::Unknown; - - return ERR_ACCOUNTINFO_BAD_APPLICATION; - } - - const auto launch_property = system.GetARPManager().GetLaunchProperty((*iter)->GetTitleID()); + // TODO(ogniK): This should be changed to reflect the target process for when we have multiple + // processes emulated. As we don't actually have pid support we should assume we're just using + // our own process + const auto& current_process = system.Kernel().CurrentProcess(); + const auto launch_property = + system.GetARPManager().GetLaunchProperty(current_process->GetTitleID()); if (launch_property.Failed()) { LOG_ERROR(Service_ACC, "Failed to get launch property"); @@ -372,10 +712,12 @@ ResultCode Module::Interface::InitializeApplicationInfoBase(u64 process_id) { case FileSys::StorageId::Host: case FileSys::StorageId::NandUser: case FileSys::StorageId::SdCard: + case FileSys::StorageId::None: // Yuzu specific, differs from hardware application_info.application_type = ApplicationType::Digital; break; default: - LOG_ERROR(Service_ACC, "Invalid game storage ID"); + LOG_ERROR(Service_ACC, "Invalid game storage ID! storage_id={}", + launch_property->base_game_storage_id); return ERR_ACCOUNTINFO_BAD_APPLICATION; } @@ -389,7 +731,7 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo LOG_DEBUG(Service_ACC, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IManagerForApplication>(); + rb.PushIpcInterface<IManagerForApplication>(profile_manager->GetLastOpenedUser()); } void Module::Interface::IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx) { @@ -400,8 +742,10 @@ void Module::Interface::IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx bool is_locked = false; if (res != Loader::ResultStatus::Success) { - FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()}; - auto nacp_unique = pm.GetControlMetadata().first; + const FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID(), + system.GetFileSystemController(), + system.GetContentProvider()}; + const auto nacp_unique = pm.GetControlMetadata().first; if (nacp_unique != nullptr) { is_locked = nacp_unique->GetUserAccountSwitchLock(); @@ -428,6 +772,37 @@ void Module::Interface::GetProfileEditor(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<IProfileEditor>(user_id, *profile_manager); } +void Module::Interface::ListQualifiedUsers(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_ACC, "called"); + + // All users should be qualified. We don't actually have parental control or anything to do with + // nintendo online currently. We're just going to assume the user running the game has access to + // the game regardless of parental control settings. + ctx.WriteBuffer(profile_manager->GetAllUsers()); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Module::Interface::LoadOpenContext(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_ACC, "(STUBBED) called"); + + // This is similar to GetBaasAccountManagerForApplication + // This command is used concurrently with ListOpenContextStoredUsers + // TODO: Find the differences between this and GetBaasAccountManagerForApplication + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IManagerForApplication>(profile_manager->GetLastOpenedUser()); +} + +void Module::Interface::ListOpenContextStoredUsers(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_ACC, "(STUBBED) called"); + + // TODO(ogniK): Handle open contexts + ctx.WriteBuffer(profile_manager->GetOpenUsers()); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_ACC, "called"); // A u8 is passed into this function which we can safely ignore. It's to determine if we have diff --git a/src/core/hle/service/acc/acc.h b/src/core/hle/service/acc/acc.h index 7a7dc9ec6..c611efd89 100644 --- a/src/core/hle/service/acc/acc.h +++ b/src/core/hle/service/acc/acc.h @@ -33,9 +33,12 @@ public: void TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx); void IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx); void GetProfileEditor(Kernel::HLERequestContext& ctx); + void ListQualifiedUsers(Kernel::HLERequestContext& ctx); + void LoadOpenContext(Kernel::HLERequestContext& ctx); + void ListOpenContextStoredUsers(Kernel::HLERequestContext& ctx); private: - ResultCode InitializeApplicationInfoBase(u64 process_id); + ResultCode InitializeApplicationInfoBase(); enum class ApplicationType : u32_le { GameCard = 0, diff --git a/src/core/hle/service/acc/acc_aa.cpp b/src/core/hle/service/acc/acc_aa.cpp index 3bac6bcd1..51f119b12 100644 --- a/src/core/hle/service/acc/acc_aa.cpp +++ b/src/core/hle/service/acc/acc_aa.cpp @@ -13,8 +13,8 @@ ACC_AA::ACC_AA(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {0, nullptr, "EnsureCacheAsync"}, {1, nullptr, "LoadCache"}, {2, nullptr, "GetDeviceAccountId"}, - {50, nullptr, "RegisterNotificationTokenAsync"}, - {51, nullptr, "UnregisterNotificationTokenAsync"}, + {50, nullptr, "RegisterNotificationTokenAsync"}, // 1.0.0 - 6.2.0 + {51, nullptr, "UnregisterNotificationTokenAsync"}, // 1.0.0 - 6.2.0 }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/acc/acc_su.cpp b/src/core/hle/service/acc/acc_su.cpp index b941c260b..d2bb8c2c8 100644 --- a/src/core/hle/service/acc/acc_su.cpp +++ b/src/core/hle/service/acc/acc_su.cpp @@ -17,26 +17,28 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {3, &ACC_SU::ListOpenUsers, "ListOpenUsers"}, {4, &ACC_SU::GetLastOpenedUser, "GetLastOpenedUser"}, {5, &ACC_SU::GetProfile, "GetProfile"}, - {6, nullptr, "GetProfileDigest"}, + {6, nullptr, "GetProfileDigest"}, // 3.0.0+ {50, &ACC_SU::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_SU::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, - {60, nullptr, "ListOpenContextStoredUsers"}, - {99, nullptr, "DebugActivateOpenContextRetention"}, + {60, &ACC_SU::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, // 5.0.0 - 5.1.0 + {99, nullptr, "DebugActivateOpenContextRetention"}, // 6.0.0+ {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, {102, nullptr, "GetBaasAccountManagerForSystemService"}, {103, nullptr, "GetBaasUserAvailabilityChangeNotifier"}, {104, nullptr, "GetProfileUpdateNotifier"}, - {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, - {106, nullptr, "GetProfileSyncNotifier"}, + {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, // 4.0.0+ + {106, nullptr, "GetProfileSyncNotifier"}, // 9.0.0+ {110, nullptr, "StoreSaveDataThumbnail"}, {111, nullptr, "ClearSaveDataThumbnail"}, {112, nullptr, "LoadSaveDataThumbnail"}, - {113, nullptr, "GetSaveDataThumbnailExistence"}, - {130, nullptr, "ActivateOpenContextRetention"}, - {140, nullptr, "ListQualifiedUsers"}, - {190, nullptr, "GetUserLastOpenedApplication"}, - {191, nullptr, "ActivateOpenContextHolder"}, + {113, nullptr, "GetSaveDataThumbnailExistence"}, // 5.0.0+ + {120, nullptr, "ListOpenUsersInApplication"}, // 10.0.0+ + {130, nullptr, "ActivateOpenContextRetention"}, // 6.0.0+ + {140, &ACC_SU::ListQualifiedUsers, "ListQualifiedUsers"}, // 6.0.0+ + {150, nullptr, "AuthenticateApplicationAsync"}, // 10.0.0+ + {190, nullptr, "GetUserLastOpenedApplication"}, // 1.0.0 - 9.2.0 + {191, nullptr, "ActivateOpenContextHolder"}, // 7.0.0+ {200, nullptr, "BeginUserRegistration"}, {201, nullptr, "CompleteUserRegistration"}, {202, nullptr, "CancelUserRegistration"}, @@ -44,15 +46,15 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {204, nullptr, "SetUserPosition"}, {205, &ACC_SU::GetProfileEditor, "GetProfileEditor"}, {206, nullptr, "CompleteUserRegistrationForcibly"}, - {210, nullptr, "CreateFloatingRegistrationRequest"}, - {211, nullptr, "CreateProcedureToRegisterUserWithNintendoAccount"}, - {212, nullptr, "ResumeProcedureToRegisterUserWithNintendoAccount"}, + {210, nullptr, "CreateFloatingRegistrationRequest"}, // 3.0.0+ + {211, nullptr, "CreateProcedureToRegisterUserWithNintendoAccount"}, // 8.0.0+ + {212, nullptr, "ResumeProcedureToRegisterUserWithNintendoAccount"}, // 8.0.0+ {230, nullptr, "AuthenticateServiceAsync"}, {250, nullptr, "GetBaasAccountAdministrator"}, {290, nullptr, "ProxyProcedureForGuestLoginWithNintendoAccount"}, - {291, nullptr, "ProxyProcedureForFloatingRegistrationWithNintendoAccount"}, + {291, nullptr, "ProxyProcedureForFloatingRegistrationWithNintendoAccount"}, // 3.0.0+ {299, nullptr, "SuspendBackgroundDaemon"}, - {997, nullptr, "DebugInvalidateTokenCacheForUser"}, + {997, nullptr, "DebugInvalidateTokenCacheForUser"}, // 3.0.0+ {998, nullptr, "DebugSetUserStateClose"}, {999, nullptr, "DebugSetUserStateOpen"}, }; diff --git a/src/core/hle/service/acc/acc_u0.cpp b/src/core/hle/service/acc/acc_u0.cpp index 0ac19f4ff..75a24f8f5 100644 --- a/src/core/hle/service/acc/acc_u0.cpp +++ b/src/core/hle/service/acc/acc_u0.cpp @@ -17,23 +17,23 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {3, &ACC_U0::ListOpenUsers, "ListOpenUsers"}, {4, &ACC_U0::GetLastOpenedUser, "GetLastOpenedUser"}, {5, &ACC_U0::GetProfile, "GetProfile"}, - {6, nullptr, "GetProfileDigest"}, + {6, nullptr, "GetProfileDigest"}, // 3.0.0+ {50, &ACC_U0::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_U0::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, - {60, nullptr, "ListOpenContextStoredUsers"}, - {99, nullptr, "DebugActivateOpenContextRetention"}, + {60, &ACC_U0::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, // 5.0.0 - 5.1.0 + {99, nullptr, "DebugActivateOpenContextRetention"}, // 6.0.0+ {100, &ACC_U0::InitializeApplicationInfo, "InitializeApplicationInfo"}, {101, &ACC_U0::GetBaasAccountManagerForApplication, "GetBaasAccountManagerForApplication"}, {102, nullptr, "AuthenticateApplicationAsync"}, - {103, nullptr, "CheckNetworkServiceAvailabilityAsync"}, + {103, nullptr, "CheckNetworkServiceAvailabilityAsync"}, // 4.0.0+ {110, nullptr, "StoreSaveDataThumbnail"}, {111, nullptr, "ClearSaveDataThumbnail"}, {120, nullptr, "CreateGuestLoginRequest"}, - {130, nullptr, "LoadOpenContext"}, - {131, nullptr, "ListOpenContextStoredUsers"}, - {140, &ACC_U0::InitializeApplicationInfoRestricted, "InitializeApplicationInfoRestricted"}, - {141, nullptr, "ListQualifiedUsers"}, - {150, &ACC_U0::IsUserAccountSwitchLocked, "IsUserAccountSwitchLocked"}, + {130, &ACC_U0::LoadOpenContext, "LoadOpenContext"}, // 5.0.0+ + {131, &ACC_U0::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, // 6.0.0+ + {140, &ACC_U0::InitializeApplicationInfoRestricted, "InitializeApplicationInfoRestricted"}, // 6.0.0+ + {141, &ACC_U0::ListQualifiedUsers, "ListQualifiedUsers"}, // 6.0.0+ + {150, &ACC_U0::IsUserAccountSwitchLocked, "IsUserAccountSwitchLocked"}, // 6.0.0+ }; // clang-format on diff --git a/src/core/hle/service/acc/acc_u1.cpp b/src/core/hle/service/acc/acc_u1.cpp index 858e91dde..a4aa5316a 100644 --- a/src/core/hle/service/acc/acc_u1.cpp +++ b/src/core/hle/service/acc/acc_u1.cpp @@ -17,27 +17,29 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {3, &ACC_U1::ListOpenUsers, "ListOpenUsers"}, {4, &ACC_U1::GetLastOpenedUser, "GetLastOpenedUser"}, {5, &ACC_U1::GetProfile, "GetProfile"}, - {6, nullptr, "GetProfileDigest"}, + {6, nullptr, "GetProfileDigest"}, // 3.0.0+ {50, &ACC_U1::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_U1::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, - {60, nullptr, "ListOpenContextStoredUsers"}, - {99, nullptr, "DebugActivateOpenContextRetention"}, + {60, &ACC_U1::ListOpenContextStoredUsers, "ListOpenContextStoredUsers"}, // 5.0.0 - 5.1.0 + {99, nullptr, "DebugActivateOpenContextRetention"}, // 6.0.0+ {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, {102, nullptr, "GetBaasAccountManagerForSystemService"}, - {103, nullptr, "GetProfileUpdateNotifier"}, - {104, nullptr, "CheckNetworkServiceAvailabilityAsync"}, - {105, nullptr, "GetBaasUserAvailabilityChangeNotifier"}, - {106, nullptr, "GetProfileSyncNotifier"}, + {103, nullptr, "GetBaasUserAvailabilityChangeNotifier"}, + {104, nullptr, "GetProfileUpdateNotifier"}, + {105, nullptr, "CheckNetworkServiceAvailabilityAsync"}, // 4.0.0+ + {106, nullptr, "GetProfileSyncNotifier"}, // 9.0.0+ {110, nullptr, "StoreSaveDataThumbnail"}, {111, nullptr, "ClearSaveDataThumbnail"}, {112, nullptr, "LoadSaveDataThumbnail"}, - {113, nullptr, "GetSaveDataThumbnailExistence"}, - {130, nullptr, "ActivateOpenContextRetention"}, - {140, nullptr, "ListQualifiedUsers"}, - {190, nullptr, "GetUserLastOpenedApplication"}, - {191, nullptr, "ActivateOpenContextHolder"}, - {997, nullptr, "DebugInvalidateTokenCacheForUser"}, + {113, nullptr, "GetSaveDataThumbnailExistence"}, // 5.0.0+ + {120, nullptr, "ListOpenUsersInApplication"}, // 10.0.0+ + {130, nullptr, "ActivateOpenContextRetention"}, // 6.0.0+ + {140, &ACC_U1::ListQualifiedUsers, "ListQualifiedUsers"}, // 6.0.0+ + {150, nullptr, "AuthenticateApplicationAsync"}, // 10.0.0+ + {190, nullptr, "GetUserLastOpenedApplication"}, // 1.0.0 - 9.2.0 + {191, nullptr, "ActivateOpenContextHolder"}, // 7.0.0+ + {997, nullptr, "DebugInvalidateTokenCacheForUser"}, // 3.0.0+ {998, nullptr, "DebugSetUserStateClose"}, {999, nullptr, "DebugSetUserStateOpen"}, }; diff --git a/src/core/hle/service/acc/profile_manager.cpp b/src/core/hle/service/acc/profile_manager.cpp index eb8c81645..9b829e957 100644 --- a/src/core/hle/service/acc/profile_manager.cpp +++ b/src/core/hle/service/acc/profile_manager.cpp @@ -13,6 +13,7 @@ namespace Service::Account { +namespace FS = Common::FS; using Common::UUID; struct UserRaw { @@ -58,7 +59,7 @@ ProfileManager::~ProfileManager() { /// internal management of the users profiles std::optional<std::size_t> ProfileManager::AddToProfiles(const ProfileInfo& profile) { if (user_count >= MAX_USERS) { - return {}; + return std::nullopt; } profiles[user_count] = profile; return user_count++; @@ -101,13 +102,14 @@ ResultCode ProfileManager::CreateNewUser(UUID uuid, const ProfileUsername& usern [&uuid](const ProfileInfo& profile) { return uuid == profile.user_uuid; })) { return ERROR_USER_ALREADY_EXISTS; } - ProfileInfo profile; - profile.user_uuid = uuid; - profile.username = username; - profile.data = {}; - profile.creation_time = 0x0; - profile.is_open = false; - return AddUser(profile); + + return AddUser({ + .user_uuid = uuid, + .username = username, + .creation_time = 0, + .data = {}, + .is_open = false, + }); } /// Creates a new user on the system. This function allows a much simpler method of registration @@ -126,7 +128,7 @@ ResultCode ProfileManager::CreateNewUser(UUID uuid, const std::string& username) std::optional<UUID> ProfileManager::GetUser(std::size_t index) const { if (index >= MAX_USERS) { - return {}; + return std::nullopt; } return profiles[index].user_uuid; @@ -135,13 +137,13 @@ std::optional<UUID> ProfileManager::GetUser(std::size_t index) const { /// Returns a users profile index based on their user id. std::optional<std::size_t> ProfileManager::GetUserIndex(const UUID& uuid) const { if (!uuid) { - return {}; + return std::nullopt; } const auto iter = std::find_if(profiles.begin(), profiles.end(), [&uuid](const ProfileInfo& p) { return p.user_uuid == uuid; }); if (iter == profiles.end()) { - return {}; + return std::nullopt; } return static_cast<std::size_t>(std::distance(profiles.begin(), iter)); @@ -317,9 +319,8 @@ bool ProfileManager::SetProfileBaseAndData(Common::UUID uuid, const ProfileBase& } void ProfileManager::ParseUserSaveFile() { - FileUtil::IOFile save(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + - ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat", - "rb"); + const FS::IOFile save( + FS::GetUserPath(FS::UserPath::NANDDir) + ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat", "rb"); if (!save.IsOpen()) { LOG_WARNING(Service_ACC, "Failed to load profile data from save data... Generating new " @@ -339,7 +340,13 @@ void ProfileManager::ParseUserSaveFile() { continue; } - AddUser({user.uuid, user.username, user.timestamp, user.extra_data, false}); + AddUser({ + .user_uuid = user.uuid, + .username = user.username, + .creation_time = user.timestamp, + .data = user.extra_data, + .is_open = false, + }); } std::stable_partition(profiles.begin(), profiles.end(), @@ -350,29 +357,31 @@ void ProfileManager::WriteUserSaveFile() { ProfileDataRaw raw{}; for (std::size_t i = 0; i < MAX_USERS; ++i) { - raw.users[i].username = profiles[i].username; - raw.users[i].uuid2 = profiles[i].user_uuid; - raw.users[i].uuid = profiles[i].user_uuid; - raw.users[i].timestamp = profiles[i].creation_time; - raw.users[i].extra_data = profiles[i].data; + raw.users[i] = { + .uuid = profiles[i].user_uuid, + .uuid2 = profiles[i].user_uuid, + .timestamp = profiles[i].creation_time, + .username = profiles[i].username, + .extra_data = profiles[i].data, + }; } - const auto raw_path = - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000010"; - if (FileUtil::Exists(raw_path) && !FileUtil::IsDirectory(raw_path)) - FileUtil::Delete(raw_path); + const auto raw_path = FS::GetUserPath(FS::UserPath::NANDDir) + "/system/save/8000000000000010"; + if (FS::Exists(raw_path) && !FS::IsDirectory(raw_path)) { + FS::Delete(raw_path); + } - const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + - ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat"; + const auto path = + FS::GetUserPath(FS::UserPath::NANDDir) + ACC_SAVE_AVATORS_BASE_PATH + "profiles.dat"; - if (!FileUtil::CreateFullPath(path)) { + if (!FS::CreateFullPath(path)) { LOG_WARNING(Service_ACC, "Failed to create full path of profiles.dat. Create the directory " "nand/system/save/8000000000000010/su/avators to mitigate this " "issue."); return; } - FileUtil::IOFile save(path, "wb"); + FS::IOFile save(path, "wb"); if (!save.IsOpen()) { LOG_WARNING(Service_ACC, "Failed to write save data to file... No changes to user data " diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index 557608e76..63421b963 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -10,6 +10,7 @@ #include "core/core.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" +#include "core/file_sys/registered_cache.h" #include "core/file_sys/savedata_factory.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/kernel.h" @@ -43,20 +44,15 @@ namespace Service::AM { -constexpr ResultCode ERR_NO_DATA_IN_CHANNEL{ErrorModule::AM, 0x2}; -constexpr ResultCode ERR_NO_MESSAGES{ErrorModule::AM, 0x3}; -constexpr ResultCode ERR_SIZE_OUT_OF_BOUNDS{ErrorModule::AM, 0x1F7}; +constexpr ResultCode ERR_NO_DATA_IN_CHANNEL{ErrorModule::AM, 2}; +constexpr ResultCode ERR_NO_MESSAGES{ErrorModule::AM, 3}; +constexpr ResultCode ERR_SIZE_OUT_OF_BOUNDS{ErrorModule::AM, 503}; enum class LaunchParameterKind : u32 { ApplicationSpecific = 1, AccountPreselectedUser = 2, }; -enum class VrMode : u8 { - Disabled = 0, - Enabled = 1, -}; - constexpr u32 LAUNCH_PARAMETER_ACCOUNT_PRESELECTED_USER_MAGIC = 0xC79497CA; struct LaunchParameterAccountPreselectedUser { @@ -73,6 +69,7 @@ IWindowController::IWindowController(Core::System& system_) static const FunctionInfo functions[] = { {0, nullptr, "CreateWindow"}, {1, &IWindowController::GetAppletResourceUserId, "GetAppletResourceUserId"}, + {2, nullptr, "GetAppletResourceUserIdOfCallerApplet"}, {10, &IWindowController::AcquireForegroundRights, "AcquireForegroundRights"}, {11, nullptr, "ReleaseForegroundRights"}, {12, nullptr, "RejectToChangeIntoBackground"}, @@ -194,8 +191,8 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController" {5, nullptr, "GetLastForegroundCaptureImageEx"}, {6, nullptr, "GetLastApplicationCaptureImageEx"}, {7, nullptr, "GetCallerAppletCaptureImageEx"}, - {8, nullptr, "TakeScreenShotOfOwnLayer"}, // 2.0.0+ - {9, nullptr, "CopyBetweenCaptureBuffers"}, // 5.0.0+ + {8, nullptr, "TakeScreenShotOfOwnLayer"}, + {9, nullptr, "CopyBetweenCaptureBuffers"}, {10, nullptr, "AcquireLastApplicationCaptureBuffer"}, {11, nullptr, "ReleaseLastApplicationCaptureBuffer"}, {12, nullptr, "AcquireLastForegroundCaptureBuffer"}, @@ -205,17 +202,14 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController" {16, nullptr, "AcquireLastApplicationCaptureBufferEx"}, {17, nullptr, "AcquireLastForegroundCaptureBufferEx"}, {18, nullptr, "AcquireCallerAppletCaptureBufferEx"}, - // 2.0.0+ {20, nullptr, "ClearCaptureBuffer"}, {21, nullptr, "ClearAppletTransitionBuffer"}, - // 4.0.0+ {22, nullptr, "AcquireLastApplicationCaptureSharedBuffer"}, {23, nullptr, "ReleaseLastApplicationCaptureSharedBuffer"}, {24, nullptr, "AcquireLastForegroundCaptureSharedBuffer"}, {25, nullptr, "ReleaseLastForegroundCaptureSharedBuffer"}, {26, nullptr, "AcquireCallerAppletCaptureSharedBuffer"}, {27, nullptr, "ReleaseCallerAppletCaptureSharedBuffer"}, - // 6.0.0+ {28, nullptr, "TakeScreenShotOfOwnLayerEx"}, }; // clang-format on @@ -230,11 +224,12 @@ IDebugFunctions::IDebugFunctions() : ServiceFramework{"IDebugFunctions"} { static const FunctionInfo functions[] = { {0, nullptr, "NotifyMessageToHomeMenuForDebug"}, {1, nullptr, "OpenMainApplication"}, - {10, nullptr, "EmulateButtonEvent"}, + {10, nullptr, "PerformSystemButtonPressing"}, {20, nullptr, "InvalidateTransitionLayer"}, {30, nullptr, "RequestLaunchApplicationWithUserAndArgumentForDebug"}, {40, nullptr, "GetAppletResourceUsageInfo"}, {100, nullptr, "SetCpuBoostModeForApplet"}, + {101, nullptr, "CancelCpuBoostModeForApplet"}, {110, nullptr, "PushToAppletBoundChannelForDebug"}, {111, nullptr, "TryPopFromAppletBoundChannelForDebug"}, {120, nullptr, "AlarmSettingNotificationEnableAppEventReserve"}, @@ -271,12 +266,14 @@ ISelfController::ISelfController(Core::System& system, {16, &ISelfController::SetOutOfFocusSuspendingEnabled, "SetOutOfFocusSuspendingEnabled"}, {17, nullptr, "SetControllerFirmwareUpdateSection"}, {18, nullptr, "SetRequiresCaptureButtonShortPressedMessage"}, - {19, &ISelfController::SetScreenShotImageOrientation, "SetScreenShotImageOrientation"}, + {19, &ISelfController::SetAlbumImageOrientation, "SetAlbumImageOrientation"}, {20, nullptr, "SetDesirableKeyboardLayout"}, {40, &ISelfController::CreateManagedDisplayLayer, "CreateManagedDisplayLayer"}, {41, nullptr, "IsSystemBufferSharingEnabled"}, {42, nullptr, "GetSystemSharedLayerHandle"}, {43, nullptr, "GetSystemSharedBufferHandle"}, + {44, &ISelfController::CreateManagedDisplaySeparableLayer, "CreateManagedDisplaySeparableLayer"}, + {45, nullptr, "SetManagedDisplayLayerSeparationMode"}, {50, &ISelfController::SetHandlesRequestToDisplay, "SetHandlesRequestToDisplay"}, {51, nullptr, "ApproveToDisplay"}, {60, nullptr, "OverrideAutoSleepTimeAndDimmingTime"}, @@ -381,7 +378,11 @@ void ISelfController::GetLibraryAppletLaunchableEvent(Kernel::HLERequestContext& } void ISelfController::SetScreenShotPermission(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_AM, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const auto permission = rp.PopEnum<ScreenshotPermission>(); + LOG_DEBUG(Service_AM, "called, permission={}", permission); + + screenshot_permission = permission; IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -445,7 +446,7 @@ void ISelfController::SetOutOfFocusSuspendingEnabled(Kernel::HLERequestContext& rb.Push(RESULT_SUCCESS); } -void ISelfController::SetScreenShotImageOrientation(Kernel::HLERequestContext& ctx) { +void ISelfController::SetAlbumImageOrientation(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_AM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2}; @@ -465,6 +466,24 @@ void ISelfController::CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx) rb.Push(*layer_id); } +void ISelfController::CreateManagedDisplaySeparableLayer(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + // TODO(Subv): Find out how AM determines the display to use, for now just + // create the layer in the Default display. + // This calls nn::vi::CreateRecordingLayer() which creates another layer. + // Currently we do not support more than 1 layer per display, output 1 layer id for now. + // Outputting 1 layer id instead of the expected 2 has not been observed to cause any adverse + // side effects. + // TODO: Support multiple layers + const auto display_id = nvflinger->OpenDisplay("Default"); + const auto layer_id = nvflinger->CreateLayer(*display_id); + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push(*layer_id); +} + void ISelfController::SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_AM, "(STUBBED) called"); @@ -609,6 +628,7 @@ ICommonStateGetter::ICommonStateGetter(Core::System& system, {20, nullptr, "PushToGeneralChannel"}, {30, nullptr, "GetHomeButtonReaderLockAccessor"}, {31, nullptr, "GetReaderLockAccessorEx"}, + {32, nullptr, "GetWriterLockAccessorEx"}, {40, nullptr, "GetCradleFwVersion"}, {50, &ICommonStateGetter::IsVrModeEnabled, "IsVrModeEnabled"}, {51, &ICommonStateGetter::SetVrModeEnabled, "SetVrModeEnabled"}, @@ -623,11 +643,15 @@ ICommonStateGetter::ICommonStateGetter(Core::System& system, {64, nullptr, "SetTvPowerStateMatchingMode"}, {65, nullptr, "GetApplicationIdByContentActionName"}, {66, &ICommonStateGetter::SetCpuBoostMode, "SetCpuBoostMode"}, + {67, nullptr, "CancelCpuBoostMode"}, {80, nullptr, "PerformSystemButtonPressingIfInFocus"}, {90, nullptr, "SetPerformanceConfigurationChangedNotification"}, {91, nullptr, "GetCurrentPerformanceConfiguration"}, + {100, nullptr, "SetHandlingHomeButtonShortPressedEnabled"}, {200, nullptr, "GetOperationModeSystemInfo"}, {300, nullptr, "GetSettingsPlatformRegion"}, + {400, nullptr, "ActivateMigrationService"}, + {401, nullptr, "DeactivateMigrationService"}, }; // clang-format on @@ -678,27 +702,21 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) { } void ICommonStateGetter::IsVrModeEnabled(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.PushEnum(VrMode::Disabled); + rb.Push(vr_mode_state); } void ICommonStateGetter::SetVrModeEnabled(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto is_vr_mode_enabled = rp.Pop<bool>(); + vr_mode_state = rp.Pop<bool>(); - LOG_WARNING(Service_AM, "(STUBBED) called. is_vr_mode_enabled={}", is_vr_mode_enabled); + LOG_WARNING(Service_AM, "VR Mode is {}", vr_mode_state ? "on" : "off"); IPC::ResponseBuilder rb{ctx, 2}; - if (!is_vr_mode_enabled) { - rb.Push(RESULT_SUCCESS); - } else { - // TODO: Find better error code for this - UNIMPLEMENTED_MSG("is_vr_mode_enabled={}", is_vr_mode_enabled); - rb.Push(RESULT_UNKNOWN); - } + rb.Push(RESULT_SUCCESS); } void ICommonStateGetter::SetLcdBacklighOffEnabled(Kernel::HLERequestContext& ctx) { @@ -733,16 +751,16 @@ void ICommonStateGetter::GetDefaultDisplayResolution(Kernel::HLERequestContext& IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); } else { rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); } } @@ -806,7 +824,7 @@ void IStorage::Open(Kernel::HLERequestContext& ctx) { } void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) { - const bool use_docked_mode{Settings::values.use_docked_mode}; + const bool use_docked_mode{Settings::values.use_docked_mode.GetValue()}; LOG_DEBUG(Service_AM, "called, use_docked_mode={}", use_docked_mode); IPC::ResponseBuilder rb{ctx, 3}; @@ -835,6 +853,7 @@ public: {25, nullptr, "Terminate"}, {30, &ILibraryAppletAccessor::GetResult, "GetResult"}, {50, nullptr, "SetOutOfFocusApplicationSuspendingEnabled"}, + {60, nullptr, "PresetLibraryAppletGpuTimeSliceZero"}, {100, &ILibraryAppletAccessor::PushInData, "PushInData"}, {101, &ILibraryAppletAccessor::PopOutData, "PopOutData"}, {102, nullptr, "PushExtraStorage"}, @@ -845,7 +864,7 @@ public: {110, nullptr, "NeedsToExitProcess"}, {120, nullptr, "GetLibraryAppletInfo"}, {150, nullptr, "RequestForAppletToGetForeground"}, - {160, nullptr, "GetIndirectLayerConsumerHandle"}, + {160, &ILibraryAppletAccessor::GetIndirectLayerConsumerHandle, "GetIndirectLayerConsumerHandle"}, }; // clang-format on @@ -903,7 +922,7 @@ private: void PopOutData(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_AM, "called"); - const auto storage = applet->GetBroker().PopNormalDataToGame(); + auto storage = applet->GetBroker().PopNormalDataToGame(); if (storage == nullptr) { LOG_ERROR(Service_AM, "storage is a nullptr. There is no data in the current normal channel"); @@ -934,7 +953,7 @@ private: void PopInteractiveOutData(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_AM, "called"); - const auto storage = applet->GetBroker().PopInteractiveDataToGame(); + auto storage = applet->GetBroker().PopInteractiveDataToGame(); if (storage == nullptr) { LOG_ERROR(Service_AM, "storage is a nullptr. There is no data in the current interactive channel"); @@ -964,6 +983,18 @@ private: rb.PushCopyObjects(applet->GetBroker().GetInteractiveDataEvent()); } + void GetIndirectLayerConsumerHandle(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + // We require a non-zero handle to be valid. Using 0xdeadbeef allows us to trace if this is + // actually used anywhere + constexpr u64 handle = 0xdeadbeef; + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push(handle); + } + std::shared_ptr<Applets::Applet> applet; }; @@ -1135,10 +1166,12 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {24, nullptr, "GetLaunchStorageInfoForDebug"}, {25, &IApplicationFunctions::ExtendSaveData, "ExtendSaveData"}, {26, &IApplicationFunctions::GetSaveDataSize, "GetSaveDataSize"}, + {27, nullptr, "CreateCacheStorage"}, {30, &IApplicationFunctions::BeginBlockingHomeButtonShortAndLongPressed, "BeginBlockingHomeButtonShortAndLongPressed"}, {31, &IApplicationFunctions::EndBlockingHomeButtonShortAndLongPressed, "EndBlockingHomeButtonShortAndLongPressed"}, {32, &IApplicationFunctions::BeginBlockingHomeButton, "BeginBlockingHomeButton"}, {33, &IApplicationFunctions::EndBlockingHomeButton, "EndBlockingHomeButton"}, + {34, nullptr, "SelectApplicationLicense"}, {40, &IApplicationFunctions::NotifyRunning, "NotifyRunning"}, {50, &IApplicationFunctions::GetPseudoDeviceId, "GetPseudoDeviceId"}, {60, nullptr, "SetMediaPlaybackStateForApplication"}, @@ -1148,6 +1181,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {68, nullptr, "RequestFlushGamePlayingMovieForDebug"}, {70, nullptr, "RequestToShutdown"}, {71, nullptr, "RequestToReboot"}, + {72, nullptr, "RequestToSleep"}, {80, nullptr, "ExitAndRequestToShowThanksMessage"}, {90, &IApplicationFunctions::EnableApplicationCrashReport, "EnableApplicationCrashReport"}, {100, &IApplicationFunctions::InitializeApplicationCopyrightFrameBuffer, "InitializeApplicationCopyrightFrameBuffer"}, @@ -1158,13 +1192,17 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {120, nullptr, "ExecuteProgram"}, {121, nullptr, "ClearUserChannel"}, {122, nullptr, "UnpopToUserChannel"}, + {123, &IApplicationFunctions::GetPreviousProgramIndex, "GetPreviousProgramIndex"}, + {124, nullptr, "EnableApplicationAllThreadDumpOnCrash"}, {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"}, - {140, nullptr, "GetFriendInvitationStorageChannelEvent"}, + {140, &IApplicationFunctions::GetFriendInvitationStorageChannelEvent, "GetFriendInvitationStorageChannelEvent"}, {141, nullptr, "TryPopFromFriendInvitationStorageChannel"}, {150, nullptr, "GetNotificationStorageChannelEvent"}, {151, nullptr, "TryPopFromNotificationStorageChannel"}, {160, nullptr, "GetHealthWarningDisappearedSystemEvent"}, {170, nullptr, "SetHdcpAuthenticationActivated"}, + {180, nullptr, "GetLaunchRequiredVersion"}, + {181, nullptr, "UpgradeLaunchRequiredVersion"}, {500, nullptr, "StartContinuousRecordingFlushForDebug"}, {1000, nullptr, "CreateMovieMaker"}, {1001, nullptr, "PrepareForJit"}, @@ -1176,6 +1214,9 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) auto& kernel = system.Kernel(); gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair( kernel, "IApplicationFunctions:GpuErrorDetectedSystemEvent"); + + friend_invitation_storage_channel_event = Kernel::WritableEvent::CreateEventPair( + kernel, "IApplicationFunctions:FriendInvitationStorageChannelEvent"); } IApplicationFunctions::~IApplicationFunctions() = default; @@ -1307,12 +1348,12 @@ void IApplicationFunctions::EnsureSaveData(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_AM, "called, uid={:016X}{:016X}", user_id[1], user_id[0]); - FileSys::SaveDataDescriptor descriptor{}; - descriptor.title_id = system.CurrentProcess()->GetTitleID(); - descriptor.user_id = user_id; - descriptor.type = FileSys::SaveDataType::SaveData; + FileSys::SaveDataAttribute attribute{}; + attribute.title_id = system.CurrentProcess()->GetTitleID(); + attribute.user_id = user_id; + attribute.type = FileSys::SaveDataType::SaveData; const auto res = system.GetFileSystemController().CreateSaveData( - FileSys::SaveDataSpaceId::NandUser, descriptor); + FileSys::SaveDataSpaceId::NandUser, attribute); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(res.Code()); @@ -1333,12 +1374,37 @@ void IApplicationFunctions::SetTerminateResult(Kernel::HLERequestContext& ctx) { } void IApplicationFunctions::GetDisplayVersion(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_AM, "(STUBBED) called"); + LOG_DEBUG(Service_AM, "called"); + + std::array<u8, 0x10> version_string{}; + + const auto res = [this] { + const auto title_id = system.CurrentProcess()->GetTitleID(); + + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; + auto res = pm.GetControlMetadata(); + if (res.first != nullptr) { + return res; + } + + const FileSys::PatchManager pm_update{FileSys::GetUpdateTitleID(title_id), + system.GetFileSystemController(), + system.GetContentProvider()}; + return pm_update.GetControlMetadata(); + }(); + + if (res.first != nullptr) { + const auto& version = res.first->GetVersionString(); + std::copy(version.begin(), version.end(), version_string.begin()); + } else { + constexpr char default_version[]{"1.0.0"}; + std::memcpy(version_string.data(), default_version, sizeof(default_version)); + } IPC::ResponseBuilder rb{ctx, 6}; rb.Push(RESULT_SUCCESS); - rb.Push<u64>(1); - rb.Push<u64>(0); + rb.PushRaw(version_string); } void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) { @@ -1348,9 +1414,23 @@ void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) { // Get supported languages from NACP, if possible // Default to 0 (all languages supported) u32 supported_languages = 0; - FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()}; - const auto res = pm.GetControlMetadata(); + const auto res = [this] { + const auto title_id = system.CurrentProcess()->GetTitleID(); + + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; + auto res = pm.GetControlMetadata(); + if (res.first != nullptr) { + return res; + } + + const FileSys::PatchManager pm_update{FileSys::GetUpdateTitleID(title_id), + system.GetFileSystemController(), + system.GetContentProvider()}; + return pm_update.GetControlMetadata(); + }(); + if (res.first != nullptr) { supported_languages = res.first->GetSupportedLanguages(); } @@ -1482,6 +1562,14 @@ void IApplicationFunctions::QueryApplicationPlayStatisticsByUid(Kernel::HLEReque rb.Push<u32>(0); } +void IApplicationFunctions::GetPreviousProgramIndex(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(previous_program_index); +} + void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_AM, "(STUBBED) called"); @@ -1490,6 +1578,14 @@ void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestCon rb.PushCopyObjects(gpu_error_detected_event.readable); } +void IApplicationFunctions::GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_AM, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushCopyObjects(friend_invitation_storage_channel_event.readable); +} + void InstallInterfaces(SM::ServiceManager& service_manager, std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) { auto message_queue = std::make_shared<AppletMessageQueue>(system.Kernel()); @@ -1504,14 +1600,15 @@ void InstallInterfaces(SM::ServiceManager& service_manager, std::make_shared<TCAP>()->InstallAsService(service_manager); } -IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions") { +IHomeMenuFunctions::IHomeMenuFunctions(Kernel::KernelCore& kernel) + : ServiceFramework("IHomeMenuFunctions"), kernel(kernel) { // clang-format off static const FunctionInfo functions[] = { {10, &IHomeMenuFunctions::RequestToGetForeground, "RequestToGetForeground"}, {11, nullptr, "LockForeground"}, {12, nullptr, "UnlockForeground"}, {20, nullptr, "PopFromGeneralChannel"}, - {21, nullptr, "GetPopFromGeneralChannelEvent"}, + {21, &IHomeMenuFunctions::GetPopFromGeneralChannelEvent, "GetPopFromGeneralChannelEvent"}, {30, nullptr, "GetHomeButtonWriterLockAccessor"}, {31, nullptr, "GetWriterLockAccessorEx"}, {100, nullptr, "PopRequestLaunchApplicationForDebug"}, @@ -1521,6 +1618,9 @@ IHomeMenuFunctions::IHomeMenuFunctions() : ServiceFramework("IHomeMenuFunctions" // clang-format on RegisterHandlers(functions); + + pop_from_general_channel_event = Kernel::WritableEvent::CreateEventPair( + kernel, "IHomeMenuFunctions:PopFromGeneralChannelEvent"); } IHomeMenuFunctions::~IHomeMenuFunctions() = default; @@ -1532,6 +1632,14 @@ void IHomeMenuFunctions::RequestToGetForeground(Kernel::HLERequestContext& ctx) rb.Push(RESULT_SUCCESS); } +void IHomeMenuFunctions::GetPopFromGeneralChannelEvent(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushCopyObjects(pop_from_general_channel_event.readable); +} + IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStateController") { // clang-format off static const FunctionInfo functions[] = { diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h index 53cfce10f..bcc06affe 100644 --- a/src/core/hle/service/am/am.h +++ b/src/core/hle/service/am/am.h @@ -138,8 +138,9 @@ private: void SetFocusHandlingMode(Kernel::HLERequestContext& ctx); void SetRestartMessageEnabled(Kernel::HLERequestContext& ctx); void SetOutOfFocusSuspendingEnabled(Kernel::HLERequestContext& ctx); - void SetScreenShotImageOrientation(Kernel::HLERequestContext& ctx); + void SetAlbumImageOrientation(Kernel::HLERequestContext& ctx); void CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx); + void CreateManagedDisplaySeparableLayer(Kernel::HLERequestContext& ctx); void SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx); void SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx); void GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx); @@ -148,6 +149,12 @@ private: void GetAccumulatedSuspendedTickValue(Kernel::HLERequestContext& ctx); void GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequestContext& ctx); + enum class ScreenshotPermission : u32 { + Inherit = 0, + Enable = 1, + Disable = 2, + }; + Core::System& system; std::shared_ptr<NVFlinger::NVFlinger> nvflinger; Kernel::EventPair launchable_event; @@ -156,6 +163,7 @@ private: u32 idle_time_detection_extension = 0; u64 num_fatal_sections_entered = 0; bool is_auto_sleep_disabled = false; + ScreenshotPermission screenshot_permission = ScreenshotPermission::Inherit; }; class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> { @@ -191,6 +199,7 @@ private: Core::System& system; std::shared_ptr<AppletMessageQueue> msg_queue; + bool vr_mode_state{}; }; class IStorageImpl { @@ -279,21 +288,29 @@ private: void SetApplicationCopyrightVisibility(Kernel::HLERequestContext& ctx); void QueryApplicationPlayStatistics(Kernel::HLERequestContext& ctx); void QueryApplicationPlayStatisticsByUid(Kernel::HLERequestContext& ctx); + void GetPreviousProgramIndex(Kernel::HLERequestContext& ctx); void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx); + void GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx); bool launch_popped_application_specific = false; bool launch_popped_account_preselect = false; + s32 previous_program_index{-1}; Kernel::EventPair gpu_error_detected_event; + Kernel::EventPair friend_invitation_storage_channel_event; Core::System& system; }; class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> { public: - IHomeMenuFunctions(); + explicit IHomeMenuFunctions(Kernel::KernelCore& kernel); ~IHomeMenuFunctions() override; private: void RequestToGetForeground(Kernel::HLERequestContext& ctx); + void GetPopFromGeneralChannelEvent(Kernel::HLERequestContext& ctx); + + Kernel::EventPair pop_from_general_channel_event; + Kernel::KernelCore& kernel; }; class IGlobalStateController final : public ServiceFramework<IGlobalStateController> { diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp index e454b77d8..be23ca747 100644 --- a/src/core/hle/service/am/applet_ae.cpp +++ b/src/core/hle/service/am/applet_ae.cpp @@ -3,8 +3,8 @@ // Refer to the license.txt file included. #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" -#include "core/hle/kernel/process.h" #include "core/hle/service/am/am.h" #include "core/hle/service/am/applet_ae.h" #include "core/hle/service/nvflinger/nvflinger.h" @@ -202,7 +202,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IHomeMenuFunctions>(); + rb.PushIpcInterface<IHomeMenuFunctions>(system.Kernel()); } void GetGlobalStateController(Kernel::HLERequestContext& ctx) { diff --git a/src/core/hle/service/am/applets/applets.cpp b/src/core/hle/service/am/applets/applets.cpp index c3261f3e6..2b626bb40 100644 --- a/src/core/hle/service/am/applets/applets.cpp +++ b/src/core/hle/service/am/applets/applets.cpp @@ -5,6 +5,7 @@ #include <cstring> #include "common/assert.h" #include "core/core.h" +#include "core/frontend/applets/controller.h" #include "core/frontend/applets/error.h" #include "core/frontend/applets/general_frontend.h" #include "core/frontend/applets/profile_select.h" @@ -15,6 +16,7 @@ #include "core/hle/kernel/writable_event.h" #include "core/hle/service/am/am.h" #include "core/hle/service/am/applets/applets.h" +#include "core/hle/service/am/applets/controller.h" #include "core/hle/service/am/applets/error.h" #include "core/hle/service/am/applets/general_backend.h" #include "core/hle/service/am/applets/profile_select.h" @@ -140,14 +142,14 @@ void Applet::Initialize() { AppletFrontendSet::AppletFrontendSet() = default; -AppletFrontendSet::AppletFrontendSet(ParentalControlsApplet parental_controls, ErrorApplet error, +AppletFrontendSet::AppletFrontendSet(ControllerApplet controller, ECommerceApplet e_commerce, + ErrorApplet error, ParentalControlsApplet parental_controls, PhotoViewer photo_viewer, ProfileSelect profile_select, - SoftwareKeyboard software_keyboard, WebBrowser web_browser, - ECommerceApplet e_commerce) - : parental_controls{std::move(parental_controls)}, error{std::move(error)}, - photo_viewer{std::move(photo_viewer)}, profile_select{std::move(profile_select)}, - software_keyboard{std::move(software_keyboard)}, web_browser{std::move(web_browser)}, - e_commerce{std::move(e_commerce)} {} + SoftwareKeyboard software_keyboard, WebBrowser web_browser) + : controller{std::move(controller)}, e_commerce{std::move(e_commerce)}, error{std::move(error)}, + parental_controls{std::move(parental_controls)}, photo_viewer{std::move(photo_viewer)}, + profile_select{std::move(profile_select)}, software_keyboard{std::move(software_keyboard)}, + web_browser{std::move(web_browser)} {} AppletFrontendSet::~AppletFrontendSet() = default; @@ -164,20 +166,37 @@ const AppletFrontendSet& AppletManager::GetAppletFrontendSet() const { } void AppletManager::SetAppletFrontendSet(AppletFrontendSet set) { - if (set.parental_controls != nullptr) - frontend.parental_controls = std::move(set.parental_controls); - if (set.error != nullptr) + if (set.controller != nullptr) { + frontend.controller = std::move(set.controller); + } + + if (set.e_commerce != nullptr) { + frontend.e_commerce = std::move(set.e_commerce); + } + + if (set.error != nullptr) { frontend.error = std::move(set.error); - if (set.photo_viewer != nullptr) + } + + if (set.parental_controls != nullptr) { + frontend.parental_controls = std::move(set.parental_controls); + } + + if (set.photo_viewer != nullptr) { frontend.photo_viewer = std::move(set.photo_viewer); - if (set.profile_select != nullptr) + } + + if (set.profile_select != nullptr) { frontend.profile_select = std::move(set.profile_select); - if (set.software_keyboard != nullptr) + } + + if (set.software_keyboard != nullptr) { frontend.software_keyboard = std::move(set.software_keyboard); - if (set.web_browser != nullptr) + } + + if (set.web_browser != nullptr) { frontend.web_browser = std::move(set.web_browser); - if (set.e_commerce != nullptr) - frontend.e_commerce = std::move(set.e_commerce); + } } void AppletManager::SetDefaultAppletFrontendSet() { @@ -186,15 +205,24 @@ void AppletManager::SetDefaultAppletFrontendSet() { } void AppletManager::SetDefaultAppletsIfMissing() { - if (frontend.parental_controls == nullptr) { - frontend.parental_controls = - std::make_unique<Core::Frontend::DefaultParentalControlsApplet>(); + if (frontend.controller == nullptr) { + frontend.controller = + std::make_unique<Core::Frontend::DefaultControllerApplet>(system.ServiceManager()); + } + + if (frontend.e_commerce == nullptr) { + frontend.e_commerce = std::make_unique<Core::Frontend::DefaultECommerceApplet>(); } if (frontend.error == nullptr) { frontend.error = std::make_unique<Core::Frontend::DefaultErrorApplet>(); } + if (frontend.parental_controls == nullptr) { + frontend.parental_controls = + std::make_unique<Core::Frontend::DefaultParentalControlsApplet>(); + } + if (frontend.photo_viewer == nullptr) { frontend.photo_viewer = std::make_unique<Core::Frontend::DefaultPhotoViewerApplet>(); } @@ -211,10 +239,6 @@ void AppletManager::SetDefaultAppletsIfMissing() { if (frontend.web_browser == nullptr) { frontend.web_browser = std::make_unique<Core::Frontend::DefaultWebBrowserApplet>(); } - - if (frontend.e_commerce == nullptr) { - frontend.e_commerce = std::make_unique<Core::Frontend::DefaultECommerceApplet>(); - } } void AppletManager::ClearAll() { @@ -225,6 +249,8 @@ std::shared_ptr<Applet> AppletManager::GetApplet(AppletId id) const { switch (id) { case AppletId::Auth: return std::make_shared<Auth>(system, *frontend.parental_controls); + case AppletId::Controller: + return std::make_shared<Controller>(system, *frontend.controller); case AppletId::Error: return std::make_shared<Error>(system, *frontend.error); case AppletId::ProfileSelect: diff --git a/src/core/hle/service/am/applets/applets.h b/src/core/hle/service/am/applets/applets.h index e75be86a2..a1f4cf897 100644 --- a/src/core/hle/service/am/applets/applets.h +++ b/src/core/hle/service/am/applets/applets.h @@ -17,6 +17,7 @@ class System; } namespace Core::Frontend { +class ControllerApplet; class ECommerceApplet; class ErrorApplet; class ParentalControlsApplet; @@ -155,19 +156,20 @@ protected: }; struct AppletFrontendSet { - using ParentalControlsApplet = std::unique_ptr<Core::Frontend::ParentalControlsApplet>; + using ControllerApplet = std::unique_ptr<Core::Frontend::ControllerApplet>; + using ECommerceApplet = std::unique_ptr<Core::Frontend::ECommerceApplet>; using ErrorApplet = std::unique_ptr<Core::Frontend::ErrorApplet>; + using ParentalControlsApplet = std::unique_ptr<Core::Frontend::ParentalControlsApplet>; using PhotoViewer = std::unique_ptr<Core::Frontend::PhotoViewerApplet>; using ProfileSelect = std::unique_ptr<Core::Frontend::ProfileSelectApplet>; using SoftwareKeyboard = std::unique_ptr<Core::Frontend::SoftwareKeyboardApplet>; using WebBrowser = std::unique_ptr<Core::Frontend::WebBrowserApplet>; - using ECommerceApplet = std::unique_ptr<Core::Frontend::ECommerceApplet>; AppletFrontendSet(); - AppletFrontendSet(ParentalControlsApplet parental_controls, ErrorApplet error, - PhotoViewer photo_viewer, ProfileSelect profile_select, - SoftwareKeyboard software_keyboard, WebBrowser web_browser, - ECommerceApplet e_commerce); + AppletFrontendSet(ControllerApplet controller, ECommerceApplet e_commerce, ErrorApplet error, + ParentalControlsApplet parental_controls, PhotoViewer photo_viewer, + ProfileSelect profile_select, SoftwareKeyboard software_keyboard, + WebBrowser web_browser); ~AppletFrontendSet(); AppletFrontendSet(const AppletFrontendSet&) = delete; @@ -176,13 +178,14 @@ struct AppletFrontendSet { AppletFrontendSet(AppletFrontendSet&&) noexcept; AppletFrontendSet& operator=(AppletFrontendSet&&) noexcept; - ParentalControlsApplet parental_controls; + ControllerApplet controller; + ECommerceApplet e_commerce; ErrorApplet error; + ParentalControlsApplet parental_controls; PhotoViewer photo_viewer; ProfileSelect profile_select; SoftwareKeyboard software_keyboard; WebBrowser web_browser; - ECommerceApplet e_commerce; }; class AppletManager { diff --git a/src/core/hle/service/am/applets/controller.cpp b/src/core/hle/service/am/applets/controller.cpp new file mode 100644 index 000000000..3ca63f020 --- /dev/null +++ b/src/core/hle/service/am/applets/controller.cpp @@ -0,0 +1,252 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <cstring> + +#include "common/assert.h" +#include "common/logging/log.h" +#include "common/string_util.h" +#include "core/core.h" +#include "core/frontend/applets/controller.h" +#include "core/hle/result.h" +#include "core/hle/service/am/am.h" +#include "core/hle/service/am/applets/controller.h" +#include "core/hle/service/hid/controllers/npad.h" + +namespace Service::AM::Applets { + +// This error code (0x183ACA) is thrown when the applet fails to initialize. +[[maybe_unused]] constexpr ResultCode ERR_CONTROLLER_APPLET_3101{ErrorModule::HID, 3101}; +// This error code (0x183CCA) is thrown when the u32 result in ControllerSupportResultInfo is 2. +[[maybe_unused]] constexpr ResultCode ERR_CONTROLLER_APPLET_3102{ErrorModule::HID, 3102}; + +static Core::Frontend::ControllerParameters ConvertToFrontendParameters( + ControllerSupportArgPrivate private_arg, ControllerSupportArgHeader header, bool enable_text, + std::vector<IdentificationColor> identification_colors, std::vector<ExplainText> text) { + HID::Controller_NPad::NpadStyleSet npad_style_set; + npad_style_set.raw = private_arg.style_set; + + return { + .min_players = std::max(s8(1), header.player_count_min), + .max_players = header.player_count_max, + .keep_controllers_connected = header.enable_take_over_connection, + .enable_single_mode = header.enable_single_mode, + .enable_border_color = header.enable_identification_color, + .border_colors = identification_colors, + .enable_explain_text = enable_text, + .explain_text = text, + .allow_pro_controller = npad_style_set.pro_controller == 1, + .allow_handheld = npad_style_set.handheld == 1, + .allow_dual_joycons = npad_style_set.joycon_dual == 1, + .allow_left_joycon = npad_style_set.joycon_left == 1, + .allow_right_joycon = npad_style_set.joycon_right == 1, + }; +} + +Controller::Controller(Core::System& system_, const Core::Frontend::ControllerApplet& frontend_) + : Applet{system_.Kernel()}, frontend(frontend_) {} + +Controller::~Controller() = default; + +void Controller::Initialize() { + Applet::Initialize(); + + LOG_INFO(Service_HID, "Initializing Controller Applet."); + + LOG_DEBUG(Service_HID, + "Initializing Applet with common_args: arg_version={}, lib_version={}, " + "play_startup_sound={}, size={}, system_tick={}, theme_color={}", + common_args.arguments_version, common_args.library_version, + common_args.play_startup_sound, common_args.size, common_args.system_tick, + common_args.theme_color); + + controller_applet_version = ControllerAppletVersion{common_args.library_version}; + + const auto private_arg_storage = broker.PopNormalDataToApplet(); + ASSERT(private_arg_storage != nullptr); + + const auto& private_arg = private_arg_storage->GetData(); + ASSERT(private_arg.size() == sizeof(ControllerSupportArgPrivate)); + + std::memcpy(&controller_private_arg, private_arg.data(), private_arg.size()); + ASSERT_MSG(controller_private_arg.arg_private_size == sizeof(ControllerSupportArgPrivate), + "Unknown ControllerSupportArgPrivate revision={} with size={}", + controller_applet_version, controller_private_arg.arg_private_size); + + // Some games such as Cave Story+ set invalid values for the ControllerSupportMode. + // Defer to arg_size to set the ControllerSupportMode. + if (controller_private_arg.mode >= ControllerSupportMode::MaxControllerSupportMode) { + switch (controller_private_arg.arg_size) { + case sizeof(ControllerSupportArgOld): + case sizeof(ControllerSupportArgNew): + controller_private_arg.mode = ControllerSupportMode::ShowControllerSupport; + break; + case sizeof(ControllerUpdateFirmwareArg): + controller_private_arg.mode = ControllerSupportMode::ShowControllerFirmwareUpdate; + break; + default: + UNIMPLEMENTED_MSG("Unknown ControllerPrivateArg mode={} with arg_size={}", + controller_private_arg.mode, controller_private_arg.arg_size); + controller_private_arg.mode = ControllerSupportMode::ShowControllerSupport; + break; + } + } + + // Some games such as Cave Story+ set invalid values for the ControllerSupportCaller. + // This is always 0 (Application) except with ShowControllerFirmwareUpdateForSystem. + if (controller_private_arg.caller >= ControllerSupportCaller::MaxControllerSupportCaller) { + if (controller_private_arg.flag_1 && + controller_private_arg.mode == ControllerSupportMode::ShowControllerFirmwareUpdate) { + controller_private_arg.caller = ControllerSupportCaller::System; + } else { + controller_private_arg.caller = ControllerSupportCaller::Application; + } + } + + switch (controller_private_arg.mode) { + case ControllerSupportMode::ShowControllerSupport: + case ControllerSupportMode::ShowControllerStrapGuide: { + const auto user_arg_storage = broker.PopNormalDataToApplet(); + ASSERT(user_arg_storage != nullptr); + + const auto& user_arg = user_arg_storage->GetData(); + switch (controller_applet_version) { + case ControllerAppletVersion::Version3: + case ControllerAppletVersion::Version4: + case ControllerAppletVersion::Version5: + ASSERT(user_arg.size() == sizeof(ControllerSupportArgOld)); + std::memcpy(&controller_user_arg_old, user_arg.data(), user_arg.size()); + break; + case ControllerAppletVersion::Version7: + ASSERT(user_arg.size() == sizeof(ControllerSupportArgNew)); + std::memcpy(&controller_user_arg_new, user_arg.data(), user_arg.size()); + break; + default: + UNIMPLEMENTED_MSG("Unknown ControllerSupportArg revision={} with size={}", + controller_applet_version, controller_private_arg.arg_size); + ASSERT(user_arg.size() >= sizeof(ControllerSupportArgNew)); + std::memcpy(&controller_user_arg_new, user_arg.data(), sizeof(ControllerSupportArgNew)); + break; + } + break; + } + case ControllerSupportMode::ShowControllerFirmwareUpdate: { + const auto update_arg_storage = broker.PopNormalDataToApplet(); + ASSERT(update_arg_storage != nullptr); + + const auto& update_arg = update_arg_storage->GetData(); + ASSERT(update_arg.size() == sizeof(ControllerUpdateFirmwareArg)); + + std::memcpy(&controller_update_arg, update_arg.data(), update_arg.size()); + break; + } + default: { + UNIMPLEMENTED_MSG("Unimplemented ControllerSupportMode={}", controller_private_arg.mode); + break; + } + } +} + +bool Controller::TransactionComplete() const { + return complete; +} + +ResultCode Controller::GetStatus() const { + return status; +} + +void Controller::ExecuteInteractive() { + UNREACHABLE_MSG("Attempted to call interactive execution on non-interactive applet."); +} + +void Controller::Execute() { + switch (controller_private_arg.mode) { + case ControllerSupportMode::ShowControllerSupport: { + const auto parameters = [this] { + switch (controller_applet_version) { + case ControllerAppletVersion::Version3: + case ControllerAppletVersion::Version4: + case ControllerAppletVersion::Version5: + return ConvertToFrontendParameters( + controller_private_arg, controller_user_arg_old.header, + controller_user_arg_old.enable_explain_text, + std::vector<IdentificationColor>( + controller_user_arg_old.identification_colors.begin(), + controller_user_arg_old.identification_colors.end()), + std::vector<ExplainText>(controller_user_arg_old.explain_text.begin(), + controller_user_arg_old.explain_text.end())); + case ControllerAppletVersion::Version7: + default: + return ConvertToFrontendParameters( + controller_private_arg, controller_user_arg_new.header, + controller_user_arg_new.enable_explain_text, + std::vector<IdentificationColor>( + controller_user_arg_new.identification_colors.begin(), + controller_user_arg_new.identification_colors.end()), + std::vector<ExplainText>(controller_user_arg_new.explain_text.begin(), + controller_user_arg_new.explain_text.end())); + } + }(); + + is_single_mode = parameters.enable_single_mode; + + LOG_DEBUG(Service_HID, + "Controller Parameters: min_players={}, max_players={}, " + "keep_controllers_connected={}, enable_single_mode={}, enable_border_color={}, " + "enable_explain_text={}, allow_pro_controller={}, allow_handheld={}, " + "allow_dual_joycons={}, allow_left_joycon={}, allow_right_joycon={}", + parameters.min_players, parameters.max_players, + parameters.keep_controllers_connected, parameters.enable_single_mode, + parameters.enable_border_color, parameters.enable_explain_text, + parameters.allow_pro_controller, parameters.allow_handheld, + parameters.allow_dual_joycons, parameters.allow_left_joycon, + parameters.allow_right_joycon); + + frontend.ReconfigureControllers([this] { ConfigurationComplete(); }, parameters); + break; + } + case ControllerSupportMode::ShowControllerStrapGuide: + case ControllerSupportMode::ShowControllerFirmwareUpdate: + UNIMPLEMENTED_MSG("ControllerSupportMode={} is not implemented", + controller_private_arg.mode); + [[fallthrough]]; + default: { + ConfigurationComplete(); + break; + } + } +} + +void Controller::ConfigurationComplete() { + ControllerSupportResultInfo result_info{}; + + const auto& players = Settings::values.players.GetValue(); + + // If enable_single_mode is enabled, player_count is 1 regardless of any other parameters. + // Otherwise, only count connected players from P1-P8. + result_info.player_count = + is_single_mode ? 1 + : static_cast<s8>(std::count_if( + players.begin(), players.end() - 2, + [](Settings::PlayerInput player) { return player.connected; })); + + result_info.selected_id = HID::Controller_NPad::IndexToNPad( + std::distance(players.begin(), + std::find_if(players.begin(), players.end(), + [](Settings::PlayerInput player) { return player.connected; }))); + + result_info.result = 0; + + LOG_DEBUG(Service_HID, "Result Info: player_count={}, selected_id={}, result={}", + result_info.player_count, result_info.selected_id, result_info.result); + + complete = true; + out_data = std::vector<u8>(sizeof(ControllerSupportResultInfo)); + std::memcpy(out_data.data(), &result_info, out_data.size()); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(out_data))); + broker.SignalStateChanged(); +} + +} // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/applets/controller.h b/src/core/hle/service/am/applets/controller.h new file mode 100644 index 000000000..a7a1f2b65 --- /dev/null +++ b/src/core/hle/service/am/applets/controller.h @@ -0,0 +1,135 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <vector> + +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "core/hle/result.h" +#include "core/hle/service/am/applets/applets.h" + +namespace Core { +class System; +} + +namespace Service::AM::Applets { + +using IdentificationColor = std::array<u8, 4>; +using ExplainText = std::array<char, 0x81>; + +enum class ControllerAppletVersion : u32_le { + Version3 = 0x3, // 1.0.0 - 2.3.0 + Version4 = 0x4, // 3.0.0 - 5.1.0 + Version5 = 0x5, // 6.0.0 - 7.0.1 + Version7 = 0x7, // 8.0.0+ +}; + +enum class ControllerSupportMode : u8 { + ShowControllerSupport, + ShowControllerStrapGuide, + ShowControllerFirmwareUpdate, + + MaxControllerSupportMode, +}; + +enum class ControllerSupportCaller : u8 { + Application, + System, + + MaxControllerSupportCaller, +}; + +struct ControllerSupportArgPrivate { + u32 arg_private_size{}; + u32 arg_size{}; + bool flag_0{}; + bool flag_1{}; + ControllerSupportMode mode{}; + ControllerSupportCaller caller{}; + u32 style_set{}; + u32 joy_hold_type{}; +}; +static_assert(sizeof(ControllerSupportArgPrivate) == 0x14, + "ControllerSupportArgPrivate has incorrect size."); + +struct ControllerSupportArgHeader { + s8 player_count_min{}; + s8 player_count_max{}; + bool enable_take_over_connection{}; + bool enable_left_justify{}; + bool enable_permit_joy_dual{}; + bool enable_single_mode{}; + bool enable_identification_color{}; +}; +static_assert(sizeof(ControllerSupportArgHeader) == 0x7, + "ControllerSupportArgHeader has incorrect size."); + +// LibraryAppletVersion 0x3, 0x4, 0x5 +struct ControllerSupportArgOld { + ControllerSupportArgHeader header{}; + std::array<IdentificationColor, 4> identification_colors{}; + bool enable_explain_text{}; + std::array<ExplainText, 4> explain_text{}; +}; +static_assert(sizeof(ControllerSupportArgOld) == 0x21C, + "ControllerSupportArgOld has incorrect size."); + +// LibraryAppletVersion 0x7 +struct ControllerSupportArgNew { + ControllerSupportArgHeader header{}; + std::array<IdentificationColor, 8> identification_colors{}; + bool enable_explain_text{}; + std::array<ExplainText, 8> explain_text{}; +}; +static_assert(sizeof(ControllerSupportArgNew) == 0x430, + "ControllerSupportArgNew has incorrect size."); + +struct ControllerUpdateFirmwareArg { + bool enable_force_update{}; + INSERT_PADDING_BYTES(3); +}; +static_assert(sizeof(ControllerUpdateFirmwareArg) == 0x4, + "ControllerUpdateFirmwareArg has incorrect size."); + +struct ControllerSupportResultInfo { + s8 player_count{}; + INSERT_PADDING_BYTES(3); + u32 selected_id{}; + u32 result{}; +}; +static_assert(sizeof(ControllerSupportResultInfo) == 0xC, + "ControllerSupportResultInfo has incorrect size."); + +class Controller final : public Applet { +public: + explicit Controller(Core::System& system_, const Core::Frontend::ControllerApplet& frontend_); + ~Controller() override; + + void Initialize() override; + + bool TransactionComplete() const override; + ResultCode GetStatus() const override; + void ExecuteInteractive() override; + void Execute() override; + + void ConfigurationComplete(); + +private: + const Core::Frontend::ControllerApplet& frontend; + + ControllerAppletVersion controller_applet_version; + ControllerSupportArgPrivate controller_private_arg; + ControllerSupportArgOld controller_user_arg_old; + ControllerSupportArgNew controller_user_arg_new; + ControllerUpdateFirmwareArg controller_update_arg; + bool complete{false}; + ResultCode status{RESULT_SUCCESS}; + bool is_single_mode{false}; + std::vector<u8> out_data; +}; + +} // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/applets/software_keyboard.cpp b/src/core/hle/service/am/applets/software_keyboard.cpp index 54e63c138..bdeb0737a 100644 --- a/src/core/hle/service/am/applets/software_keyboard.cpp +++ b/src/core/hle/service/am/applets/software_keyboard.cpp @@ -13,11 +13,23 @@ namespace Service::AM::Applets { +namespace { +enum class Request : u32 { + Finalize = 0x4, + SetUserWordInfo = 0x6, + SetCustomizeDic = 0x7, + Calc = 0xa, + SetCustomizedDictionaries = 0xb, + UnsetCustomizedDictionaries = 0xc, + UnknownD = 0xd, + UnknownE = 0xe, +}; +constexpr std::size_t SWKBD_INLINE_INIT_SIZE = 0x8; constexpr std::size_t SWKBD_OUTPUT_BUFFER_SIZE = 0x7D8; constexpr std::size_t SWKBD_OUTPUT_INTERACTIVE_BUFFER_SIZE = 0x7D4; constexpr std::size_t DEFAULT_MAX_LENGTH = 500; constexpr bool INTERACTIVE_STATUS_OK = false; - +} // Anonymous namespace static Core::Frontend::SoftwareKeyboardParameters ConvertToFrontendParameters( KeyboardConfig config, std::u16string initial_text) { Core::Frontend::SoftwareKeyboardParameters params{}; @@ -30,7 +42,7 @@ static Core::Frontend::SoftwareKeyboardParameters ConvertToFrontendParameters( config.sub_text.size()); params.guide_text = Common::UTF16StringFromFixedZeroTerminatedBuffer(config.guide_text.data(), config.guide_text.size()); - params.initial_text = initial_text; + params.initial_text = std::move(initial_text); params.max_length = config.length_limit == 0 ? DEFAULT_MAX_LENGTH : config.length_limit; params.password = static_cast<bool>(config.is_password); params.cursor_at_beginning = static_cast<bool>(config.initial_cursor_position); @@ -47,6 +59,7 @@ SoftwareKeyboard::~SoftwareKeyboard() = default; void SoftwareKeyboard::Initialize() { complete = false; + is_inline = false; initial_text.clear(); final_data.clear(); @@ -56,11 +69,16 @@ void SoftwareKeyboard::Initialize() { ASSERT(keyboard_config_storage != nullptr); const auto& keyboard_config = keyboard_config_storage->GetData(); + if (keyboard_config.size() == SWKBD_INLINE_INIT_SIZE) { + is_inline = true; + return; + } + ASSERT(keyboard_config.size() >= sizeof(KeyboardConfig)); std::memcpy(&config, keyboard_config.data(), sizeof(KeyboardConfig)); const auto work_buffer_storage = broker.PopNormalDataToApplet(); - ASSERT(work_buffer_storage != nullptr); + ASSERT_OR_EXECUTE(work_buffer_storage != nullptr, { return; }); const auto& work_buffer = work_buffer_storage->GetData(); if (config.initial_string_size == 0) @@ -87,16 +105,31 @@ void SoftwareKeyboard::ExecuteInteractive() { const auto storage = broker.PopInteractiveDataToApplet(); ASSERT(storage != nullptr); const auto data = storage->GetData(); - const auto status = static_cast<bool>(data[0]); - - if (status == INTERACTIVE_STATUS_OK) { - complete = true; + if (!is_inline) { + const auto status = static_cast<bool>(data[0]); + if (status == INTERACTIVE_STATUS_OK) { + complete = true; + } else { + std::array<char16_t, SWKBD_OUTPUT_INTERACTIVE_BUFFER_SIZE / 2 - 2> string; + std::memcpy(string.data(), data.data() + 4, string.size() * 2); + frontend.SendTextCheckDialog( + Common::UTF16StringFromFixedZeroTerminatedBuffer(string.data(), string.size()), + [this] { broker.SignalStateChanged(); }); + } } else { - std::array<char16_t, SWKBD_OUTPUT_INTERACTIVE_BUFFER_SIZE / 2 - 2> string; - std::memcpy(string.data(), data.data() + 4, string.size() * 2); - frontend.SendTextCheckDialog( - Common::UTF16StringFromFixedZeroTerminatedBuffer(string.data(), string.size()), - [this] { broker.SignalStateChanged(); }); + Request request{}; + std::memcpy(&request, data.data(), sizeof(Request)); + + switch (request) { + case Request::Calc: { + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::vector<u8>{1})); + broker.SignalStateChanged(); + break; + } + default: + UNIMPLEMENTED_MSG("Request {:X} is not implemented", request); + break; + } } } @@ -108,9 +141,10 @@ void SoftwareKeyboard::Execute() { } const auto parameters = ConvertToFrontendParameters(config, initial_text); - - frontend.RequestText([this](std::optional<std::u16string> text) { WriteText(text); }, - parameters); + if (!is_inline) { + frontend.RequestText( + [this](std::optional<std::u16string> text) { WriteText(std::move(text)); }, parameters); + } } void SoftwareKeyboard::WriteText(std::optional<std::u16string> text) { diff --git a/src/core/hle/service/am/applets/software_keyboard.h b/src/core/hle/service/am/applets/software_keyboard.h index ef4801fc6..5a3824b5a 100644 --- a/src/core/hle/service/am/applets/software_keyboard.h +++ b/src/core/hle/service/am/applets/software_keyboard.h @@ -78,6 +78,7 @@ private: KeyboardConfig config; std::u16string initial_text; bool complete = false; + bool is_inline = false; std::vector<u8> final_data; }; diff --git a/src/core/hle/service/am/applets/web_browser.cpp b/src/core/hle/service/am/applets/web_browser.cpp index 9f30e167d..efe595c4f 100644 --- a/src/core/hle/service/am/applets/web_browser.cpp +++ b/src/core/hle/service/am/applets/web_browser.cpp @@ -293,8 +293,8 @@ void WebBrowser::Finalize() { broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(data))); broker.SignalStateChanged(); - if (!temporary_dir.empty() && FileUtil::IsDirectory(temporary_dir)) { - FileUtil::DeleteDirRecursively(temporary_dir); + if (!temporary_dir.empty() && Common::FS::IsDirectory(temporary_dir)) { + Common::FS::DeleteDirRecursively(temporary_dir); } } @@ -452,10 +452,10 @@ void WebBrowser::InitializeOffline() { }; temporary_dir = - FileUtil::SanitizePath(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + "web_applet_" + - WEB_SOURCE_NAMES[static_cast<u32>(source) - 1], - FileUtil::DirectorySeparator::PlatformDefault); - FileUtil::DeleteDirRecursively(temporary_dir); + Common::FS::SanitizePath(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + + "web_applet_" + WEB_SOURCE_NAMES[static_cast<u32>(source) - 1], + Common::FS::DirectorySeparator::PlatformDefault); + Common::FS::DeleteDirRecursively(temporary_dir); u64 title_id = 0; // 0 corresponds to current process ASSERT(args[WebArgTLVType::ApplicationID].size() >= 0x8); @@ -492,8 +492,8 @@ void WebBrowser::InitializeOffline() { } filename = - FileUtil::SanitizePath(temporary_dir + path_additional_directory + DIR_SEP + filename, - FileUtil::DirectorySeparator::PlatformDefault); + Common::FS::SanitizePath(temporary_dir + path_additional_directory + DIR_SEP + filename, + Common::FS::DirectorySeparator::PlatformDefault); } void WebBrowser::ExecuteShop() { @@ -551,7 +551,8 @@ void WebBrowser::ExecuteShop() { } void WebBrowser::ExecuteOffline() { - frontend.OpenPageLocal(filename, [this] { UnpackRomFS(); }, [this] { Finalize(); }); + frontend.OpenPageLocal( + filename, [this] { UnpackRomFS(); }, [this] { Finalize(); }); } } // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/spsm.cpp b/src/core/hle/service/am/spsm.cpp index 003ee8667..f27729ce7 100644 --- a/src/core/hle/service/am/spsm.cpp +++ b/src/core/hle/service/am/spsm.cpp @@ -10,17 +10,17 @@ SPSM::SPSM() : ServiceFramework{"spsm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetState"}, - {1, nullptr, "SleepSystemAndWaitAwake"}, - {2, nullptr, "Unknown1"}, - {3, nullptr, "Unknown2"}, + {1, nullptr, "EnterSleep"}, + {2, nullptr, "GetLastWakeReason"}, + {3, nullptr, "Shutdown"}, {4, nullptr, "GetNotificationMessageEventHandle"}, - {5, nullptr, "Unknown3"}, - {6, nullptr, "Unknown4"}, - {7, nullptr, "Unknown5"}, + {5, nullptr, "ReceiveNotificationMessage"}, + {6, nullptr, "AnalyzeLogForLastSleepWakeSequence"}, + {7, nullptr, "ResetEventLog"}, {8, nullptr, "AnalyzePerformanceLogForLastSleepWakeSequence"}, {9, nullptr, "ChangeHomeButtonLongPressingTime"}, - {10, nullptr, "Unknown6"}, - {11, nullptr, "Unknown7"}, + {10, nullptr, "PutErrorState"}, + {11, nullptr, "InvalidateCurrentHomeButtonPressing"}, }; // clang-format on diff --git a/src/core/hle/service/aoc/aoc_u.cpp b/src/core/hle/service/aoc/aoc_u.cpp index 4227a4adf..173b36da4 100644 --- a/src/core/hle/service/aoc/aoc_u.cpp +++ b/src/core/hle/service/aoc/aoc_u.cpp @@ -6,6 +6,7 @@ #include <numeric> #include <vector> #include "common/logging/log.h" +#include "core/core.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/nca_metadata.h" @@ -60,6 +61,7 @@ AOC_U::AOC_U(Core::System& system) {6, nullptr, "PrepareAddOnContentByApplicationId"}, {7, &AOC_U::PrepareAddOnContent, "PrepareAddOnContent"}, {8, &AOC_U::GetAddOnContentListChangedEvent, "GetAddOnContentListChangedEvent"}, + {9, nullptr, "GetAddOnContentLostErrorCode"}, {100, nullptr, "CreateEcPurchasedEventManager"}, {101, nullptr, "CreatePermanentEcPurchasedEventManager"}, }; @@ -162,7 +164,8 @@ void AOC_U::GetAddOnContentBaseId(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); const auto title_id = system.CurrentProcess()->GetTitleID(); - FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto res = pm.GetControlMetadata(); if (res.first == nullptr) { diff --git a/src/core/hle/service/apm/apm.cpp b/src/core/hle/service/apm/apm.cpp index 85bbf5988..e2d8f0027 100644 --- a/src/core/hle/service/apm/apm.cpp +++ b/src/core/hle/service/apm/apm.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/service/apm/apm.h" #include "core/hle/service/apm/interface.h" diff --git a/src/core/hle/service/apm/controller.cpp b/src/core/hle/service/apm/controller.cpp index 25a886238..ce993bad3 100644 --- a/src/core/hle/service/apm/controller.cpp +++ b/src/core/hle/service/apm/controller.cpp @@ -69,7 +69,8 @@ void Controller::SetFromCpuBoostMode(CpuBoostMode mode) { } PerformanceMode Controller::GetCurrentPerformanceMode() const { - return Settings::values.use_docked_mode ? PerformanceMode::Docked : PerformanceMode::Handheld; + return Settings::values.use_docked_mode.GetValue() ? PerformanceMode::Docked + : PerformanceMode::Handheld; } PerformanceConfiguration Controller::GetCurrentPerformanceConfiguration(PerformanceMode mode) { diff --git a/src/core/hle/service/audio/audctl.cpp b/src/core/hle/service/audio/audctl.cpp index 9e08e5346..6ddb547fb 100644 --- a/src/core/hle/service/audio/audctl.cpp +++ b/src/core/hle/service/audio/audctl.cpp @@ -39,6 +39,8 @@ AudCtl::AudCtl() : ServiceFramework{"audctl"} { {25, nullptr, "GetAudioVolumeDataForPlayReport"}, {26, nullptr, "UpdateHeadphoneSettings"}, {27, nullptr, "SetVolumeMappingTableForDev"}, + {28, nullptr, "GetAudioOutputChannelCountForPlayReport"}, + {29, nullptr, "BindAudioOutputChannelCountUpdateEventForPlayReport"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp index d7f1d348d..3e2299426 100644 --- a/src/core/hle/service/audio/audin_u.cpp +++ b/src/core/hle/service/audio/audin_u.cpp @@ -2,6 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" +#include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/audio/audin_u.h" namespace Service::Audio { @@ -36,11 +39,12 @@ public: AudInU::AudInU() : ServiceFramework("audin:u") { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "ListAudioIns"}, - {1, nullptr, "OpenAudioIn"}, - {2, nullptr, "Unknown"}, - {3, nullptr, "OpenAudioInAuto"}, - {4, nullptr, "ListAudioInsAuto"}, + {0, &AudInU::ListAudioIns, "ListAudioIns"}, + {1, &AudInU::OpenAudioIn, "OpenAudioIn"}, + {2, &AudInU::ListAudioIns, "ListAudioInsAuto"}, + {3, &AudInU::OpenAudioIn, "OpenAudioInAuto"}, + {4, &AudInU::ListAudioInsAutoFiltered, "ListAudioInsAutoFiltered"}, + {5, &AudInU::OpenAudioInProtocolSpecified, "OpenAudioInProtocolSpecified"}, }; // clang-format on @@ -49,4 +53,60 @@ AudInU::AudInU() : ServiceFramework("audin:u") { AudInU::~AudInU() = default; +void AudInU::ListAudioIns(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Audio, "called"); + const std::size_t count = ctx.GetWriteBufferSize() / sizeof(AudioInDeviceName); + + const std::size_t device_count = std::min(count, audio_device_names.size()); + std::vector<AudioInDeviceName> device_names; + device_names.reserve(device_count); + + for (std::size_t i = 0; i < device_count; i++) { + const auto& device_name = audio_device_names[i]; + auto& entry = device_names.emplace_back(); + device_name.copy(entry.data(), device_name.size()); + } + + ctx.WriteBuffer(device_names); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(static_cast<u32>(device_names.size())); +} + +void AudInU::ListAudioInsAutoFiltered(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Audio, "called"); + constexpr u32 device_count = 0; + + // Since we don't actually use any other audio input devices, we return 0 devices. Filtered + // device listing just omits the default input device + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(static_cast<u32>(device_count)); +} + +void AudInU::OpenInOutImpl(Kernel::HLERequestContext& ctx) { + AudInOutParams params{}; + params.channel_count = 2; + params.sample_format = SampleFormat::PCM16; + params.sample_rate = 48000; + params.state = State::Started; + + IPC::ResponseBuilder rb{ctx, 6, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw<AudInOutParams>(params); + rb.PushIpcInterface<IAudioIn>(); +} + +void AudInU::OpenAudioIn(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_Audio, "(STUBBED) called"); + OpenInOutImpl(ctx); +} + +void AudInU::OpenAudioInProtocolSpecified(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_Audio, "(STUBBED) called"); + OpenInOutImpl(ctx); +} + } // namespace Service::Audio diff --git a/src/core/hle/service/audio/audin_u.h b/src/core/hle/service/audio/audin_u.h index 0538b9560..a599f4a64 100644 --- a/src/core/hle/service/audio/audin_u.h +++ b/src/core/hle/service/audio/audin_u.h @@ -16,6 +16,35 @@ class AudInU final : public ServiceFramework<AudInU> { public: explicit AudInU(); ~AudInU() override; + +private: + enum class SampleFormat : u32_le { + PCM16 = 2, + }; + + enum class State : u32_le { + Started = 0, + Stopped = 1, + }; + + struct AudInOutParams { + u32_le sample_rate{}; + u32_le channel_count{}; + SampleFormat sample_format{}; + State state{}; + }; + static_assert(sizeof(AudInOutParams) == 0x10, "AudInOutParams is an invalid size"); + + using AudioInDeviceName = std::array<char, 256>; + static constexpr std::array<std::string_view, 1> audio_device_names{{ + "BuiltInHeadset", + }}; + + void ListAudioIns(Kernel::HLERequestContext& ctx); + void ListAudioInsAutoFiltered(Kernel::HLERequestContext& ctx); + void OpenInOutImpl(Kernel::HLERequestContext& ctx); + void OpenAudioIn(Kernel::HLERequestContext& ctx); + void OpenAudioInProtocolSpecified(Kernel::HLERequestContext& ctx); }; } // namespace Service::Audio diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp index 4fb2cbc4b..9b4910e53 100644 --- a/src/core/hle/service/audio/audout_u.cpp +++ b/src/core/hle/service/audio/audout_u.cpp @@ -71,7 +71,7 @@ public: stream = audio_core.OpenStream(system.CoreTiming(), audio_params.sample_rate, audio_params.channel_count, std::move(unique_name), - [=]() { buffer_event.writable->Signal(); }); + [this] { buffer_event.writable->Signal(); }); } private: @@ -206,11 +206,11 @@ private: AudioCore::StreamPtr stream; std::string device_name; - [[maybe_unused]] AudoutParams audio_params {}; + [[maybe_unused]] AudoutParams audio_params{}; /// This is the event handle used to check if the audio buffer was released Kernel::EventPair buffer_event; - Memory::Memory& main_memory; + Core::Memory::Memory& main_memory; }; AudOutU::AudOutU(Core::System& system_) : ServiceFramework("audout:u"), system{system_} { diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index 82a5dbf14..a2d3ded7b 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -26,7 +26,7 @@ namespace Service::Audio { class IAudioRenderer final : public ServiceFramework<IAudioRenderer> { public: - explicit IAudioRenderer(Core::System& system, AudioCore::AudioRendererParameter audren_params, + explicit IAudioRenderer(Core::System& system, AudioCommon::AudioRendererParameter audren_params, const std::size_t instance_number) : ServiceFramework("IAudioRenderer") { // clang-format off @@ -92,11 +92,17 @@ private: } void RequestUpdateImpl(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_Audio, "(STUBBED) called"); + LOG_DEBUG(Service_Audio, "(STUBBED) called"); + + std::vector<u8> output_params(ctx.GetWriteBufferSize()); + auto result = renderer->UpdateAudioRenderer(ctx.ReadBuffer(), output_params); + + if (result.IsSuccess()) { + ctx.WriteBuffer(output_params); + } - ctx.WriteBuffer(renderer->UpdateAudioRenderer(ctx.ReadBuffer())); IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); + rb.Push(result); } void Start(Kernel::HLERequestContext& ctx) { @@ -129,7 +135,7 @@ private: LOG_DEBUG(Service_Audio, "called. rendering_time_limit_percent={}", rendering_time_limit_percent); - ASSERT(rendering_time_limit_percent >= 0 && rendering_time_limit_percent <= 100); + ASSERT(rendering_time_limit_percent <= 100); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -252,8 +258,6 @@ private: } void GetAudioDeviceOutputVolume(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto device_name_buffer = ctx.ReadBuffer(); const std::string name = Common::StringFromBuffer(device_name_buffer); @@ -343,7 +347,7 @@ void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) { OpenAudioRendererImpl(ctx); } -static u64 CalculateNumPerformanceEntries(const AudioCore::AudioRendererParameter& params) { +static u64 CalculateNumPerformanceEntries(const AudioCommon::AudioRendererParameter& params) { // +1 represents the final mix. return u64{params.effect_count} + params.submix_count + params.sink_count + params.voice_count + 1; @@ -372,7 +376,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { constexpr u64 upsampler_manager_size = 0x48; // Calculates the part of the size that relates to mix buffers. - const auto calculate_mix_buffer_sizes = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_mix_buffer_sizes = [](const AudioCommon::AudioRendererParameter& params) { // As of 8.0.0 this is the maximum on voice channels. constexpr u64 max_voice_channels = 6; @@ -394,7 +398,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { }; // Calculates the portion of the size related to the mix data (and the sorting thereof). - const auto calculate_mix_info_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_mix_info_size = [](const AudioCommon::AudioRendererParameter& params) { // The size of the mixing info data structure. constexpr u64 mix_info_size = 0x940; @@ -444,7 +448,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { }; // Calculates the part of the size related to voice channel info. - const auto calculate_voice_info_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_voice_info_size = [](const AudioCommon::AudioRendererParameter& params) { constexpr u64 voice_info_size = 0x220; constexpr u64 voice_resource_size = 0xD0; @@ -458,7 +462,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { }; // Calculates the part of the size related to memory pools. - const auto calculate_memory_pools_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_memory_pools_size = [](const AudioCommon::AudioRendererParameter& params) { const u64 num_memory_pools = sizeof(s32) * (u64{params.effect_count} + params.voice_count); const u64 memory_pool_info_size = 0x20; return Common::AlignUp(num_memory_pools * memory_pool_info_size, info_field_alignment_size); @@ -466,7 +470,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { // Calculates the part of the size related to the splitter context. const auto calculate_splitter_context_size = - [](const AudioCore::AudioRendererParameter& params) -> u64 { + [](const AudioCommon::AudioRendererParameter& params) -> u64 { if (!IsFeatureSupported(AudioFeatures::Splitter, params.revision)) { return 0; } @@ -485,27 +489,29 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { }; // Calculates the part of the size related to the upsampler info. - const auto calculate_upsampler_info_size = [](const AudioCore::AudioRendererParameter& params) { - constexpr u64 upsampler_info_size = 0x280; - // Yes, using the buffer size over info alignment size is intentional here. - return Common::AlignUp(upsampler_info_size * (u64{params.submix_count} + params.sink_count), - buffer_alignment_size); - }; + const auto calculate_upsampler_info_size = + [](const AudioCommon::AudioRendererParameter& params) { + constexpr u64 upsampler_info_size = 0x280; + // Yes, using the buffer size over info alignment size is intentional here. + return Common::AlignUp(upsampler_info_size * + (u64{params.submix_count} + params.sink_count), + buffer_alignment_size); + }; // Calculates the part of the size related to effect info. - const auto calculate_effect_info_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_effect_info_size = [](const AudioCommon::AudioRendererParameter& params) { constexpr u64 effect_info_size = 0x2B0; return Common::AlignUp(effect_info_size * params.effect_count, info_field_alignment_size); }; // Calculates the part of the size related to audio sink info. - const auto calculate_sink_info_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_sink_info_size = [](const AudioCommon::AudioRendererParameter& params) { const u64 sink_info_size = 0x170; return Common::AlignUp(sink_info_size * params.sink_count, info_field_alignment_size); }; // Calculates the part of the size related to voice state info. - const auto calculate_voice_state_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_voice_state_size = [](const AudioCommon::AudioRendererParameter& params) { const u64 voice_state_size = 0x100; const u64 additional_size = buffer_alignment_size - 1; return Common::AlignUp(voice_state_size * params.voice_count + additional_size, @@ -513,7 +519,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { }; // Calculates the part of the size related to performance statistics. - const auto calculate_perf_size = [](const AudioCore::AudioRendererParameter& params) { + const auto calculate_perf_size = [](const AudioCommon::AudioRendererParameter& params) { // Extra size value appended to the end of the calculation. constexpr u64 appended = 128; @@ -540,79 +546,81 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { }; // Calculates the part of the size that relates to the audio command buffer. - const auto calculate_command_buffer_size = [](const AudioCore::AudioRendererParameter& params) { - constexpr u64 alignment = (buffer_alignment_size - 1) * 2; + const auto calculate_command_buffer_size = + [](const AudioCommon::AudioRendererParameter& params) { + constexpr u64 alignment = (buffer_alignment_size - 1) * 2; - if (!IsFeatureSupported(AudioFeatures::VariadicCommandBuffer, params.revision)) { - constexpr u64 command_buffer_size = 0x18000; + if (!IsFeatureSupported(AudioFeatures::VariadicCommandBuffer, params.revision)) { + constexpr u64 command_buffer_size = 0x18000; - return command_buffer_size + alignment; - } + return command_buffer_size + alignment; + } - // When the variadic command buffer is supported, this means - // the command generator for the audio renderer can issue commands - // that are (as one would expect), variable in size. So what we need to do - // is determine the maximum possible size for a few command data structures - // then multiply them by the amount of present commands indicated by the given - // respective audio parameters. + // When the variadic command buffer is supported, this means + // the command generator for the audio renderer can issue commands + // that are (as one would expect), variable in size. So what we need to do + // is determine the maximum possible size for a few command data structures + // then multiply them by the amount of present commands indicated by the given + // respective audio parameters. - constexpr u64 max_biquad_filters = 2; - constexpr u64 max_mix_buffers = 24; + constexpr u64 max_biquad_filters = 2; + constexpr u64 max_mix_buffers = 24; - constexpr u64 biquad_filter_command_size = 0x2C; + constexpr u64 biquad_filter_command_size = 0x2C; - constexpr u64 depop_mix_command_size = 0x24; - constexpr u64 depop_setup_command_size = 0x50; + constexpr u64 depop_mix_command_size = 0x24; + constexpr u64 depop_setup_command_size = 0x50; - constexpr u64 effect_command_max_size = 0x540; + constexpr u64 effect_command_max_size = 0x540; - constexpr u64 mix_command_size = 0x1C; - constexpr u64 mix_ramp_command_size = 0x24; - constexpr u64 mix_ramp_grouped_command_size = 0x13C; + constexpr u64 mix_command_size = 0x1C; + constexpr u64 mix_ramp_command_size = 0x24; + constexpr u64 mix_ramp_grouped_command_size = 0x13C; - constexpr u64 perf_command_size = 0x28; + constexpr u64 perf_command_size = 0x28; - constexpr u64 sink_command_size = 0x130; + constexpr u64 sink_command_size = 0x130; - constexpr u64 submix_command_max_size = - depop_mix_command_size + (mix_command_size * max_mix_buffers) * max_mix_buffers; + constexpr u64 submix_command_max_size = + depop_mix_command_size + (mix_command_size * max_mix_buffers) * max_mix_buffers; - constexpr u64 volume_command_size = 0x1C; - constexpr u64 volume_ramp_command_size = 0x20; + constexpr u64 volume_command_size = 0x1C; + constexpr u64 volume_ramp_command_size = 0x20; - constexpr u64 voice_biquad_filter_command_size = - biquad_filter_command_size * max_biquad_filters; - constexpr u64 voice_data_command_size = 0x9C; - const u64 voice_command_max_size = - (params.splitter_count * depop_setup_command_size) + - (voice_data_command_size + voice_biquad_filter_command_size + volume_ramp_command_size + - mix_ramp_grouped_command_size); + constexpr u64 voice_biquad_filter_command_size = + biquad_filter_command_size * max_biquad_filters; + constexpr u64 voice_data_command_size = 0x9C; + const u64 voice_command_max_size = + (params.splitter_count * depop_setup_command_size) + + (voice_data_command_size + voice_biquad_filter_command_size + + volume_ramp_command_size + mix_ramp_grouped_command_size); - // Now calculate the individual elements that comprise the size and add them together. - const u64 effect_commands_size = params.effect_count * effect_command_max_size; + // Now calculate the individual elements that comprise the size and add them together. + const u64 effect_commands_size = params.effect_count * effect_command_max_size; - const u64 final_mix_commands_size = - depop_mix_command_size + volume_command_size * max_mix_buffers; + const u64 final_mix_commands_size = + depop_mix_command_size + volume_command_size * max_mix_buffers; - const u64 perf_commands_size = - perf_command_size * (CalculateNumPerformanceEntries(params) + max_perf_detail_entries); + const u64 perf_commands_size = + perf_command_size * + (CalculateNumPerformanceEntries(params) + max_perf_detail_entries); - const u64 sink_commands_size = params.sink_count * sink_command_size; + const u64 sink_commands_size = params.sink_count * sink_command_size; - const u64 splitter_commands_size = - params.num_splitter_send_channels * max_mix_buffers * mix_ramp_command_size; + const u64 splitter_commands_size = + params.num_splitter_send_channels * max_mix_buffers * mix_ramp_command_size; - const u64 submix_commands_size = params.submix_count * submix_command_max_size; + const u64 submix_commands_size = params.submix_count * submix_command_max_size; - const u64 voice_commands_size = params.voice_count * voice_command_max_size; + const u64 voice_commands_size = params.voice_count * voice_command_max_size; - return effect_commands_size + final_mix_commands_size + perf_commands_size + - sink_commands_size + splitter_commands_size + submix_commands_size + - voice_commands_size + alignment; - }; + return effect_commands_size + final_mix_commands_size + perf_commands_size + + sink_commands_size + splitter_commands_size + submix_commands_size + + voice_commands_size + alignment; + }; IPC::RequestParser rp{ctx}; - const auto params = rp.PopRaw<AudioCore::AudioRendererParameter>(); + const auto params = rp.PopRaw<AudioCommon::AudioRendererParameter>(); u64 size = 0; size += calculate_mix_buffer_sizes(params); @@ -678,7 +686,7 @@ void AudRenU::GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& c void AudRenU::OpenAudioRendererImpl(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto params = rp.PopRaw<AudioCore::AudioRendererParameter>(); + const auto params = rp.PopRaw<AudioCommon::AudioRendererParameter>(); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp index d19513cbb..f1d81602c 100644 --- a/src/core/hle/service/audio/hwopus.cpp +++ b/src/core/hle/service/audio/hwopus.cpp @@ -92,7 +92,7 @@ private: if (performance) { rb.Push<u64>(*performance); } - ctx.WriteBuffer(samples.data(), samples.size() * sizeof(s16)); + ctx.WriteBuffer(samples); } bool DecodeOpusData(u32& consumed, u32& sample_count, const std::vector<u8>& input, diff --git a/src/core/hle/service/bcat/backend/backend.cpp b/src/core/hle/service/bcat/backend/backend.cpp index def3410cc..174388445 100644 --- a/src/core/hle/service/bcat/backend/backend.cpp +++ b/src/core/hle/service/bcat/backend/backend.cpp @@ -84,7 +84,7 @@ void ProgressServiceBackend::FinishDownload(ResultCode result) { void ProgressServiceBackend::SignalUpdate() const { if (need_hle_lock) { - std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock); + std::lock_guard lock(HLE::g_hle_lock); event.writable->Signal(); } else { event.writable->Signal(); diff --git a/src/core/hle/service/bcat/backend/boxcat.cpp b/src/core/hle/service/bcat/backend/boxcat.cpp index f589864ee..3b6f7498e 100644 --- a/src/core/hle/service/bcat/backend/boxcat.cpp +++ b/src/core/hle/service/bcat/backend/boxcat.cpp @@ -4,8 +4,8 @@ #include <fmt/ostream.h> #include <httplib.h> -#include <json.hpp> #include <mbedtls/sha256.h> +#include <nlohmann/json.hpp> #include "common/hex_util.h" #include "common/logging/backend.h" #include "common/logging/log.h" @@ -18,6 +18,7 @@ #include "core/hle/service/bcat/backend/boxcat.h" #include "core/settings.h" +namespace Service::BCAT { namespace { // Prevents conflicts with windows macro called CreateFile @@ -30,10 +31,6 @@ bool VfsDeleteFileWrap(FileSys::VirtualDir dir, std::string_view name) { return dir->DeleteFile(name); } -} // Anonymous namespace - -namespace Service::BCAT { - constexpr ResultCode ERROR_GENERAL_BCAT_FAILURE{ErrorModule::BCAT, 1}; constexpr char BOXCAT_HOSTNAME[] = "api.yuzu-emu.org"; @@ -90,16 +87,14 @@ constexpr u32 PORT = 443; constexpr u32 TIMEOUT_SECONDS = 30; [[maybe_unused]] constexpr u64 VFS_COPY_BLOCK_SIZE = 1ULL << 24; // 4MB -namespace { - std::string GetBINFilePath(u64 title_id) { return fmt::format("{}bcat/{:016X}/launchparam.bin", - FileUtil::GetUserPath(FileUtil::UserPath::CacheDir), title_id); + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir), title_id); } std::string GetZIPFilePath(u64 title_id) { return fmt::format("{}bcat/{:016X}/data.zip", - FileUtil::GetUserPath(FileUtil::UserPath::CacheDir), title_id); + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir), title_id); } // If the error is something the user should know about (build ID mismatch, bad client version), @@ -201,7 +196,9 @@ private: const std::string& content_type_name) { if (client == nullptr) { client = std::make_unique<httplib::SSLClient>(BOXCAT_HOSTNAME, PORT); - client->set_timeout_sec(timeout_seconds); + client->set_connection_timeout(timeout_seconds); + client->set_read_timeout(timeout_seconds); + client->set_write_timeout(timeout_seconds); } httplib::Headers headers{ @@ -210,8 +207,8 @@ private: {std::string("Game-Build-Id"), fmt::format("{:016X}", build_id)}, }; - if (FileUtil::Exists(path)) { - FileUtil::IOFile file{path, "rb"}; + if (Common::FS::Exists(path)) { + Common::FS::IOFile file{path, "rb"}; if (file.IsOpen()) { std::vector<u8> bytes(file.GetSize()); file.ReadBytes(bytes.data(), bytes.size()); @@ -241,8 +238,8 @@ private: return DownloadResult::InvalidContentType; } - FileUtil::CreateFullPath(path); - FileUtil::IOFile file{path, "wb"}; + Common::FS::CreateFullPath(path); + Common::FS::IOFile file{path, "wb"}; if (!file.IsOpen()) return DownloadResult::GeneralFSError; if (!file.Resize(response->body.size())) @@ -260,7 +257,7 @@ private: return out; } - std::unique_ptr<httplib::Client> client; + std::unique_ptr<httplib::SSLClient> client; std::string path; u64 title_id; u64 build_id; @@ -295,7 +292,7 @@ void SynchronizeInternal(AM::Applets::AppletManager& applet_manager, DirectoryGe LOG_ERROR(Service_BCAT, "Boxcat synchronization failed with error '{}'!", res); if (res == DownloadResult::NoMatchBuildId || res == DownloadResult::NoMatchTitleId) { - FileUtil::Delete(zip_path); + Common::FS::Delete(zip_path); } HandleDownloadDisplayResult(applet_manager, res); @@ -305,7 +302,7 @@ void SynchronizeInternal(AM::Applets::AppletManager& applet_manager, DirectoryGe progress.StartProcessingDataList(); - FileUtil::IOFile zip{zip_path, "rb"}; + Common::FS::IOFile zip{zip_path, "rb"}; const auto size = zip.GetSize(); std::vector<u8> bytes(size); if (!zip.IsOpen() || size == 0 || zip.ReadBytes(bytes.data(), bytes.size()) != bytes.size()) { @@ -370,8 +367,7 @@ bool Boxcat::Synchronize(TitleIDVersion title, ProgressServiceBackend& progress) std::thread([this, title, &progress] { SynchronizeInternal(applet_manager, dir_getter, title, progress); - }) - .detach(); + }).detach(); return true; } @@ -382,8 +378,7 @@ bool Boxcat::SynchronizeDirectory(TitleIDVersion title, std::string name, std::thread([this, title, name, &progress] { SynchronizeInternal(applet_manager, dir_getter, title, progress, name); - }) - .detach(); + }).detach(); return true; } @@ -427,7 +422,7 @@ std::optional<std::vector<u8>> Boxcat::GetLaunchParameter(TitleIDVersion title) LOG_ERROR(Service_BCAT, "Boxcat synchronization failed with error '{}'!", res); if (res == DownloadResult::NoMatchBuildId || res == DownloadResult::NoMatchTitleId) { - FileUtil::Delete(path); + Common::FS::Delete(path); } HandleDownloadDisplayResult(applet_manager, res); @@ -435,7 +430,7 @@ std::optional<std::vector<u8>> Boxcat::GetLaunchParameter(TitleIDVersion title) } } - FileUtil::IOFile bin{path, "rb"}; + Common::FS::IOFile bin{path, "rb"}; const auto size = bin.GetSize(); std::vector<u8> bytes(size); if (!bin.IsOpen() || size == 0 || bin.ReadBytes(bytes.data(), bytes.size()) != bytes.size()) { @@ -450,13 +445,25 @@ std::optional<std::vector<u8>> Boxcat::GetLaunchParameter(TitleIDVersion title) Boxcat::StatusResult Boxcat::GetStatus(std::optional<std::string>& global, std::map<std::string, EventStatus>& games) { httplib::SSLClient client{BOXCAT_HOSTNAME, static_cast<int>(PORT)}; - client.set_timeout_sec(static_cast<int>(TIMEOUT_SECONDS)); + client.set_connection_timeout(static_cast<int>(TIMEOUT_SECONDS)); + client.set_read_timeout(static_cast<int>(TIMEOUT_SECONDS)); + client.set_write_timeout(static_cast<int>(TIMEOUT_SECONDS)); httplib::Headers headers{ {std::string("Game-Assets-API-Version"), std::string(BOXCAT_API_VERSION)}, {std::string("Boxcat-Client-Type"), std::string(BOXCAT_CLIENT_TYPE)}, }; + if (!client.is_valid()) { + LOG_ERROR(Service_BCAT, "Client is invalid, going offline!"); + return StatusResult::Offline; + } + + if (!client.is_socket_open()) { + LOG_ERROR(Service_BCAT, "Failed to open socket, going offline!"); + return StatusResult::Offline; + } + const auto response = client.Get(BOXCAT_PATHNAME_EVENTS, headers); if (response == nullptr) return StatusResult::Offline; diff --git a/src/core/hle/service/bcat/bcat.cpp b/src/core/hle/service/bcat/bcat.cpp index 8bb2528c9..b31766212 100644 --- a/src/core/hle/service/bcat/bcat.cpp +++ b/src/core/hle/service/bcat/bcat.cpp @@ -14,6 +14,8 @@ BCAT::BCAT(Core::System& system, std::shared_ptr<Module> module, {0, &BCAT::CreateBcatService, "CreateBcatService"}, {1, &BCAT::CreateDeliveryCacheStorageService, "CreateDeliveryCacheStorageService"}, {2, &BCAT::CreateDeliveryCacheStorageServiceWithApplicationId, "CreateDeliveryCacheStorageServiceWithApplicationId"}, + {3, nullptr, "CreateDeliveryCacheProgressService"}, + {4, nullptr, "CreateDeliveryCacheProgressServiceWithApplicationId"}, }; // clang-format on RegisterHandlers(functions); diff --git a/src/core/hle/service/bcat/module.cpp b/src/core/hle/service/bcat/module.cpp index 7ada67130..68deb0600 100644 --- a/src/core/hle/service/bcat/module.cpp +++ b/src/core/hle/service/bcat/module.cpp @@ -8,6 +8,7 @@ #include "common/hex_util.h" #include "common/logging/log.h" #include "common/string_util.h" +#include "core/core.h" #include "core/file_sys/vfs.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/process.h" @@ -112,7 +113,7 @@ private: void GetImpl(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_BCAT, "called"); - ctx.WriteBuffer(&impl, sizeof(DeliveryCacheProgressImpl)); + ctx.WriteBuffer(impl); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -141,11 +142,15 @@ public: {20301, nullptr, "RequestSuspendDeliveryTask"}, {20400, nullptr, "RegisterSystemApplicationDeliveryTask"}, {20401, nullptr, "UnregisterSystemApplicationDeliveryTask"}, + {20410, nullptr, "SetSystemApplicationDeliveryTaskTimer"}, {30100, &IBcatService::SetPassphrase, "SetPassphrase"}, + {30101, nullptr, "Unknown"}, + {30102, nullptr, "Unknown2"}, {30200, nullptr, "RegisterBackgroundDeliveryTask"}, {30201, nullptr, "UnregisterBackgroundDeliveryTask"}, {30202, nullptr, "BlockDeliveryTask"}, {30203, nullptr, "UnblockDeliveryTask"}, + {30210, nullptr, "SetDeliveryTaskTimer"}, {30300, nullptr, "RegisterSystemApplicationDeliveryTasks"}, {90100, nullptr, "EnumerateBackgroundDeliveryTask"}, {90200, nullptr, "GetDeliveryList"}, diff --git a/src/core/hle/service/bpc/bpc.cpp b/src/core/hle/service/bpc/bpc.cpp index 1c1ecdb60..fac6b2f9c 100644 --- a/src/core/hle/service/bpc/bpc.cpp +++ b/src/core/hle/service/bpc/bpc.cpp @@ -23,9 +23,14 @@ public: {5, nullptr, "GetBoardPowerControlEvent"}, {6, nullptr, "GetSleepButtonState"}, {7, nullptr, "GetPowerEvent"}, - {8, nullptr, "Unknown1"}, - {9, nullptr, "Unknown2"}, - {10, nullptr, "Unknown3"}, + {8, nullptr, "CreateWakeupTimer"}, + {9, nullptr, "CancelWakeupTimer"}, + {10, nullptr, "EnableWakeupTimerOnDevice"}, + {11, nullptr, "CreateWakeupTimerEx"}, + {12, nullptr, "GetLastEnabledWakeupTimerType"}, + {13, nullptr, "CleanAllWakeupTimers"}, + {14, nullptr, "Unknown"}, + {15, nullptr, "Unknown2"}, }; // clang-format on @@ -38,10 +43,11 @@ public: explicit BPC_R() : ServiceFramework{"bpc:r"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "GetExternalRtcValue"}, - {1, nullptr, "SetExternalRtcValue"}, - {2, nullptr, "ReadExternalRtcResetFlag"}, - {3, nullptr, "ClearExternalRtcResetFlag"}, + {0, nullptr, "GetRtcTime"}, + {1, nullptr, "SetRtcTime"}, + {2, nullptr, "GetRtcResetDetected"}, + {3, nullptr, "ClearRtcResetDetected"}, + {4, nullptr, "SetUpRtcResetOnShutdown"}, }; // clang-format on diff --git a/src/core/hle/service/btdrv/btdrv.cpp b/src/core/hle/service/btdrv/btdrv.cpp index 40a06c9fd..d4f0dd1ab 100644 --- a/src/core/hle/service/btdrv/btdrv.cpp +++ b/src/core/hle/service/btdrv/btdrv.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/kernel.h" @@ -58,102 +59,103 @@ public: {1, nullptr, "InitializeBluetooth"}, {2, nullptr, "EnableBluetooth"}, {3, nullptr, "DisableBluetooth"}, - {4, nullptr, "CleanupBluetooth"}, + {4, nullptr, "FinalizeBluetooth"}, {5, nullptr, "GetAdapterProperties"}, {6, nullptr, "GetAdapterProperty"}, {7, nullptr, "SetAdapterProperty"}, - {8, nullptr, "StartDiscovery"}, - {9, nullptr, "CancelDiscovery"}, + {8, nullptr, "StartInquiry"}, + {9, nullptr, "StopInquiry"}, {10, nullptr, "CreateBond"}, {11, nullptr, "RemoveBond"}, {12, nullptr, "CancelBond"}, - {13, nullptr, "PinReply"}, - {14, nullptr, "SspReply"}, + {13, nullptr, "RespondToPinRequest"}, + {14, nullptr, "RespondToSspRequest"}, {15, nullptr, "GetEventInfo"}, {16, nullptr, "InitializeHid"}, - {17, nullptr, "HidConnect"}, - {18, nullptr, "HidDisconnect"}, - {19, nullptr, "HidSendData"}, - {20, nullptr, "HidSendData2"}, - {21, nullptr, "HidSetReport"}, - {22, nullptr, "HidGetReport"}, - {23, nullptr, "HidWakeController"}, - {24, nullptr, "HidAddPairedDevice"}, - {25, nullptr, "HidGetPairedDevice"}, - {26, nullptr, "CleanupHid"}, - {27, nullptr, "HidGetEventInfo"}, - {28, nullptr, "ExtSetTsi"}, - {29, nullptr, "ExtSetBurstMode"}, - {30, nullptr, "ExtSetZeroRetran"}, - {31, nullptr, "ExtSetMcMode"}, - {32, nullptr, "ExtStartLlrMode"}, - {33, nullptr, "ExtExitLlrMode"}, - {34, nullptr, "ExtSetRadio"}, - {35, nullptr, "ExtSetVisibility"}, - {36, nullptr, "ExtSetTbfcScan"}, + {17, nullptr, "OpenHidConnection"}, + {18, nullptr, "CloseHidConnection"}, + {19, nullptr, "WriteHidData"}, + {20, nullptr, "WriteHidData2"}, + {21, nullptr, "SetHidReport"}, + {22, nullptr, "GetHidReport"}, + {23, nullptr, "TriggerConnection"}, + {24, nullptr, "AddPairedDeviceInfo"}, + {25, nullptr, "GetPairedDeviceInfo"}, + {26, nullptr, "FinalizeHid"}, + {27, nullptr, "GetHidEventInfo"}, + {28, nullptr, "SetTsi"}, + {29, nullptr, "EnableBurstMode"}, + {30, nullptr, "SetZeroRetransmission"}, + {31, nullptr, "EnableMcMode"}, + {32, nullptr, "EnableLlrScan"}, + {33, nullptr, "DisableLlrScan"}, + {34, nullptr, "EnableRadio"}, + {35, nullptr, "SetVisibility"}, + {36, nullptr, "EnableTbfcScan"}, {37, nullptr, "RegisterHidReportEvent"}, - {38, nullptr, "HidGetReportEventInfo"}, + {38, nullptr, "GetHidReportEventInfo"}, {39, nullptr, "GetLatestPlr"}, - {40, nullptr, "ExtGetPendingConnections"}, + {40, nullptr, "GetPendingConnections"}, {41, nullptr, "GetChannelMap"}, - {42, nullptr, "EnableBluetoothBoostSetting"}, - {43, nullptr, "IsBluetoothBoostSettingEnabled"}, - {44, nullptr, "EnableBluetoothAfhSetting"}, - {45, nullptr, "IsBluetoothAfhSettingEnabled"}, - {46, nullptr, "InitializeBluetoothLe"}, - {47, nullptr, "EnableBluetoothLe"}, - {48, nullptr, "DisableBluetoothLe"}, - {49, nullptr, "CleanupBluetoothLe"}, - {50, nullptr, "SetLeVisibility"}, - {51, nullptr, "SetLeConnectionParameter"}, - {52, nullptr, "SetLeDefaultConnectionParameter"}, - {53, nullptr, "SetLeAdvertiseData"}, - {54, nullptr, "SetLeAdvertiseParameter"}, - {55, nullptr, "StartLeScan"}, - {56, nullptr, "StopLeScan"}, - {57, nullptr, "AddLeScanFilterCondition"}, - {58, nullptr, "DeleteLeScanFilterCondition"}, - {59, nullptr, "DeleteLeScanFilter"}, - {60, nullptr, "ClearLeScanFilters"}, - {61, nullptr, "EnableLeScanFilter"}, - {62, nullptr, "RegisterLeClient"}, - {63, nullptr, "UnregisterLeClient"}, - {64, nullptr, "UnregisterLeClientAll"}, - {65, nullptr, "LeClientConnect"}, - {66, nullptr, "LeClientCancelConnection"}, - {67, nullptr, "LeClientDisconnect"}, - {68, nullptr, "LeClientGetAttributes"}, - {69, nullptr, "LeClientDiscoverService"}, - {70, nullptr, "LeClientConfigureMtu"}, - {71, nullptr, "RegisterLeServer"}, - {72, nullptr, "UnregisterLeServer"}, - {73, nullptr, "LeServerConnect"}, - {74, nullptr, "LeServerDisconnect"}, - {75, nullptr, "CreateLeService"}, - {76, nullptr, "StartLeService"}, - {77, nullptr, "AddLeCharacteristic"}, - {78, nullptr, "AddLeDescriptor"}, - {79, nullptr, "GetLeCoreEventInfo"}, - {80, nullptr, "LeGetFirstCharacteristic"}, - {81, nullptr, "LeGetNextCharacteristic"}, - {82, nullptr, "LeGetFirstDescriptor"}, - {83, nullptr, "LeGetNextDescriptor"}, - {84, nullptr, "RegisterLeCoreDataPath"}, - {85, nullptr, "UnregisterLeCoreDataPath"}, - {86, nullptr, "RegisterLeHidDataPath"}, - {87, nullptr, "UnregisterLeHidDataPath"}, - {88, nullptr, "RegisterLeDataPath"}, - {89, nullptr, "UnregisterLeDataPath"}, - {90, nullptr, "LeClientReadCharacteristic"}, - {91, nullptr, "LeClientReadDescriptor"}, - {92, nullptr, "LeClientWriteCharacteristic"}, - {93, nullptr, "LeClientWriteDescriptor"}, - {94, nullptr, "LeClientRegisterNotification"}, - {95, nullptr, "LeClientDeregisterNotification"}, + {42, nullptr, "EnableTxPowerBoostSetting"}, + {43, nullptr, "IsTxPowerBoostSettingEnabled"}, + {44, nullptr, "EnableAfhSetting"}, + {45, nullptr, "IsAfhSettingEnabled"}, + {46, nullptr, "InitializeBle"}, + {47, nullptr, "EnableBle"}, + {48, nullptr, "DisableBle"}, + {49, nullptr, "FinalizeBle"}, + {50, nullptr, "SetBleVisibility"}, + {51, nullptr, "SetBleConnectionParameter"}, + {52, nullptr, "SetBleDefaultConnectionParameter"}, + {53, nullptr, "SetBleAdvertiseData"}, + {54, nullptr, "SetBleAdvertiseParameter"}, + {55, nullptr, "StartBleScan"}, + {56, nullptr, "StopBleScan"}, + {57, nullptr, "AddBleScanFilterCondition"}, + {58, nullptr, "DeleteBleScanFilterCondition"}, + {59, nullptr, "DeleteBleScanFilter"}, + {60, nullptr, "ClearBleScanFilters"}, + {61, nullptr, "EnableBleScanFilter"}, + {62, nullptr, "RegisterGattClient"}, + {63, nullptr, "UnregisterGattClient"}, + {64, nullptr, "UnregisterAllGattClients"}, + {65, nullptr, "ConnectGattServer"}, + {66, nullptr, "CancelConnectGattServer"}, + {67, nullptr, "DisconnectGattServer"}, + {68, nullptr, "GetGattAttribute"}, + {69, nullptr, "GetGattService"}, + {70, nullptr, "ConfigureAttMtu"}, + {71, nullptr, "RegisterGattServer"}, + {72, nullptr, "UnregisterGattServer"}, + {73, nullptr, "ConnectGattClient"}, + {74, nullptr, "DisconnectGattClient"}, + {75, nullptr, "AddGattService"}, + {76, nullptr, "EnableGattService"}, + {77, nullptr, "AddGattCharacteristic"}, + {78, nullptr, "AddGattDescriptor"}, + {79, nullptr, "GetBleManagedEventInfo"}, + {80, nullptr, "GetGattFirstCharacteristic"}, + {81, nullptr, "GetGattNextCharacteristic"}, + {82, nullptr, "GetGattFirstDescriptor"}, + {83, nullptr, "GetGattNextDescriptor"}, + {84, nullptr, "RegisterGattManagedDataPath"}, + {85, nullptr, "UnregisterGattManagedDataPath"}, + {86, nullptr, "RegisterGattHidDataPath"}, + {87, nullptr, "UnregisterGattHidDataPath"}, + {88, nullptr, "RegisterGattDataPath"}, + {89, nullptr, "UnregisterGattDataPath"}, + {90, nullptr, "ReadGattCharacteristic"}, + {91, nullptr, "ReadGattDescriptor"}, + {92, nullptr, "WriteGattCharacteristic"}, + {93, nullptr, "WriteGattDescriptor"}, + {94, nullptr, "RegisterGattNotification"}, + {95, nullptr, "UnregisterGattNotification"}, {96, nullptr, "GetLeHidEventInfo"}, {97, nullptr, "RegisterBleHidEvent"}, - {98, nullptr, "SetLeScanParameter"}, - {256, nullptr, "GetIsManufacturingMode"}, + {98, nullptr, "SetBleScanParameter"}, + {99, nullptr, "MoveToSecondaryPiconet"}, + {256, nullptr, "IsManufacturingMode"}, {257, nullptr, "EmulateBluetoothCrash"}, {258, nullptr, "GetBleChannelMap"}, }; diff --git a/src/core/hle/service/btm/btm.cpp b/src/core/hle/service/btm/btm.cpp index 251b3c9df..c8f8ddbd5 100644 --- a/src/core/hle/service/btm/btm.cpp +++ b/src/core/hle/service/btm/btm.cpp @@ -5,6 +5,7 @@ #include <memory> #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/kernel.h" @@ -132,66 +133,71 @@ public: explicit BTM() : ServiceFramework{"btm"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, - {1, nullptr, "Unknown2"}, - {2, nullptr, "RegisterSystemEventForConnectedDeviceCondition"}, - {3, nullptr, "Unknown3"}, - {4, nullptr, "Unknown4"}, - {5, nullptr, "Unknown5"}, - {6, nullptr, "Unknown6"}, - {7, nullptr, "Unknown7"}, - {8, nullptr, "RegisterSystemEventForRegisteredDeviceInfo"}, - {9, nullptr, "Unknown8"}, - {10, nullptr, "Unknown9"}, - {11, nullptr, "Unknown10"}, - {12, nullptr, "Unknown11"}, - {13, nullptr, "Unknown12"}, + {0, nullptr, "GetState"}, + {1, nullptr, "GetHostDeviceProperty"}, + {2, nullptr, "AcquireDeviceConditionEvent"}, + {3, nullptr, "GetDeviceCondition"}, + {4, nullptr, "SetBurstMode"}, + {5, nullptr, "SetSlotMode"}, + {6, nullptr, "SetBluetoothMode"}, + {7, nullptr, "SetWlanMode"}, + {8, nullptr, "AcquireDeviceInfoEvent"}, + {9, nullptr, "GetDeviceInfo"}, + {10, nullptr, "AddDeviceInfo"}, + {11, nullptr, "RemoveDeviceInfo"}, + {12, nullptr, "IncreaseDeviceInfoOrder"}, + {13, nullptr, "LlrNotify"}, {14, nullptr, "EnableRadio"}, {15, nullptr, "DisableRadio"}, - {16, nullptr, "Unknown13"}, - {17, nullptr, "Unknown14"}, - {18, nullptr, "Unknown15"}, - {19, nullptr, "Unknown16"}, - {20, nullptr, "Unknown17"}, - {21, nullptr, "Unknown18"}, - {22, nullptr, "Unknown19"}, - {23, nullptr, "Unknown20"}, - {24, nullptr, "Unknown21"}, - {25, nullptr, "Unknown22"}, - {26, nullptr, "Unknown23"}, - {27, nullptr, "Unknown24"}, - {28, nullptr, "Unknown25"}, - {29, nullptr, "Unknown26"}, - {30, nullptr, "Unknown27"}, - {31, nullptr, "Unknown28"}, - {32, nullptr, "Unknown29"}, - {33, nullptr, "Unknown30"}, - {34, nullptr, "Unknown31"}, - {35, nullptr, "Unknown32"}, - {36, nullptr, "Unknown33"}, - {37, nullptr, "Unknown34"}, - {38, nullptr, "Unknown35"}, - {39, nullptr, "Unknown36"}, - {40, nullptr, "Unknown37"}, - {41, nullptr, "Unknown38"}, - {42, nullptr, "Unknown39"}, - {43, nullptr, "Unknown40"}, - {44, nullptr, "Unknown41"}, - {45, nullptr, "Unknown42"}, - {46, nullptr, "Unknown43"}, - {47, nullptr, "Unknown44"}, - {48, nullptr, "Unknown45"}, - {49, nullptr, "Unknown46"}, - {50, nullptr, "Unknown47"}, - {51, nullptr, "Unknown48"}, - {52, nullptr, "Unknown49"}, - {53, nullptr, "Unknown50"}, - {54, nullptr, "Unknown51"}, - {55, nullptr, "Unknown52"}, - {56, nullptr, "Unknown53"}, - {57, nullptr, "Unknown54"}, - {58, nullptr, "Unknown55"}, - {59, nullptr, "Unknown56"}, + {16, nullptr, "HidDisconnect"}, + {17, nullptr, "HidSetRetransmissionMode"}, + {18, nullptr, "AcquireAwakeReqEvent"}, + {19, nullptr, "AcquireLlrStateEvent"}, + {20, nullptr, "IsLlrStarted"}, + {21, nullptr, "EnableSlotSaving"}, + {22, nullptr, "ProtectDeviceInfo"}, + {23, nullptr, "AcquireBleScanEvent"}, + {24, nullptr, "GetBleScanParameterGeneral"}, + {25, nullptr, "GetBleScanParameterSmartDevice"}, + {26, nullptr, "StartBleScanForGeneral"}, + {27, nullptr, "StopBleScanForGeneral"}, + {28, nullptr, "GetBleScanResultsForGeneral"}, + {29, nullptr, "StartBleScanForPairedDevice"}, + {30, nullptr, "StopBleScanForPairedDevice"}, + {31, nullptr, "StartBleScanForSmartDevice"}, + {32, nullptr, "StopBleScanForSmartDevice"}, + {33, nullptr, "GetBleScanResultsForSmartDevice"}, + {34, nullptr, "AcquireBleConnectionEvent"}, + {35, nullptr, "BleConnect"}, + {36, nullptr, "BleOverrideConnection"}, + {37, nullptr, "BleDisconnect"}, + {38, nullptr, "BleGetConnectionState"}, + {39, nullptr, "BleGetGattClientConditionList"}, + {40, nullptr, "AcquireBlePairingEvent"}, + {41, nullptr, "BlePairDevice"}, + {42, nullptr, "BleUnpairDeviceOnBoth"}, + {43, nullptr, "BleUnpairDevice"}, + {44, nullptr, "BleGetPairedAddresses"}, + {45, nullptr, "AcquireBleServiceDiscoveryEvent"}, + {46, nullptr, "GetGattServices"}, + {47, nullptr, "GetGattService"}, + {48, nullptr, "GetGattIncludedServices"}, + {49, nullptr, "GetBelongingService"}, + {50, nullptr, "GetGattCharacteristics"}, + {51, nullptr, "GetGattDescriptors"}, + {52, nullptr, "AcquireBleMtuConfigEvent"}, + {53, nullptr, "ConfigureBleMtu"}, + {54, nullptr, "GetBleMtu"}, + {55, nullptr, "RegisterBleGattDataPath"}, + {56, nullptr, "UnregisterBleGattDataPath"}, + {57, nullptr, "RegisterAppletResourceUserId"}, + {58, nullptr, "UnregisterAppletResourceUserId"}, + {59, nullptr, "SetAppletResourceUserId"}, + {60, nullptr, "Unknown60"}, + {61, nullptr, "Unknown61"}, + {62, nullptr, "Unknown62"}, + {63, nullptr, "Unknown63"}, + {64, nullptr, "Unknown64"}, }; // clang-format on @@ -204,19 +210,19 @@ public: explicit BTM_DBG() : ServiceFramework{"btm:dbg"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "RegisterSystemEventForDiscovery"}, - {1, nullptr, "Unknown1"}, - {2, nullptr, "Unknown2"}, - {3, nullptr, "Unknown3"}, - {4, nullptr, "Unknown4"}, - {5, nullptr, "Unknown5"}, - {6, nullptr, "Unknown6"}, - {7, nullptr, "Unknown7"}, - {8, nullptr, "Unknown8"}, - {9, nullptr, "Unknown9"}, - {10, nullptr, "Unknown10"}, - {11, nullptr, "Unknown11"}, - {12, nullptr, "Unknown11"}, + {0, nullptr, "AcquireDiscoveryEvent"}, + {1, nullptr, "StartDiscovery"}, + {2, nullptr, "CancelDiscovery"}, + {3, nullptr, "GetDeviceProperty"}, + {4, nullptr, "CreateBond"}, + {5, nullptr, "CancelBond"}, + {6, nullptr, "SetTsiMode"}, + {7, nullptr, "GeneralTest"}, + {8, nullptr, "HidConnect"}, + {9, nullptr, "GeneralGet"}, + {10, nullptr, "GetGattClientDisconnectionReason"}, + {11, nullptr, "GetBleConnectionParameter"}, + {12, nullptr, "GetBleConnectionParameterRequest"}, }; // clang-format on diff --git a/src/core/hle/service/caps/caps.cpp b/src/core/hle/service/caps/caps.cpp index 26c8a7081..ba5749b84 100644 --- a/src/core/hle/service/caps/caps.cpp +++ b/src/core/hle/service/caps/caps.cpp @@ -1,4 +1,4 @@ -// Copyright 2018 yuzu emulator team +// Copyright 2018 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps.h b/src/core/hle/service/caps/caps.h index fc70a4c27..b8c67b6e2 100644 --- a/src/core/hle/service/caps/caps.h +++ b/src/core/hle/service/caps/caps.h @@ -1,4 +1,4 @@ -// Copyright 2018 yuzu emulator team +// Copyright 2018 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -12,73 +12,79 @@ class ServiceManager; namespace Service::Capture { -enum AlbumImageOrientation { +enum class AlbumImageOrientation { Orientation0 = 0, Orientation1 = 1, Orientation2 = 2, Orientation3 = 3, }; -enum AlbumReportOption { +enum class AlbumReportOption { Disable = 0, Enable = 1, }; -enum ContentType : u8 { +enum class ContentType : u8 { Screenshot = 0, Movie = 1, ExtraMovie = 3, }; -enum AlbumStorage : u8 { +enum class AlbumStorage : u8 { NAND = 0, SD = 1, }; struct AlbumFileDateTime { - u16 year; - u8 month; - u8 day; - u8 hour; - u8 minute; - u8 second; - u8 uid; + s16 year{}; + s8 month{}; + s8 day{}; + s8 hour{}; + s8 minute{}; + s8 second{}; + s8 uid{}; }; +static_assert(sizeof(AlbumFileDateTime) == 0x8, "AlbumFileDateTime has incorrect size."); struct AlbumEntry { - u64 size; - u64 application_id; - AlbumFileDateTime datetime; - AlbumStorage storage; - ContentType content; - u8 padding[6]; + u64 size{}; + u64 application_id{}; + AlbumFileDateTime datetime{}; + AlbumStorage storage{}; + ContentType content{}; + INSERT_PADDING_BYTES(6); }; +static_assert(sizeof(AlbumEntry) == 0x20, "AlbumEntry has incorrect size."); struct AlbumFileEntry { - u64 size; - u64 hash; - AlbumFileDateTime datetime; - AlbumStorage storage; - ContentType content; - u8 padding[5]; - u8 unknown; + u64 size{}; // Size of the entry + u64 hash{}; // AES256 with hardcoded key over AlbumEntry + AlbumFileDateTime datetime{}; + AlbumStorage storage{}; + ContentType content{}; + INSERT_PADDING_BYTES(5); + u8 unknown{1}; // Set to 1 on official SW }; +static_assert(sizeof(AlbumFileEntry) == 0x20, "AlbumFileEntry has incorrect size."); struct ApplicationAlbumEntry { - u64 size; - u64 hash; - AlbumFileDateTime datetime; - AlbumStorage storage; - ContentType content; - u8 padding[5]; - u8 unknown; + u64 size{}; // Size of the entry + u64 hash{}; // AES256 with hardcoded key over AlbumEntry + AlbumFileDateTime datetime{}; + AlbumStorage storage{}; + ContentType content{}; + INSERT_PADDING_BYTES(5); + u8 unknown{1}; // Set to 1 on official SW }; +static_assert(sizeof(ApplicationAlbumEntry) == 0x20, "ApplicationAlbumEntry has incorrect size."); struct ApplicationAlbumFileEntry { - ApplicationAlbumEntry entry; - AlbumFileDateTime datetime; - u64 unknown; + ApplicationAlbumEntry entry{}; + AlbumFileDateTime datetime{}; + u64 unknown{}; }; +static_assert(sizeof(ApplicationAlbumFileEntry) == 0x30, + "ApplicationAlbumFileEntry has incorrect size."); /// Registers all Capture services with the specified service manager. void InstallInterfaces(SM::ServiceManager& sm); diff --git a/src/core/hle/service/caps/caps_a.cpp b/src/core/hle/service/caps/caps_a.cpp index 88a3fdc05..a0a3b2ae3 100644 --- a/src/core/hle/service/caps/caps_a.cpp +++ b/src/core/hle/service/caps/caps_a.cpp @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps_a.h b/src/core/hle/service/caps/caps_a.h index 8de832491..cb93aad5b 100644 --- a/src/core/hle/service/caps/caps_a.h +++ b/src/core/hle/service/caps/caps_a.h @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps_c.cpp b/src/core/hle/service/caps/caps_c.cpp index ea6452ffa..a0ee116fa 100644 --- a/src/core/hle/service/caps/caps_c.cpp +++ b/src/core/hle/service/caps/caps_c.cpp @@ -1,7 +1,9 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" +#include "core/hle/ipc_helpers.h" #include "core/hle/service/caps/caps_c.h" namespace Service::Capture { @@ -47,7 +49,7 @@ CAPS_C::CAPS_C() : ServiceFramework("caps:c") { static const FunctionInfo functions[] = { {1, nullptr, "CaptureRawImage"}, {2, nullptr, "CaptureRawImageWithTimeout"}, - {33, nullptr, "Unknown33"}, + {33, &CAPS_C::SetShimLibraryVersion, "SetShimLibraryVersion"}, {1001, nullptr, "RequestTakingScreenShot"}, {1002, nullptr, "RequestTakingScreenShotWithTimeout"}, {1011, nullptr, "NotifyTakingScreenShotRefused"}, @@ -72,4 +74,16 @@ CAPS_C::CAPS_C() : ServiceFramework("caps:c") { CAPS_C::~CAPS_C() = default; +void CAPS_C::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto library_version{rp.Pop<u64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}", + library_version, applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_c.h b/src/core/hle/service/caps/caps_c.h index d07cdb441..b110301d4 100644 --- a/src/core/hle/service/caps/caps_c.h +++ b/src/core/hle/service/caps/caps_c.h @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -16,6 +16,9 @@ class CAPS_C final : public ServiceFramework<CAPS_C> { public: explicit CAPS_C(); ~CAPS_C() override; + +private: + void SetShimLibraryVersion(Kernel::HLERequestContext& ctx); }; } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_sc.cpp b/src/core/hle/service/caps/caps_sc.cpp index d01a8a58e..822ee96c8 100644 --- a/src/core/hle/service/caps/caps_sc.cpp +++ b/src/core/hle/service/caps/caps_sc.cpp @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps_sc.h b/src/core/hle/service/caps/caps_sc.h index 9ba372f7a..ac3e929ca 100644 --- a/src/core/hle/service/caps/caps_sc.h +++ b/src/core/hle/service/caps/caps_sc.h @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps_ss.cpp b/src/core/hle/service/caps/caps_ss.cpp index eaa3a7494..24dc716e7 100644 --- a/src/core/hle/service/caps/caps_ss.cpp +++ b/src/core/hle/service/caps/caps_ss.cpp @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps_ss.h b/src/core/hle/service/caps/caps_ss.h index e258a6925..450686e4f 100644 --- a/src/core/hle/service/caps/caps_ss.h +++ b/src/core/hle/service/caps/caps_ss.h @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. diff --git a/src/core/hle/service/caps/caps_su.cpp b/src/core/hle/service/caps/caps_su.cpp index 2b4c2d808..e386470f7 100644 --- a/src/core/hle/service/caps/caps_su.cpp +++ b/src/core/hle/service/caps/caps_su.cpp @@ -1,7 +1,9 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" +#include "core/hle/ipc_helpers.h" #include "core/hle/service/caps/caps_su.h" namespace Service::Capture { @@ -9,8 +11,11 @@ namespace Service::Capture { CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") { // clang-format off static const FunctionInfo functions[] = { + {32, &CAPS_SU::SetShimLibraryVersion, "SetShimLibraryVersion"}, {201, nullptr, "SaveScreenShot"}, {203, nullptr, "SaveScreenShotEx0"}, + {205, nullptr, "SaveScreenShotEx1"}, + {210, nullptr, "SaveScreenShotEx2"}, }; // clang-format on @@ -19,4 +24,16 @@ CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") { CAPS_SU::~CAPS_SU() = default; +void CAPS_SU::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto library_version{rp.Pop<u64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}", + library_version, applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_su.h b/src/core/hle/service/caps/caps_su.h index cb11f7c9a..62c9603a9 100644 --- a/src/core/hle/service/caps/caps_su.h +++ b/src/core/hle/service/caps/caps_su.h @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -16,6 +16,9 @@ class CAPS_SU final : public ServiceFramework<CAPS_SU> { public: explicit CAPS_SU(); ~CAPS_SU() override; + +private: + void SetShimLibraryVersion(Kernel::HLERequestContext& ctx); }; } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_u.cpp b/src/core/hle/service/caps/caps_u.cpp index 78bab6ed8..f9479bdb3 100644 --- a/src/core/hle/service/caps/caps_u.cpp +++ b/src/core/hle/service/caps/caps_u.cpp @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -31,8 +31,7 @@ public: CAPS_U::CAPS_U() : ServiceFramework("caps:u") { // clang-format off static const FunctionInfo functions[] = { - {31, nullptr, "GetShimLibraryVersion"}, - {32, nullptr, "SetShimLibraryVersion"}, + {32, &CAPS_U::SetShimLibraryVersion, "SetShimLibraryVersion"}, {102, &CAPS_U::GetAlbumContentsFileListForApplication, "GetAlbumContentsFileListForApplication"}, {103, nullptr, "DeleteAlbumContentsFileForApplication"}, {104, nullptr, "GetAlbumContentsFileSizeForApplication"}, @@ -42,7 +41,7 @@ CAPS_U::CAPS_U() : ServiceFramework("caps:u") { {130, nullptr, "PrecheckToCreateContentsForApplication"}, {140, nullptr, "GetAlbumFileList1AafeAruidDeprecated"}, {141, nullptr, "GetAlbumFileList2AafeUidAruidDeprecated"}, - {142, nullptr, "GetAlbumFileList3AaeAruid"}, + {142, &CAPS_U::GetAlbumFileList3AaeAruid, "GetAlbumFileList3AaeAruid"}, {143, nullptr, "GetAlbumFileList4AaeUidAruid"}, {60002, nullptr, "OpenAccessorSessionForApplication"}, }; @@ -53,24 +52,49 @@ CAPS_U::CAPS_U() : ServiceFramework("caps:u") { CAPS_U::~CAPS_U() = default; +void CAPS_U::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto library_version{rp.Pop<u64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}", + library_version, applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + void CAPS_U::GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& ctx) { // Takes a type-0x6 output buffer containing an array of ApplicationAlbumFileEntry, a PID, an // u8 ContentType, two s64s, and an u64 AppletResourceUserId. Returns an output u64 for total // output entries (which is copied to a s32 by official SW). IPC::RequestParser rp{ctx}; - [[maybe_unused]] const auto application_album_file_entries = rp.PopRaw<std::array<u8, 0x30>>(); - const auto pid = rp.Pop<s32>(); - const auto content_type = rp.PopRaw<ContentType>(); - [[maybe_unused]] const auto start_datetime = rp.PopRaw<AlbumFileDateTime>(); - [[maybe_unused]] const auto end_datetime = rp.PopRaw<AlbumFileDateTime>(); - const auto applet_resource_user_id = rp.Pop<u64>(); - LOG_WARNING(Service_Capture, - "(STUBBED) called. pid={}, content_type={}, applet_resource_user_id={}", pid, - content_type, applet_resource_user_id); - - IPC::ResponseBuilder rb{ctx, 3}; + const auto pid{rp.Pop<s32>()}; + const auto content_type{rp.PopEnum<ContentType>()}; + const auto start_posix_time{rp.Pop<s64>()}; + const auto end_posix_time{rp.Pop<s64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + // TODO: Update this when we implement the album. + // Currently we do not have a method of accessing album entries, set this to 0 for now. + constexpr u32 total_entries_1{}; + constexpr u32 total_entries_2{}; + + LOG_WARNING( + Service_Capture, + "(STUBBED) called. pid={}, content_type={}, start_posix_time={}, " + "end_posix_time={}, applet_resource_user_id={}, total_entries_1={}, total_entries_2={}", + pid, content_type, start_posix_time, end_posix_time, applet_resource_user_id, + total_entries_1, total_entries_2); + + IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<s32>(0); + rb.Push(total_entries_1); + rb.Push(total_entries_2); +} + +void CAPS_U::GetAlbumFileList3AaeAruid(Kernel::HLERequestContext& ctx) { + GetAlbumContentsFileListForApplication(ctx); } } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_u.h b/src/core/hle/service/caps/caps_u.h index e6e0716ff..4b80f3156 100644 --- a/src/core/hle/service/caps/caps_u.h +++ b/src/core/hle/service/caps/caps_u.h @@ -1,4 +1,4 @@ -// Copyright 2020 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -18,7 +18,9 @@ public: ~CAPS_U() override; private: + void SetShimLibraryVersion(Kernel::HLERequestContext& ctx); void GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& ctx); + void GetAlbumFileList3AaeAruid(Kernel::HLERequestContext& ctx); }; } // namespace Service::Capture diff --git a/src/core/hle/service/es/es.cpp b/src/core/hle/service/es/es.cpp index df00ae625..c2737a365 100644 --- a/src/core/hle/service/es/es.cpp +++ b/src/core/hle/service/es/es.cpp @@ -4,6 +4,7 @@ #include "core/crypto/key_manager.h" #include "core/hle/ipc_helpers.h" +#include "core/hle/service/es/es.h" #include "core/hle/service/service.h" namespace Service::ES { @@ -26,8 +27,8 @@ public: {8, &ETicket::GetTitleKey, "GetTitleKey"}, {9, &ETicket::CountCommonTicket, "CountCommonTicket"}, {10, &ETicket::CountPersonalizedTicket, "CountPersonalizedTicket"}, - {11, &ETicket::ListCommonTicket, "ListCommonTicket"}, - {12, &ETicket::ListPersonalizedTicket, "ListPersonalizedTicket"}, + {11, &ETicket::ListCommonTicketRightsIds, "ListCommonTicketRightsIds"}, + {12, &ETicket::ListPersonalizedTicketRightsIds, "ListPersonalizedTicketRightsIds"}, {13, nullptr, "ListMissingPersonalizedTicket"}, {14, &ETicket::GetCommonTicketSize, "GetCommonTicketSize"}, {15, &ETicket::GetPersonalizedTicketSize, "GetPersonalizedTicketSize"}, @@ -54,7 +55,46 @@ public: {36, nullptr, "DeleteAllInactiveELicenseRequiredPersonalizedTicket"}, {37, nullptr, "OwnTicket2"}, {38, nullptr, "OwnTicket3"}, + {501, nullptr, "Unknown501"}, + {502, nullptr, "Unknown502"}, {503, nullptr, "GetTitleKey"}, + {504, nullptr, "Unknown504"}, + {508, nullptr, "Unknown508"}, + {509, nullptr, "Unknown509"}, + {510, nullptr, "Unknown510"}, + {511, nullptr, "Unknown511"}, + {1001, nullptr, "Unknown1001"}, + {1002, nullptr, "Unknown1001"}, + {1003, nullptr, "Unknown1003"}, + {1004, nullptr, "Unknown1004"}, + {1005, nullptr, "Unknown1005"}, + {1006, nullptr, "Unknown1006"}, + {1007, nullptr, "Unknown1007"}, + {1009, nullptr, "Unknown1009"}, + {1010, nullptr, "Unknown1010"}, + {1011, nullptr, "Unknown1011"}, + {1012, nullptr, "Unknown1012"}, + {1013, nullptr, "Unknown1013"}, + {1014, nullptr, "Unknown1014"}, + {1015, nullptr, "Unknown1015"}, + {1016, nullptr, "Unknown1016"}, + {1017, nullptr, "Unknown1017"}, + {1018, nullptr, "Unknown1018"}, + {1019, nullptr, "Unknown1019"}, + {1020, nullptr, "Unknown1020"}, + {1021, nullptr, "Unknown1021"}, + {1501, nullptr, "Unknown1501"}, + {1502, nullptr, "Unknown1502"}, + {1503, nullptr, "Unknown1503"}, + {1504, nullptr, "Unknown1504"}, + {1505, nullptr, "Unknown1505"}, + {2000, nullptr, "Unknown2000"}, + {2001, nullptr, "Unknown2001"}, + {2100, nullptr, "Unknown2100"}, + {2501, nullptr, "Unknown2501"}, + {2502, nullptr, "Unknown2502"}, + {3001, nullptr, "Unknown3001"}, + {3002, nullptr, "Unknown3002"}, }; // clang-format on RegisterHandlers(functions); @@ -76,7 +116,6 @@ private: } void ImportTicket(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; const auto ticket = ctx.ReadBuffer(); const auto cert = ctx.ReadBuffer(1); @@ -121,7 +160,7 @@ private: return; } - ctx.WriteBuffer(key.data(), key.size()); + ctx.WriteBuffer(key); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -147,7 +186,7 @@ private: rb.Push<u32>(count); } - void ListCommonTicket(Kernel::HLERequestContext& ctx) { + void ListCommonTicketRightsIds(Kernel::HLERequestContext& ctx) { u32 out_entries; if (keys.GetCommonTickets().empty()) out_entries = 0; @@ -170,7 +209,7 @@ private: rb.Push<u32>(out_entries); } - void ListPersonalizedTicket(Kernel::HLERequestContext& ctx) { + void ListPersonalizedTicketRightsIds(Kernel::HLERequestContext& ctx) { u32 out_entries; if (keys.GetPersonalizedTickets().empty()) out_entries = 0; @@ -263,7 +302,7 @@ private: rb.Push<u64>(write_size); } - Core::Crypto::KeyManager keys; + Core::Crypto::KeyManager& keys = Core::Crypto::KeyManager::Instance(); }; void InstallInterfaces(SM::ServiceManager& service_manager) { diff --git a/src/core/hle/service/eupld/eupld.cpp b/src/core/hle/service/eupld/eupld.cpp index 2df30acee..0d6d244f4 100644 --- a/src/core/hle/service/eupld/eupld.cpp +++ b/src/core/hle/service/eupld/eupld.cpp @@ -19,6 +19,7 @@ public: {1, nullptr, "ImportCrt"}, {2, nullptr, "ImportPki"}, {3, nullptr, "SetAutoUpload"}, + {4, nullptr, "GetAutoUpload"}, }; // clang-format on diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp index 102017d73..2e53cae5b 100644 --- a/src/core/hle/service/filesystem/filesystem.cpp +++ b/src/core/hle/service/filesystem/filesystem.cpp @@ -36,7 +36,7 @@ constexpr u64 SUFFICIENT_SAVE_DATA_SIZE = 0xF0000000; static FileSys::VirtualDir GetDirectoryRelativeWrapped(FileSys::VirtualDir base, std::string_view dir_name_) { - std::string dir_name(FileUtil::SanitizePath(dir_name_)); + std::string dir_name(Common::FS::SanitizePath(dir_name_)); if (dir_name.empty() || dir_name == "." || dir_name == "/" || dir_name == "\\") return base; @@ -53,9 +53,13 @@ std::string VfsDirectoryServiceWrapper::GetName() const { } ResultCode VfsDirectoryServiceWrapper::CreateFile(const std::string& path_, u64 size) const { - std::string path(FileUtil::SanitizePath(path_)); - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path)); - auto file = dir->CreateFile(FileUtil::GetFilename(path)); + std::string path(Common::FS::SanitizePath(path_)); + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); + // dir can be nullptr if path contains subdirectories, create those prior to creating the file. + if (dir == nullptr) { + dir = backing->CreateSubdirectory(Common::FS::GetParentPath(path)); + } + auto file = dir->CreateFile(Common::FS::GetFilename(path)); if (file == nullptr) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; @@ -68,17 +72,17 @@ ResultCode VfsDirectoryServiceWrapper::CreateFile(const std::string& path_, u64 } ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) const { - std::string path(FileUtil::SanitizePath(path_)); + std::string path(Common::FS::SanitizePath(path_)); if (path.empty()) { // TODO(DarkLordZach): Why do games call this and what should it do? Works as is but... return RESULT_SUCCESS; } - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path)); - if (dir->GetFile(FileUtil::GetFilename(path)) == nullptr) { + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); + if (dir == nullptr || dir->GetFile(Common::FS::GetFilename(path)) == nullptr) { return FileSys::ERROR_PATH_NOT_FOUND; } - if (!dir->DeleteFile(FileUtil::GetFilename(path))) { + if (!dir->DeleteFile(Common::FS::GetFilename(path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -87,11 +91,12 @@ ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) cons } ResultCode VfsDirectoryServiceWrapper::CreateDirectory(const std::string& path_) const { - std::string path(FileUtil::SanitizePath(path_)); - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path)); - if (dir == nullptr && FileUtil::GetFilename(FileUtil::GetParentPath(path)).empty()) + std::string path(Common::FS::SanitizePath(path_)); + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); + if (dir == nullptr || Common::FS::GetFilename(Common::FS::GetParentPath(path)).empty()) { dir = backing; - auto new_dir = dir->CreateSubdirectory(FileUtil::GetFilename(path)); + } + auto new_dir = dir->CreateSubdirectory(Common::FS::GetFilename(path)); if (new_dir == nullptr) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; @@ -100,9 +105,9 @@ ResultCode VfsDirectoryServiceWrapper::CreateDirectory(const std::string& path_) } ResultCode VfsDirectoryServiceWrapper::DeleteDirectory(const std::string& path_) const { - std::string path(FileUtil::SanitizePath(path_)); - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path)); - if (!dir->DeleteSubdirectory(FileUtil::GetFilename(path))) { + std::string path(Common::FS::SanitizePath(path_)); + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); + if (!dir->DeleteSubdirectory(Common::FS::GetFilename(path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -110,9 +115,9 @@ ResultCode VfsDirectoryServiceWrapper::DeleteDirectory(const std::string& path_) } ResultCode VfsDirectoryServiceWrapper::DeleteDirectoryRecursively(const std::string& path_) const { - std::string path(FileUtil::SanitizePath(path_)); - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path)); - if (!dir->DeleteSubdirectoryRecursive(FileUtil::GetFilename(path))) { + std::string path(Common::FS::SanitizePath(path_)); + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); + if (!dir->DeleteSubdirectoryRecursive(Common::FS::GetFilename(path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -120,10 +125,10 @@ ResultCode VfsDirectoryServiceWrapper::DeleteDirectoryRecursively(const std::str } ResultCode VfsDirectoryServiceWrapper::CleanDirectoryRecursively(const std::string& path) const { - const std::string sanitized_path(FileUtil::SanitizePath(path)); - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(sanitized_path)); + const std::string sanitized_path(Common::FS::SanitizePath(path)); + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(sanitized_path)); - if (!dir->CleanSubdirectoryRecursive(FileUtil::GetFilename(sanitized_path))) { + if (!dir->CleanSubdirectoryRecursive(Common::FS::GetFilename(sanitized_path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -133,14 +138,14 @@ ResultCode VfsDirectoryServiceWrapper::CleanDirectoryRecursively(const std::stri ResultCode VfsDirectoryServiceWrapper::RenameFile(const std::string& src_path_, const std::string& dest_path_) const { - std::string src_path(FileUtil::SanitizePath(src_path_)); - std::string dest_path(FileUtil::SanitizePath(dest_path_)); + std::string src_path(Common::FS::SanitizePath(src_path_)); + std::string dest_path(Common::FS::SanitizePath(dest_path_)); auto src = backing->GetFileRelative(src_path); - if (FileUtil::GetParentPath(src_path) == FileUtil::GetParentPath(dest_path)) { + if (Common::FS::GetParentPath(src_path) == Common::FS::GetParentPath(dest_path)) { // Use more-optimized vfs implementation rename. if (src == nullptr) return FileSys::ERROR_PATH_NOT_FOUND; - if (!src->Rename(FileUtil::GetFilename(dest_path))) { + if (!src->Rename(Common::FS::GetFilename(dest_path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -158,7 +163,7 @@ ResultCode VfsDirectoryServiceWrapper::RenameFile(const std::string& src_path_, ASSERT_MSG(dest->WriteBytes(src->ReadAllBytes()) == src->GetSize(), "Could not write all of the bytes but everything else has succeded."); - if (!src->GetContainingDirectory()->DeleteFile(FileUtil::GetFilename(src_path))) { + if (!src->GetContainingDirectory()->DeleteFile(Common::FS::GetFilename(src_path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -168,14 +173,14 @@ ResultCode VfsDirectoryServiceWrapper::RenameFile(const std::string& src_path_, ResultCode VfsDirectoryServiceWrapper::RenameDirectory(const std::string& src_path_, const std::string& dest_path_) const { - std::string src_path(FileUtil::SanitizePath(src_path_)); - std::string dest_path(FileUtil::SanitizePath(dest_path_)); + std::string src_path(Common::FS::SanitizePath(src_path_)); + std::string dest_path(Common::FS::SanitizePath(dest_path_)); auto src = GetDirectoryRelativeWrapped(backing, src_path); - if (FileUtil::GetParentPath(src_path) == FileUtil::GetParentPath(dest_path)) { + if (Common::FS::GetParentPath(src_path) == Common::FS::GetParentPath(dest_path)) { // Use more-optimized vfs implementation rename. if (src == nullptr) return FileSys::ERROR_PATH_NOT_FOUND; - if (!src->Rename(FileUtil::GetFilename(dest_path))) { + if (!src->Rename(Common::FS::GetFilename(dest_path))) { // TODO(DarkLordZach): Find a better error code for this return RESULT_UNKNOWN; } @@ -194,7 +199,7 @@ ResultCode VfsDirectoryServiceWrapper::RenameDirectory(const std::string& src_pa ResultVal<FileSys::VirtualFile> VfsDirectoryServiceWrapper::OpenFile(const std::string& path_, FileSys::Mode mode) const { - const std::string path(FileUtil::SanitizePath(path_)); + const std::string path(Common::FS::SanitizePath(path_)); std::string_view npath = path; while (!npath.empty() && (npath[0] == '/' || npath[0] == '\\')) { npath.remove_prefix(1); @@ -214,7 +219,7 @@ ResultVal<FileSys::VirtualFile> VfsDirectoryServiceWrapper::OpenFile(const std:: } ResultVal<FileSys::VirtualDir> VfsDirectoryServiceWrapper::OpenDirectory(const std::string& path_) { - std::string path(FileUtil::SanitizePath(path_)); + std::string path(Common::FS::SanitizePath(path_)); auto dir = GetDirectoryRelativeWrapped(backing, path); if (dir == nullptr) { // TODO(DarkLordZach): Find a better error code for this @@ -225,11 +230,11 @@ ResultVal<FileSys::VirtualDir> VfsDirectoryServiceWrapper::OpenDirectory(const s ResultVal<FileSys::EntryType> VfsDirectoryServiceWrapper::GetEntryType( const std::string& path_) const { - std::string path(FileUtil::SanitizePath(path_)); - auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path)); + std::string path(Common::FS::SanitizePath(path_)); + auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); if (dir == nullptr) return FileSys::ERROR_PATH_NOT_FOUND; - auto filename = FileUtil::GetFilename(path); + auto filename = Common::FS::GetFilename(path); // TODO(Subv): Some games use the '/' path, find out what this means. if (filename.empty()) return MakeResult(FileSys::EntryType::Directory); @@ -307,7 +312,7 @@ ResultVal<FileSys::VirtualFile> FileSystemController::OpenRomFS( } ResultVal<FileSys::VirtualDir> FileSystemController::CreateSaveData( - FileSys::SaveDataSpaceId space, const FileSys::SaveDataDescriptor& save_struct) const { + FileSys::SaveDataSpaceId space, const FileSys::SaveDataAttribute& save_struct) const { LOG_TRACE(Service_FS, "Creating Save Data for space_id={:01X}, save_struct={}", static_cast<u8>(space), save_struct.DebugInfo()); @@ -319,15 +324,15 @@ ResultVal<FileSys::VirtualDir> FileSystemController::CreateSaveData( } ResultVal<FileSys::VirtualDir> FileSystemController::OpenSaveData( - FileSys::SaveDataSpaceId space, const FileSys::SaveDataDescriptor& descriptor) const { + FileSys::SaveDataSpaceId space, const FileSys::SaveDataAttribute& attribute) const { LOG_TRACE(Service_FS, "Opening Save Data for space_id={:01X}, save_struct={}", - static_cast<u8>(space), descriptor.DebugInfo()); + static_cast<u8>(space), attribute.DebugInfo()); if (save_data_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; } - return save_data_factory->Open(space, descriptor); + return save_data_factory->Open(space, attribute); } ResultVal<FileSys::VirtualDir> FileSystemController::OpenSaveDataSpace( @@ -375,7 +380,7 @@ ResultVal<FileSys::VirtualFile> FileSystemController::OpenBISPartitionStorage( return FileSys::ERROR_ENTITY_NOT_FOUND; } - auto part = bis_factory->OpenPartitionStorage(id); + auto part = bis_factory->OpenPartitionStorage(id, system.GetFilesystem()); if (part == nullptr) { return FileSys::ERROR_INVALID_ARGUMENT; } @@ -450,8 +455,11 @@ FileSys::SaveDataSize FileSystemController::ReadSaveDataSize(FileSys::SaveDataTy const auto res = system.GetAppLoader().ReadControlData(nacp); if (res != Loader::ResultStatus::Success) { - FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()}; - auto [nacp_unique, discard] = pm.GetControlMetadata(); + const FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID(), + system.GetFileSystemController(), + system.GetContentProvider()}; + const auto metadata = pm.GetControlMetadata(); + const auto& nacp_unique = metadata.first; if (nacp_unique != nullptr) { new_size = {nacp_unique->GetDefaultNormalSaveSize(), @@ -690,13 +698,13 @@ void FileSystemController::CreateFactories(FileSys::VfsFilesystem& vfs, bool ove sdmc_factory = nullptr; } - auto nand_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir), + auto nand_directory = vfs.OpenDirectory(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir), FileSys::Mode::ReadWrite); - auto sd_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir), + auto sd_directory = vfs.OpenDirectory(Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir), FileSys::Mode::ReadWrite); - auto load_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir), + auto load_directory = vfs.OpenDirectory(Common::FS::GetUserPath(Common::FS::UserPath::LoadDir), FileSys::Mode::ReadWrite); - auto dump_directory = vfs.OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::DumpDir), + auto dump_directory = vfs.OpenDirectory(Common::FS::GetUserPath(Common::FS::UserPath::DumpDir), FileSys::Mode::ReadWrite); if (bis_factory == nullptr) { @@ -722,7 +730,8 @@ void FileSystemController::CreateFactories(FileSys::VfsFilesystem& vfs, bool ove void InstallInterfaces(Core::System& system) { std::make_shared<FSP_LDR>()->InstallAsService(system.ServiceManager()); std::make_shared<FSP_PR>()->InstallAsService(system.ServiceManager()); - std::make_shared<FSP_SRV>(system.GetFileSystemController(), system.GetReporter()) + std::make_shared<FSP_SRV>(system.GetFileSystemController(), system.GetContentProvider(), + system.GetReporter()) ->InstallAsService(system.ServiceManager()); } diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h index 1b0a6a949..6dbbf0b2b 100644 --- a/src/core/hle/service/filesystem/filesystem.h +++ b/src/core/hle/service/filesystem/filesystem.h @@ -31,7 +31,7 @@ enum class SaveDataSpaceId : u8; enum class SaveDataType : u8; enum class StorageId : u8; -struct SaveDataDescriptor; +struct SaveDataAttribute; struct SaveDataSize; } // namespace FileSys @@ -69,9 +69,9 @@ public: ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id, FileSys::ContentRecordType type) const; ResultVal<FileSys::VirtualDir> CreateSaveData( - FileSys::SaveDataSpaceId space, const FileSys::SaveDataDescriptor& save_struct) const; + FileSys::SaveDataSpaceId space, const FileSys::SaveDataAttribute& save_struct) const; ResultVal<FileSys::VirtualDir> OpenSaveData( - FileSys::SaveDataSpaceId space, const FileSys::SaveDataDescriptor& save_struct) const; + FileSys::SaveDataSpaceId space, const FileSys::SaveDataAttribute& save_struct) const; ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space) const; ResultVal<FileSys::VirtualDir> OpenSDMC() const; ResultVal<FileSys::VirtualDir> OpenBISPartition(FileSys::BisPartitionId id) const; diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index e6811d5b5..031c6dbf6 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -316,8 +316,8 @@ public: {8, &IFileSystem::OpenFile, "OpenFile"}, {9, &IFileSystem::OpenDirectory, "OpenDirectory"}, {10, &IFileSystem::Commit, "Commit"}, - {11, nullptr, "GetFreeSpaceSize"}, - {12, nullptr, "GetTotalSpaceSize"}, + {11, &IFileSystem::GetFreeSpaceSize, "GetFreeSpaceSize"}, + {12, &IFileSystem::GetTotalSpaceSize, "GetTotalSpaceSize"}, {13, &IFileSystem::CleanDirectoryRecursively, "CleanDirectoryRecursively"}, {14, nullptr, "GetFileTimeStampRaw"}, {15, nullptr, "QueryEntry"}, @@ -575,6 +575,7 @@ private: 0, user_id->GetSize(), {}, + {}, }); continue; @@ -595,6 +596,7 @@ private: stoull_be(title_id->GetName()), title_id->GetSize(), {}, + {}, }); } } @@ -619,6 +621,7 @@ private: stoull_be(title_id->GetName()), title_id->GetSize(), {}, + {}, }); } } @@ -647,8 +650,10 @@ private: u64 next_entry_index = 0; }; -FSP_SRV::FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter) - : ServiceFramework("fsp-srv"), fsc(fsc), reporter(reporter) { +FSP_SRV::FSP_SRV(FileSystemController& fsc_, const FileSys::ContentProvider& content_provider_, + const Core::Reporter& reporter_) + : ServiceFramework("fsp-srv"), fsc(fsc_), content_provider{content_provider_}, + reporter(reporter_) { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "OpenFileSystem"}, @@ -693,13 +698,15 @@ FSP_SRV::FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter) {67, nullptr, "FindSaveDataWithFilter"}, {68, nullptr, "OpenSaveDataInfoReaderBySaveDataFilter"}, {69, nullptr, "ReadSaveDataFileSystemExtraDataBySaveDataAttribute"}, - {70, nullptr, "WriteSaveDataFileSystemExtraDataBySaveDataAttribute"}, + {70, &FSP_SRV::WriteSaveDataFileSystemExtraDataBySaveDataAttribute, "WriteSaveDataFileSystemExtraDataBySaveDataAttribute"}, + {71, &FSP_SRV::ReadSaveDataFileSystemExtraDataWithMaskBySaveDataAttribute, "ReadSaveDataFileSystemExtraDataWithMaskBySaveDataAttribute"}, {80, nullptr, "OpenSaveDataMetaFile"}, {81, nullptr, "OpenSaveDataTransferManager"}, {82, nullptr, "OpenSaveDataTransferManagerVersion2"}, {83, nullptr, "OpenSaveDataTransferProhibiterForCloudBackUp"}, {84, nullptr, "ListApplicationAccessibleSaveDataOwnerId"}, {85, nullptr, "OpenSaveDataTransferManagerForSaveDataRepair"}, + {86, nullptr, "OpenSaveDataMover"}, {100, nullptr, "OpenImageDirectoryFileSystem"}, {110, nullptr, "OpenContentStorageFileSystem"}, {120, nullptr, "OpenCloudBackupWorkStorageFileSystem"}, @@ -759,9 +766,11 @@ FSP_SRV::FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter) {1011, &FSP_SRV::GetAccessLogVersionInfo, "GetAccessLogVersionInfo"}, {1012, nullptr, "GetFsStackUsage"}, {1013, nullptr, "UnsetSaveDataRootPath"}, + {1014, nullptr, "OutputMultiProgramTagAccessLog"}, {1100, nullptr, "OverrideSaveDataTransferTokenSignVerificationKey"}, {1110, nullptr, "CorruptSaveDataFileSystemBySaveDataSpaceId2"}, - {1200, nullptr, "OpenMultiCommitManager"}, + {1200, &FSP_SRV::OpenMultiCommitManager, "OpenMultiCommitManager"}, + {1300, nullptr, "OpenBisWiper"}, }; // clang-format on RegisterHandlers(functions); @@ -805,7 +814,7 @@ void FSP_SRV::OpenSdCardFileSystem(Kernel::HLERequestContext& ctx) { void FSP_SRV::CreateSaveDataFileSystem(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto save_struct = rp.PopRaw<FileSys::SaveDataDescriptor>(); + auto save_struct = rp.PopRaw<FileSys::SaveDataAttribute>(); [[maybe_unused]] auto save_create_struct = rp.PopRaw<std::array<u8, 0x40>>(); u128 uid = rp.PopRaw<u128>(); @@ -819,31 +828,40 @@ void FSP_SRV::CreateSaveDataFileSystem(Kernel::HLERequestContext& ctx) { } void FSP_SRV::OpenSaveDataFileSystem(Kernel::HLERequestContext& ctx) { - LOG_INFO(Service_FS, "called."); + IPC::RequestParser rp{ctx}; struct Parameters { - FileSys::SaveDataSpaceId save_data_space_id; - FileSys::SaveDataDescriptor descriptor; + FileSys::SaveDataSpaceId space_id; + FileSys::SaveDataAttribute attribute; }; - IPC::RequestParser rp{ctx}; const auto parameters = rp.PopRaw<Parameters>(); - auto dir = fsc.OpenSaveData(parameters.save_data_space_id, parameters.descriptor); + LOG_INFO(Service_FS, "called."); + + auto dir = fsc.OpenSaveData(parameters.space_id, parameters.attribute); if (dir.Failed()) { IPC::ResponseBuilder rb{ctx, 2, 0, 0}; rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND); return; } - FileSys::StorageId id; - if (parameters.save_data_space_id == FileSys::SaveDataSpaceId::NandUser) { + FileSys::StorageId id{}; + switch (parameters.space_id) { + case FileSys::SaveDataSpaceId::NandUser: id = FileSys::StorageId::NandUser; - } else if (parameters.save_data_space_id == FileSys::SaveDataSpaceId::SdCardSystem || - parameters.save_data_space_id == FileSys::SaveDataSpaceId::SdCardUser) { + break; + case FileSys::SaveDataSpaceId::SdCardSystem: + case FileSys::SaveDataSpaceId::SdCardUser: id = FileSys::StorageId::SdCard; - } else { + break; + case FileSys::SaveDataSpaceId::NandSystem: id = FileSys::StorageId::NandSystem; + break; + case FileSys::SaveDataSpaceId::TemporaryStorage: + case FileSys::SaveDataSpaceId::ProperSystem: + case FileSys::SaveDataSpaceId::SafeMode: + UNREACHABLE(); } auto filesystem = @@ -869,22 +887,38 @@ void FSP_SRV::OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext& rb.PushIpcInterface<ISaveDataInfoReader>(std::make_shared<ISaveDataInfoReader>(space, fsc)); } -void FSP_SRV::SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - log_mode = rp.PopEnum<LogMode>(); - - LOG_DEBUG(Service_FS, "called, log_mode={:08X}", static_cast<u32>(log_mode)); +void FSP_SRV::WriteSaveDataFileSystemExtraDataBySaveDataAttribute(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_FS, "(STUBBED) called."); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_FS, "called"); +void FSP_SRV::ReadSaveDataFileSystemExtraDataWithMaskBySaveDataAttribute( + Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + struct Parameters { + FileSys::SaveDataSpaceId space_id; + FileSys::SaveDataAttribute attribute; + }; + + const auto parameters = rp.PopRaw<Parameters>(); + // Stub this to None for now, backend needs an impl to read/write the SaveDataExtraData + constexpr auto flags = static_cast<u32>(FileSys::SaveDataFlags::None); + + LOG_WARNING(Service_FS, + "(STUBBED) called, flags={}, space_id={}, attribute.title_id={:016X}\n" + "attribute.user_id={:016X}{:016X}, attribute.save_id={:016X}\n" + "attribute.type={}, attribute.rank={}, attribute.index={}", + flags, static_cast<u32>(parameters.space_id), parameters.attribute.title_id, + parameters.attribute.user_id[1], parameters.attribute.user_id[0], + parameters.attribute.save_id, static_cast<u32>(parameters.attribute.type), + static_cast<u32>(parameters.attribute.rank), parameters.attribute.index); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.PushEnum(log_mode); + rb.Push(flags); } void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) { @@ -936,7 +970,7 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) { return; } - FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, fsc, content_provider}; auto storage = std::make_shared<IStorage>( pm.PatchRomFS(std::move(data.Unwrap()), 0, FileSys::ContentRecordType::Data)); @@ -959,6 +993,24 @@ void FSP_SRV::OpenPatchDataStorageByCurrentProcess(Kernel::HLERequestContext& ct rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND); } +void FSP_SRV::SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + log_mode = rp.PopEnum<LogMode>(); + + LOG_DEBUG(Service_FS, "called, log_mode={:08X}", static_cast<u32>(log_mode)); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_FS, "called"); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(log_mode); +} + void FSP_SRV::OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx) { const auto raw = ctx.ReadBuffer(); auto log = Common::StringFromFixedZeroTerminatedBuffer( @@ -981,4 +1033,40 @@ void FSP_SRV::GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx) { rb.Push(access_log_program_index); } +class IMultiCommitManager final : public ServiceFramework<IMultiCommitManager> { +public: + explicit IMultiCommitManager() : ServiceFramework("IMultiCommitManager") { + static const FunctionInfo functions[] = { + {1, &IMultiCommitManager::Add, "Add"}, + {2, &IMultiCommitManager::Commit, "Commit"}, + }; + RegisterHandlers(functions); + } + +private: + FileSys::VirtualFile backend; + + void Add(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_FS, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void Commit(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_FS, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } +}; + +void FSP_SRV::OpenMultiCommitManager(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_FS, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IMultiCommitManager>(std::make_shared<IMultiCommitManager>()); +} + } // namespace Service::FileSystem diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h index d52b55999..6c7239e6a 100644 --- a/src/core/hle/service/filesystem/fsp_srv.h +++ b/src/core/hle/service/filesystem/fsp_srv.h @@ -12,8 +12,9 @@ class Reporter; } namespace FileSys { +class ContentProvider; class FileSystemBackend; -} +} // namespace FileSys namespace Service::FileSystem { @@ -32,7 +33,8 @@ enum class LogMode : u32 { class FSP_SRV final : public ServiceFramework<FSP_SRV> { public: - explicit FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter); + explicit FSP_SRV(FileSystemController& fsc_, const FileSys::ContentProvider& content_provider_, + const Core::Reporter& reporter_); ~FSP_SRV() override; private: @@ -43,15 +45,19 @@ private: void OpenSaveDataFileSystem(Kernel::HLERequestContext& ctx); void OpenReadOnlySaveDataFileSystem(Kernel::HLERequestContext& ctx); void OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext& ctx); - void SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); - void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); + void WriteSaveDataFileSystemExtraDataBySaveDataAttribute(Kernel::HLERequestContext& ctx); + void ReadSaveDataFileSystemExtraDataWithMaskBySaveDataAttribute(Kernel::HLERequestContext& ctx); void OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx); void OpenDataStorageByDataId(Kernel::HLERequestContext& ctx); void OpenPatchDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx); + void SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); + void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); void OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx); void GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx); + void OpenMultiCommitManager(Kernel::HLERequestContext& ctx); FileSystemController& fsc; + const FileSys::ContentProvider& content_provider; FileSys::VirtualFile romfs; u64 current_process_id = 0; diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp index 7938b4b80..ebb323da2 100644 --- a/src/core/hle/service/friend/friend.cpp +++ b/src/core/hle/service/friend/friend.cpp @@ -5,6 +5,7 @@ #include <queue> #include "common/logging/log.h" #include "common/uuid.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/writable_event.h" @@ -25,9 +26,13 @@ public: {10101, &IFriendService::GetFriendList, "GetFriendList"}, {10102, nullptr, "UpdateFriendInfo"}, {10110, nullptr, "GetFriendProfileImage"}, + {10120, nullptr, "Unknown10120"}, + {10121, nullptr, "Unknown10121"}, {10200, nullptr, "SendFriendRequestForApplication"}, {10211, nullptr, "AddFacedFriendRequestForApplication"}, {10400, &IFriendService::GetBlockedUserListIds, "GetBlockedUserListIds"}, + {10420, nullptr, "Unknown10420"}, + {10421, nullptr, "Unknown10421"}, {10500, nullptr, "GetProfileList"}, {10600, nullptr, "DeclareOpenOnlinePlaySession"}, {10601, &IFriendService::DeclareCloseOnlinePlaySession, "DeclareCloseOnlinePlaySession"}, @@ -96,6 +101,9 @@ public: {30830, nullptr, "ClearPlayLog"}, {30900, nullptr, "SendFriendInvitation"}, {30910, nullptr, "ReadFriendInvitation"}, + {30911, nullptr, "ReadAllFriendInvitations"}, + {40100, nullptr, "Unknown40100"}, + {40400, nullptr, "Unknown40400"}, {49900, nullptr, "DeleteNetworkServiceAccountCache"}, }; // clang-format on diff --git a/src/core/hle/service/glue/arp.cpp b/src/core/hle/service/glue/arp.cpp index b591ce31b..c6252ff89 100644 --- a/src/core/hle/service/glue/arp.cpp +++ b/src/core/hle/service/glue/arp.cpp @@ -5,6 +5,7 @@ #include <memory> #include "common/logging/log.h" +#include "core/core.h" #include "core/file_sys/control_metadata.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" diff --git a/src/core/hle/service/glue/errors.h b/src/core/hle/service/glue/errors.h index c2874c585..f6647f724 100644 --- a/src/core/hle/service/glue/errors.h +++ b/src/core/hle/service/glue/errors.h @@ -8,9 +8,9 @@ namespace Service::Glue { -constexpr ResultCode ERR_INVALID_RESOURCE{ErrorModule::ARP, 0x1E}; -constexpr ResultCode ERR_INVALID_PROCESS_ID{ErrorModule::ARP, 0x1F}; -constexpr ResultCode ERR_INVALID_ACCESS{ErrorModule::ARP, 0x2A}; -constexpr ResultCode ERR_NOT_REGISTERED{ErrorModule::ARP, 0x66}; +constexpr ResultCode ERR_INVALID_RESOURCE{ErrorModule::ARP, 30}; +constexpr ResultCode ERR_INVALID_PROCESS_ID{ErrorModule::ARP, 31}; +constexpr ResultCode ERR_INVALID_ACCESS{ErrorModule::ARP, 42}; +constexpr ResultCode ERR_NOT_REGISTERED{ErrorModule::ARP, 102}; } // namespace Service::Glue diff --git a/src/core/hle/service/grc/grc.cpp b/src/core/hle/service/grc/grc.cpp index 24910ac6c..401e0b208 100644 --- a/src/core/hle/service/grc/grc.cpp +++ b/src/core/hle/service/grc/grc.cpp @@ -17,6 +17,9 @@ public: static const FunctionInfo functions[] = { {1, nullptr, "OpenContinuousRecorder"}, {2, nullptr, "OpenGameMovieTrimmer"}, + {3, nullptr, "OpenOffscreenRecorder"}, + {101, nullptr, "CreateMovieMaker"}, + {9903, nullptr, "SetOffscreenRecordingMarker"} }; // clang-format on diff --git a/src/core/hle/service/hid/controllers/controller_base.h b/src/core/hle/service/hid/controllers/controller_base.h index 8bc69c372..f47a9e61c 100644 --- a/src/core/hle/service/hid/controllers/controller_base.h +++ b/src/core/hle/service/hid/controllers/controller_base.h @@ -31,6 +31,10 @@ public: virtual void OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) = 0; + // When the controller is requesting a motion update for the shared memory + virtual void OnMotionUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, + std::size_t size) {} + // Called when input devices should be loaded virtual void OnLoadInputDevices() = 0; diff --git a/src/core/hle/service/hid/controllers/debug_pad.cpp b/src/core/hle/service/hid/controllers/debug_pad.cpp index 1f2131ec8..ad251ed4a 100644 --- a/src/core/hle/service/hid/controllers/debug_pad.cpp +++ b/src/core/hle/service/hid/controllers/debug_pad.cpp @@ -23,7 +23,7 @@ void Controller_DebugPad::OnRelease() {} void Controller_DebugPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) { - shared_memory.header.timestamp = core_timing.GetTicks(); + shared_memory.header.timestamp = core_timing.GetCPUTicks(); shared_memory.header.total_entry_count = 17; if (!IsControllerActivated()) { @@ -39,33 +39,36 @@ void Controller_DebugPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, cur_entry.sampling_number = last_entry.sampling_number + 1; cur_entry.sampling_number2 = cur_entry.sampling_number; - cur_entry.attribute.connected.Assign(1); - auto& pad = cur_entry.pad_state; - using namespace Settings::NativeButton; - pad.a.Assign(buttons[A - BUTTON_HID_BEGIN]->GetStatus()); - pad.b.Assign(buttons[B - BUTTON_HID_BEGIN]->GetStatus()); - pad.x.Assign(buttons[X - BUTTON_HID_BEGIN]->GetStatus()); - pad.y.Assign(buttons[Y - BUTTON_HID_BEGIN]->GetStatus()); - pad.l.Assign(buttons[L - BUTTON_HID_BEGIN]->GetStatus()); - pad.r.Assign(buttons[R - BUTTON_HID_BEGIN]->GetStatus()); - pad.zl.Assign(buttons[ZL - BUTTON_HID_BEGIN]->GetStatus()); - pad.zr.Assign(buttons[ZR - BUTTON_HID_BEGIN]->GetStatus()); - pad.plus.Assign(buttons[Plus - BUTTON_HID_BEGIN]->GetStatus()); - pad.minus.Assign(buttons[Minus - BUTTON_HID_BEGIN]->GetStatus()); - pad.d_left.Assign(buttons[DLeft - BUTTON_HID_BEGIN]->GetStatus()); - pad.d_up.Assign(buttons[DUp - BUTTON_HID_BEGIN]->GetStatus()); - pad.d_right.Assign(buttons[DRight - BUTTON_HID_BEGIN]->GetStatus()); - pad.d_down.Assign(buttons[DDown - BUTTON_HID_BEGIN]->GetStatus()); + if (Settings::values.debug_pad_enabled) { + cur_entry.attribute.connected.Assign(1); + auto& pad = cur_entry.pad_state; - const auto [stick_l_x_f, stick_l_y_f] = - analogs[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetStatus(); - const auto [stick_r_x_f, stick_r_y_f] = - analogs[static_cast<std::size_t>(JoystickId::Joystick_Right)]->GetStatus(); - cur_entry.l_stick.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX); - cur_entry.l_stick.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); - cur_entry.r_stick.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX); - cur_entry.r_stick.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX); + using namespace Settings::NativeButton; + pad.a.Assign(buttons[A - BUTTON_HID_BEGIN]->GetStatus()); + pad.b.Assign(buttons[B - BUTTON_HID_BEGIN]->GetStatus()); + pad.x.Assign(buttons[X - BUTTON_HID_BEGIN]->GetStatus()); + pad.y.Assign(buttons[Y - BUTTON_HID_BEGIN]->GetStatus()); + pad.l.Assign(buttons[L - BUTTON_HID_BEGIN]->GetStatus()); + pad.r.Assign(buttons[R - BUTTON_HID_BEGIN]->GetStatus()); + pad.zl.Assign(buttons[ZL - BUTTON_HID_BEGIN]->GetStatus()); + pad.zr.Assign(buttons[ZR - BUTTON_HID_BEGIN]->GetStatus()); + pad.plus.Assign(buttons[Plus - BUTTON_HID_BEGIN]->GetStatus()); + pad.minus.Assign(buttons[Minus - BUTTON_HID_BEGIN]->GetStatus()); + pad.d_left.Assign(buttons[DLeft - BUTTON_HID_BEGIN]->GetStatus()); + pad.d_up.Assign(buttons[DUp - BUTTON_HID_BEGIN]->GetStatus()); + pad.d_right.Assign(buttons[DRight - BUTTON_HID_BEGIN]->GetStatus()); + pad.d_down.Assign(buttons[DDown - BUTTON_HID_BEGIN]->GetStatus()); + + const auto [stick_l_x_f, stick_l_y_f] = + analogs[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetStatus(); + const auto [stick_r_x_f, stick_r_y_f] = + analogs[static_cast<std::size_t>(JoystickId::Joystick_Right)]->GetStatus(); + cur_entry.l_stick.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX); + cur_entry.l_stick.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); + cur_entry.r_stick.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX); + cur_entry.r_stick.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX); + } std::memcpy(data, &shared_memory, sizeof(SharedMemory)); } diff --git a/src/core/hle/service/hid/controllers/gesture.cpp b/src/core/hle/service/hid/controllers/gesture.cpp index 6e990dd00..b7b7bfeae 100644 --- a/src/core/hle/service/hid/controllers/gesture.cpp +++ b/src/core/hle/service/hid/controllers/gesture.cpp @@ -19,7 +19,7 @@ void Controller_Gesture::OnRelease() {} void Controller_Gesture::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) { - shared_memory.header.timestamp = core_timing.GetTicks(); + shared_memory.header.timestamp = core_timing.GetCPUTicks(); shared_memory.header.total_entry_count = 17; if (!IsControllerActivated()) { diff --git a/src/core/hle/service/hid/controllers/keyboard.cpp b/src/core/hle/service/hid/controllers/keyboard.cpp index 358cb9329..59b694cd4 100644 --- a/src/core/hle/service/hid/controllers/keyboard.cpp +++ b/src/core/hle/service/hid/controllers/keyboard.cpp @@ -21,7 +21,7 @@ void Controller_Keyboard::OnRelease() {} void Controller_Keyboard::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) { - shared_memory.header.timestamp = core_timing.GetTicks(); + shared_memory.header.timestamp = core_timing.GetCPUTicks(); shared_memory.header.total_entry_count = 17; if (!IsControllerActivated()) { @@ -38,16 +38,18 @@ void Controller_Keyboard::OnUpdate(const Core::Timing::CoreTiming& core_timing, cur_entry.sampling_number = last_entry.sampling_number + 1; cur_entry.sampling_number2 = cur_entry.sampling_number; - for (std::size_t i = 0; i < keyboard_keys.size(); ++i) { - for (std::size_t k = 0; k < KEYS_PER_BYTE; ++k) { - cur_entry.key[i / KEYS_PER_BYTE] |= (keyboard_keys[i]->GetStatus() << k); + cur_entry.key.fill(0); + cur_entry.modifier = 0; + if (Settings::values.keyboard_enabled) { + for (std::size_t i = 0; i < keyboard_keys.size(); ++i) { + auto& entry = cur_entry.key[i / KEYS_PER_BYTE]; + entry = static_cast<u8>(entry | (keyboard_keys[i]->GetStatus() << (i % KEYS_PER_BYTE))); } - } - for (std::size_t i = 0; i < keyboard_mods.size(); ++i) { - cur_entry.modifier |= (keyboard_mods[i]->GetStatus() << i); + for (std::size_t i = 0; i < keyboard_mods.size(); ++i) { + cur_entry.modifier |= (keyboard_mods[i]->GetStatus() << i); + } } - std::memcpy(data + SHARED_MEMORY_OFFSET, &shared_memory, sizeof(SharedMemory)); } diff --git a/src/core/hle/service/hid/controllers/mouse.cpp b/src/core/hle/service/hid/controllers/mouse.cpp index 93d88ea50..ac40989c5 100644 --- a/src/core/hle/service/hid/controllers/mouse.cpp +++ b/src/core/hle/service/hid/controllers/mouse.cpp @@ -19,7 +19,7 @@ void Controller_Mouse::OnRelease() {} void Controller_Mouse::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) { - shared_memory.header.timestamp = core_timing.GetTicks(); + shared_memory.header.timestamp = core_timing.GetCPUTicks(); shared_memory.header.total_entry_count = 17; if (!IsControllerActivated()) { diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp index c1e32b28c..e2539ded8 100644 --- a/src/core/hle/service/hid/controllers/npad.cpp +++ b/src/core/hle/service/hid/controllers/npad.cpp @@ -24,6 +24,7 @@ constexpr s32 HID_JOYSTICK_MAX = 0x7fff; constexpr std::size_t NPAD_OFFSET = 0x9A00; constexpr u32 BATTERY_FULL = 2; constexpr u32 MAX_NPAD_ID = 7; +constexpr std::size_t HANDHELD_INDEX = 8; constexpr std::array<u32, 10> npad_id_list{ 0, 1, 2, 3, 4, 5, 6, 7, NPAD_HANDHELD, NPAD_UNKNOWN, }; @@ -33,19 +34,41 @@ enum class JoystickId : std::size_t { Joystick_Right, }; -static Controller_NPad::NPadControllerType MapSettingsTypeToNPad(Settings::ControllerType type) { +Controller_NPad::NPadControllerType Controller_NPad::MapSettingsTypeToNPad( + Settings::ControllerType type) { switch (type) { case Settings::ControllerType::ProController: - return Controller_NPad::NPadControllerType::ProController; - case Settings::ControllerType::DualJoycon: - return Controller_NPad::NPadControllerType::JoyDual; + return NPadControllerType::ProController; + case Settings::ControllerType::DualJoyconDetached: + return NPadControllerType::JoyDual; case Settings::ControllerType::LeftJoycon: - return Controller_NPad::NPadControllerType::JoyLeft; + return NPadControllerType::JoyLeft; case Settings::ControllerType::RightJoycon: - return Controller_NPad::NPadControllerType::JoyRight; + return NPadControllerType::JoyRight; + case Settings::ControllerType::Handheld: + return NPadControllerType::Handheld; default: UNREACHABLE(); - return Controller_NPad::NPadControllerType::JoyDual; + return NPadControllerType::ProController; + } +} + +Settings::ControllerType Controller_NPad::MapNPadToSettingsType( + Controller_NPad::NPadControllerType type) { + switch (type) { + case NPadControllerType::ProController: + return Settings::ControllerType::ProController; + case NPadControllerType::JoyDual: + return Settings::ControllerType::DualJoyconDetached; + case NPadControllerType::JoyLeft: + return Settings::ControllerType::LeftJoycon; + case NPadControllerType::JoyRight: + return Settings::ControllerType::RightJoycon; + case NPadControllerType::Handheld: + return Settings::ControllerType::Handheld; + default: + UNREACHABLE(); + return Settings::ControllerType::ProController; } } @@ -60,9 +83,9 @@ std::size_t Controller_NPad::NPadIdToIndex(u32 npad_id) { case 6: case 7: return npad_id; - case 8: + case HANDHELD_INDEX: case NPAD_HANDHELD: - return 8; + return HANDHELD_INDEX; case 9: case NPAD_UNKNOWN: return 9; @@ -83,37 +106,51 @@ u32 Controller_NPad::IndexToNPad(std::size_t index) { case 6: case 7: return static_cast<u32>(index); - case 8: + case HANDHELD_INDEX: return NPAD_HANDHELD; case 9: return NPAD_UNKNOWN; default: UNIMPLEMENTED_MSG("Unknown npad index {}", index); return 0; - }; + } } Controller_NPad::Controller_NPad(Core::System& system) : ControllerBase(system), system(system) {} -Controller_NPad::~Controller_NPad() = default; -void Controller_NPad::InitNewlyAddedControler(std::size_t controller_idx) { +Controller_NPad::~Controller_NPad() { + OnRelease(); +} + +void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { const auto controller_type = connected_controllers[controller_idx].type; auto& controller = shared_memory_entries[controller_idx]; if (controller_type == NPadControllerType::None) { + styleset_changed_events[controller_idx].writable->Signal(); return; } controller.joy_styles.raw = 0; // Zero out controller.device_type.raw = 0; + controller.properties.raw = 0; switch (controller_type) { case NPadControllerType::None: UNREACHABLE(); + break; + case NPadControllerType::ProController: + controller.joy_styles.pro_controller.Assign(1); + controller.device_type.pro_controller.Assign(1); + controller.properties.is_vertical.Assign(1); + controller.properties.use_plus.Assign(1); + controller.properties.use_minus.Assign(1); + controller.pad_assignment = NpadAssignments::Single; + break; case NPadControllerType::Handheld: controller.joy_styles.handheld.Assign(1); controller.device_type.handheld.Assign(1); - controller.pad_assignment = NPadAssignments::Dual; controller.properties.is_vertical.Assign(1); controller.properties.use_plus.Assign(1); controller.properties.use_minus.Assign(1); + controller.pad_assignment = NpadAssignments::Dual; break; case NPadControllerType::JoyDual: controller.joy_styles.joycon_dual.Assign(1); @@ -122,34 +159,26 @@ void Controller_NPad::InitNewlyAddedControler(std::size_t controller_idx) { controller.properties.is_vertical.Assign(1); controller.properties.use_plus.Assign(1); controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Dual; + controller.pad_assignment = NpadAssignments::Dual; break; case NPadControllerType::JoyLeft: controller.joy_styles.joycon_left.Assign(1); controller.device_type.joycon_left.Assign(1); controller.properties.is_horizontal.Assign(1); controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; case NPadControllerType::JoyRight: controller.joy_styles.joycon_right.Assign(1); controller.device_type.joycon_right.Assign(1); controller.properties.is_horizontal.Assign(1); controller.properties.use_plus.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; case NPadControllerType::Pokeball: controller.joy_styles.pokeball.Assign(1); controller.device_type.pokeball.Assign(1); - controller.pad_assignment = NPadAssignments::Single; - break; - case NPadControllerType::ProController: - controller.joy_styles.pro_controller.Assign(1); - controller.device_type.pro_controller.Assign(1); - controller.properties.is_vertical.Assign(1); - controller.properties.use_plus.Assign(1); - controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; } @@ -158,21 +187,25 @@ void Controller_NPad::InitNewlyAddedControler(std::size_t controller_idx) { controller.single_color.button_color = 0; controller.dual_color_error = ColorReadError::ReadOk; - controller.left_color.body_color = Settings::values.players[controller_idx].body_color_left; - controller.left_color.button_color = Settings::values.players[controller_idx].button_color_left; - controller.right_color.body_color = Settings::values.players[controller_idx].body_color_right; + controller.left_color.body_color = + Settings::values.players.GetValue()[controller_idx].body_color_left; + controller.left_color.button_color = + Settings::values.players.GetValue()[controller_idx].button_color_left; + controller.right_color.body_color = + Settings::values.players.GetValue()[controller_idx].body_color_right; controller.right_color.button_color = - Settings::values.players[controller_idx].button_color_right; + Settings::values.players.GetValue()[controller_idx].button_color_right; controller.battery_level[0] = BATTERY_FULL; controller.battery_level[1] = BATTERY_FULL; controller.battery_level[2] = BATTERY_FULL; - styleset_changed_events[controller_idx].writable->Signal(); + + SignalStyleSetChangedEvent(IndexToNPad(controller_idx)); } void Controller_NPad::OnInit() { auto& kernel = system.Kernel(); - for (std::size_t i = 0; i < styleset_changed_events.size(); i++) { + for (std::size_t i = 0; i < styleset_changed_events.size(); ++i) { styleset_changed_events[i] = Kernel::WritableEvent::CreateEventPair( kernel, fmt::format("npad:NpadStyleSetChanged_{}", i)); } @@ -181,6 +214,8 @@ void Controller_NPad::OnInit() { return; } + OnLoadInputDevices(); + if (style.raw == 0) { // We want to support all controllers style.handheld.Assign(1); @@ -191,42 +226,46 @@ void Controller_NPad::OnInit() { style.pokeball.Assign(1); } - std::transform( - Settings::values.players.begin(), Settings::values.players.end(), - connected_controllers.begin(), [](const Settings::PlayerInput& player) { - return ControllerHolder{MapSettingsTypeToNPad(player.type), player.connected}; - }); + std::transform(Settings::values.players.GetValue().begin(), + Settings::values.players.GetValue().end(), connected_controllers.begin(), + [](const Settings::PlayerInput& player) { + return ControllerHolder{MapSettingsTypeToNPad(player.controller_type), + player.connected}; + }); - std::stable_partition(connected_controllers.begin(), connected_controllers.begin() + 8, - [](const ControllerHolder& holder) { return holder.is_connected; }); + // Connect the Player 1 or Handheld controller if none are connected. + if (std::none_of(connected_controllers.begin(), connected_controllers.end(), + [](const ControllerHolder& controller) { return controller.is_connected; })) { + const auto controller = + MapSettingsTypeToNPad(Settings::values.players.GetValue()[0].controller_type); + if (controller == NPadControllerType::Handheld) { + Settings::values.players.GetValue()[HANDHELD_INDEX].connected = true; + connected_controllers[HANDHELD_INDEX] = {controller, true}; + } else { + Settings::values.players.GetValue()[0].connected = true; + connected_controllers[0] = {controller, true}; + } + } // Account for handheld - if (connected_controllers[8].is_connected) - connected_controllers[8].type = NPadControllerType::Handheld; + if (connected_controllers[HANDHELD_INDEX].is_connected) { + connected_controllers[HANDHELD_INDEX].type = NPadControllerType::Handheld; + } supported_npad_id_types.resize(npad_id_list.size()); std::memcpy(supported_npad_id_types.data(), npad_id_list.data(), npad_id_list.size() * sizeof(u32)); - // Add a default dual joycon controller if none are present. - if (std::none_of(connected_controllers.begin(), connected_controllers.end(), - [](const ControllerHolder& controller) { return controller.is_connected; })) { - supported_npad_id_types.resize(npad_id_list.size()); - std::memcpy(supported_npad_id_types.data(), npad_id_list.data(), - npad_id_list.size() * sizeof(u32)); - AddNewController(NPadControllerType::JoyDual); - } - for (std::size_t i = 0; i < connected_controllers.size(); ++i) { const auto& controller = connected_controllers[i]; if (controller.is_connected) { - AddNewControllerAt(controller.type, IndexToNPad(i)); + AddNewControllerAt(controller.type, i); } } } void Controller_NPad::OnLoadInputDevices() { - const auto& players = Settings::values.players; + const auto& players = Settings::values.players.GetValue(); for (std::size_t i = 0; i < players.size(); ++i) { std::transform(players[i].buttons.begin() + Settings::NativeButton::BUTTON_HID_BEGIN, players[i].buttons.begin() + Settings::NativeButton::BUTTON_HID_END, @@ -234,14 +273,30 @@ void Controller_NPad::OnLoadInputDevices() { std::transform(players[i].analogs.begin() + Settings::NativeAnalog::STICK_HID_BEGIN, players[i].analogs.begin() + Settings::NativeAnalog::STICK_HID_END, sticks[i].begin(), Input::CreateDevice<Input::AnalogDevice>); + std::transform(players[i].vibrations.begin() + + Settings::NativeVibration::VIBRATION_HID_BEGIN, + players[i].vibrations.begin() + Settings::NativeVibration::VIBRATION_HID_END, + vibrations[i].begin(), Input::CreateDevice<Input::VibrationDevice>); + std::transform(players[i].motions.begin() + Settings::NativeMotion::MOTION_HID_BEGIN, + players[i].motions.begin() + Settings::NativeMotion::MOTION_HID_END, + motions[i].begin(), Input::CreateDevice<Input::MotionDevice>); + for (std::size_t device_idx = 0; device_idx < vibrations[i].size(); ++device_idx) { + InitializeVibrationDeviceAtIndex(i, device_idx); + } } } -void Controller_NPad::OnRelease() {} +void Controller_NPad::OnRelease() { + for (std::size_t npad_idx = 0; npad_idx < vibrations.size(); ++npad_idx) { + for (std::size_t device_idx = 0; device_idx < vibrations[npad_idx].size(); ++device_idx) { + VibrateControllerAtIndex(npad_idx, device_idx, {}); + } + } +} void Controller_NPad::RequestPadStateUpdate(u32 npad_id) { const auto controller_idx = NPadIdToIndex(npad_id); - [[maybe_unused]] const auto controller_type = connected_controllers[controller_idx].type; + const auto controller_type = connected_controllers[controller_idx].type; if (!connected_controllers[controller_idx].is_connected) { return; } @@ -256,61 +311,71 @@ void Controller_NPad::RequestPadStateUpdate(u32 npad_id) { analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)]->GetStatus(); using namespace Settings::NativeButton; - pad_state.a.Assign(button_state[A - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.b.Assign(button_state[B - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.x.Assign(button_state[X - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.y.Assign(button_state[Y - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.l_stick.Assign(button_state[LStick - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.r_stick.Assign(button_state[RStick - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.l.Assign(button_state[L - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.r.Assign(button_state[R - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.zl.Assign(button_state[ZL - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.zr.Assign(button_state[ZR - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.plus.Assign(button_state[Plus - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.minus.Assign(button_state[Minus - BUTTON_HID_BEGIN]->GetStatus()); - - pad_state.d_left.Assign(button_state[DLeft - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.d_up.Assign(button_state[DUp - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.d_right.Assign(button_state[DRight - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.d_down.Assign(button_state[DDown - BUTTON_HID_BEGIN]->GetStatus()); - - pad_state.l_stick_right.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::RIGHT)); - pad_state.l_stick_left.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::LEFT)); - pad_state.l_stick_up.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::UP)); - pad_state.l_stick_down.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::DOWN)); - - pad_state.r_stick_right.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::RIGHT)); - pad_state.r_stick_left.Assign(analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::LEFT)); - pad_state.r_stick_up.Assign(analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::UP)); - pad_state.r_stick_down.Assign(analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::DOWN)); - - pad_state.left_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.left_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus()); - - lstick_entry.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX); - lstick_entry.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); - rstick_entry.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX); - rstick_entry.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX); + if (controller_type != NPadControllerType::JoyLeft) { + pad_state.a.Assign(button_state[A - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.b.Assign(button_state[B - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.x.Assign(button_state[X - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.y.Assign(button_state[Y - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.r_stick.Assign(button_state[RStick - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.r.Assign(button_state[R - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.zr.Assign(button_state[ZR - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.plus.Assign(button_state[Plus - BUTTON_HID_BEGIN]->GetStatus()); + + pad_state.r_stick_right.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::RIGHT)); + pad_state.r_stick_left.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::LEFT)); + pad_state.r_stick_up.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::UP)); + pad_state.r_stick_down.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::DOWN)); + rstick_entry.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX); + rstick_entry.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX); + } + + if (controller_type != NPadControllerType::JoyRight) { + pad_state.d_left.Assign(button_state[DLeft - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.d_up.Assign(button_state[DUp - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.d_right.Assign(button_state[DRight - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.d_down.Assign(button_state[DDown - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.l_stick.Assign(button_state[LStick - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.l.Assign(button_state[L - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.zl.Assign(button_state[ZL - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.minus.Assign(button_state[Minus - BUTTON_HID_BEGIN]->GetStatus()); + + pad_state.l_stick_right.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::RIGHT)); + pad_state.l_stick_left.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::LEFT)); + pad_state.l_stick_up.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::UP)); + pad_state.l_stick_down.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::DOWN)); + lstick_entry.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX); + lstick_entry.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); + } + + if (controller_type == NPadControllerType::JoyLeft || + controller_type == NPadControllerType::JoyRight) { + pad_state.left_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.left_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus()); + } } void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t data_len) { - if (!IsControllerActivated()) + if (!IsControllerActivated()) { return; - for (std::size_t i = 0; i < shared_memory_entries.size(); i++) { + } + for (std::size_t i = 0; i < shared_memory_entries.size(); ++i) { auto& npad = shared_memory_entries[i]; const std::array<NPadGeneric*, 7> controller_npads{&npad.main_controller_states, &npad.handheld_states, @@ -327,7 +392,7 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* const auto& last_entry = main_controller->npad[main_controller->common.last_entry_index]; - main_controller->common.timestamp = core_timing.GetTicks(); + main_controller->common.timestamp = core_timing.GetCPUTicks(); main_controller->common.last_entry_index = (main_controller->common.last_entry_index + 1) % 17; @@ -343,6 +408,7 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* continue; } const u32 npad_index = static_cast<u32>(i); + RequestPadStateUpdate(npad_index); auto& pad_state = npad_pad_states[npad_index]; @@ -359,12 +425,25 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* auto& libnx_entry = npad.libnx.npad[npad.libnx.common.last_entry_index]; libnx_entry.connection_status.raw = 0; + libnx_entry.connection_status.IsConnected.Assign(1); switch (controller_type) { case NPadControllerType::None: UNREACHABLE(); + break; + case NPadControllerType::ProController: + main_controller.connection_status.raw = 0; + main_controller.connection_status.IsConnected.Assign(1); + main_controller.connection_status.IsWired.Assign(1); + main_controller.pad.pad_states.raw = pad_state.pad_states.raw; + main_controller.pad.l_stick = pad_state.l_stick; + main_controller.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsWired.Assign(1); + break; case NPadControllerType::Handheld: handheld_entry.connection_status.raw = 0; + handheld_entry.connection_status.IsConnected.Assign(1); handheld_entry.connection_status.IsWired.Assign(1); handheld_entry.connection_status.IsLeftJoyConnected.Assign(1); handheld_entry.connection_status.IsRightJoyConnected.Assign(1); @@ -373,57 +452,52 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* handheld_entry.pad.pad_states.raw = pad_state.pad_states.raw; handheld_entry.pad.l_stick = pad_state.l_stick; handheld_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsWired.Assign(1); + libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); + libnx_entry.connection_status.IsRightJoyConnected.Assign(1); + libnx_entry.connection_status.IsLeftJoyWired.Assign(1); + libnx_entry.connection_status.IsRightJoyWired.Assign(1); break; case NPadControllerType::JoyDual: dual_entry.connection_status.raw = 0; - + dual_entry.connection_status.IsConnected.Assign(1); dual_entry.connection_status.IsLeftJoyConnected.Assign(1); dual_entry.connection_status.IsRightJoyConnected.Assign(1); - dual_entry.connection_status.IsConnected.Assign(1); - - libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); - libnx_entry.connection_status.IsRightJoyConnected.Assign(1); - libnx_entry.connection_status.IsConnected.Assign(1); - dual_entry.pad.pad_states.raw = pad_state.pad_states.raw; dual_entry.pad.l_stick = pad_state.l_stick; dual_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); + libnx_entry.connection_status.IsRightJoyConnected.Assign(1); break; case NPadControllerType::JoyLeft: left_entry.connection_status.raw = 0; - left_entry.connection_status.IsConnected.Assign(1); + left_entry.connection_status.IsLeftJoyConnected.Assign(1); left_entry.pad.pad_states.raw = pad_state.pad_states.raw; left_entry.pad.l_stick = pad_state.l_stick; left_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); break; case NPadControllerType::JoyRight: right_entry.connection_status.raw = 0; - right_entry.connection_status.IsConnected.Assign(1); + right_entry.connection_status.IsRightJoyConnected.Assign(1); right_entry.pad.pad_states.raw = pad_state.pad_states.raw; right_entry.pad.l_stick = pad_state.l_stick; right_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsRightJoyConnected.Assign(1); break; case NPadControllerType::Pokeball: pokeball_entry.connection_status.raw = 0; - pokeball_entry.connection_status.IsConnected.Assign(1); - pokeball_entry.connection_status.IsWired.Assign(1); - pokeball_entry.pad.pad_states.raw = pad_state.pad_states.raw; pokeball_entry.pad.l_stick = pad_state.l_stick; pokeball_entry.pad.r_stick = pad_state.r_stick; break; - case NPadControllerType::ProController: - main_controller.connection_status.raw = 0; - - main_controller.connection_status.IsConnected.Assign(1); - main_controller.connection_status.IsWired.Assign(1); - main_controller.pad.pad_states.raw = pad_state.pad_states.raw; - main_controller.pad.l_stick = pad_state.l_stick; - main_controller.pad.r_stick = pad_state.r_stick; - break; } // LibNX exclusively uses this section, so we always update it since LibNX doesn't activate @@ -438,39 +512,144 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* shared_memory_entries.size() * sizeof(NPadEntry)); } -void Controller_NPad::SetSupportedStyleSet(NPadType style_set) { +void Controller_NPad::OnMotionUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, + std::size_t data_len) { + if (!IsControllerActivated()) { + return; + } + for (std::size_t i = 0; i < shared_memory_entries.size(); ++i) { + auto& npad = shared_memory_entries[i]; + + const auto& controller_type = connected_controllers[i].type; + + if (controller_type == NPadControllerType::None || !connected_controllers[i].is_connected) { + continue; + } + + const std::array<SixAxisGeneric*, 6> controller_sixaxes{ + &npad.sixaxis_full, &npad.sixaxis_handheld, &npad.sixaxis_dual_left, + &npad.sixaxis_dual_right, &npad.sixaxis_left, &npad.sixaxis_right, + }; + + for (auto* sixaxis_sensor : controller_sixaxes) { + sixaxis_sensor->common.entry_count = 16; + sixaxis_sensor->common.total_entry_count = 17; + + const auto& last_entry = + sixaxis_sensor->sixaxis[sixaxis_sensor->common.last_entry_index]; + + sixaxis_sensor->common.timestamp = core_timing.GetCPUTicks(); + sixaxis_sensor->common.last_entry_index = + (sixaxis_sensor->common.last_entry_index + 1) % 17; + + auto& cur_entry = sixaxis_sensor->sixaxis[sixaxis_sensor->common.last_entry_index]; + + cur_entry.timestamp = last_entry.timestamp + 1; + cur_entry.timestamp2 = cur_entry.timestamp; + } + + // Try to read sixaxis sensor states + std::array<MotionDevice, 2> motion_devices; + + if (sixaxis_sensors_enabled && Settings::values.motion_enabled.GetValue()) { + sixaxis_at_rest = true; + for (std::size_t e = 0; e < motion_devices.size(); ++e) { + const auto& device = motions[i][e]; + if (device) { + std::tie(motion_devices[e].accel, motion_devices[e].gyro, + motion_devices[e].rotation, motion_devices[e].orientation) = + device->GetStatus(); + sixaxis_at_rest = sixaxis_at_rest && motion_devices[e].gyro.Length2() < 0.0001f; + } + } + } + + auto& full_sixaxis_entry = + npad.sixaxis_full.sixaxis[npad.sixaxis_full.common.last_entry_index]; + auto& handheld_sixaxis_entry = + npad.sixaxis_handheld.sixaxis[npad.sixaxis_handheld.common.last_entry_index]; + auto& dual_left_sixaxis_entry = + npad.sixaxis_dual_left.sixaxis[npad.sixaxis_dual_left.common.last_entry_index]; + auto& dual_right_sixaxis_entry = + npad.sixaxis_dual_right.sixaxis[npad.sixaxis_dual_right.common.last_entry_index]; + auto& left_sixaxis_entry = + npad.sixaxis_left.sixaxis[npad.sixaxis_left.common.last_entry_index]; + auto& right_sixaxis_entry = + npad.sixaxis_right.sixaxis[npad.sixaxis_right.common.last_entry_index]; + + switch (controller_type) { + case NPadControllerType::None: + UNREACHABLE(); + break; + case NPadControllerType::ProController: + if (sixaxis_sensors_enabled && motions[i][0]) { + full_sixaxis_entry.accel = motion_devices[0].accel; + full_sixaxis_entry.gyro = motion_devices[0].gyro; + full_sixaxis_entry.rotation = motion_devices[0].rotation; + full_sixaxis_entry.orientation = motion_devices[0].orientation; + } + break; + case NPadControllerType::Handheld: + if (sixaxis_sensors_enabled && motions[i][0]) { + handheld_sixaxis_entry.accel = motion_devices[0].accel; + handheld_sixaxis_entry.gyro = motion_devices[0].gyro; + handheld_sixaxis_entry.rotation = motion_devices[0].rotation; + handheld_sixaxis_entry.orientation = motion_devices[0].orientation; + } + break; + case NPadControllerType::JoyDual: + if (sixaxis_sensors_enabled && motions[i][0]) { + // Set motion for the left joycon + dual_left_sixaxis_entry.accel = motion_devices[0].accel; + dual_left_sixaxis_entry.gyro = motion_devices[0].gyro; + dual_left_sixaxis_entry.rotation = motion_devices[0].rotation; + dual_left_sixaxis_entry.orientation = motion_devices[0].orientation; + } + if (sixaxis_sensors_enabled && motions[i][1]) { + // Set motion for the right joycon + dual_right_sixaxis_entry.accel = motion_devices[1].accel; + dual_right_sixaxis_entry.gyro = motion_devices[1].gyro; + dual_right_sixaxis_entry.rotation = motion_devices[1].rotation; + dual_right_sixaxis_entry.orientation = motion_devices[1].orientation; + } + break; + case NPadControllerType::JoyLeft: + if (sixaxis_sensors_enabled && motions[i][0]) { + left_sixaxis_entry.accel = motion_devices[0].accel; + left_sixaxis_entry.gyro = motion_devices[0].gyro; + left_sixaxis_entry.rotation = motion_devices[0].rotation; + left_sixaxis_entry.orientation = motion_devices[0].orientation; + } + break; + case NPadControllerType::JoyRight: + if (sixaxis_sensors_enabled && motions[i][1]) { + right_sixaxis_entry.accel = motion_devices[1].accel; + right_sixaxis_entry.gyro = motion_devices[1].gyro; + right_sixaxis_entry.rotation = motion_devices[1].rotation; + right_sixaxis_entry.orientation = motion_devices[1].orientation; + } + break; + case NPadControllerType::Pokeball: + break; + } + } + std::memcpy(data + NPAD_OFFSET, shared_memory_entries.data(), + shared_memory_entries.size() * sizeof(NPadEntry)); +} + +void Controller_NPad::SetSupportedStyleSet(NpadStyleSet style_set) { style.raw = style_set.raw; } -Controller_NPad::NPadType Controller_NPad::GetSupportedStyleSet() const { +Controller_NPad::NpadStyleSet Controller_NPad::GetSupportedStyleSet() const { return style; } -void Controller_NPad::SetSupportedNPadIdTypes(u8* data, std::size_t length) { +void Controller_NPad::SetSupportedNpadIdTypes(u8* data, std::size_t length) { ASSERT(length > 0 && (length % sizeof(u32)) == 0); supported_npad_id_types.clear(); supported_npad_id_types.resize(length / sizeof(u32)); std::memcpy(supported_npad_id_types.data(), data, length); - for (std::size_t i = 0; i < connected_controllers.size(); i++) { - auto& controller = connected_controllers[i]; - if (!controller.is_connected) { - continue; - } - const auto requested_controller = - i <= MAX_NPAD_ID ? MapSettingsTypeToNPad(Settings::values.players[i].type) - : NPadControllerType::Handheld; - if (!IsControllerSupported(requested_controller)) { - const auto is_handheld = requested_controller == NPadControllerType::Handheld; - if (is_handheld) { - controller.type = NPadControllerType::None; - controller.is_connected = false; - AddNewController(requested_controller); - } else { - controller.type = requested_controller; - InitNewlyAddedControler(i); - } - } - } } void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) { @@ -478,7 +657,7 @@ void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) std::memcpy(data, supported_npad_id_types.data(), supported_npad_id_types.size()); } -std::size_t Controller_NPad::GetSupportedNPadIdTypesSize() const { +std::size_t Controller_NPad::GetSupportedNpadIdTypesSize() const { return supported_npad_id_types.size(); } @@ -490,7 +669,15 @@ Controller_NPad::NpadHoldType Controller_NPad::GetHoldType() const { return hold_type; } -void Controller_NPad::SetNpadMode(u32 npad_id, NPadAssignments assignment_mode) { +void Controller_NPad::SetNpadHandheldActivationMode(NpadHandheldActivationMode activation_mode) { + handheld_activation_mode = activation_mode; +} + +Controller_NPad::NpadHandheldActivationMode Controller_NPad::GetNpadHandheldActivationMode() const { + return handheld_activation_mode; +} + +void Controller_NPad::SetNpadMode(u32 npad_id, NpadAssignments assignment_mode) { const std::size_t npad_index = NPadIdToIndex(npad_id); ASSERT(npad_index < shared_memory_entries.size()); if (shared_memory_entries[npad_index].pad_assignment != assignment_mode) { @@ -498,70 +685,230 @@ void Controller_NPad::SetNpadMode(u32 npad_id, NPadAssignments assignment_mode) } } -void Controller_NPad::VibrateController(const std::vector<u32>& controller_ids, - const std::vector<Vibration>& vibrations) { - LOG_WARNING(Service_HID, "(STUBBED) called"); +bool Controller_NPad::VibrateControllerAtIndex(std::size_t npad_index, std::size_t device_index, + const VibrationValue& vibration_value) { + if (!connected_controllers[npad_index].is_connected || !vibrations[npad_index][device_index]) { + return false; + } + + const auto& player = Settings::values.players.GetValue()[npad_index]; - if (!can_controllers_vibrate) { - return; + if (!player.vibration_enabled) { + if (latest_vibration_values[npad_index][device_index].amp_low != 0.0f || + latest_vibration_values[npad_index][device_index].amp_high != 0.0f) { + // Send an empty vibration to stop any vibrations. + vibrations[npad_index][device_index]->SetRumblePlay(0.0f, 160.0f, 0.0f, 320.0f); + // Then reset the vibration value to its default value. + latest_vibration_values[npad_index][device_index] = {}; + } + + return false; } - for (std::size_t i = 0; i < controller_ids.size(); i++) { - std::size_t controller_pos = NPadIdToIndex(static_cast<u32>(i)); - if (connected_controllers[controller_pos].is_connected) { - // TODO(ogniK): Vibrate the physical controller + + if (!Settings::values.enable_accurate_vibrations.GetValue()) { + using std::chrono::duration_cast; + using std::chrono::milliseconds; + using std::chrono::steady_clock; + + const auto now = steady_clock::now(); + + // Filter out non-zero vibrations that are within 10ms of each other. + if ((vibration_value.amp_low != 0.0f || vibration_value.amp_high != 0.0f) && + duration_cast<milliseconds>(now - last_vibration_timepoints[npad_index][device_index]) < + milliseconds(10)) { + return false; } + + last_vibration_timepoints[npad_index][device_index] = now; } - last_processed_vibration = vibrations.back(); + + auto& vibration = vibrations[npad_index][device_index]; + const auto player_vibration_strength = static_cast<f32>(player.vibration_strength); + const auto amp_low = + std::min(vibration_value.amp_low * player_vibration_strength / 100.0f, 1.0f); + const auto amp_high = + std::min(vibration_value.amp_high * player_vibration_strength / 100.0f, 1.0f); + return vibration->SetRumblePlay(amp_low, vibration_value.freq_low, amp_high, + vibration_value.freq_high); +} + +void Controller_NPad::VibrateController(const DeviceHandle& vibration_device_handle, + const VibrationValue& vibration_value) { + if (!Settings::values.vibration_enabled.GetValue() && !permit_vibration_session_enabled) { + return; + } + + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + + if (!vibration_devices_mounted[npad_index][device_index] || + !connected_controllers[npad_index].is_connected) { + return; + } + + if (vibration_device_handle.device_index == DeviceIndex::None) { + UNREACHABLE_MSG("DeviceIndex should never be None!"); + return; + } + + // Some games try to send mismatched parameters in the device handle, block these. + if ((connected_controllers[npad_index].type == NPadControllerType::JoyLeft && + (vibration_device_handle.npad_type == NpadType::JoyconRight || + vibration_device_handle.device_index == DeviceIndex::Right)) || + (connected_controllers[npad_index].type == NPadControllerType::JoyRight && + (vibration_device_handle.npad_type == NpadType::JoyconLeft || + vibration_device_handle.device_index == DeviceIndex::Left))) { + return; + } + + // Filter out vibrations with equivalent values to reduce unnecessary state changes. + if (vibration_value.amp_low == latest_vibration_values[npad_index][device_index].amp_low && + vibration_value.amp_high == latest_vibration_values[npad_index][device_index].amp_high) { + return; + } + + if (VibrateControllerAtIndex(npad_index, device_index, vibration_value)) { + latest_vibration_values[npad_index][device_index] = vibration_value; + } +} + +void Controller_NPad::VibrateControllers(const std::vector<DeviceHandle>& vibration_device_handles, + const std::vector<VibrationValue>& vibration_values) { + if (!Settings::values.vibration_enabled.GetValue() && !permit_vibration_session_enabled) { + return; + } + + ASSERT_OR_EXECUTE_MSG( + vibration_device_handles.size() == vibration_values.size(), { return; }, + "The amount of device handles does not match with the amount of vibration values," + "this is undefined behavior!"); + + for (std::size_t i = 0; i < vibration_device_handles.size(); ++i) { + VibrateController(vibration_device_handles[i], vibration_values[i]); + } +} + +Controller_NPad::VibrationValue Controller_NPad::GetLastVibration( + const DeviceHandle& vibration_device_handle) const { + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + return latest_vibration_values[npad_index][device_index]; +} + +void Controller_NPad::InitializeVibrationDevice(const DeviceHandle& vibration_device_handle) { + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + InitializeVibrationDeviceAtIndex(npad_index, device_index); +} + +void Controller_NPad::InitializeVibrationDeviceAtIndex(std::size_t npad_index, + std::size_t device_index) { + if (vibrations[npad_index][device_index]) { + vibration_devices_mounted[npad_index][device_index] = + vibrations[npad_index][device_index]->GetStatus() == 1; + } else { + vibration_devices_mounted[npad_index][device_index] = false; + } +} + +void Controller_NPad::SetPermitVibrationSession(bool permit_vibration_session) { + permit_vibration_session_enabled = permit_vibration_session; +} + +bool Controller_NPad::IsVibrationDeviceMounted(const DeviceHandle& vibration_device_handle) const { + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + return vibration_devices_mounted[npad_index][device_index]; } std::shared_ptr<Kernel::ReadableEvent> Controller_NPad::GetStyleSetChangedEvent(u32 npad_id) const { - // TODO(ogniK): Figure out the best time to signal this event. This event seems that it should - // be signalled at least once, and signaled after a new controller is connected? const auto& styleset_event = styleset_changed_events[NPadIdToIndex(npad_id)]; return styleset_event.readable; } -Controller_NPad::Vibration Controller_NPad::GetLastVibration() const { - return last_processed_vibration; +void Controller_NPad::SignalStyleSetChangedEvent(u32 npad_id) const { + styleset_changed_events[NPadIdToIndex(npad_id)].writable->Signal(); } -void Controller_NPad::AddNewController(NPadControllerType controller) { - controller = DecideBestController(controller); - if (controller == NPadControllerType::Handheld) { - connected_controllers[8] = {controller, true}; - InitNewlyAddedControler(8); +void Controller_NPad::AddNewControllerAt(NPadControllerType controller, std::size_t npad_index) { + UpdateControllerAt(controller, npad_index, true); +} + +void Controller_NPad::UpdateControllerAt(NPadControllerType controller, std::size_t npad_index, + bool connected) { + if (!connected) { + DisconnectNpadAtIndex(npad_index); return; } - const auto pos = - std::find_if(connected_controllers.begin(), connected_controllers.end() - 2, - [](const ControllerHolder& holder) { return !holder.is_connected; }); - if (pos == connected_controllers.end() - 2) { - LOG_ERROR(Service_HID, "Cannot connect any more controllers!"); + + if (controller == NPadControllerType::Handheld) { + Settings::values.players.GetValue()[HANDHELD_INDEX].controller_type = + MapNPadToSettingsType(controller); + Settings::values.players.GetValue()[HANDHELD_INDEX].connected = true; + connected_controllers[HANDHELD_INDEX] = {controller, true}; + InitNewlyAddedController(HANDHELD_INDEX); return; } - const auto controller_id = std::distance(connected_controllers.begin(), pos); - connected_controllers[controller_id] = {controller, true}; - InitNewlyAddedControler(controller_id); + + Settings::values.players.GetValue()[npad_index].controller_type = + MapNPadToSettingsType(controller); + Settings::values.players.GetValue()[npad_index].connected = true; + connected_controllers[npad_index] = {controller, true}; + InitNewlyAddedController(npad_index); } -void Controller_NPad::AddNewControllerAt(NPadControllerType controller, u32 npad_id) { - controller = DecideBestController(controller); - if (controller == NPadControllerType::Handheld) { - connected_controllers[NPadIdToIndex(NPAD_HANDHELD)] = {controller, true}; - InitNewlyAddedControler(NPadIdToIndex(NPAD_HANDHELD)); - return; +void Controller_NPad::DisconnectNpad(u32 npad_id) { + DisconnectNpadAtIndex(NPadIdToIndex(npad_id)); +} + +void Controller_NPad::DisconnectNpadAtIndex(std::size_t npad_index) { + for (std::size_t device_idx = 0; device_idx < vibrations[npad_index].size(); ++device_idx) { + // Send an empty vibration to stop any vibrations. + VibrateControllerAtIndex(npad_index, device_idx, {}); + vibration_devices_mounted[npad_index][device_idx] = false; } - connected_controllers[NPadIdToIndex(npad_id)] = {controller, true}; - InitNewlyAddedControler(NPadIdToIndex(npad_id)); + Settings::values.players.GetValue()[npad_index].connected = false; + connected_controllers[npad_index].is_connected = false; + + auto& controller = shared_memory_entries[npad_index]; + controller.joy_styles.raw = 0; // Zero out + controller.device_type.raw = 0; + controller.properties.raw = 0; + + SignalStyleSetChangedEvent(IndexToNPad(npad_index)); +} + +void Controller_NPad::SetGyroscopeZeroDriftMode(GyroscopeZeroDriftMode drift_mode) { + gyroscope_zero_drift_mode = drift_mode; +} + +Controller_NPad::GyroscopeZeroDriftMode Controller_NPad::GetGyroscopeZeroDriftMode() const { + return gyroscope_zero_drift_mode; +} + +bool Controller_NPad::IsSixAxisSensorAtRest() const { + return sixaxis_at_rest; } -void Controller_NPad::ConnectNPad(u32 npad_id) { - connected_controllers[NPadIdToIndex(npad_id)].is_connected = true; +void Controller_NPad::SetSixAxisEnabled(bool six_axis_status) { + sixaxis_sensors_enabled = six_axis_status; } -void Controller_NPad::DisconnectNPad(u32 npad_id) { - connected_controllers[NPadIdToIndex(npad_id)].is_connected = false; +void Controller_NPad::MergeSingleJoyAsDualJoy(u32 npad_id_1, u32 npad_id_2) { + const auto npad_index_1 = NPadIdToIndex(npad_id_1); + const auto npad_index_2 = NPadIdToIndex(npad_id_2); + + // If the controllers at both npad indices form a pair of left and right joycons, merge them. + // Otherwise, do nothing. + if ((connected_controllers[npad_index_1].type == NPadControllerType::JoyLeft && + connected_controllers[npad_index_2].type == NPadControllerType::JoyRight) || + (connected_controllers[npad_index_2].type == NPadControllerType::JoyLeft && + connected_controllers[npad_index_1].type == NPadControllerType::JoyRight)) { + // Disconnect the joycon at the second id and connect the dual joycon at the first index. + DisconnectNpad(npad_id_2); + AddNewControllerAt(NPadControllerType::JoyDual, npad_index_1); + } } void Controller_NPad::StartLRAssignmentMode() { @@ -589,8 +936,8 @@ bool Controller_NPad::SwapNpadAssignment(u32 npad_id_1, u32 npad_id_2) { std::swap(connected_controllers[npad_index_1].type, connected_controllers[npad_index_2].type); - InitNewlyAddedControler(npad_index_1); - InitNewlyAddedControler(npad_index_2); + AddNewControllerAt(connected_controllers[npad_index_1].type, npad_index_1); + AddNewControllerAt(connected_controllers[npad_index_2].type, npad_index_2); return true; } @@ -604,11 +951,11 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) { case 0: return LedPattern{1, 0, 0, 0}; case 1: - return LedPattern{0, 1, 0, 0}; + return LedPattern{1, 1, 0, 0}; case 2: - return LedPattern{0, 0, 1, 0}; + return LedPattern{1, 1, 1, 0}; case 3: - return LedPattern{0, 0, 0, 1}; + return LedPattern{1, 1, 1, 1}; case 4: return LedPattern{1, 0, 0, 1}; case 5: @@ -618,17 +965,17 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) { case 7: return LedPattern{0, 1, 1, 0}; default: - UNIMPLEMENTED_MSG("Unhandled npad_id {}", npad_id); return LedPattern{0, 0, 0, 0}; - }; + } } -void Controller_NPad::SetVibrationEnabled(bool can_vibrate) { - can_controllers_vibrate = can_vibrate; +bool Controller_NPad::IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const { + return unintended_home_button_input_protection[NPadIdToIndex(npad_id)]; } -bool Controller_NPad::IsVibrationEnabled() const { - return can_controllers_vibrate; +void Controller_NPad::SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled, + u32 npad_id) { + unintended_home_button_input_protection[NPadIdToIndex(npad_id)] = is_protection_enabled; } void Controller_NPad::ClearAllConnectedControllers() { @@ -641,13 +988,13 @@ void Controller_NPad::ClearAllConnectedControllers() { } void Controller_NPad::DisconnectAllConnectedControllers() { - for (ControllerHolder& controller : connected_controllers) { + for (auto& controller : connected_controllers) { controller.is_connected = false; } } void Controller_NPad::ConnectAllDisconnectedControllers() { - for (ControllerHolder& controller : connected_controllers) { + for (auto& controller : connected_controllers) { if (controller.type != NPadControllerType::None && !controller.is_connected) { controller.is_connected = true; } @@ -655,7 +1002,7 @@ void Controller_NPad::ConnectAllDisconnectedControllers() { } void Controller_NPad::ClearAllControllers() { - for (ControllerHolder& controller : connected_controllers) { + for (auto& controller : connected_controllers) { controller.type = NPadControllerType::None; controller.is_connected = false; } @@ -675,7 +1022,7 @@ bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const return false; } // Handheld should not be supported in docked mode - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { return false; } @@ -703,92 +1050,4 @@ bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const return false; } -Controller_NPad::NPadControllerType Controller_NPad::DecideBestController( - NPadControllerType priority) const { - if (IsControllerSupported(priority)) { - return priority; - } - const auto is_docked = Settings::values.use_docked_mode; - if (is_docked && priority == NPadControllerType::Handheld) { - priority = NPadControllerType::JoyDual; - if (IsControllerSupported(priority)) { - return priority; - } - } - std::vector<NPadControllerType> priority_list; - switch (priority) { - case NPadControllerType::ProController: - priority_list.push_back(NPadControllerType::JoyDual); - if (!is_docked) { - priority_list.push_back(NPadControllerType::Handheld); - } - priority_list.push_back(NPadControllerType::JoyLeft); - priority_list.push_back(NPadControllerType::JoyRight); - priority_list.push_back(NPadControllerType::Pokeball); - break; - case NPadControllerType::Handheld: - priority_list.push_back(NPadControllerType::JoyDual); - priority_list.push_back(NPadControllerType::ProController); - priority_list.push_back(NPadControllerType::JoyLeft); - priority_list.push_back(NPadControllerType::JoyRight); - priority_list.push_back(NPadControllerType::Pokeball); - break; - case NPadControllerType::JoyDual: - if (!is_docked) { - priority_list.push_back(NPadControllerType::Handheld); - } - priority_list.push_back(NPadControllerType::ProController); - priority_list.push_back(NPadControllerType::JoyLeft); - priority_list.push_back(NPadControllerType::JoyRight); - priority_list.push_back(NPadControllerType::Pokeball); - break; - case NPadControllerType::JoyLeft: - priority_list.push_back(NPadControllerType::JoyRight); - priority_list.push_back(NPadControllerType::JoyDual); - if (!is_docked) { - priority_list.push_back(NPadControllerType::Handheld); - } - priority_list.push_back(NPadControllerType::ProController); - priority_list.push_back(NPadControllerType::Pokeball); - break; - case NPadControllerType::JoyRight: - priority_list.push_back(NPadControllerType::JoyLeft); - priority_list.push_back(NPadControllerType::JoyDual); - if (!is_docked) { - priority_list.push_back(NPadControllerType::Handheld); - } - priority_list.push_back(NPadControllerType::ProController); - priority_list.push_back(NPadControllerType::Pokeball); - break; - case NPadControllerType::Pokeball: - priority_list.push_back(NPadControllerType::JoyLeft); - priority_list.push_back(NPadControllerType::JoyRight); - priority_list.push_back(NPadControllerType::JoyDual); - if (!is_docked) { - priority_list.push_back(NPadControllerType::Handheld); - } - priority_list.push_back(NPadControllerType::ProController); - break; - default: - priority_list.push_back(NPadControllerType::JoyDual); - if (!is_docked) { - priority_list.push_back(NPadControllerType::Handheld); - } - priority_list.push_back(NPadControllerType::ProController); - priority_list.push_back(NPadControllerType::JoyLeft); - priority_list.push_back(NPadControllerType::JoyRight); - priority_list.push_back(NPadControllerType::JoyDual); - break; - } - - const auto iter = std::find_if(priority_list.begin(), priority_list.end(), - [this](auto type) { return IsControllerSupported(type); }); - if (iter == priority_list.end()) { - UNIMPLEMENTED_MSG("Could not find supported controller!"); - return priority; - } - - return *iter; -} - } // namespace Service::HID diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h index 931f03430..160dcbbe3 100644 --- a/src/core/hle/service/hid/controllers/npad.h +++ b/src/core/hle/service/hid/controllers/npad.h @@ -32,51 +32,90 @@ public: // When the controller is requesting an update for the shared memory void OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) override; + // When the controller is requesting a motion update for the shared memory + void OnMotionUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, + std::size_t size) override; + // Called when input devices should be loaded void OnLoadInputDevices() override; - struct NPadType { - union { - u32_le raw{}; + enum class NPadControllerType { + None, + ProController, + Handheld, + JoyDual, + JoyLeft, + JoyRight, + Pokeball, + }; - BitField<0, 1, u32> pro_controller; - BitField<1, 1, u32> handheld; - BitField<2, 1, u32> joycon_dual; - BitField<3, 1, u32> joycon_left; - BitField<4, 1, u32> joycon_right; + enum class NpadType : u8 { + ProController = 3, + Handheld = 4, + JoyconDual = 5, + JoyconLeft = 6, + JoyconRight = 7, + Pokeball = 9, + }; - BitField<6, 1, u32> pokeball; // TODO(ogniK): Confirm when possible - }; + enum class DeviceIndex : u8 { + Left = 0, + Right = 1, + None = 2, }; - static_assert(sizeof(NPadType) == 4, "NPadType is an invalid size"); - struct Vibration { - f32 amp_low; - f32 freq_low; - f32 amp_high; - f32 freq_high; + enum class GyroscopeZeroDriftMode : u32 { + Loose = 0, + Standard = 1, + Tight = 2, }; - static_assert(sizeof(Vibration) == 0x10, "Vibration is an invalid size"); enum class NpadHoldType : u64 { Vertical = 0, Horizontal = 1, }; - enum class NPadAssignments : u32_le { + enum class NpadAssignments : u32 { Dual = 0, Single = 1, }; - enum class NPadControllerType { - None, - ProController, - Handheld, - JoyDual, - JoyLeft, - JoyRight, - Pokeball, + enum class NpadHandheldActivationMode : u64 { + Dual = 0, + Single = 1, + None = 2, + }; + + struct DeviceHandle { + NpadType npad_type{}; + u8 npad_id{}; + DeviceIndex device_index{}; + INSERT_PADDING_BYTES(1); }; + static_assert(sizeof(DeviceHandle) == 4, "DeviceHandle is an invalid size"); + + struct NpadStyleSet { + union { + u32_le raw{}; + + BitField<0, 1, u32> pro_controller; + BitField<1, 1, u32> handheld; + BitField<2, 1, u32> joycon_dual; + BitField<3, 1, u32> joycon_left; + BitField<4, 1, u32> joycon_right; + + BitField<6, 1, u32> pokeball; // TODO(ogniK): Confirm when possible + }; + }; + static_assert(sizeof(NpadStyleSet) == 4, "NpadStyleSet is an invalid size"); + + struct VibrationValue { + f32 amp_low{0.0f}; + f32 freq_low{160.0f}; + f32 amp_high{0.0f}; + f32 freq_high{320.0f}; + }; + static_assert(sizeof(VibrationValue) == 0x10, "Vibration is an invalid size"); struct LedPattern { explicit LedPattern(u64 light1, u64 light2, u64 light3, u64 light4) { @@ -94,37 +133,64 @@ public: }; }; - void SetSupportedStyleSet(NPadType style_set); - NPadType GetSupportedStyleSet() const; + void SetSupportedStyleSet(NpadStyleSet style_set); + NpadStyleSet GetSupportedStyleSet() const; - void SetSupportedNPadIdTypes(u8* data, std::size_t length); + void SetSupportedNpadIdTypes(u8* data, std::size_t length); void GetSupportedNpadIdTypes(u32* data, std::size_t max_length); - std::size_t GetSupportedNPadIdTypesSize() const; + std::size_t GetSupportedNpadIdTypesSize() const; void SetHoldType(NpadHoldType joy_hold_type); NpadHoldType GetHoldType() const; - void SetNpadMode(u32 npad_id, NPadAssignments assignment_mode); + void SetNpadHandheldActivationMode(NpadHandheldActivationMode activation_mode); + NpadHandheldActivationMode GetNpadHandheldActivationMode() const; + + void SetNpadMode(u32 npad_id, NpadAssignments assignment_mode); - void VibrateController(const std::vector<u32>& controller_ids, - const std::vector<Vibration>& vibrations); + bool VibrateControllerAtIndex(std::size_t npad_index, std::size_t device_index, + const VibrationValue& vibration_value); + + void VibrateController(const DeviceHandle& vibration_device_handle, + const VibrationValue& vibration_value); + + void VibrateControllers(const std::vector<DeviceHandle>& vibration_device_handles, + const std::vector<VibrationValue>& vibration_values); + + VibrationValue GetLastVibration(const DeviceHandle& vibration_device_handle) const; + + void InitializeVibrationDevice(const DeviceHandle& vibration_device_handle); + + void InitializeVibrationDeviceAtIndex(std::size_t npad_index, std::size_t device_index); + + void SetPermitVibrationSession(bool permit_vibration_session); + + bool IsVibrationDeviceMounted(const DeviceHandle& vibration_device_handle) const; std::shared_ptr<Kernel::ReadableEvent> GetStyleSetChangedEvent(u32 npad_id) const; - Vibration GetLastVibration() const; + void SignalStyleSetChangedEvent(u32 npad_id) const; + + // Adds a new controller at an index. + void AddNewControllerAt(NPadControllerType controller, std::size_t npad_index); + // Adds a new controller at an index with connection status. + void UpdateControllerAt(NPadControllerType controller, std::size_t npad_index, bool connected); - void AddNewController(NPadControllerType controller); - void AddNewControllerAt(NPadControllerType controller, u32 npad_id); + void DisconnectNpad(u32 npad_id); + void DisconnectNpadAtIndex(std::size_t index); - void ConnectNPad(u32 npad_id); - void DisconnectNPad(u32 npad_id); + void SetGyroscopeZeroDriftMode(GyroscopeZeroDriftMode drift_mode); + GyroscopeZeroDriftMode GetGyroscopeZeroDriftMode() const; + bool IsSixAxisSensorAtRest() const; + void SetSixAxisEnabled(bool six_axis_status); LedPattern GetLedPattern(u32 npad_id); - void SetVibrationEnabled(bool can_vibrate); - bool IsVibrationEnabled() const; + bool IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const; + void SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled, u32 npad_id); void ClearAllConnectedControllers(); void DisconnectAllConnectedControllers(); void ConnectAllDisconnectedControllers(); void ClearAllControllers(); + void MergeSingleJoyAsDualJoy(u32 npad_id_1, u32 npad_id_2); void StartLRAssignmentMode(); void StopLRAssignmentMode(); bool SwapNpadAssignment(u32 npad_id_1, u32 npad_id_2); @@ -133,6 +199,8 @@ public: // Specifically for cheat engine and other features. u32 GetAndResetPressState(); + static Controller_NPad::NPadControllerType MapSettingsTypeToNPad(Settings::ControllerType type); + static Settings::ControllerType MapNPadToSettingsType(Controller_NPad::NPadControllerType type); static std::size_t NPadIdToIndex(u32 npad_id); static u32 IndexToNPad(std::size_t index); @@ -236,6 +304,24 @@ private: }; static_assert(sizeof(NPadGeneric) == 0x350, "NPadGeneric is an invalid size"); + struct SixAxisStates { + s64_le timestamp{}; + INSERT_PADDING_WORDS(2); + s64_le timestamp2{}; + Common::Vec3f accel{}; + Common::Vec3f gyro{}; + Common::Vec3f rotation{}; + std::array<Common::Vec3f, 3> orientation{}; + s64_le always_one{1}; + }; + static_assert(sizeof(SixAxisStates) == 0x68, "SixAxisStates is an invalid size"); + + struct SixAxisGeneric { + CommonHeader common{}; + std::array<SixAxisStates, 17> sixaxis{}; + }; + static_assert(sizeof(SixAxisGeneric) == 0x708, "SixAxisGeneric is an invalid size"); + enum class ColorReadError : u32_le { ReadOk = 0, ColorDoesntExist = 1, @@ -265,9 +351,16 @@ private: }; }; + struct MotionDevice { + Common::Vec3f accel; + Common::Vec3f gyro; + Common::Vec3f rotation; + std::array<Common::Vec3f, 3> orientation; + }; + struct NPadEntry { - NPadType joy_styles; - NPadAssignments pad_assignment; + NpadStyleSet joy_styles; + NpadAssignments pad_assignment; ColorReadError single_color_error; ControllerColor single_color; @@ -284,9 +377,12 @@ private: NPadGeneric pokeball_states; NPadGeneric libnx; // TODO(ogniK): Find out what this actually is, libnx seems to only be // relying on this for the time being - INSERT_PADDING_BYTES( - 0x708 * - 6); // TODO(ogniK): SixAxis states, require more information before implementation + SixAxisGeneric sixaxis_full; + SixAxisGeneric sixaxis_handheld; + SixAxisGeneric sixaxis_dual_left; + SixAxisGeneric sixaxis_dual_right; + SixAxisGeneric sixaxis_left; + SixAxisGeneric sixaxis_right; NPadDevice device_type; NPadProperties properties; INSERT_PADDING_WORDS(1); @@ -301,31 +397,44 @@ private: bool is_connected; }; - void InitNewlyAddedControler(std::size_t controller_idx); + void InitNewlyAddedController(std::size_t controller_idx); bool IsControllerSupported(NPadControllerType controller) const; - NPadControllerType DecideBestController(NPadControllerType priority) const; void RequestPadStateUpdate(u32 npad_id); u32 press_state{}; - NPadType style{}; + NpadStyleSet style{}; std::array<NPadEntry, 10> shared_memory_entries{}; - std::array< + using ButtonArray = std::array< std::array<std::unique_ptr<Input::ButtonDevice>, Settings::NativeButton::NUM_BUTTONS_HID>, - 10> - buttons; - std::array< + 10>; + using StickArray = std::array< std::array<std::unique_ptr<Input::AnalogDevice>, Settings::NativeAnalog::NUM_STICKS_HID>, - 10> - sticks; + 10>; + using VibrationArray = std::array<std::array<std::unique_ptr<Input::VibrationDevice>, + Settings::NativeVibration::NUM_VIBRATIONS_HID>, + 10>; + using MotionArray = std::array< + std::array<std::unique_ptr<Input::MotionDevice>, Settings::NativeMotion::NUM_MOTIONS_HID>, + 10>; + ButtonArray buttons; + StickArray sticks; + VibrationArray vibrations; + MotionArray motions; std::vector<u32> supported_npad_id_types{}; NpadHoldType hold_type{NpadHoldType::Vertical}; + NpadHandheldActivationMode handheld_activation_mode{NpadHandheldActivationMode::Dual}; // Each controller should have their own styleset changed event std::array<Kernel::EventPair, 10> styleset_changed_events; - Vibration last_processed_vibration{}; + std::array<std::array<std::chrono::steady_clock::time_point, 2>, 10> last_vibration_timepoints; + std::array<std::array<VibrationValue, 2>, 10> latest_vibration_values{}; + bool permit_vibration_session_enabled{false}; + std::array<std::array<bool, 2>, 10> vibration_devices_mounted{}; std::array<ControllerHolder, 10> connected_controllers{}; - bool can_controllers_vibrate{true}; - + std::array<bool, 10> unintended_home_button_input_protection{}; + GyroscopeZeroDriftMode gyroscope_zero_drift_mode{GyroscopeZeroDriftMode::Standard}; + bool sixaxis_sensors_enabled{true}; + bool sixaxis_at_rest{true}; std::array<ControllerPad, 10> npad_pad_states{}; bool is_in_lr_assignment_mode{false}; Core::System& system; diff --git a/src/core/hle/service/hid/controllers/stubbed.cpp b/src/core/hle/service/hid/controllers/stubbed.cpp index 9e527d176..e7483bfa2 100644 --- a/src/core/hle/service/hid/controllers/stubbed.cpp +++ b/src/core/hle/service/hid/controllers/stubbed.cpp @@ -23,7 +23,7 @@ void Controller_Stubbed::OnUpdate(const Core::Timing::CoreTiming& core_timing, u } CommonHeader header{}; - header.timestamp = core_timing.GetTicks(); + header.timestamp = core_timing.GetCPUTicks(); header.total_entry_count = 17; header.entry_count = 0; header.last_entry_index = 0; diff --git a/src/core/hle/service/hid/controllers/touchscreen.cpp b/src/core/hle/service/hid/controllers/touchscreen.cpp index 1c6e55566..0df395e85 100644 --- a/src/core/hle/service/hid/controllers/touchscreen.cpp +++ b/src/core/hle/service/hid/controllers/touchscreen.cpp @@ -22,7 +22,7 @@ void Controller_Touchscreen::OnRelease() {} void Controller_Touchscreen::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) { - shared_memory.header.timestamp = core_timing.GetTicks(); + shared_memory.header.timestamp = core_timing.GetCPUTicks(); shared_memory.header.total_entry_count = 17; if (!IsControllerActivated()) { @@ -40,16 +40,21 @@ void Controller_Touchscreen::OnUpdate(const Core::Timing::CoreTiming& core_timin cur_entry.sampling_number = last_entry.sampling_number + 1; cur_entry.sampling_number2 = cur_entry.sampling_number; - const auto [x, y, pressed] = touch_device->GetStatus(); + bool pressed = false; + float x, y; + std::tie(x, y, pressed) = touch_device->GetStatus(); auto& touch_entry = cur_entry.states[0]; touch_entry.attribute.raw = 0; + if (!pressed && touch_btn_device) { + std::tie(x, y, pressed) = touch_btn_device->GetStatus(); + } if (pressed && Settings::values.touchscreen.enabled) { touch_entry.x = static_cast<u16>(x * Layout::ScreenUndocked::Width); touch_entry.y = static_cast<u16>(y * Layout::ScreenUndocked::Height); touch_entry.diameter_x = Settings::values.touchscreen.diameter_x; touch_entry.diameter_y = Settings::values.touchscreen.diameter_y; touch_entry.rotation_angle = Settings::values.touchscreen.rotation_angle; - const u64 tick = core_timing.GetTicks(); + const u64 tick = core_timing.GetCPUTicks(); touch_entry.delta_time = tick - last_touch; last_touch = tick; touch_entry.finger = Settings::values.touchscreen.finger; @@ -63,5 +68,10 @@ void Controller_Touchscreen::OnUpdate(const Core::Timing::CoreTiming& core_timin void Controller_Touchscreen::OnLoadInputDevices() { touch_device = Input::CreateDevice<Input::TouchDevice>(Settings::values.touchscreen.device); + if (Settings::values.use_touch_from_button) { + touch_btn_device = Input::CreateDevice<Input::TouchDevice>("engine:touch_from_button"); + } else { + touch_btn_device.reset(); + } } } // namespace Service::HID diff --git a/src/core/hle/service/hid/controllers/touchscreen.h b/src/core/hle/service/hid/controllers/touchscreen.h index a1d97269e..4d9042adc 100644 --- a/src/core/hle/service/hid/controllers/touchscreen.h +++ b/src/core/hle/service/hid/controllers/touchscreen.h @@ -68,6 +68,7 @@ private: "TouchScreenSharedMemory is an invalid size"); TouchScreenSharedMemory shared_memory{}; std::unique_ptr<Input::TouchDevice> touch_device; + std::unique_ptr<Input::TouchDevice> touch_btn_device; s64_le last_touch{}; }; } // namespace Service::HID diff --git a/src/core/hle/service/hid/controllers/xpad.cpp b/src/core/hle/service/hid/controllers/xpad.cpp index 27511b27b..2503ef241 100644 --- a/src/core/hle/service/hid/controllers/xpad.cpp +++ b/src/core/hle/service/hid/controllers/xpad.cpp @@ -20,7 +20,7 @@ void Controller_XPad::OnRelease() {} void Controller_XPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) { for (auto& xpad_entry : shared_memory.shared_memory_entries) { - xpad_entry.header.timestamp = core_timing.GetTicks(); + xpad_entry.header.timestamp = core_timing.GetCPUTicks(); xpad_entry.header.total_entry_count = 17; if (!IsControllerActivated()) { diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index d6ed5f304..902516b29 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -14,6 +14,7 @@ #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/shared_memory.h" #include "core/hle/kernel/writable_event.h" @@ -37,12 +38,10 @@ namespace Service::HID { // Updating period for each HID device. -// TODO(ogniK): Find actual polling rate of hid -constexpr s64 pad_update_ticks = static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 66); -[[maybe_unused]] constexpr s64 accelerometer_update_ticks = - static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 100); -[[maybe_unused]] constexpr s64 gyroscope_update_ticks = - static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 100); +// HID is polled every 15ms, this value was derived from +// https://github.com/dekuNukem/Nintendo_Switch_Reverse_Engineering#joy-con-status-data-packet +constexpr auto pad_update_ns = std::chrono::nanoseconds{1000 * 1000}; // (1ms, 1000Hz) +constexpr auto motion_update_ns = std::chrono::nanoseconds{15 * 1000 * 1000}; // (15ms, 66.666Hz) constexpr std::size_t SHARED_MEMORY_SIZE = 0x40000; IAppletResource::IAppletResource(Core::System& system) @@ -53,9 +52,7 @@ IAppletResource::IAppletResource(Core::System& system) RegisterHandlers(functions); auto& kernel = system.Kernel(); - shared_mem = Kernel::SharedMemory::Create( - kernel, nullptr, SHARED_MEMORY_SIZE, Kernel::MemoryPermission::ReadWrite, - Kernel::MemoryPermission::Read, 0, Kernel::MemoryRegion::BASE, "HID:SharedMemory"); + shared_mem = SharedFrom(&kernel.GetHidSharedMem()); MakeController<Controller_DebugPad>(HidController::DebugPad); MakeController<Controller_Touchscreen>(HidController::Touchscreen); @@ -78,14 +75,19 @@ IAppletResource::IAppletResource(Core::System& system) GetController<Controller_Stubbed>(HidController::Unknown3).SetCommonHeaderOffset(0x5000); // Register update callbacks - pad_update_event = - Core::Timing::CreateEvent("HID::UpdatePadCallback", [this](u64 userdata, s64 cycles_late) { - UpdateControllers(userdata, cycles_late); + pad_update_event = Core::Timing::CreateEvent( + "HID::UpdatePadCallback", + [this](std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + UpdateControllers(user_data, ns_late); + }); + motion_update_event = Core::Timing::CreateEvent( + "HID::MotionPadCallback", + [this](std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + UpdateMotion(user_data, ns_late); }); - // TODO(shinyquagsire23): Other update callbacks? (accel, gyro?) - - system.CoreTiming().ScheduleEvent(pad_update_ticks, pad_update_event); + system.CoreTiming().ScheduleEvent(pad_update_ns, pad_update_event); + system.CoreTiming().ScheduleEvent(motion_update_ns, motion_update_event); ReloadInputDevices(); } @@ -110,7 +112,8 @@ void IAppletResource::GetSharedMemoryHandle(Kernel::HLERequestContext& ctx) { rb.PushCopyObjects(shared_mem); } -void IAppletResource::UpdateControllers(u64 userdata, s64 cycles_late) { +void IAppletResource::UpdateControllers(std::uintptr_t user_data, + std::chrono::nanoseconds ns_late) { auto& core_timing = system.CoreTiming(); const bool should_reload = Settings::values.is_device_reload_pending.exchange(false); @@ -121,25 +124,49 @@ void IAppletResource::UpdateControllers(u64 userdata, s64 cycles_late) { controller->OnUpdate(core_timing, shared_mem->GetPointer(), SHARED_MEMORY_SIZE); } - core_timing.ScheduleEvent(pad_update_ticks - cycles_late, pad_update_event); + core_timing.ScheduleEvent(pad_update_ns - ns_late, pad_update_event); +} + +void IAppletResource::UpdateMotion(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + auto& core_timing = system.CoreTiming(); + + for (const auto& controller : controllers) { + controller->OnMotionUpdate(core_timing, shared_mem->GetPointer(), SHARED_MEMORY_SIZE); + } + + core_timing.ScheduleEvent(motion_update_ns - ns_late, motion_update_event); } class IActiveVibrationDeviceList final : public ServiceFramework<IActiveVibrationDeviceList> { public: - IActiveVibrationDeviceList() : ServiceFramework("IActiveVibrationDeviceList") { + explicit IActiveVibrationDeviceList(std::shared_ptr<IAppletResource> applet_resource_) + : ServiceFramework("IActiveVibrationDeviceList"), applet_resource(applet_resource_) { + // clang-format off static const FunctionInfo functions[] = { - {0, &IActiveVibrationDeviceList::ActivateVibrationDevice, "ActivateVibrationDevice"}, + {0, &IActiveVibrationDeviceList::InitializeVibrationDevice, "InitializeVibrationDevice"}, }; + // clang-format on + RegisterHandlers(functions); } private: - void ActivateVibrationDevice(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_HID, "(STUBBED) called"); + void InitializeVibrationDevice(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto vibration_device_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .InitializeVibrationDevice(vibration_device_handle); + + LOG_DEBUG(Service_HID, "called, npad_type={}, npad_id={}, device_index={}", + vibration_device_handle.npad_type, vibration_device_handle.npad_id, + vibration_device_handle.device_index); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } + + std::shared_ptr<IAppletResource> applet_resource; }; std::shared_ptr<IAppletResource> Hid::GetAppletResource() { @@ -158,16 +185,16 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {11, &Hid::ActivateTouchScreen, "ActivateTouchScreen"}, {21, &Hid::ActivateMouse, "ActivateMouse"}, {31, &Hid::ActivateKeyboard, "ActivateKeyboard"}, - {32, nullptr, "SendKeyboardLockKeyEvent"}, + {32, &Hid::SendKeyboardLockKeyEvent, "SendKeyboardLockKeyEvent"}, {40, nullptr, "AcquireXpadIdEventHandle"}, {41, nullptr, "ReleaseXpadIdEventHandle"}, {51, &Hid::ActivateXpad, "ActivateXpad"}, - {55, nullptr, "GetXpadIds"}, + {55, &Hid::GetXpadIDs, "GetXpadIds"}, {56, nullptr, "ActivateJoyXpad"}, {58, nullptr, "GetJoyXpadLifoHandle"}, {59, nullptr, "GetJoyXpadIds"}, - {60, nullptr, "ActivateSixAxisSensor"}, - {61, nullptr, "DeactivateSixAxisSensor"}, + {60, &Hid::ActivateSixAxisSensor, "ActivateSixAxisSensor"}, + {61, &Hid::DeactivateSixAxisSensor, "DeactivateSixAxisSensor"}, {62, nullptr, "GetSixAxisSensorLifoHandle"}, {63, nullptr, "ActivateJoySixAxisSensor"}, {64, nullptr, "DeactivateJoySixAxisSensor"}, @@ -175,7 +202,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {66, &Hid::StartSixAxisSensor, "StartSixAxisSensor"}, {67, &Hid::StopSixAxisSensor, "StopSixAxisSensor"}, {68, nullptr, "IsSixAxisSensorFusionEnabled"}, - {69, nullptr, "EnableSixAxisSensorFusion"}, + {69, &Hid::EnableSixAxisSensorFusion, "EnableSixAxisSensorFusion"}, {70, nullptr, "SetSixAxisSensorFusionParameters"}, {71, nullptr, "GetSixAxisSensorFusionParameters"}, {72, nullptr, "ResetSixAxisSensorFusionParameters"}, @@ -186,8 +213,8 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {77, nullptr, "GetAccelerometerPlayMode"}, {78, nullptr, "ResetAccelerometerPlayMode"}, {79, &Hid::SetGyroscopeZeroDriftMode, "SetGyroscopeZeroDriftMode"}, - {80, nullptr, "GetGyroscopeZeroDriftMode"}, - {81, nullptr, "ResetGyroscopeZeroDriftMode"}, + {80, &Hid::GetGyroscopeZeroDriftMode, "GetGyroscopeZeroDriftMode"}, + {81, &Hid::ResetGyroscopeZeroDriftMode, "ResetGyroscopeZeroDriftMode"}, {82, &Hid::IsSixAxisSensorAtRest, "IsSixAxisSensorAtRest"}, {83, nullptr, "IsFirmwareUpdateAvailableForSixAxisSensor"}, {91, &Hid::ActivateGesture, "ActivateGesture"}, @@ -211,8 +238,8 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {128, &Hid::SetNpadHandheldActivationMode, "SetNpadHandheldActivationMode"}, {129, &Hid::GetNpadHandheldActivationMode, "GetNpadHandheldActivationMode"}, {130, &Hid::SwapNpadAssignment, "SwapNpadAssignment"}, - {131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"}, - {132, nullptr, "EnableUnintendedHomeButtonInputProtection"}, + {131, &Hid::IsUnintendedHomeButtonInputProtectionEnabled, "IsUnintendedHomeButtonInputProtectionEnabled"}, + {132, &Hid::EnableUnintendedHomeButtonInputProtection, "EnableUnintendedHomeButtonInputProtection"}, {133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"}, {134, nullptr, "SetNpadAnalogStickUseCenterClamp"}, {135, nullptr, "SetNpadCaptureButtonAssignment"}, @@ -228,18 +255,18 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {208, nullptr, "GetActualVibrationGcErmCommand"}, {209, &Hid::BeginPermitVibrationSession, "BeginPermitVibrationSession"}, {210, &Hid::EndPermitVibrationSession, "EndPermitVibrationSession"}, - {211, nullptr, "IsVibrationDeviceMounted"}, + {211, &Hid::IsVibrationDeviceMounted, "IsVibrationDeviceMounted"}, {300, &Hid::ActivateConsoleSixAxisSensor, "ActivateConsoleSixAxisSensor"}, {301, &Hid::StartConsoleSixAxisSensor, "StartConsoleSixAxisSensor"}, - {302, nullptr, "StopConsoleSixAxisSensor"}, - {303, nullptr, "ActivateSevenSixAxisSensor"}, - {304, nullptr, "StartSevenSixAxisSensor"}, - {305, nullptr, "StopSevenSixAxisSensor"}, + {302, &Hid::StopConsoleSixAxisSensor, "StopConsoleSixAxisSensor"}, + {303, &Hid::ActivateSevenSixAxisSensor, "ActivateSevenSixAxisSensor"}, + {304, &Hid::StartSevenSixAxisSensor, "StartSevenSixAxisSensor"}, + {305, &Hid::StopSevenSixAxisSensor, "StopSevenSixAxisSensor"}, {306, &Hid::InitializeSevenSixAxisSensor, "InitializeSevenSixAxisSensor"}, - {307, nullptr, "FinalizeSevenSixAxisSensor"}, + {307, &Hid::FinalizeSevenSixAxisSensor, "FinalizeSevenSixAxisSensor"}, {308, nullptr, "SetSevenSixAxisSensorFusionStrength"}, {309, nullptr, "GetSevenSixAxisSensorFusionStrength"}, - {310, nullptr, "ResetSevenSixAxisSensorTimestamp"}, + {310, &Hid::ResetSevenSixAxisSensorTimestamp, "ResetSevenSixAxisSensorTimestamp"}, {400, nullptr, "IsUsbFullKeyControllerEnabled"}, {401, nullptr, "EnableUsbFullKeyController"}, {402, nullptr, "IsUsbFullKeyControllerConnected"}, @@ -247,7 +274,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {404, nullptr, "HasLeftRightBattery"}, {405, nullptr, "GetNpadInterfaceType"}, {406, nullptr, "GetNpadLeftRightInterfaceType"}, - {407, nullptr, "GetNpadOfHighestBatteryLevelForJoyLeft"}, + {407, nullptr, "GetNpadOfHighestBatteryLevel"}, {408, nullptr, "GetNpadOfHighestBatteryLevelForJoyRight"}, {500, nullptr, "GetPalmaConnectionHandle"}, {501, nullptr, "InitializePalma"}, @@ -283,6 +310,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {1001, nullptr, "GetNpadCommunicationMode"}, {1002, nullptr, "SetTouchScreenConfiguration"}, {1003, nullptr, "IsFirmwareUpdateNeededForNotification"}, + {2000, nullptr, "ActivateDigitizer"}, }; // clang-format on @@ -306,97 +334,195 @@ void Hid::CreateAppletResource(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<IAppletResource>(applet_resource); } -void Hid::ActivateXpad(Kernel::HLERequestContext& ctx) { +void Hid::ActivateDebugPad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto basic_xpad_id{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, basic_xpad_id={}, applet_resource_user_id={}", basic_xpad_id, - applet_resource_user_id); + applet_resource->ActivateController(HidController::DebugPad); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::XPad); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateDebugPad(Kernel::HLERequestContext& ctx) { +void Hid::ActivateTouchScreen(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->ActivateController(HidController::Touchscreen); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::DebugPad); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateTouchScreen(Kernel::HLERequestContext& ctx) { +void Hid::ActivateMouse(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->ActivateController(HidController::Mouse); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::Touchscreen); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateMouse(Kernel::HLERequestContext& ctx) { +void Hid::ActivateKeyboard(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->ActivateController(HidController::Keyboard); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::Mouse); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateKeyboard(Kernel::HLERequestContext& ctx) { +void Hid::SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + const auto flags{rp.Pop<u32>()}; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + LOG_WARNING(Service_HID, "(STUBBED) called. flags={}", flags); - applet_resource->ActivateController(HidController::Keyboard); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateGesture(Kernel::HLERequestContext& ctx) { +void Hid::ActivateXpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto unknown{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 basic_xpad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", unknown, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->ActivateController(HidController::XPad); + + LOG_DEBUG(Service_HID, "called, basic_xpad_id={}, applet_resource_user_id={}", + parameters.basic_xpad_id, parameters.applet_resource_user_id); - applet_resource->ActivateController(HidController::Gesture); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) { - // Should have no effect with how our npad sets up the data +void Hid::GetXpadIDs(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto unknown{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", unknown, - applet_resource_user_id); + LOG_DEBUG(Service_HID, "(STUBBED) called, applet_resource_user_id={}", applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(0); +} + +void Hid::ActivateSixAxisSensor(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(true); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::DeactivateSixAxisSensor(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(false); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); - applet_resource->ActivateController(HidController::NPad); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } void Hid::StartSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(true); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(false); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::EnableSixAxisSensorFusion(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + bool enable_sixaxis_sensor_fusion{}; + INSERT_PADDING_BYTES(3); + Controller_NPad::DeviceHandle sixaxis_handle{}; + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_WARNING(Service_HID, + "(STUBBED) called, enable_sixaxis_sensor_fusion={}, npad_type={}, npad_id={}, " + "device_index={}, applet_resource_user_id={}", + parameters.enable_sixaxis_sensor_fusion, parameters.sixaxis_handle.npad_type, + parameters.sixaxis_handle.npad_id, parameters.sixaxis_handle.device_index, + parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -404,13 +530,61 @@ void Hid::StartSixAxisSensor(Kernel::HLERequestContext& ctx) { void Hid::SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto drift_mode{rp.Pop<u32>()}; + const auto sixaxis_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; + const auto drift_mode{rp.PopEnum<Controller_NPad::GyroscopeZeroDriftMode>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_WARNING(Service_HID, - "(STUBBED) called, handle={}, drift_mode={}, applet_resource_user_id={}", handle, - drift_mode, applet_resource_user_id); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetGyroscopeZeroDriftMode(drift_mode); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, drift_mode={}, " + "applet_resource_user_id={}", + sixaxis_handle.npad_type, sixaxis_handle.npad_id, sixaxis_handle.device_index, + drift_mode, applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::GetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetGyroscopeZeroDriftMode()); +} + +void Hid::ResetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetGyroscopeZeroDriftMode(Controller_NPad::GyroscopeZeroDriftMode::Standard); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -418,27 +592,53 @@ void Hid::SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { void Hid::IsSixAxisSensorAtRest(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - // TODO (Hexagon12): Properly implement reading gyroscope values from controllers. - rb.Push(true); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .IsSixAxisSensorAtRest()); +} + +void Hid::ActivateGesture(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + u32 unknown{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->ActivateController(HidController::Gesture); + + LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", parameters.unknown, + parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); } void Hid::SetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto supported_styleset{rp.Pop<u32>()}; - LOG_DEBUG(Service_HID, "called, supported_styleset={}", supported_styleset); - applet_resource->GetController<Controller_NPad>(HidController::NPad) .SetSupportedStyleSet({supported_styleset}); + LOG_DEBUG(Service_HID, "called, supported_styleset={}", supported_styleset); + IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -449,21 +649,22 @@ void Hid::GetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(controller.GetSupportedStyleSet().raw); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetSupportedStyleSet() + .raw); } void Hid::SetSupportedNpadIdType(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetSupportedNpadIdTypes(ctx.ReadBuffer().data(), ctx.GetReadBufferSize()); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->GetController<Controller_NPad>(HidController::NPad) - .SetSupportedNPadIdTypes(ctx.ReadBuffer().data(), ctx.GetReadBufferSize()); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -472,48 +673,62 @@ void Hid::ActivateNpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->ActivateController(HidController::NPad); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - applet_resource->ActivateController(HidController::NPad); } void Hid::DeactivateNpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->DeactivateController(HidController::NPad); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - applet_resource->DeactivateController(HidController::NPad); } void Hid::AcquireNpadStyleSetUpdateEventHandle(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto unknown{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + u64 unknown{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; - LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}, unknown={}", npad_id, - applet_resource_user_id, unknown); + LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}, unknown={}", + parameters.npad_id, parameters.applet_resource_user_id, parameters.unknown); IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); rb.PushCopyObjects(applet_resource->GetController<Controller_NPad>(HidController::NPad) - .GetStyleSetChangedEvent(npad_id)); + .GetStyleSetChangedEvent(parameters.npad_id)); } void Hid::DisconnectNpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", npad_id, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .DisconnectNpad(parameters.npad_id); + + LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", parameters.npad_id, + parameters.applet_resource_user_id); - applet_resource->GetController<Controller_NPad>(HidController::NPad).DisconnectNPad(npad_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -526,22 +741,41 @@ void Hid::GetPlayerLedPattern(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<u64>(applet_resource->GetController<Controller_NPad>(HidController::NPad) - .GetLedPattern(npad_id) - .raw); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetLedPattern(npad_id) + .raw); +} + +void Hid::ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) { + // Should have no effect with how our npad sets up the data + IPC::RequestParser rp{ctx}; + struct Parameters { + u32 unknown{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->ActivateController(HidController::NPad); + + LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", parameters.unknown, + parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); } void Hid::SetNpadJoyHoldType(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto hold_type{rp.Pop<u64>()}; + const auto hold_type{rp.PopEnum<Controller_NPad::NpadHoldType>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetHoldType(hold_type); LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, hold_type={}", applet_resource_user_id, hold_type); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetHoldType(Controller_NPad::NpadHoldType{hold_type}); - IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -552,22 +786,26 @@ void Hid::GetNpadJoyHoldType(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - const auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u64>(static_cast<u64>(controller.GetHoldType())); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad).GetHoldType()); } void Hid::SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", npad_id, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Single); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadMode(parameters.npad_id, Controller_NPad::NpadAssignments::Single); + + LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", + parameters.npad_id, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -576,16 +814,22 @@ void Hid::SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx void Hid::SetNpadJoyAssignmentModeSingle(Kernel::HLERequestContext& ctx) { // TODO: Check the differences between this and SetNpadJoyAssignmentModeSingleByDefault IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto npad_joy_device_type{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + u64 npad_joy_device_type{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadMode(parameters.npad_id, Controller_NPad::NpadAssignments::Single); LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}, npad_joy_device_type={}", - npad_id, applet_resource_user_id, npad_joy_device_type); - - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Single); + parameters.npad_id, parameters.applet_resource_user_id, + parameters.npad_joy_device_type); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -593,14 +837,19 @@ void Hid::SetNpadJoyAssignmentModeSingle(Kernel::HLERequestContext& ctx) { void Hid::SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", npad_id, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Dual); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadMode(parameters.npad_id, Controller_NPad::NpadAssignments::Dual); + + LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", + parameters.npad_id, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -608,13 +857,15 @@ void Hid::SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx) { void Hid::MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto unknown_1{rp.Pop<u32>()}; - const auto unknown_2{rp.Pop<u32>()}; + const auto npad_id_1{rp.Pop<u32>()}; + const auto npad_id_2{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_WARNING(Service_HID, - "(STUBBED) called, unknown_1={}, unknown_2={}, applet_resource_user_id={}", - unknown_1, unknown_2, applet_resource_user_id); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .MergeSingleJoyAsDualJoy(npad_id_1, npad_id_2); + + LOG_DEBUG(Service_HID, "called, npad_id_1={}, npad_id_2={}, applet_resource_user_id={}", + npad_id_1, npad_id_2, applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -624,9 +875,9 @@ void Hid::StartLrAssignmentMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad).StartLRAssignmentMode(); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.StartLRAssignmentMode(); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -636,9 +887,9 @@ void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad).StopLRAssignmentMode(); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.StopLRAssignmentMode(); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -647,10 +898,13 @@ void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) { void Hid::SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto mode{rp.Pop<u64>()}; + const auto activation_mode{rp.PopEnum<Controller_NPad::NpadHandheldActivationMode>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadHandheldActivationMode(activation_mode); - LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}, mode={}", - applet_resource_user_id, mode); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, activation_mode={}", + applet_resource_user_id, activation_mode); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -660,25 +914,28 @@ void Hid::GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", - applet_resource_user_id); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetNpadHandheldActivationMode()); } void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_1{rp.Pop<u32>()}; - const auto npad_2{rp.Pop<u32>()}; + const auto npad_id_1{rp.Pop<u32>()}; + const auto npad_id_2{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, npad_1={}, npad_2={}", - applet_resource_user_id, npad_1, npad_2); + const bool res = applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SwapNpadAssignment(npad_id_1, npad_id_2); + + LOG_DEBUG(Service_HID, "called, npad_id_1={}, npad_id_2={}, applet_resource_user_id={}", + npad_id_1, npad_id_2, applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); IPC::ResponseBuilder rb{ctx, 2}; - if (controller.SwapNpadAssignment(npad_1, npad_2)) { + if (res) { rb.Push(RESULT_SUCCESS); } else { LOG_ERROR(Service_HID, "Npads are not connected!"); @@ -686,119 +943,275 @@ void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) { } } -void Hid::BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) { +void Hid::IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; - applet_resource->GetController<Controller_NPad>(HidController::NPad).SetVibrationEnabled(true); - IPC::ResponseBuilder rb{ctx, 2}; + LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", + parameters.npad_id, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .IsUnintendedHomeButtonInputProtectionEnabled(parameters.npad_id)); } -void Hid::EndPermitVibrationSession(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_HID, "called"); +void Hid::EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + bool unintended_home_button_input_protection{}; + INSERT_PADDING_BYTES(3); + u32 npad_id{}; + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetUnintendedHomeButtonInputProtectionEnabled( + parameters.unintended_home_button_input_protection, parameters.npad_id); + + LOG_WARNING(Service_HID, + "(STUBBED) called, unintended_home_button_input_protection={}, npad_id={}," + "applet_resource_user_id={}", + parameters.unintended_home_button_input_protection, parameters.npad_id, + parameters.applet_resource_user_id); - applet_resource->GetController<Controller_NPad>(HidController::NPad).SetVibrationEnabled(false); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } +void Hid::GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto vibration_device_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; + + VibrationDeviceInfo vibration_device_info; + + vibration_device_info.type = VibrationDeviceType::LinearResonantActuator; + + switch (vibration_device_handle.device_index) { + case Controller_NPad::DeviceIndex::Left: + vibration_device_info.position = VibrationDevicePosition::Left; + break; + case Controller_NPad::DeviceIndex::Right: + vibration_device_info.position = VibrationDevicePosition::Right; + break; + case Controller_NPad::DeviceIndex::None: + default: + UNREACHABLE_MSG("DeviceIndex should never be None!"); + vibration_device_info.position = VibrationDevicePosition::None; + break; + } + + LOG_DEBUG(Service_HID, "called, vibration_device_type={}, vibration_device_position={}", + vibration_device_info.type, vibration_device_info.position); + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw(vibration_device_info); +} + void Hid::SendVibrationValue(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto controller_id{rp.Pop<u32>()}; - const auto vibration_values{rp.PopRaw<Controller_NPad::Vibration>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle vibration_device_handle{}; + Controller_NPad::VibrationValue vibration_value{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .VibrateController(parameters.vibration_device_handle, parameters.vibration_value); - LOG_DEBUG(Service_HID, "called, controller_id={}, applet_resource_user_id={}", controller_id, - applet_resource_user_id); + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.vibration_device_handle.npad_type, + parameters.vibration_device_handle.npad_id, + parameters.vibration_device_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); +} - applet_resource->GetController<Controller_NPad>(HidController::NPad) - .VibrateController({controller_id}, {vibration_values}); +void Hid::GetActualVibrationValue(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle vibration_device_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.vibration_device_handle.npad_type, + parameters.vibration_device_handle.npad_id, + parameters.vibration_device_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 6}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetLastVibration(parameters.vibration_device_handle)); +} + +void Hid::CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_HID, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IActiveVibrationDeviceList>(applet_resource); +} + +void Hid::PermitVibration(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto can_vibrate{rp.Pop<bool>()}; + + Settings::values.vibration_enabled.SetValue(can_vibrate); + + LOG_DEBUG(Service_HID, "called, can_vibrate={}", can_vibrate); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::IsVibrationPermitted(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_HID, "called"); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(Settings::values.vibration_enabled.GetValue()); } void Hid::SendVibrationValues(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - - const auto controllers = ctx.ReadBuffer(0); + const auto handles = ctx.ReadBuffer(0); const auto vibrations = ctx.ReadBuffer(1); - std::vector<u32> controller_list(controllers.size() / sizeof(u32)); - std::vector<Controller_NPad::Vibration> vibration_list(vibrations.size() / - sizeof(Controller_NPad::Vibration)); + std::vector<Controller_NPad::DeviceHandle> vibration_device_handles( + handles.size() / sizeof(Controller_NPad::DeviceHandle)); + std::vector<Controller_NPad::VibrationValue> vibration_values( + vibrations.size() / sizeof(Controller_NPad::VibrationValue)); - std::memcpy(controller_list.data(), controllers.data(), controllers.size()); - std::memcpy(vibration_list.data(), vibrations.data(), vibrations.size()); - std::transform(controller_list.begin(), controller_list.end(), controller_list.begin(), - [](u32 controller_id) { return controller_id - 3; }); + std::memcpy(vibration_device_handles.data(), handles.data(), handles.size()); + std::memcpy(vibration_values.data(), vibrations.data(), vibrations.size()); applet_resource->GetController<Controller_NPad>(HidController::NPad) - .VibrateController(controller_list, vibration_list); + .VibrateControllers(vibration_device_handles, vibration_values); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::GetActualVibrationValue(Kernel::HLERequestContext& ctx) { +void Hid::BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto controller_id{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, controller_id={}, applet_resource_user_id={}", controller_id, - applet_resource_user_id); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetPermitVibrationSession(true); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - IPC::ResponseBuilder rb{ctx, 6}; + IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<Controller_NPad::Vibration>( - applet_resource->GetController<Controller_NPad>(HidController::NPad).GetLastVibration()); } -void Hid::GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) { +void Hid::EndPermitVibrationSession(Kernel::HLERequestContext& ctx) { + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetPermitVibrationSession(false); + LOG_DEBUG(Service_HID, "called"); - IPC::ResponseBuilder rb{ctx, 4}; + IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(1); - rb.Push<u32>(0); } -void Hid::CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_HID, "called"); +void Hid::IsVibrationDeviceMounted(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle vibration_device_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.vibration_device_handle.npad_type, + parameters.vibration_device_handle.npad_id, + parameters.vibration_device_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IActiveVibrationDeviceList>(); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .IsVibrationDeviceMounted(parameters.vibration_device_handle)); } -void Hid::PermitVibration(Kernel::HLERequestContext& ctx) { +void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto can_vibrate{rp.Pop<bool>()}; - applet_resource->GetController<Controller_NPad>(HidController::NPad) - .SetVibrationEnabled(can_vibrate); + const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, can_vibrate={}", can_vibrate); + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", + applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::IsVibrationPermitted(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_HID, "called"); +void Hid::StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - IPC::ResponseBuilder rb{ctx, 3}; + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_WARNING( + Service_HID, + "(STUBBED) called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - rb.Push( - applet_resource->GetController<Controller_NPad>(HidController::NPad).IsVibrationEnabled()); } -void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { +void Hid::StopConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_WARNING( + Service_HID, + "(STUBBED) called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::ActivateSevenSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; @@ -809,52 +1222,75 @@ void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } -void Hid::StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { +void Hid::StartSevenSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) { +void Hid::StopSevenSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}", handle); + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", + applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx) { +void Hid::InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_HID, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::FinalizeSevenSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto unknown{rp.Pop<u32>()}; - LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}, unknown={}", - applet_resource_user_id, unknown); + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", + applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) { +void Hid::ResetSevenSixAxisSensorTimestamp(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto unknown{rp.Pop<u32>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_WARNING(Service_HID, "(STUBBED) called, unknown={}", unknown); + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", + applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_HID, "(STUBBED) called"); +void Hid::SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + const auto is_palma_all_connectable{rp.Pop<bool>()}; + + LOG_WARNING(Service_HID, + "(STUBBED) called, applet_resource_user_id={}, is_palma_all_connectable={}", + applet_resource_user_id, is_palma_all_connectable); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto palma_boost_mode{rp.Pop<bool>()}; + + LOG_WARNING(Service_HID, "(STUBBED) called, palma_boost_mode={}", palma_boost_mode); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -871,6 +1307,7 @@ public: {10, nullptr, "DeactivateTouchScreen"}, {11, nullptr, "SetTouchScreenAutoPilotState"}, {12, nullptr, "UnsetTouchScreenAutoPilotState"}, + {13, nullptr, "GetTouchScreenConfiguration"}, {20, nullptr, "DeactivateMouse"}, {21, nullptr, "SetMouseAutoPilotState"}, {22, nullptr, "UnsetMouseAutoPilotState"}, @@ -880,7 +1317,9 @@ public: {50, nullptr, "DeactivateXpad"}, {51, nullptr, "SetXpadAutoPilotState"}, {52, nullptr, "UnsetXpadAutoPilotState"}, - {60, nullptr, "DeactivateJoyXpad"}, + {60, nullptr, "ClearNpadSystemCommonPolicy"}, + {61, nullptr, "DeactivateNpad"}, + {62, nullptr, "ForceDisconnectNpad"}, {91, nullptr, "DeactivateGesture"}, {110, nullptr, "DeactivateHomeButton"}, {111, nullptr, "SetHomeButtonAutoPilotState"}, @@ -900,6 +1339,15 @@ public: {141, nullptr, "GetConsoleSixAxisSensorSamplingFrequency"}, {142, nullptr, "DeactivateSevenSixAxisSensor"}, {143, nullptr, "GetConsoleSixAxisSensorCountStates"}, + {144, nullptr, "GetAccelerometerFsr"}, + {145, nullptr, "SetAccelerometerFsr"}, + {146, nullptr, "GetAccelerometerOdr"}, + {147, nullptr, "SetAccelerometerOdr"}, + {148, nullptr, "GetGyroscopeFsr"}, + {149, nullptr, "SetGyroscopeFsr"}, + {150, nullptr, "GetGyroscopeOdr"}, + {151, nullptr, "SetGyroscopeOdr"}, + {152, nullptr, "GetWhoAmI"}, {201, nullptr, "ActivateFirmwareUpdate"}, {202, nullptr, "DeactivateFirmwareUpdate"}, {203, nullptr, "StartFirmwareUpdate"}, @@ -928,6 +1376,17 @@ public: {233, nullptr, "ClearPairingInfo"}, {234, nullptr, "GetUniquePadDeviceTypeSetInternal"}, {235, nullptr, "EnableAnalogStickPower"}, + {236, nullptr, "RequestKuinaUartClockCal"}, + {237, nullptr, "GetKuinaUartClockCal"}, + {238, nullptr, "SetKuinaUartClockTrim"}, + {239, nullptr, "KuinaLoopbackTest"}, + {240, nullptr, "RequestBatteryVoltage"}, + {241, nullptr, "GetBatteryVoltage"}, + {242, nullptr, "GetUniquePadPowerInfo"}, + {243, nullptr, "RebootUniquePad"}, + {244, nullptr, "RequestKuinaFirmwareVersion"}, + {245, nullptr, "GetKuinaFirmwareVersion"}, + {246, nullptr, "GetVidPid"}, {301, nullptr, "GetAbstractedPadHandles"}, {302, nullptr, "GetAbstractedPadState"}, {303, nullptr, "GetAbstractedPadsState"}, @@ -946,6 +1405,17 @@ public: {350, nullptr, "AddRegisteredDevice"}, {400, nullptr, "DisableExternalMcuOnNxDevice"}, {401, nullptr, "DisableRailDeviceFiltering"}, + {402, nullptr, "EnableWiredPairing"}, + {403, nullptr, "EnableShipmentModeAutoClear"}, + {500, nullptr, "SetFactoryInt"}, + {501, nullptr, "IsFactoryBootEnabled"}, + {550, nullptr, "SetAnalogStickModelDataTemporarily"}, + {551, nullptr, "GetAnalogStickModelData"}, + {552, nullptr, "ResetAnalogStickModelData"}, + {600, nullptr, "ConvertPadState"}, + {2000, nullptr, "DeactivateDigitizer"}, + {2001, nullptr, "SetDigitizerAutoPilotState"}, + {2002, nullptr, "UnsetDigitizerAutoPilotState"}, }; // clang-format on diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h index 039c38b58..c8e4a4b55 100644 --- a/src/core/hle/service/hid/hid.h +++ b/src/core/hle/service/hid/hid.h @@ -4,10 +4,9 @@ #pragma once -#include "core/hle/service/hid/controllers/controller_base.h" -#include "core/hle/service/service.h" +#include <chrono> -#include "controllers/controller_base.h" +#include "core/hle/service/hid/controllers/controller_base.h" #include "core/hle/service/service.h" namespace Core::Timing { @@ -65,11 +64,13 @@ private: } void GetSharedMemoryHandle(Kernel::HLERequestContext& ctx); - void UpdateControllers(u64 userdata, s64 cycles_late); + void UpdateControllers(std::uintptr_t user_data, std::chrono::nanoseconds ns_late); + void UpdateMotion(std::uintptr_t user_data, std::chrono::nanoseconds ns_late); std::shared_ptr<Kernel::SharedMemory> shared_mem; std::shared_ptr<Core::Timing::EventType> pad_update_event; + std::shared_ptr<Core::Timing::EventType> motion_update_event; Core::System& system; std::array<std::unique_ptr<ControllerBase>, static_cast<size_t>(HidController::MaxControllers)> @@ -85,16 +86,23 @@ public: private: void CreateAppletResource(Kernel::HLERequestContext& ctx); - void ActivateXpad(Kernel::HLERequestContext& ctx); void ActivateDebugPad(Kernel::HLERequestContext& ctx); void ActivateTouchScreen(Kernel::HLERequestContext& ctx); void ActivateMouse(Kernel::HLERequestContext& ctx); void ActivateKeyboard(Kernel::HLERequestContext& ctx); - void ActivateGesture(Kernel::HLERequestContext& ctx); - void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx); + void SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx); + void ActivateXpad(Kernel::HLERequestContext& ctx); + void GetXpadIDs(Kernel::HLERequestContext& ctx); + void ActivateSixAxisSensor(Kernel::HLERequestContext& ctx); + void DeactivateSixAxisSensor(Kernel::HLERequestContext& ctx); void StartSixAxisSensor(Kernel::HLERequestContext& ctx); + void StopSixAxisSensor(Kernel::HLERequestContext& ctx); + void EnableSixAxisSensorFusion(Kernel::HLERequestContext& ctx); void SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx); + void GetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx); + void ResetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx); void IsSixAxisSensorAtRest(Kernel::HLERequestContext& ctx); + void ActivateGesture(Kernel::HLERequestContext& ctx); void SetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx); void GetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx); void SetSupportedNpadIdType(Kernel::HLERequestContext& ctx); @@ -103,6 +111,7 @@ private: void AcquireNpadStyleSetUpdateEventHandle(Kernel::HLERequestContext& ctx); void DisconnectNpad(Kernel::HLERequestContext& ctx); void GetPlayerLedPattern(Kernel::HLERequestContext& ctx); + void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx); void SetNpadJoyHoldType(Kernel::HLERequestContext& ctx); void GetNpadJoyHoldType(Kernel::HLERequestContext& ctx); void SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx); @@ -114,21 +123,45 @@ private: void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx); void GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx); void SwapNpadAssignment(Kernel::HLERequestContext& ctx); - void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx); - void EndPermitVibrationSession(Kernel::HLERequestContext& ctx); + void IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx); + void EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx); + void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx); void SendVibrationValue(Kernel::HLERequestContext& ctx); - void SendVibrationValues(Kernel::HLERequestContext& ctx); void GetActualVibrationValue(Kernel::HLERequestContext& ctx); - void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx); void CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx); void PermitVibration(Kernel::HLERequestContext& ctx); void IsVibrationPermitted(Kernel::HLERequestContext& ctx); + void SendVibrationValues(Kernel::HLERequestContext& ctx); + void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx); + void EndPermitVibrationSession(Kernel::HLERequestContext& ctx); + void IsVibrationDeviceMounted(Kernel::HLERequestContext& ctx); void ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx); void StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx); - void StopSixAxisSensor(Kernel::HLERequestContext& ctx); + void StopConsoleSixAxisSensor(Kernel::HLERequestContext& ctx); + void ActivateSevenSixAxisSensor(Kernel::HLERequestContext& ctx); + void StartSevenSixAxisSensor(Kernel::HLERequestContext& ctx); + void StopSevenSixAxisSensor(Kernel::HLERequestContext& ctx); + void InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx); + void FinalizeSevenSixAxisSensor(Kernel::HLERequestContext& ctx); + void ResetSevenSixAxisSensorTimestamp(Kernel::HLERequestContext& ctx); void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx); void SetPalmaBoostMode(Kernel::HLERequestContext& ctx); - void InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx); + + enum class VibrationDeviceType : u32 { + LinearResonantActuator = 1, + }; + + enum class VibrationDevicePosition : u32 { + None = 0, + Left = 1, + Right = 2, + }; + + struct VibrationDeviceInfo { + VibrationDeviceType type{}; + VibrationDevicePosition position{}; + }; + static_assert(sizeof(VibrationDeviceInfo) == 0x8, "VibrationDeviceInfo has incorrect size."); std::shared_ptr<IAppletResource> applet_resource; Core::System& system; diff --git a/src/core/hle/service/hid/irs.cpp b/src/core/hle/service/hid/irs.cpp index 5e79e2c1a..e82fd031b 100644 --- a/src/core/hle/service/hid/irs.cpp +++ b/src/core/hle/service/hid/irs.cpp @@ -6,6 +6,7 @@ #include "core/core.h" #include "core/core_timing.h" #include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/shared_memory.h" #include "core/hle/service/hid/irs.h" @@ -38,9 +39,8 @@ IRS::IRS(Core::System& system) : ServiceFramework{"irs"}, system(system) { RegisterHandlers(functions); auto& kernel = system.Kernel(); - shared_mem = Kernel::SharedMemory::Create( - kernel, nullptr, 0x8000, Kernel::MemoryPermission::ReadWrite, - Kernel::MemoryPermission::Read, 0, Kernel::MemoryRegion::BASE, "IRS:SharedMemory"); + + shared_mem = SharedFrom(&kernel.GetIrsSharedMem()); } void IRS::ActivateIrsensor(Kernel::HLERequestContext& ctx) { @@ -98,7 +98,7 @@ void IRS::GetImageTransferProcessorState(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 5}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<u64>(system.CoreTiming().GetTicks()); + rb.PushRaw<u64>(system.CoreTiming().GetCPUTicks()); rb.PushRaw<u32>(0); } diff --git a/src/core/hle/service/lbl/lbl.cpp b/src/core/hle/service/lbl/lbl.cpp index e8f9f2d29..17350b403 100644 --- a/src/core/hle/service/lbl/lbl.cpp +++ b/src/core/hle/service/lbl/lbl.cpp @@ -47,6 +47,7 @@ public: {26, &LBL::EnableVrMode, "EnableVrMode"}, {27, &LBL::DisableVrMode, "DisableVrMode"}, {28, &LBL::IsVrModeEnabled, "IsVrModeEnabled"}, + {29, nullptr, "IsAutoBrightnessControlSupported"}, }; // clang-format on diff --git a/src/core/hle/service/ldn/ldn.cpp b/src/core/hle/service/ldn/ldn.cpp index 92adde6d4..49972cd69 100644 --- a/src/core/hle/service/ldn/ldn.cpp +++ b/src/core/hle/service/ldn/ldn.cpp @@ -69,6 +69,7 @@ public: {101, nullptr, "GetNetworkInfoLatestUpdate"}, {102, nullptr, "Scan"}, {103, nullptr, "ScanPrivate"}, + {104, nullptr, "SetWirelessControllerRestriction"}, {200, nullptr, "OpenAccessPoint"}, {201, nullptr, "CloseAccessPoint"}, {202, nullptr, "CreateNetwork"}, diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp index 647943020..65c209725 100644 --- a/src/core/hle/service/ldr/ldr.cpp +++ b/src/core/hle/service/ldr/ldr.cpp @@ -8,15 +8,23 @@ #include "common/alignment.h" #include "common/hex_util.h" +#include "common/scope_exit.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/errors.h" +#include "core/hle/kernel/memory/page_table.h" +#include "core/hle/kernel/memory/system_control.h" #include "core/hle/kernel/process.h" #include "core/hle/service/ldr/ldr.h" #include "core/hle/service/service.h" #include "core/loader/nro.h" +#include "core/memory.h" namespace Service::LDR { -constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51}; +constexpr ResultCode ERROR_INSUFFICIENT_ADDRESS_SPACE{ErrorModule::RO, 2}; + +[[maybe_unused]] constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51}; constexpr ResultCode ERROR_INVALID_NRO{ErrorModule::Loader, 52}; constexpr ResultCode ERROR_INVALID_NRR{ErrorModule::Loader, 53}; constexpr ResultCode ERROR_MISSING_NRR_HASH{ErrorModule::Loader, 54}; @@ -26,10 +34,84 @@ constexpr ResultCode ERROR_ALREADY_LOADED{ErrorModule::Loader, 57}; constexpr ResultCode ERROR_INVALID_ALIGNMENT{ErrorModule::Loader, 81}; constexpr ResultCode ERROR_INVALID_SIZE{ErrorModule::Loader, 82}; constexpr ResultCode ERROR_INVALID_NRO_ADDRESS{ErrorModule::Loader, 84}; -constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85}; +[[maybe_unused]] constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85}; constexpr ResultCode ERROR_NOT_INITIALIZED{ErrorModule::Loader, 87}; -constexpr u64 MAXIMUM_LOADED_RO = 0x40; +constexpr std::size_t MAXIMUM_LOADED_RO{0x40}; +constexpr std::size_t MAXIMUM_MAP_RETRIES{0x200}; + +constexpr std::size_t TEXT_INDEX{0}; +constexpr std::size_t RO_INDEX{1}; +constexpr std::size_t DATA_INDEX{2}; + +struct NRRCertification { + u64_le application_id_mask; + u64_le application_id_pattern; + INSERT_PADDING_BYTES(0x10); + std::array<u8, 0x100> public_key; // Also known as modulus + std::array<u8, 0x100> signature; +}; +static_assert(sizeof(NRRCertification) == 0x220, "NRRCertification has invalid size."); + +struct NRRHeader { + u32_le magic; + u32_le certification_signature_key_generation; // 9.0.0+ + INSERT_PADDING_WORDS(2); + NRRCertification certification; + std::array<u8, 0x100> signature; + u64_le application_id; + u32_le size; + u8 nrr_kind; // 7.0.0+ + INSERT_PADDING_BYTES(3); + u32_le hash_offset; + u32_le hash_count; + INSERT_PADDING_WORDS(2); +}; +static_assert(sizeof(NRRHeader) == 0x350, "NRRHeader has invalid size."); + +struct SegmentHeader { + u32_le memory_offset; + u32_le memory_size; +}; +static_assert(sizeof(SegmentHeader) == 0x8, "SegmentHeader has invalid size."); + +struct NROHeader { + // Switchbrew calls this "Start" (0x10) + INSERT_PADDING_WORDS(1); + u32_le mod_offset; + INSERT_PADDING_WORDS(2); + + // Switchbrew calls this "Header" (0x70) + u32_le magic; + u32_le version; + u32_le nro_size; + u32_le flags; + // .text, .ro, .data + std::array<SegmentHeader, 3> segment_headers; + u32_le bss_size; + INSERT_PADDING_WORDS(1); + std::array<u8, 0x20> build_id; + u32_le dso_handle_offset; + INSERT_PADDING_WORDS(1); + // .apiInfo, .dynstr, .dynsym + std::array<SegmentHeader, 3> segment_headers_2; +}; +static_assert(sizeof(NROHeader) == 0x80, "NROHeader has invalid size."); + +using SHA256Hash = std::array<u8, 0x20>; + +struct NROInfo { + SHA256Hash hash{}; + VAddr nro_address{}; + std::size_t nro_size{}; + VAddr bss_address{}; + std::size_t bss_size{}; + std::size_t text_size{}; + std::size_t ro_size{}; + std::size_t data_size{}; + VAddr src_addr{}; +}; +static_assert(sizeof(NROInfo) == 0x60, "NROInfo has invalid size."); class DebugMonitor final : public ServiceFramework<DebugMonitor> { public: @@ -55,6 +137,7 @@ public: {1, nullptr, "GetProgramInfo"}, {2, nullptr, "RegisterTitle"}, {3, nullptr, "UnregisterTitle"}, + {4, nullptr, "SetEnabledProgramVerification"}, }; // clang-format on @@ -164,11 +247,11 @@ public: return; } - if (system.CurrentProcess()->GetTitleID() != header.title_id) { + if (system.CurrentProcess()->GetTitleID() != header.application_id) { LOG_ERROR(Service_LDR, "Attempting to load NRR with title ID other than current process. (actual " "{:016X})!", - header.title_id); + header.application_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_NRR); return; @@ -191,45 +274,140 @@ public: } void UnloadNrr(Kernel::HLERequestContext& ctx) { - if (!initialized) { - LOG_ERROR(Service_LDR, "LDR:RO not initialized before use!"); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_NOT_INITIALIZED); - return; + IPC::RequestParser rp{ctx}; + const auto pid = rp.Pop<u64>(); + const auto nrr_address = rp.Pop<VAddr>(); + + LOG_DEBUG(Service_LDR, "called with pid={}, nrr_address={:016X}", pid, nrr_address); + + nrr.erase(nrr_address); + + IPC::ResponseBuilder rb{ctx, 2}; + + rb.Push(RESULT_SUCCESS); + } + + bool ValidateRegionForMap(Kernel::Memory::PageTable& page_table, VAddr start, + std::size_t size) const { + constexpr std::size_t padding_size{4 * Kernel::Memory::PageSize}; + const auto start_info{page_table.QueryInfo(start - 1)}; + + if (start_info.state != Kernel::Memory::MemoryState::Free) { + return {}; } - struct Parameters { - u64_le process_id; - u64_le nrr_address; - }; + if (start_info.GetAddress() > (start - padding_size)) { + return {}; + } - IPC::RequestParser rp{ctx}; - const auto [process_id, nrr_address] = rp.PopRaw<Parameters>(); + const auto end_info{page_table.QueryInfo(start + size)}; - LOG_DEBUG(Service_LDR, "called with process_id={:016X}, nrr_addr={:016X}", process_id, - nrr_address); + if (end_info.state != Kernel::Memory::MemoryState::Free) { + return {}; + } - if (!Common::Is4KBAligned(nrr_address)) { - LOG_ERROR(Service_LDR, "NRR Address has invalid alignment (actual {:016X})!", - nrr_address); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_ALIGNMENT); - return; + return (start + size + padding_size) <= (end_info.GetAddress() + end_info.GetSize()); + } + + VAddr GetRandomMapRegion(const Kernel::Memory::PageTable& page_table, std::size_t size) const { + VAddr addr{}; + const std::size_t end_pages{(page_table.GetAliasCodeRegionSize() - size) >> + Kernel::Memory::PageBits}; + do { + addr = page_table.GetAliasCodeRegionStart() + + (Kernel::Memory::SystemControl::GenerateRandomRange(0, end_pages) + << Kernel::Memory::PageBits); + } while (!page_table.IsInsideAddressSpace(addr, size) || + page_table.IsInsideHeapRegion(addr, size) || + page_table.IsInsideAliasRegion(addr, size)); + return addr; + } + + ResultVal<VAddr> MapProcessCodeMemory(Kernel::Process* process, VAddr baseAddress, + u64 size) const { + for (std::size_t retry = 0; retry < MAXIMUM_MAP_RETRIES; retry++) { + auto& page_table{process->PageTable()}; + const VAddr addr{GetRandomMapRegion(page_table, size)}; + const ResultCode result{page_table.MapProcessCodeMemory(addr, baseAddress, size)}; + + if (result == Kernel::ERR_INVALID_ADDRESS_STATE) { + continue; + } + + CASCADE_CODE(result); + + if (ValidateRegionForMap(page_table, addr, size)) { + return MakeResult<VAddr>(addr); + } } - const auto iter = nrr.find(nrr_address); - if (iter == nrr.end()) { - LOG_ERROR(Service_LDR, - "Attempting to unload NRR which has not been loaded! (addr={:016X})", - nrr_address); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_NRR_ADDRESS); - return; + return ERROR_INSUFFICIENT_ADDRESS_SPACE; + } + + ResultVal<VAddr> MapNro(Kernel::Process* process, VAddr nro_addr, std::size_t nro_size, + VAddr bss_addr, std::size_t bss_size, std::size_t size) const { + for (std::size_t retry = 0; retry < MAXIMUM_MAP_RETRIES; retry++) { + auto& page_table{process->PageTable()}; + VAddr addr{}; + + CASCADE_RESULT(addr, MapProcessCodeMemory(process, nro_addr, nro_size)); + + if (bss_size) { + auto block_guard = detail::ScopeExit([&] { + page_table.UnmapProcessCodeMemory(addr + nro_size, bss_addr, bss_size); + page_table.UnmapProcessCodeMemory(addr, nro_addr, nro_size); + }); + + const ResultCode result{ + page_table.MapProcessCodeMemory(addr + nro_size, bss_addr, bss_size)}; + + if (result == Kernel::ERR_INVALID_ADDRESS_STATE) { + continue; + } + + if (result.IsError()) { + return result; + } + + block_guard.Cancel(); + } + + if (ValidateRegionForMap(page_table, addr, size)) { + return MakeResult<VAddr>(addr); + } } - nrr.erase(iter); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); + return ERROR_INSUFFICIENT_ADDRESS_SPACE; + } + + ResultCode LoadNro(Kernel::Process* process, const NROHeader& nro_header, VAddr nro_addr, + VAddr start) const { + const VAddr text_start{start + nro_header.segment_headers[TEXT_INDEX].memory_offset}; + const VAddr ro_start{start + nro_header.segment_headers[RO_INDEX].memory_offset}; + const VAddr data_start{start + nro_header.segment_headers[DATA_INDEX].memory_offset}; + const VAddr bss_start{data_start + nro_header.segment_headers[DATA_INDEX].memory_size}; + const VAddr bss_end_addr{ + Common::AlignUp(bss_start + nro_header.bss_size, Kernel::Memory::PageSize)}; + + auto CopyCode{[&](VAddr src_addr, VAddr dst_addr, u64 size) { + std::vector<u8> source_data(size); + system.Memory().ReadBlock(src_addr, source_data.data(), source_data.size()); + system.Memory().WriteBlock(dst_addr, source_data.data(), source_data.size()); + }}; + CopyCode(nro_addr + nro_header.segment_headers[TEXT_INDEX].memory_offset, text_start, + nro_header.segment_headers[TEXT_INDEX].memory_size); + CopyCode(nro_addr + nro_header.segment_headers[RO_INDEX].memory_offset, ro_start, + nro_header.segment_headers[RO_INDEX].memory_size); + CopyCode(nro_addr + nro_header.segment_headers[DATA_INDEX].memory_offset, data_start, + nro_header.segment_headers[DATA_INDEX].memory_size); + + CASCADE_CODE(process->PageTable().SetCodeMemoryPermission( + text_start, ro_start - text_start, Kernel::Memory::MemoryPermission::ReadAndExecute)); + CASCADE_CODE(process->PageTable().SetCodeMemoryPermission( + ro_start, data_start - ro_start, Kernel::Memory::MemoryPermission::Read)); + + return process->PageTable().SetCodeMemoryPermission( + data_start, bss_end_addr - data_start, Kernel::Memory::MemoryPermission::ReadAndWrite); } void LoadNro(Kernel::HLERequestContext& ctx) { @@ -317,9 +495,9 @@ public: return; } - NROHeader header; + // Load and validate the NRO header + NROHeader header{}; std::memcpy(&header, nro_data.data(), sizeof(NROHeader)); - if (!IsValidNRO(header, nro_size, bss_size)) { LOG_ERROR(Service_LDR, "NRO was invalid!"); IPC::ResponseBuilder rb{ctx, 2}; @@ -327,62 +505,50 @@ public: return; } - // Load NRO as new executable module - auto* process = system.CurrentProcess(); - auto& vm_manager = process->VMManager(); - auto map_address = vm_manager.FindFreeRegion(nro_size + bss_size); - - if (!map_address.Succeeded() || - *map_address + nro_size + bss_size > vm_manager.GetAddressSpaceEndAddress()) { - - LOG_ERROR(Service_LDR, - "General error while allocation memory or no available memory to allocate!"); + // Map memory for the NRO + const auto map_result{MapNro(system.CurrentProcess(), nro_address, nro_size, bss_address, + bss_size, nro_size + bss_size)}; + if (map_result.Failed()) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_MEMORY_STATE); - return; + rb.Push(map_result.Code()); } - // Mark text and read-only region as ModuleCode - ASSERT(vm_manager - .MirrorMemory(*map_address, nro_address, header.text_size + header.ro_size, - Kernel::MemoryState::ModuleCode) - .IsSuccess()); - // Mark read/write region as ModuleCodeData, which is necessary if this region is used for - // TransferMemory (e.g. Final Fantasy VIII Remastered does this) - ASSERT(vm_manager - .MirrorMemory(*map_address + header.rw_offset, nro_address + header.rw_offset, - header.rw_size, Kernel::MemoryState::ModuleCodeData) - .IsSuccess()); - // Revoke permissions from the old memory region - ASSERT(vm_manager.ReprotectRange(nro_address, nro_size, Kernel::VMAPermission::None) - .IsSuccess()); - - if (bss_size > 0) { - // Mark BSS region as ModuleCodeData, which is necessary if this region is used for - // TransferMemory (e.g. Final Fantasy VIII Remastered does this) - ASSERT(vm_manager - .MirrorMemory(*map_address + nro_size, bss_address, bss_size, - Kernel::MemoryState::ModuleCodeData) - .IsSuccess()); - ASSERT(vm_manager.ReprotectRange(bss_address, bss_size, Kernel::VMAPermission::None) - .IsSuccess()); + // Load the NRO into the mapped memory + if (const auto result{LoadNro(system.CurrentProcess(), header, nro_address, *map_result)}; + result.IsError()) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(map_result.Code()); } - vm_manager.ReprotectRange(*map_address, header.text_size, - Kernel::VMAPermission::ReadExecute); - vm_manager.ReprotectRange(*map_address + header.ro_offset, header.ro_size, - Kernel::VMAPermission::Read); - vm_manager.ReprotectRange(*map_address + header.rw_offset, header.rw_size, - Kernel::VMAPermission::ReadWrite); + // Track the loaded NRO + nro.insert_or_assign(*map_result, + NROInfo{hash, *map_result, nro_size, bss_address, bss_size, + header.segment_headers[TEXT_INDEX].memory_size, + header.segment_headers[RO_INDEX].memory_size, + header.segment_headers[DATA_INDEX].memory_size, nro_address}); + // Invalidate JIT caches for the newly mapped process code system.InvalidateCpuInstructionCaches(); - nro.insert_or_assign(*map_address, - NROInfo{hash, nro_address, nro_size, bss_address, bss_size}); - IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push(*map_address); + rb.Push(*map_result); + } + + ResultCode UnmapNro(const NROInfo& info) { + // Each region must be unmapped separately to validate memory state + auto& page_table{system.CurrentProcess()->PageTable()}; + CASCADE_CODE(page_table.UnmapProcessCodeMemory(info.nro_address + info.text_size + + info.ro_size + info.data_size, + info.bss_address, info.bss_size)); + CASCADE_CODE(page_table.UnmapProcessCodeMemory( + info.nro_address + info.text_size + info.ro_size, + info.src_addr + info.text_size + info.ro_size, info.data_size)); + CASCADE_CODE(page_table.UnmapProcessCodeMemory( + info.nro_address + info.text_size, info.src_addr + info.text_size, info.ro_size)); + CASCADE_CODE( + page_table.UnmapProcessCodeMemory(info.nro_address, info.src_addr, info.text_size)); + return RESULT_SUCCESS; } void UnloadNro(Kernel::HLERequestContext& ctx) { @@ -422,30 +588,15 @@ public: return; } - auto& vm_manager = system.CurrentProcess()->VMManager(); - const auto& nro_info = iter->second; - - // Unmap the mirrored memory - ASSERT( - vm_manager.UnmapRange(nro_address, nro_info.nro_size + nro_info.bss_size).IsSuccess()); - - // Reprotect the source memory - ASSERT(vm_manager - .ReprotectRange(nro_info.nro_address, nro_info.nro_size, - Kernel::VMAPermission::ReadWrite) - .IsSuccess()); - if (nro_info.bss_size > 0) { - ASSERT(vm_manager - .ReprotectRange(nro_info.bss_address, nro_info.bss_size, - Kernel::VMAPermission::ReadWrite) - .IsSuccess()); - } + const auto result{UnmapNro(iter->second)}; system.InvalidateCpuInstructionCaches(); nro.erase(iter); + IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); + + rb.Push(result); } void Initialize(Kernel::HLERequestContext& ctx) { @@ -458,56 +609,7 @@ public: } private: - using SHA256Hash = std::array<u8, 0x20>; - - struct NROHeader { - INSERT_PADDING_WORDS(1); - u32_le mod_offset; - INSERT_PADDING_WORDS(2); - u32_le magic; - u32_le version; - u32_le nro_size; - u32_le flags; - u32_le text_offset; - u32_le text_size; - u32_le ro_offset; - u32_le ro_size; - u32_le rw_offset; - u32_le rw_size; - u32_le bss_size; - INSERT_PADDING_WORDS(1); - std::array<u8, 0x20> build_id; - INSERT_PADDING_BYTES(0x20); - }; - static_assert(sizeof(NROHeader) == 0x80, "NROHeader has invalid size."); - - struct NRRHeader { - u32_le magic; - INSERT_PADDING_BYTES(12); - u64_le title_id_mask; - u64_le title_id_pattern; - INSERT_PADDING_BYTES(16); - std::array<u8, 0x100> modulus; - std::array<u8, 0x100> signature_1; - std::array<u8, 0x100> signature_2; - u64_le title_id; - u32_le size; - INSERT_PADDING_BYTES(4); - u32_le hash_offset; - u32_le hash_count; - INSERT_PADDING_BYTES(8); - }; - static_assert(sizeof(NRRHeader) == 0x350, "NRRHeader has incorrect size."); - - struct NROInfo { - SHA256Hash hash; - VAddr nro_address; - u64 nro_size; - VAddr bss_address; - u64 bss_size; - }; - - bool initialized = false; + bool initialized{}; std::map<VAddr, NROInfo> nro; std::map<VAddr, std::vector<SHA256Hash>> nrr; @@ -521,11 +623,21 @@ private: static bool IsValidNRO(const NROHeader& header, u64 nro_size, u64 bss_size) { return header.magic == Common::MakeMagic('N', 'R', 'O', '0') && header.nro_size == nro_size && header.bss_size == bss_size && - header.ro_offset == header.text_offset + header.text_size && - header.rw_offset == header.ro_offset + header.ro_size && - nro_size == header.rw_offset + header.rw_size && - Common::Is4KBAligned(header.text_size) && Common::Is4KBAligned(header.ro_size) && - Common::Is4KBAligned(header.rw_size); + + header.segment_headers[RO_INDEX].memory_offset == + header.segment_headers[TEXT_INDEX].memory_offset + + header.segment_headers[TEXT_INDEX].memory_size && + + header.segment_headers[DATA_INDEX].memory_offset == + header.segment_headers[RO_INDEX].memory_offset + + header.segment_headers[RO_INDEX].memory_size && + + nro_size == header.segment_headers[DATA_INDEX].memory_offset + + header.segment_headers[DATA_INDEX].memory_size && + + Common::Is4KBAligned(header.segment_headers[TEXT_INDEX].memory_size) && + Common::Is4KBAligned(header.segment_headers[RO_INDEX].memory_size) && + Common::Is4KBAligned(header.segment_headers[DATA_INDEX].memory_size); } Core::System& system; }; diff --git a/src/core/hle/service/lm/lm.cpp b/src/core/hle/service/lm/lm.cpp index 346c8f899..49a42a9c9 100644 --- a/src/core/hle/service/lm/lm.cpp +++ b/src/core/hle/service/lm/lm.cpp @@ -7,6 +7,7 @@ #include "common/logging/log.h" #include "common/scope_exit.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/service/lm/lm.h" #include "core/hle/service/lm/manager.h" @@ -17,7 +18,7 @@ namespace Service::LM { class ILogger final : public ServiceFramework<ILogger> { public: - explicit ILogger(Manager& manager_, Memory::Memory& memory_) + explicit ILogger(Manager& manager_, Core::Memory::Memory& memory_) : ServiceFramework("ILogger"), manager{manager_}, memory{memory_} { static const FunctionInfo functions[] = { {0, &ILogger::Log, "Log"}, @@ -75,12 +76,12 @@ private: } Manager& manager; - Memory::Memory& memory; + Core::Memory::Memory& memory; }; class LM final : public ServiceFramework<LM> { public: - explicit LM(Manager& manager_, Memory::Memory& memory_) + explicit LM(Manager& manager_, Core::Memory::Memory& memory_) : ServiceFramework{"lm"}, manager{manager_}, memory{memory_} { // clang-format off static const FunctionInfo functions[] = { @@ -101,7 +102,7 @@ private: } Manager& manager; - Memory::Memory& memory; + Core::Memory::Memory& memory; }; void InstallInterfaces(Core::System& system) { diff --git a/src/core/hle/service/lm/manager.cpp b/src/core/hle/service/lm/manager.cpp index b67081b86..3ee2374e7 100644 --- a/src/core/hle/service/lm/manager.cpp +++ b/src/core/hle/service/lm/manager.cpp @@ -86,7 +86,8 @@ std::string FormatField(Field type, const std::vector<u8>& data) { return Common::StringFromFixedZeroTerminatedBuffer( reinterpret_cast<const char*>(data.data()), data.size()); default: - UNIMPLEMENTED(); + UNIMPLEMENTED_MSG("Unimplemented field type={}", type); + return ""; } } diff --git a/src/core/hle/service/mig/mig.cpp b/src/core/hle/service/mig/mig.cpp index d16367f2c..113a4665c 100644 --- a/src/core/hle/service/mig/mig.cpp +++ b/src/core/hle/service/mig/mig.cpp @@ -20,6 +20,12 @@ public: {101, nullptr, "ResumeServer"}, {200, nullptr, "CreateClient"}, {201, nullptr, "ResumeClient"}, + {1001, nullptr, "Unknown1001"}, + {1010, nullptr, "Unknown1010"}, + {1100, nullptr, "Unknown1100"}, + {1101, nullptr, "Unknown1101"}, + {1200, nullptr, "Unknown1200"}, + {1201, nullptr, "Unknown1201"} }; // clang-format on diff --git a/src/core/hle/service/mii/manager.cpp b/src/core/hle/service/mii/manager.cpp new file mode 100644 index 000000000..d73b90015 --- /dev/null +++ b/src/core/hle/service/mii/manager.cpp @@ -0,0 +1,484 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> +#include <random> + +#include "common/assert.h" +#include "common/file_util.h" +#include "common/logging/log.h" +#include "common/string_util.h" + +#include "core/hle/service/acc/profile_manager.h" +#include "core/hle/service/mii/manager.h" +#include "core/hle/service/mii/raw_data.h" +#include "core/hle/service/mii/types.h" + +namespace Service::Mii { + +namespace { + +constexpr ResultCode ERROR_CANNOT_FIND_ENTRY{ErrorModule::Mii, 4}; + +constexpr std::size_t DefaultMiiCount{sizeof(RawData::DefaultMii) / sizeof(DefaultMii)}; + +constexpr MiiStoreData::Name DefaultMiiName{u'y', u'u', u'z', u'u'}; +constexpr std::array<u8, 8> HairColorLookup{8, 1, 2, 3, 4, 5, 6, 7}; +constexpr std::array<u8, 6> EyeColorLookup{8, 9, 10, 11, 12, 13}; +constexpr std::array<u8, 5> MouthColorLookup{19, 20, 21, 22, 23}; +constexpr std::array<u8, 7> GlassesColorLookup{8, 14, 15, 16, 17, 18, 0}; +constexpr std::array<u8, 62> EyeRotateLookup{ + {0x03, 0x04, 0x04, 0x04, 0x03, 0x04, 0x04, 0x04, 0x03, 0x04, 0x04, 0x04, 0x04, 0x03, 0x03, 0x04, + 0x04, 0x04, 0x03, 0x03, 0x04, 0x03, 0x04, 0x03, 0x03, 0x04, 0x03, 0x04, 0x04, 0x03, 0x04, 0x04, + 0x04, 0x03, 0x03, 0x03, 0x04, 0x04, 0x03, 0x03, 0x03, 0x04, 0x04, 0x03, 0x03, 0x03, 0x03, 0x03, + 0x03, 0x03, 0x03, 0x03, 0x04, 0x04, 0x04, 0x04, 0x03, 0x04, 0x04, 0x03, 0x04, 0x04}}; +constexpr std::array<u8, 24> EyebrowRotateLookup{{0x06, 0x06, 0x05, 0x07, 0x06, 0x07, 0x06, 0x07, + 0x04, 0x07, 0x06, 0x08, 0x05, 0x05, 0x06, 0x06, + 0x07, 0x07, 0x06, 0x06, 0x05, 0x06, 0x07, 0x05}}; + +template <typename T, std::size_t SourceArraySize, std::size_t DestArraySize> +std::array<T, DestArraySize> ResizeArray(const std::array<T, SourceArraySize>& in) { + std::array<T, DestArraySize> out{}; + std::memcpy(out.data(), in.data(), sizeof(T) * std::min(SourceArraySize, DestArraySize)); + return out; +} + +MiiInfo ConvertStoreDataToInfo(const MiiStoreData& data) { + MiiStoreBitFields bf; + std::memcpy(&bf, data.data.data.data(), sizeof(MiiStoreBitFields)); + + return { + .uuid = data.data.uuid, + .name = ResizeArray<char16_t, 10, 11>(data.data.name), + .font_region = static_cast<u8>(bf.font_region.Value()), + .favorite_color = static_cast<u8>(bf.favorite_color.Value()), + .gender = static_cast<u8>(bf.gender.Value()), + .height = static_cast<u8>(bf.height.Value()), + .build = static_cast<u8>(bf.build.Value()), + .type = static_cast<u8>(bf.type.Value()), + .region_move = static_cast<u8>(bf.region_move.Value()), + .faceline_type = static_cast<u8>(bf.faceline_type.Value()), + .faceline_color = static_cast<u8>(bf.faceline_color.Value()), + .faceline_wrinkle = static_cast<u8>(bf.faceline_wrinkle.Value()), + .faceline_make = static_cast<u8>(bf.faceline_makeup.Value()), + .hair_type = static_cast<u8>(bf.hair_type.Value()), + .hair_color = static_cast<u8>(bf.hair_color.Value()), + .hair_flip = static_cast<u8>(bf.hair_flip.Value()), + .eye_type = static_cast<u8>(bf.eye_type.Value()), + .eye_color = static_cast<u8>(bf.eye_color.Value()), + .eye_scale = static_cast<u8>(bf.eye_scale.Value()), + .eye_aspect = static_cast<u8>(bf.eye_aspect.Value()), + .eye_rotate = static_cast<u8>(bf.eye_rotate.Value()), + .eye_x = static_cast<u8>(bf.eye_x.Value()), + .eye_y = static_cast<u8>(bf.eye_y.Value()), + .eyebrow_type = static_cast<u8>(bf.eyebrow_type.Value()), + .eyebrow_color = static_cast<u8>(bf.eyebrow_color.Value()), + .eyebrow_scale = static_cast<u8>(bf.eyebrow_scale.Value()), + .eyebrow_aspect = static_cast<u8>(bf.eyebrow_aspect.Value()), + .eyebrow_rotate = static_cast<u8>(bf.eyebrow_rotate.Value()), + .eyebrow_x = static_cast<u8>(bf.eyebrow_x.Value()), + .eyebrow_y = static_cast<u8>(bf.eyebrow_y.Value() + 3), + .nose_type = static_cast<u8>(bf.nose_type.Value()), + .nose_scale = static_cast<u8>(bf.nose_scale.Value()), + .nose_y = static_cast<u8>(bf.nose_y.Value()), + .mouth_type = static_cast<u8>(bf.mouth_type.Value()), + .mouth_color = static_cast<u8>(bf.mouth_color.Value()), + .mouth_scale = static_cast<u8>(bf.mouth_scale.Value()), + .mouth_aspect = static_cast<u8>(bf.mouth_aspect.Value()), + .mouth_y = static_cast<u8>(bf.mouth_y.Value()), + .beard_color = static_cast<u8>(bf.beard_color.Value()), + .beard_type = static_cast<u8>(bf.beard_type.Value()), + .mustache_type = static_cast<u8>(bf.mustache_type.Value()), + .mustache_scale = static_cast<u8>(bf.mustache_scale.Value()), + .mustache_y = static_cast<u8>(bf.mustache_y.Value()), + .glasses_type = static_cast<u8>(bf.glasses_type.Value()), + .glasses_color = static_cast<u8>(bf.glasses_color.Value()), + .glasses_scale = static_cast<u8>(bf.glasses_scale.Value()), + .glasses_y = static_cast<u8>(bf.glasses_y.Value()), + .mole_type = static_cast<u8>(bf.mole_type.Value()), + .mole_scale = static_cast<u8>(bf.mole_scale.Value()), + .mole_x = static_cast<u8>(bf.mole_x.Value()), + .mole_y = static_cast<u8>(bf.mole_y.Value()), + }; +} + +u16 GenerateCrc16(const void* data, std::size_t size) { + s32 crc{}; + for (std::size_t i = 0; i < size; i++) { + crc ^= static_cast<const u8*>(data)[i] << 8; + for (std::size_t j = 0; j < 8; j++) { + crc <<= 1; + if ((crc & 0x10000) != 0) { + crc = (crc ^ 0x1021) & 0xFFFF; + } + } + } + return Common::swap16(static_cast<u16>(crc)); +} + +Common::UUID GenerateValidUUID() { + auto uuid{Common::UUID::Generate()}; + + // Bit 7 must be set, and bit 6 unset for the UUID to be valid + uuid.uuid[1] &= 0xFFFFFFFFFFFFFF3FULL; + uuid.uuid[1] |= 0x0000000000000080ULL; + + return uuid; +} + +template <typename T> +T GetRandomValue(T min, T max) { + std::random_device device; + std::mt19937 gen(device()); + std::uniform_int_distribution<u64> distribution(static_cast<u64>(min), static_cast<u64>(max)); + return static_cast<T>(distribution(gen)); +} + +template <typename T> +T GetRandomValue(T max) { + return GetRandomValue<T>({}, max); +} + +template <typename T> +T GetArrayValue(const u8* data, std::size_t index) { + T result{}; + std::memcpy(&result, &data[index * sizeof(T)], sizeof(T)); + return result; +} + +MiiStoreData BuildRandomStoreData(Age age, Gender gender, Race race, const Common::UUID& user_id) { + MiiStoreBitFields bf{}; + + if (gender == Gender::All) { + gender = GetRandomValue<Gender>(Gender::Maximum); + } + + bf.gender.Assign(gender); + bf.favorite_color.Assign(GetRandomValue<u8>(11)); + bf.region_move.Assign(0); + bf.font_region.Assign(FontRegion::Standard); + bf.type.Assign(0); + bf.height.Assign(64); + bf.build.Assign(64); + + if (age == Age::All) { + const auto temp{GetRandomValue<int>(10)}; + if (temp >= 8) { + age = Age::Old; + } else if (temp >= 4) { + age = Age::Normal; + } else { + age = Age::Young; + } + } + + if (race == Race::All) { + const auto temp{GetRandomValue<int>(10)}; + if (temp >= 8) { + race = Race::Black; + } else if (temp >= 4) { + race = Race::White; + } else { + race = Race::Asian; + } + } + + u32 axis_y{}; + if (gender == Gender::Female && age == Age::Young) { + axis_y = GetRandomValue<u32>(3); + } + + const std::size_t index{3 * static_cast<std::size_t>(age) + + 9 * static_cast<std::size_t>(gender) + static_cast<std::size_t>(race)}; + + const auto faceline_type_info{ + GetArrayValue<RandomMiiData4>(&RawData::RandomMiiFaceline[0], index)}; + const auto faceline_color_info{GetArrayValue<RandomMiiData3>( + RawData::RandomMiiFacelineColor.data(), + 3 * static_cast<std::size_t>(gender) + static_cast<std::size_t>(race))}; + const auto faceline_wrinkle_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiFacelineWrinkle.data(), index)}; + const auto faceline_makeup_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiFacelineMakeup.data(), index)}; + const auto hair_type_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiHairType.data(), index)}; + const auto hair_color_info{GetArrayValue<RandomMiiData3>(RawData::RandomMiiHairColor.data(), + 3 * static_cast<std::size_t>(race) + + static_cast<std::size_t>(age))}; + const auto eye_type_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiEyeType.data(), index)}; + const auto eye_color_info{GetArrayValue<RandomMiiData2>(RawData::RandomMiiEyeColor.data(), + static_cast<std::size_t>(race))}; + const auto eyebrow_type_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiEyebrowType.data(), index)}; + const auto nose_type_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiNoseType.data(), index)}; + const auto mouth_type_info{ + GetArrayValue<RandomMiiData4>(RawData::RandomMiiMouthType.data(), index)}; + const auto glasses_type_info{GetArrayValue<RandomMiiData2>(RawData::RandomMiiGlassType.data(), + static_cast<std::size_t>(age))}; + + bf.faceline_type.Assign( + faceline_type_info.values[GetRandomValue<std::size_t>(faceline_type_info.values_count)]); + bf.faceline_color.Assign( + faceline_color_info.values[GetRandomValue<std::size_t>(faceline_color_info.values_count)]); + bf.faceline_wrinkle.Assign( + faceline_wrinkle_info + .values[GetRandomValue<std::size_t>(faceline_wrinkle_info.values_count)]); + bf.faceline_makeup.Assign( + faceline_makeup_info + .values[GetRandomValue<std::size_t>(faceline_makeup_info.values_count)]); + + bf.hair_type.Assign( + hair_type_info.values[GetRandomValue<std::size_t>(hair_type_info.values_count)]); + bf.hair_color.Assign( + HairColorLookup[hair_color_info + .values[GetRandomValue<std::size_t>(hair_color_info.values_count)]]); + bf.hair_flip.Assign(GetRandomValue<HairFlip>(HairFlip::Maximum)); + + bf.eye_type.Assign( + eye_type_info.values[GetRandomValue<std::size_t>(eye_type_info.values_count)]); + + const auto eye_rotate_1{gender != Gender::Male ? 4 : 2}; + const auto eye_rotate_2{gender != Gender::Male ? 3 : 4}; + const auto eye_rotate_offset{32 - EyeRotateLookup[eye_rotate_1] + eye_rotate_2}; + const auto eye_rotate{32 - EyeRotateLookup[bf.eye_type]}; + + bf.eye_color.Assign( + EyeColorLookup[eye_color_info + .values[GetRandomValue<std::size_t>(eye_color_info.values_count)]]); + bf.eye_scale.Assign(4); + bf.eye_aspect.Assign(3); + bf.eye_rotate.Assign(eye_rotate_offset - eye_rotate); + bf.eye_x.Assign(2); + bf.eye_y.Assign(axis_y + 12); + + bf.eyebrow_type.Assign( + eyebrow_type_info.values[GetRandomValue<std::size_t>(eyebrow_type_info.values_count)]); + + const auto eyebrow_rotate_1{race == Race::Asian ? 6 : 0}; + const auto eyebrow_y{race == Race::Asian ? 9 : 10}; + const auto eyebrow_rotate_offset{32 - EyebrowRotateLookup[eyebrow_rotate_1] + 6}; + const auto eyebrow_rotate{ + 32 - EyebrowRotateLookup[static_cast<std::size_t>(bf.eyebrow_type.Value())]}; + + bf.eyebrow_color.Assign(bf.hair_color); + bf.eyebrow_scale.Assign(4); + bf.eyebrow_aspect.Assign(3); + bf.eyebrow_rotate.Assign(eyebrow_rotate_offset - eyebrow_rotate); + bf.eyebrow_x.Assign(2); + bf.eyebrow_y.Assign(axis_y + eyebrow_y); + + const auto nose_scale{gender == Gender::Female ? 3 : 4}; + + bf.nose_type.Assign( + nose_type_info.values[GetRandomValue<std::size_t>(nose_type_info.values_count)]); + bf.nose_scale.Assign(nose_scale); + bf.nose_y.Assign(axis_y + 9); + + const auto mouth_color{gender == Gender::Female ? GetRandomValue<int>(4) : 0}; + + bf.mouth_type.Assign( + mouth_type_info.values[GetRandomValue<std::size_t>(mouth_type_info.values_count)]); + bf.mouth_color.Assign(MouthColorLookup[mouth_color]); + bf.mouth_scale.Assign(4); + bf.mouth_aspect.Assign(3); + bf.mouth_y.Assign(axis_y + 13); + + bf.beard_color.Assign(bf.hair_color); + bf.mustache_scale.Assign(4); + + if (gender == Gender::Male && age != Age::Young && GetRandomValue<int>(10) < 2) { + const auto mustache_and_beard_flag{ + GetRandomValue<BeardAndMustacheFlag>(BeardAndMustacheFlag::All)}; + + auto beard_type{BeardType::None}; + auto mustache_type{MustacheType::None}; + + if ((mustache_and_beard_flag & BeardAndMustacheFlag::Beard) == + BeardAndMustacheFlag::Beard) { + beard_type = GetRandomValue<BeardType>(BeardType::Beard1, BeardType::Beard5); + } + + if ((mustache_and_beard_flag & BeardAndMustacheFlag::Mustache) == + BeardAndMustacheFlag::Mustache) { + mustache_type = + GetRandomValue<MustacheType>(MustacheType::Mustache1, MustacheType::Mustache5); + } + + bf.mustache_type.Assign(mustache_type); + bf.beard_type.Assign(beard_type); + bf.mustache_y.Assign(10); + } else { + bf.mustache_type.Assign(MustacheType::None); + bf.beard_type.Assign(BeardType::None); + bf.mustache_y.Assign(axis_y + 10); + } + + const auto glasses_type_start{GetRandomValue<std::size_t>(100)}; + u8 glasses_type{}; + while (glasses_type_start < glasses_type_info.values[glasses_type]) { + if (++glasses_type >= glasses_type_info.values_count) { + UNREACHABLE(); + break; + } + } + + bf.glasses_type.Assign(glasses_type); + bf.glasses_color.Assign(GlassesColorLookup[0]); + bf.glasses_scale.Assign(4); + bf.glasses_y.Assign(axis_y + 10); + + bf.mole_type.Assign(0); + bf.mole_scale.Assign(4); + bf.mole_x.Assign(2); + bf.mole_y.Assign(20); + + return {DefaultMiiName, bf, user_id}; +} + +MiiStoreData BuildDefaultStoreData(const DefaultMii& info, const Common::UUID& user_id) { + MiiStoreBitFields bf{}; + + bf.font_region.Assign(info.font_region); + bf.favorite_color.Assign(info.favorite_color); + bf.gender.Assign(info.gender); + bf.height.Assign(info.height); + bf.build.Assign(info.weight); + bf.type.Assign(info.type); + bf.region_move.Assign(info.region); + bf.faceline_type.Assign(info.face_type); + bf.faceline_color.Assign(info.face_color); + bf.faceline_wrinkle.Assign(info.face_wrinkle); + bf.faceline_makeup.Assign(info.face_makeup); + bf.hair_type.Assign(info.hair_type); + bf.hair_color.Assign(HairColorLookup[info.hair_color]); + bf.hair_flip.Assign(static_cast<HairFlip>(info.hair_flip)); + bf.eye_type.Assign(info.eye_type); + bf.eye_color.Assign(EyeColorLookup[info.eye_color]); + bf.eye_scale.Assign(info.eye_scale); + bf.eye_aspect.Assign(info.eye_aspect); + bf.eye_rotate.Assign(info.eye_rotate); + bf.eye_x.Assign(info.eye_x); + bf.eye_y.Assign(info.eye_y); + bf.eyebrow_type.Assign(info.eyebrow_type); + bf.eyebrow_color.Assign(HairColorLookup[info.eyebrow_color]); + bf.eyebrow_scale.Assign(info.eyebrow_scale); + bf.eyebrow_aspect.Assign(info.eyebrow_aspect); + bf.eyebrow_rotate.Assign(info.eyebrow_rotate); + bf.eyebrow_x.Assign(info.eyebrow_x); + bf.eyebrow_y.Assign(info.eyebrow_y - 3); + bf.nose_type.Assign(info.nose_type); + bf.nose_scale.Assign(info.nose_scale); + bf.nose_y.Assign(info.nose_y); + bf.mouth_type.Assign(info.mouth_type); + bf.mouth_color.Assign(MouthColorLookup[info.mouth_color]); + bf.mouth_scale.Assign(info.mouth_scale); + bf.mouth_aspect.Assign(info.mouth_aspect); + bf.mouth_y.Assign(info.mouth_y); + bf.beard_color.Assign(HairColorLookup[info.beard_color]); + bf.beard_type.Assign(static_cast<BeardType>(info.beard_type)); + bf.mustache_type.Assign(static_cast<MustacheType>(info.mustache_type)); + bf.mustache_scale.Assign(info.mustache_scale); + bf.mustache_y.Assign(info.mustache_y); + bf.glasses_type.Assign(info.glasses_type); + bf.glasses_color.Assign(GlassesColorLookup[info.glasses_color]); + bf.glasses_scale.Assign(info.glasses_scale); + bf.glasses_y.Assign(info.glasses_y); + bf.mole_type.Assign(info.mole_type); + bf.mole_scale.Assign(info.mole_scale); + bf.mole_x.Assign(info.mole_x); + bf.mole_y.Assign(info.mole_y); + + return {DefaultMiiName, bf, user_id}; +} + +} // namespace + +MiiStoreData::MiiStoreData() = default; + +MiiStoreData::MiiStoreData(const MiiStoreData::Name& name, const MiiStoreBitFields& bit_fields, + const Common::UUID& user_id) { + data.name = name; + data.uuid = GenerateValidUUID(); + + std::memcpy(data.data.data(), &bit_fields, sizeof(MiiStoreBitFields)); + data_crc = GenerateCrc16(data.data.data(), sizeof(data)); + device_crc = GenerateCrc16(&user_id, sizeof(Common::UUID)); +} + +MiiManager::MiiManager() : user_id{Service::Account::ProfileManager().GetLastOpenedUser()} {} + +bool MiiManager::CheckAndResetUpdateCounter(SourceFlag source_flag, u64& current_update_counter) { + if ((source_flag & SourceFlag::Database) == SourceFlag::None) { + return false; + } + + const bool result{current_update_counter != update_counter}; + + current_update_counter = update_counter; + + return result; +} + +bool MiiManager::IsFullDatabase() const { + // TODO(bunnei): We don't implement the Mii database, so it cannot be full + return false; +} + +u32 MiiManager::GetCount(SourceFlag source_flag) const { + std::size_t count{}; + if ((source_flag & SourceFlag::Database) != SourceFlag::None) { + // TODO(bunnei): We don't implement the Mii database, but when we do, update this + count += 0; + } + if ((source_flag & SourceFlag::Default) != SourceFlag::None) { + count += DefaultMiiCount; + } + return static_cast<u32>(count); +} + +ResultVal<MiiInfo> MiiManager::UpdateLatest([[maybe_unused]] const MiiInfo& info, + SourceFlag source_flag) { + if ((source_flag & SourceFlag::Database) == SourceFlag::None) { + return ERROR_CANNOT_FIND_ENTRY; + } + + // TODO(bunnei): We don't implement the Mii database, so we can't have an entry + return ERROR_CANNOT_FIND_ENTRY; +} + +MiiInfo MiiManager::BuildRandom(Age age, Gender gender, Race race) { + return ConvertStoreDataToInfo(BuildRandomStoreData(age, gender, race, user_id)); +} + +MiiInfo MiiManager::BuildDefault(std::size_t index) { + return ConvertStoreDataToInfo(BuildDefaultStoreData( + GetArrayValue<DefaultMii>(RawData::DefaultMii.data(), index), user_id)); +} + +ResultVal<std::vector<MiiInfoElement>> MiiManager::GetDefault(SourceFlag source_flag) { + std::vector<MiiInfoElement> result; + + if ((source_flag & SourceFlag::Default) == SourceFlag::None) { + return MakeResult(std::move(result)); + } + + for (std::size_t index = 0; index < DefaultMiiCount; index++) { + result.emplace_back(BuildDefault(index), Source::Default); + } + + return MakeResult(std::move(result)); +} + +ResultCode MiiManager::GetIndex([[maybe_unused]] const MiiInfo& info, u32& index) { + constexpr u32 INVALID_INDEX{0xFFFFFFFF}; + + index = INVALID_INDEX; + + // TODO(bunnei): We don't implement the Mii database, so we can't have an index + return ERROR_CANNOT_FIND_ENTRY; +} + +} // namespace Service::Mii diff --git a/src/core/hle/service/mii/manager.h b/src/core/hle/service/mii/manager.h new file mode 100644 index 000000000..927451dea --- /dev/null +++ b/src/core/hle/service/mii/manager.h @@ -0,0 +1,331 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/bit_field.h" +#include "common/common_funcs.h" +#include "common/uuid.h" +#include "core/hle/result.h" +#include "core/hle/service/mii/types.h" + +namespace Service::Mii { + +enum class Source : u32 { + Database = 0, + Default = 1, + Account = 2, + Friend = 3, +}; + +enum class SourceFlag : u32 { + None = 0, + Database = 1 << 0, + Default = 1 << 1, +}; +DECLARE_ENUM_FLAG_OPERATORS(SourceFlag); + +struct MiiInfo { + Common::UUID uuid{Common::INVALID_UUID}; + std::array<char16_t, 11> name{}; + u8 font_region{}; + u8 favorite_color{}; + u8 gender{}; + u8 height{}; + u8 build{}; + u8 type{}; + u8 region_move{}; + u8 faceline_type{}; + u8 faceline_color{}; + u8 faceline_wrinkle{}; + u8 faceline_make{}; + u8 hair_type{}; + u8 hair_color{}; + u8 hair_flip{}; + u8 eye_type{}; + u8 eye_color{}; + u8 eye_scale{}; + u8 eye_aspect{}; + u8 eye_rotate{}; + u8 eye_x{}; + u8 eye_y{}; + u8 eyebrow_type{}; + u8 eyebrow_color{}; + u8 eyebrow_scale{}; + u8 eyebrow_aspect{}; + u8 eyebrow_rotate{}; + u8 eyebrow_x{}; + u8 eyebrow_y{}; + u8 nose_type{}; + u8 nose_scale{}; + u8 nose_y{}; + u8 mouth_type{}; + u8 mouth_color{}; + u8 mouth_scale{}; + u8 mouth_aspect{}; + u8 mouth_y{}; + u8 beard_color{}; + u8 beard_type{}; + u8 mustache_type{}; + u8 mustache_scale{}; + u8 mustache_y{}; + u8 glasses_type{}; + u8 glasses_color{}; + u8 glasses_scale{}; + u8 glasses_y{}; + u8 mole_type{}; + u8 mole_scale{}; + u8 mole_x{}; + u8 mole_y{}; + INSERT_PADDING_BYTES(1); + + std::u16string Name() const; +}; +static_assert(sizeof(MiiInfo) == 0x58, "MiiInfo has incorrect size."); +static_assert(std::has_unique_object_representations_v<MiiInfo>, + "All bits of MiiInfo must contribute to its value."); + +#pragma pack(push, 4) + +struct MiiInfoElement { + MiiInfoElement(const MiiInfo& info, Source source) : info{info}, source{source} {} + + MiiInfo info{}; + Source source{}; +}; +static_assert(sizeof(MiiInfoElement) == 0x5c, "MiiInfoElement has incorrect size."); + +struct MiiStoreBitFields { + union { + u32 word_0{}; + + BitField<0, 8, u32> hair_type; + BitField<8, 7, u32> height; + BitField<15, 1, u32> mole_type; + BitField<16, 7, u32> build; + BitField<23, 1, HairFlip> hair_flip; + BitField<24, 7, u32> hair_color; + BitField<31, 1, u32> type; + }; + + union { + u32 word_1{}; + + BitField<0, 7, u32> eye_color; + BitField<7, 1, Gender> gender; + BitField<8, 7, u32> eyebrow_color; + BitField<16, 7, u32> mouth_color; + BitField<24, 7, u32> beard_color; + }; + + union { + u32 word_2{}; + + BitField<0, 7, u32> glasses_color; + BitField<8, 6, u32> eye_type; + BitField<14, 2, u32> region_move; + BitField<16, 6, u32> mouth_type; + BitField<22, 2, FontRegion> font_region; + BitField<24, 5, u32> eye_y; + BitField<29, 3, u32> glasses_scale; + }; + + union { + u32 word_3{}; + + BitField<0, 5, u32> eyebrow_type; + BitField<5, 3, MustacheType> mustache_type; + BitField<8, 5, u32> nose_type; + BitField<13, 3, BeardType> beard_type; + BitField<16, 5, u32> nose_y; + BitField<21, 3, u32> mouth_aspect; + BitField<24, 5, u32> mouth_y; + BitField<29, 3, u32> eyebrow_aspect; + }; + + union { + u32 word_4{}; + + BitField<0, 5, u32> mustache_y; + BitField<5, 3, u32> eye_rotate; + BitField<8, 5, u32> glasses_y; + BitField<13, 3, u32> eye_aspect; + BitField<16, 5, u32> mole_x; + BitField<21, 3, u32> eye_scale; + BitField<24, 5, u32> mole_y; + }; + + union { + u32 word_5{}; + + BitField<0, 5, u32> glasses_type; + BitField<8, 4, u32> favorite_color; + BitField<12, 4, u32> faceline_type; + BitField<16, 4, u32> faceline_color; + BitField<20, 4, u32> faceline_wrinkle; + BitField<24, 4, u32> faceline_makeup; + BitField<28, 4, u32> eye_x; + }; + + union { + u32 word_6{}; + + BitField<0, 4, u32> eyebrow_scale; + BitField<4, 4, u32> eyebrow_rotate; + BitField<8, 4, u32> eyebrow_x; + BitField<12, 4, u32> eyebrow_y; + BitField<16, 4, u32> nose_scale; + BitField<20, 4, u32> mouth_scale; + BitField<24, 4, u32> mustache_scale; + BitField<28, 4, u32> mole_scale; + }; +}; +static_assert(sizeof(MiiStoreBitFields) == 0x1c, "MiiStoreBitFields has incorrect size."); +static_assert(std::is_trivially_copyable_v<MiiStoreBitFields>, + "MiiStoreBitFields is not trivially copyable."); + +struct MiiStoreData { + using Name = std::array<char16_t, 10>; + + MiiStoreData(); + MiiStoreData(const Name& name, const MiiStoreBitFields& bit_fields, + const Common::UUID& user_id); + + // This corresponds to the above structure MiiStoreBitFields. I did it like this because the + // BitField<> type makes this (and any thing that contains it) not trivially copyable, which is + // not suitable for our uses. + struct { + std::array<u8, 0x1C> data{}; + static_assert(sizeof(MiiStoreBitFields) == sizeof(data), "data field has incorrect size."); + + Name name{}; + Common::UUID uuid{Common::INVALID_UUID}; + } data; + + u16 data_crc{}; + u16 device_crc{}; +}; +static_assert(sizeof(MiiStoreData) == 0x44, "MiiStoreData has incorrect size."); + +struct MiiStoreDataElement { + MiiStoreData data{}; + Source source{}; +}; +static_assert(sizeof(MiiStoreDataElement) == 0x48, "MiiStoreDataElement has incorrect size."); + +struct MiiDatabase { + u32 magic{}; // 'NFDB' + std::array<MiiStoreData, 0x64> miis{}; + INSERT_PADDING_BYTES(1); + u8 count{}; + u16 crc{}; +}; +static_assert(sizeof(MiiDatabase) == 0x1A98, "MiiDatabase has incorrect size."); + +struct RandomMiiValues { + std::array<u8, 0xbc> values{}; +}; +static_assert(sizeof(RandomMiiValues) == 0xbc, "RandomMiiValues has incorrect size."); + +struct RandomMiiData4 { + Gender gender{}; + Age age{}; + Race race{}; + u32 values_count{}; + std::array<u8, 0xbc> values{}; +}; +static_assert(sizeof(RandomMiiData4) == 0xcc, "RandomMiiData4 has incorrect size."); + +struct RandomMiiData3 { + u32 arg_1; + u32 arg_2; + u32 values_count; + std::array<u8, 0xbc> values{}; +}; +static_assert(sizeof(RandomMiiData3) == 0xc8, "RandomMiiData3 has incorrect size."); + +struct RandomMiiData2 { + u32 arg_1; + u32 values_count; + std::array<u8, 0xbc> values{}; +}; +static_assert(sizeof(RandomMiiData2) == 0xc4, "RandomMiiData2 has incorrect size."); + +struct DefaultMii { + u32 face_type{}; + u32 face_color{}; + u32 face_wrinkle{}; + u32 face_makeup{}; + u32 hair_type{}; + u32 hair_color{}; + u32 hair_flip{}; + u32 eye_type{}; + u32 eye_color{}; + u32 eye_scale{}; + u32 eye_aspect{}; + u32 eye_rotate{}; + u32 eye_x{}; + u32 eye_y{}; + u32 eyebrow_type{}; + u32 eyebrow_color{}; + u32 eyebrow_scale{}; + u32 eyebrow_aspect{}; + u32 eyebrow_rotate{}; + u32 eyebrow_x{}; + u32 eyebrow_y{}; + u32 nose_type{}; + u32 nose_scale{}; + u32 nose_y{}; + u32 mouth_type{}; + u32 mouth_color{}; + u32 mouth_scale{}; + u32 mouth_aspect{}; + u32 mouth_y{}; + u32 mustache_type{}; + u32 beard_type{}; + u32 beard_color{}; + u32 mustache_scale{}; + u32 mustache_y{}; + u32 glasses_type{}; + u32 glasses_color{}; + u32 glasses_scale{}; + u32 glasses_y{}; + u32 mole_type{}; + u32 mole_scale{}; + u32 mole_x{}; + u32 mole_y{}; + u32 height{}; + u32 weight{}; + Gender gender{}; + u32 favorite_color{}; + u32 region{}; + FontRegion font_region{}; + u32 type{}; + INSERT_PADDING_WORDS(5); +}; +static_assert(sizeof(DefaultMii) == 0xd8, "MiiStoreData has incorrect size."); + +#pragma pack(pop) + +// The Mii manager is responsible for loading and storing the Miis to the database in NAND along +// with providing an easy interface for HLE emulation of the mii service. +class MiiManager { +public: + MiiManager(); + + bool CheckAndResetUpdateCounter(SourceFlag source_flag, u64& current_update_counter); + bool IsFullDatabase() const; + u32 GetCount(SourceFlag source_flag) const; + ResultVal<MiiInfo> UpdateLatest(const MiiInfo& info, SourceFlag source_flag); + MiiInfo BuildRandom(Age age, Gender gender, Race race); + MiiInfo BuildDefault(std::size_t index); + ResultVal<std::vector<MiiInfoElement>> GetDefault(SourceFlag source_flag); + ResultCode GetIndex(const MiiInfo& info, u32& index); + +private: + const Common::UUID user_id; + u64 update_counter{}; +}; + +}; // namespace Service::Mii diff --git a/src/core/hle/service/mii/mii.cpp b/src/core/hle/service/mii/mii.cpp index a128edb43..d7080b715 100644 --- a/src/core/hle/service/mii/mii.cpp +++ b/src/core/hle/service/mii/mii.cpp @@ -4,22 +4,17 @@ #include <memory> -#include <fmt/ostream.h> - #include "common/logging/log.h" -#include "common/string_util.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" +#include "core/hle/service/mii/manager.h" #include "core/hle/service/mii/mii.h" -#include "core/hle/service/mii/mii_manager.h" #include "core/hle/service/service.h" #include "core/hle/service/sm/sm.h" namespace Service::Mii { constexpr ResultCode ERROR_INVALID_ARGUMENT{ErrorModule::Mii, 1}; -constexpr ResultCode ERROR_CANNOT_FIND_ENTRY{ErrorModule::Mii, 4}; -constexpr ResultCode ERROR_NOT_IN_TEST_MODE{ErrorModule::Mii, 99}; class IDatabaseService final : public ServiceFramework<IDatabaseService> { public: @@ -31,19 +26,19 @@ public: {2, &IDatabaseService::GetCount, "GetCount"}, {3, &IDatabaseService::Get, "Get"}, {4, &IDatabaseService::Get1, "Get1"}, - {5, nullptr, "UpdateLatest"}, + {5, &IDatabaseService::UpdateLatest, "UpdateLatest"}, {6, &IDatabaseService::BuildRandom, "BuildRandom"}, {7, &IDatabaseService::BuildDefault, "BuildDefault"}, - {8, &IDatabaseService::Get2, "Get2"}, - {9, &IDatabaseService::Get3, "Get3"}, + {8, nullptr, "Get2"}, + {9, nullptr, "Get3"}, {10, nullptr, "UpdateLatest1"}, - {11, &IDatabaseService::FindIndex, "FindIndex"}, - {12, &IDatabaseService::Move, "Move"}, - {13, &IDatabaseService::AddOrReplace, "AddOrReplace"}, - {14, &IDatabaseService::Delete, "Delete"}, - {15, &IDatabaseService::DestroyFile, "DestroyFile"}, - {16, &IDatabaseService::DeleteFile, "DeleteFile"}, - {17, &IDatabaseService::Format, "Format"}, + {11, nullptr, "FindIndex"}, + {12, nullptr, "Move"}, + {13, nullptr, "AddOrReplace"}, + {14, nullptr, "Delete"}, + {15, nullptr, "DestroyFile"}, + {16, nullptr, "DeleteFile"}, + {17, nullptr, "Format"}, {18, nullptr, "Import"}, {19, nullptr, "Export"}, {20, nullptr, "IsBrokenDatabaseWithClearFlag"}, @@ -52,6 +47,7 @@ public: {23, nullptr, "Convert"}, {24, nullptr, "ConvertCoreDataToCharInfo"}, {25, nullptr, "ConvertCharInfoToCoreData"}, + {26, nullptr, "Append"}, }; // clang-format on @@ -59,31 +55,26 @@ public: } private: - template <typename OutType> - std::vector<u8> SerializeArray(OutType (MiiManager::*getter)(u32) const, u32 offset, - u32 requested_size, u32& read_size) { - read_size = std::min(requested_size, db.Size() - offset); - - std::vector<u8> out(read_size * sizeof(OutType)); - - for (u32 i = 0; i < read_size; ++i) { - const auto obj = (db.*getter)(offset + i); - std::memcpy(out.data() + i * sizeof(OutType), &obj, sizeof(OutType)); + template <typename T> + std::vector<u8> SerializeArray(const std::vector<T>& values) { + std::vector<u8> out(values.size() * sizeof(T)); + std::size_t offset{}; + for (const auto& value : values) { + std::memcpy(out.data() + offset, &value, sizeof(T)); + offset += sizeof(T); } - return out; } void IsUpdated(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto source{rp.PopRaw<Source>()}; + const auto source_flag{rp.PopRaw<SourceFlag>()}; - LOG_DEBUG(Service_Mii, "called with source={}", source); + LOG_DEBUG(Service_Mii, "called with source_flag={}", source_flag); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push(db.CheckUpdatedFlag()); - db.ResetUpdatedFlag(); + rb.Push(manager.CheckAndResetUpdateCounter(source_flag, current_update_counter)); } void IsFullDatabase(Kernel::HLERequestContext& ctx) { @@ -91,264 +82,149 @@ private: IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push(db.Full()); + rb.Push(manager.IsFullDatabase()); } void GetCount(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto source{rp.PopRaw<Source>()}; + const auto source_flag{rp.PopRaw<SourceFlag>()}; - LOG_DEBUG(Service_Mii, "called with source={}", source); + LOG_DEBUG(Service_Mii, "called with source_flag={}", source_flag); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(db.Size()); + rb.Push<u32>(manager.GetCount(source_flag)); } - // Gets Miis from database at offset and index in format MiiInfoElement void Get(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto size{rp.PopRaw<u32>()}; - const auto source{rp.PopRaw<Source>()}; - - LOG_DEBUG(Service_Mii, "called with size={:08X}, offset={:08X}, source={}", size, - offsets[0], source); - - u32 read_size{}; - ctx.WriteBuffer(SerializeArray(&MiiManager::GetInfoElement, offsets[0], size, read_size)); - offsets[0] += read_size; - - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(read_size); - } + const auto source_flag{rp.PopRaw<SourceFlag>()}; - // Gets Miis from database at offset and index in format MiiInfo - void Get1(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto size{rp.PopRaw<u32>()}; - const auto source{rp.PopRaw<Source>()}; + LOG_DEBUG(Service_Mii, "called with source_flag={}", source_flag); - LOG_DEBUG(Service_Mii, "called with size={:08X}, offset={:08X}, source={}", size, - offsets[1], source); + const auto result{manager.GetDefault(source_flag)}; + if (result.Failed()) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(result.Code()); + return; + } - u32 read_size{}; - ctx.WriteBuffer(SerializeArray(&MiiManager::GetInfo, offsets[1], size, read_size)); - offsets[1] += read_size; + if (result->size() > 0) { + ctx.WriteBuffer(SerializeArray(*result)); + } IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(read_size); + rb.Push<u32>(static_cast<u32>(result->size())); } - void BuildRandom(Kernel::HLERequestContext& ctx) { + void Get1(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto [unknown1, unknown2, unknown3] = rp.PopRaw<RandomParameters>(); + const auto source_flag{rp.PopRaw<SourceFlag>()}; - if (unknown1 > 3) { - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_ARGUMENT); - LOG_ERROR(Service_Mii, "Invalid unknown1 value: {}", unknown1); - return; - } + LOG_DEBUG(Service_Mii, "called with source_flag={}", source_flag); - if (unknown2 > 2) { + const auto result{manager.GetDefault(source_flag)}; + if (result.Failed()) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_ARGUMENT); - LOG_ERROR(Service_Mii, "Invalid unknown2 value: {}", unknown2); + rb.Push(result.Code()); return; } - if (unknown3 > 3) { - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_ARGUMENT); - LOG_ERROR(Service_Mii, "Invalid unknown3 value: {}", unknown3); - return; + std::vector<MiiInfo> values; + for (const auto& element : *result) { + values.emplace_back(element.info); } - LOG_DEBUG(Service_Mii, "called with param_1={:08X}, param_2={:08X}, param_3={:08X}", - unknown1, unknown2, unknown3); + ctx.WriteBuffer(SerializeArray(values)); - const auto info = db.CreateRandom({unknown1, unknown2, unknown3}); - IPC::ResponseBuilder rb{ctx, 2 + sizeof(MiiInfo) / sizeof(u32)}; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<MiiInfo>(info); + rb.Push<u32>(static_cast<u32>(result->size())); } - void BuildDefault(Kernel::HLERequestContext& ctx) { + void UpdateLatest(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto index{rp.PopRaw<u32>()}; + const auto info{rp.PopRaw<MiiInfo>()}; + const auto source_flag{rp.PopRaw<SourceFlag>()}; - if (index > 5) { - LOG_ERROR(Service_Mii, "invalid argument, index cannot be greater than 5 but is {:08X}", - index); + LOG_DEBUG(Service_Mii, "called with source_flag={}", source_flag); + + const auto result{manager.UpdateLatest(info, source_flag)}; + if (result.Failed()) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_INVALID_ARGUMENT); + rb.Push(result.Code()); return; } - LOG_DEBUG(Service_Mii, "called with index={:08X}", index); - - const auto info = db.CreateDefault(index); IPC::ResponseBuilder rb{ctx, 2 + sizeof(MiiInfo) / sizeof(u32)}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<MiiInfo>(info); + rb.PushRaw<MiiInfo>(*result); } - // Gets Miis from database at offset and index in format MiiStoreDataElement - void Get2(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto size{rp.PopRaw<u32>()}; - const auto source{rp.PopRaw<Source>()}; - - LOG_DEBUG(Service_Mii, "called with size={:08X}, offset={:08X}, source={}", size, - offsets[2], source); - - u32 read_size{}; - ctx.WriteBuffer( - SerializeArray(&MiiManager::GetStoreDataElement, offsets[2], size, read_size)); - offsets[2] += read_size; - - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(read_size); - } - - // Gets Miis from database at offset and index in format MiiStoreData - void Get3(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto size{rp.PopRaw<u32>()}; - const auto source{rp.PopRaw<Source>()}; - - LOG_DEBUG(Service_Mii, "called with size={:08X}, offset={:08X}, source={}", size, - offsets[3], source); - - u32 read_size{}; - ctx.WriteBuffer(SerializeArray(&MiiManager::GetStoreData, offsets[3], size, read_size)); - offsets[3] += read_size; - - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(read_size); - } - - void FindIndex(Kernel::HLERequestContext& ctx) { + void BuildRandom(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto uuid{rp.PopRaw<Common::UUID>()}; - const auto unknown{rp.PopRaw<bool>()}; - LOG_DEBUG(Service_Mii, "called with uuid={}, unknown={}", uuid.FormatSwitch(), unknown); + const auto age{rp.PopRaw<Age>()}; + const auto gender{rp.PopRaw<Gender>()}; + const auto race{rp.PopRaw<Race>()}; - IPC::ResponseBuilder rb{ctx, 3}; + LOG_DEBUG(Service_Mii, "called with age={}, gender={}, race={}", age, gender, race); - const auto index = db.IndexOf(uuid); - if (index > MAX_MIIS) { - // TODO(DarkLordZach): Find a better error code - rb.Push(RESULT_UNKNOWN); - rb.Push(index); - } else { - rb.Push(RESULT_SUCCESS); - rb.Push(index); + if (age > Age::All) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(ERROR_INVALID_ARGUMENT); + LOG_ERROR(Service_Mii, "invalid age={}", age); + return; } - } - void Move(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto uuid{rp.PopRaw<Common::UUID>()}; - const auto index{rp.PopRaw<s32>()}; - - if (index < 0) { - LOG_ERROR(Service_Mii, "Index cannot be negative but is {:08X}!", index); + if (gender > Gender::All) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_ARGUMENT); + LOG_ERROR(Service_Mii, "invalid gender={}", gender); return; } - LOG_DEBUG(Service_Mii, "called with uuid={}, index={:08X}", uuid.FormatSwitch(), index); - - const auto success = db.Move(uuid, index); - - IPC::ResponseBuilder rb{ctx, 2}; - // TODO(DarkLordZach): Find a better error code - rb.Push(success ? RESULT_SUCCESS : RESULT_UNKNOWN); - } - - void AddOrReplace(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto data{rp.PopRaw<MiiStoreData>()}; - - LOG_DEBUG(Service_Mii, "called with Mii data uuid={}, name={}", data.uuid.FormatSwitch(), - Common::UTF16ToUTF8(data.Name())); - - const auto success = db.AddOrReplace(data); - - IPC::ResponseBuilder rb{ctx, 2}; - // TODO(DarkLordZach): Find a better error code - rb.Push(success ? RESULT_SUCCESS : RESULT_UNKNOWN); - } - - void Delete(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto uuid{rp.PopRaw<Common::UUID>()}; - - LOG_DEBUG(Service_Mii, "called with uuid={}", uuid.FormatSwitch()); - - const auto success = db.Remove(uuid); - - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(success ? RESULT_SUCCESS : ERROR_CANNOT_FIND_ENTRY); - } - - void DestroyFile(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_Mii, "called"); - - if (!db.IsTestModeEnabled()) { - LOG_ERROR(Service_Mii, "Database is not in test mode -- cannot destory database file."); + if (race > Race::All) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_NOT_IN_TEST_MODE); + rb.Push(ERROR_INVALID_ARGUMENT); + LOG_ERROR(Service_Mii, "invalid race={}", race); return; } - IPC::ResponseBuilder rb{ctx, 3}; + IPC::ResponseBuilder rb{ctx, 2 + sizeof(MiiInfo) / sizeof(u32)}; rb.Push(RESULT_SUCCESS); - rb.Push(db.DestroyFile()); + rb.PushRaw<MiiInfo>(manager.BuildRandom(age, gender, race)); } - void DeleteFile(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_Mii, "called"); + void BuildDefault(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto index{rp.Pop<u32>()}; + + LOG_DEBUG(Service_Mii, "called with index={}", index); - if (!db.IsTestModeEnabled()) { - LOG_ERROR(Service_Mii, "Database is not in test mode -- cannot delete database file."); + if (index > 5) { + LOG_ERROR(Service_Mii, "invalid argument, index cannot be greater than 5 but is {:08X}", + index); IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_NOT_IN_TEST_MODE); + rb.Push(ERROR_INVALID_ARGUMENT); return; } - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push(db.DeleteFile()); - } - - void Format(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_Mii, "called"); - - db.Clear(); - - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 2 + sizeof(MiiInfo) / sizeof(u32)}; rb.Push(RESULT_SUCCESS); + rb.PushRaw<MiiInfo>(manager.BuildDefault(index)); } void GetIndex(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto info{rp.PopRaw<MiiInfo>()}; - LOG_DEBUG(Service_Mii, "called with Mii info uuid={}, name={}", info.uuid.FormatSwitch(), - Common::UTF16ToUTF8(info.Name())); - - const auto index = db.IndexOf(info); + LOG_DEBUG(Service_Mii, "called"); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); + u32 index{}; + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(manager.GetIndex(info, index)); rb.Push(index); } @@ -364,12 +240,14 @@ private: rb.Push(RESULT_SUCCESS); } - MiiManager db; + constexpr bool IsInterfaceVersionSupported(u32 interface_version) const { + return current_interface_version >= interface_version; + } - u32 current_interface_version = 0; + MiiManager manager; - // Last read offsets of Get functions - std::array<u32, 4> offsets{}; + u32 current_interface_version{}; + u64 current_update_counter{}; }; class MiiDBModule final : public ServiceFramework<MiiDBModule> { diff --git a/src/core/hle/service/mii/mii_manager.cpp b/src/core/hle/service/mii/mii_manager.cpp deleted file mode 100644 index 8d0353075..000000000 --- a/src/core/hle/service/mii/mii_manager.cpp +++ /dev/null @@ -1,420 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <cstring> -#include "common/assert.h" -#include "common/file_util.h" -#include "common/logging/log.h" -#include "common/string_util.h" -#include "core/hle/service/mii/mii_manager.h" - -namespace Service::Mii { - -namespace { - -constexpr char MII_SAVE_DATABASE_PATH[] = "/system/save/8000000000000030/MiiDatabase.dat"; -constexpr std::array<char16_t, 11> DEFAULT_MII_NAME = {u'y', u'u', u'z', u'u', u'\0'}; - -// This value was retrieved from HW test -constexpr MiiStoreData DEFAULT_MII = { - { - 0x21, 0x40, 0x40, 0x01, 0x08, 0x01, 0x13, 0x08, 0x08, 0x02, 0x17, 0x8C, 0x06, 0x01, - 0x69, 0x6D, 0x8A, 0x6A, 0x82, 0x14, 0x00, 0x00, 0x00, 0x20, 0x64, 0x72, 0x44, 0x44, - }, - {'y', 'u', 'z', 'u', '\0'}, - Common::UUID{1, 0}, - 0, - 0, -}; - -// Default values taken from multiple real databases -const MiiDatabase DEFAULT_MII_DATABASE{Common::MakeMagic('N', 'F', 'D', 'B'), {}, {1}, 0, 0}; - -constexpr std::array<const char*, 4> SOURCE_NAMES{ - "Database", - "Default", - "Account", - "Friend", -}; - -template <typename T, std::size_t SourceArraySize, std::size_t DestArraySize> -std::array<T, DestArraySize> ResizeArray(const std::array<T, SourceArraySize>& in) { - std::array<T, DestArraySize> out{}; - std::memcpy(out.data(), in.data(), sizeof(T) * std::min(SourceArraySize, DestArraySize)); - return out; -} - -MiiInfo ConvertStoreDataToInfo(const MiiStoreData& data) { - MiiStoreBitFields bf{}; - std::memcpy(&bf, data.data.data(), sizeof(MiiStoreBitFields)); - return { - data.uuid, - ResizeArray<char16_t, 10, 11>(data.name), - static_cast<u8>(bf.font_region.Value()), - static_cast<u8>(bf.favorite_color.Value()), - static_cast<u8>(bf.gender.Value()), - static_cast<u8>(bf.height.Value()), - static_cast<u8>(bf.weight.Value()), - static_cast<u8>(bf.mii_type.Value()), - static_cast<u8>(bf.mii_region.Value()), - static_cast<u8>(bf.face_type.Value()), - static_cast<u8>(bf.face_color.Value()), - static_cast<u8>(bf.face_wrinkle.Value()), - static_cast<u8>(bf.face_makeup.Value()), - static_cast<u8>(bf.hair_type.Value()), - static_cast<u8>(bf.hair_color.Value()), - static_cast<bool>(bf.hair_flip.Value()), - static_cast<u8>(bf.eye_type.Value()), - static_cast<u8>(bf.eye_color.Value()), - static_cast<u8>(bf.eye_scale.Value()), - static_cast<u8>(bf.eye_aspect.Value()), - static_cast<u8>(bf.eye_rotate.Value()), - static_cast<u8>(bf.eye_x.Value()), - static_cast<u8>(bf.eye_y.Value()), - static_cast<u8>(bf.eyebrow_type.Value()), - static_cast<u8>(bf.eyebrow_color.Value()), - static_cast<u8>(bf.eyebrow_scale.Value()), - static_cast<u8>(bf.eyebrow_aspect.Value()), - static_cast<u8>(bf.eyebrow_rotate.Value()), - static_cast<u8>(bf.eyebrow_x.Value()), - static_cast<u8>(bf.eyebrow_y.Value()), - static_cast<u8>(bf.nose_type.Value()), - static_cast<u8>(bf.nose_scale.Value()), - static_cast<u8>(bf.nose_y.Value()), - static_cast<u8>(bf.mouth_type.Value()), - static_cast<u8>(bf.mouth_color.Value()), - static_cast<u8>(bf.mouth_scale.Value()), - static_cast<u8>(bf.mouth_aspect.Value()), - static_cast<u8>(bf.mouth_y.Value()), - static_cast<u8>(bf.facial_hair_color.Value()), - static_cast<u8>(bf.beard_type.Value()), - static_cast<u8>(bf.mustache_type.Value()), - static_cast<u8>(bf.mustache_scale.Value()), - static_cast<u8>(bf.mustache_y.Value()), - static_cast<u8>(bf.glasses_type.Value()), - static_cast<u8>(bf.glasses_color.Value()), - static_cast<u8>(bf.glasses_scale.Value()), - static_cast<u8>(bf.glasses_y.Value()), - static_cast<u8>(bf.mole_type.Value()), - static_cast<u8>(bf.mole_scale.Value()), - static_cast<u8>(bf.mole_x.Value()), - static_cast<u8>(bf.mole_y.Value()), - 0x00, - }; -} -MiiStoreData ConvertInfoToStoreData(const MiiInfo& info) { - MiiStoreData out{}; - out.name = ResizeArray<char16_t, 11, 10>(info.name); - out.uuid = info.uuid; - - MiiStoreBitFields bf{}; - - bf.hair_type.Assign(info.hair_type); - bf.mole_type.Assign(info.mole_type); - bf.height.Assign(info.height); - bf.hair_flip.Assign(info.hair_flip); - bf.weight.Assign(info.weight); - bf.hair_color.Assign(info.hair_color); - - bf.gender.Assign(info.gender); - bf.eye_color.Assign(info.eye_color); - bf.eyebrow_color.Assign(info.eyebrow_color); - bf.mouth_color.Assign(info.mouth_color); - bf.facial_hair_color.Assign(info.facial_hair_color); - - bf.mii_type.Assign(info.mii_type); - bf.glasses_color.Assign(info.glasses_color); - bf.font_region.Assign(info.font_region); - bf.eye_type.Assign(info.eye_type); - bf.mii_region.Assign(info.mii_region); - bf.mouth_type.Assign(info.mouth_type); - bf.glasses_scale.Assign(info.glasses_scale); - bf.eye_y.Assign(info.eye_y); - - bf.mustache_type.Assign(info.mustache_type); - bf.eyebrow_type.Assign(info.eyebrow_type); - bf.beard_type.Assign(info.beard_type); - bf.nose_type.Assign(info.nose_type); - bf.mouth_aspect.Assign(info.mouth_aspect_ratio); - bf.nose_y.Assign(info.nose_y); - bf.eyebrow_aspect.Assign(info.eyebrow_aspect_ratio); - bf.mouth_y.Assign(info.mouth_y); - - bf.eye_rotate.Assign(info.eye_rotate); - bf.mustache_y.Assign(info.mustache_y); - bf.eye_aspect.Assign(info.eye_aspect_ratio); - bf.glasses_y.Assign(info.glasses_y); - bf.eye_scale.Assign(info.eye_scale); - bf.mole_x.Assign(info.mole_x); - bf.mole_y.Assign(info.mole_y); - - bf.glasses_type.Assign(info.glasses_type); - bf.face_type.Assign(info.face_type); - bf.favorite_color.Assign(info.favorite_color); - bf.face_wrinkle.Assign(info.face_wrinkle); - bf.face_color.Assign(info.face_color); - bf.eye_x.Assign(info.eye_x); - bf.face_makeup.Assign(info.face_makeup); - - bf.eyebrow_rotate.Assign(info.eyebrow_rotate); - bf.eyebrow_scale.Assign(info.eyebrow_scale); - bf.eyebrow_y.Assign(info.eyebrow_y); - bf.eyebrow_x.Assign(info.eyebrow_x); - bf.mouth_scale.Assign(info.mouth_scale); - bf.nose_scale.Assign(info.nose_scale); - bf.mole_scale.Assign(info.mole_scale); - bf.mustache_scale.Assign(info.mustache_scale); - - std::memcpy(out.data.data(), &bf, sizeof(MiiStoreBitFields)); - - return out; -} - -} // namespace - -std::ostream& operator<<(std::ostream& os, Source source) { - if (static_cast<std::size_t>(source) >= SOURCE_NAMES.size()) { - return os << "[UNKNOWN SOURCE]"; - } - - os << SOURCE_NAMES.at(static_cast<std::size_t>(source)); - return os; -} - -std::u16string MiiInfo::Name() const { - return Common::UTF16StringFromFixedZeroTerminatedBuffer(name.data(), name.size()); -} - -bool operator==(const MiiInfo& lhs, const MiiInfo& rhs) { - return std::memcmp(&lhs, &rhs, sizeof(MiiInfo)) == 0; -} - -bool operator!=(const MiiInfo& lhs, const MiiInfo& rhs) { - return !operator==(lhs, rhs); -} - -std::u16string MiiStoreData::Name() const { - return Common::UTF16StringFromFixedZeroTerminatedBuffer(name.data(), name.size()); -} - -MiiManager::MiiManager() = default; - -MiiManager::~MiiManager() = default; - -MiiInfo MiiManager::CreateRandom(RandomParameters params) { - LOG_WARNING(Service_Mii, - "(STUBBED) called with params={:08X}{:08X}{:08X}, returning default Mii", - params.unknown_1, params.unknown_2, params.unknown_3); - - return ConvertStoreDataToInfo(CreateMiiWithUniqueUUID()); -} - -MiiInfo MiiManager::CreateDefault(u32 index) { - const auto new_mii = CreateMiiWithUniqueUUID(); - - database.miis.at(index) = new_mii; - - EnsureDatabasePartition(); - return ConvertStoreDataToInfo(new_mii); -} - -bool MiiManager::CheckUpdatedFlag() const { - return updated_flag; -} - -void MiiManager::ResetUpdatedFlag() { - updated_flag = false; -} - -bool MiiManager::IsTestModeEnabled() const { - return is_test_mode_enabled; -} - -bool MiiManager::Empty() const { - return Size() == 0; -} - -bool MiiManager::Full() const { - return Size() == MAX_MIIS; -} - -void MiiManager::Clear() { - updated_flag = true; - std::fill(database.miis.begin(), database.miis.end(), MiiStoreData{}); -} - -u32 MiiManager::Size() const { - return static_cast<u32>(std::count_if(database.miis.begin(), database.miis.end(), - [](const MiiStoreData& elem) { return elem.uuid; })); -} - -MiiInfo MiiManager::GetInfo(u32 index) const { - return ConvertStoreDataToInfo(GetStoreData(index)); -} - -MiiInfoElement MiiManager::GetInfoElement(u32 index) const { - return {GetInfo(index), Source::Database}; -} - -MiiStoreData MiiManager::GetStoreData(u32 index) const { - return database.miis.at(index); -} - -MiiStoreDataElement MiiManager::GetStoreDataElement(u32 index) const { - return {GetStoreData(index), Source::Database}; -} - -bool MiiManager::Remove(Common::UUID uuid) { - const auto iter = std::find_if(database.miis.begin(), database.miis.end(), - [uuid](const MiiStoreData& elem) { return elem.uuid == uuid; }); - - if (iter == database.miis.end()) - return false; - - updated_flag = true; - *iter = MiiStoreData{}; - EnsureDatabasePartition(); - return true; -} - -u32 MiiManager::IndexOf(Common::UUID uuid) const { - const auto iter = std::find_if(database.miis.begin(), database.miis.end(), - [uuid](const MiiStoreData& elem) { return elem.uuid == uuid; }); - - if (iter == database.miis.end()) - return INVALID_INDEX; - - return static_cast<u32>(std::distance(database.miis.begin(), iter)); -} - -u32 MiiManager::IndexOf(const MiiInfo& info) const { - const auto iter = - std::find_if(database.miis.begin(), database.miis.end(), [&info](const MiiStoreData& elem) { - return ConvertStoreDataToInfo(elem) == info; - }); - - if (iter == database.miis.end()) - return INVALID_INDEX; - - return static_cast<u32>(std::distance(database.miis.begin(), iter)); -} - -bool MiiManager::Move(Common::UUID uuid, u32 new_index) { - const auto index = IndexOf(uuid); - - if (index == INVALID_INDEX || new_index >= MAX_MIIS) - return false; - - updated_flag = true; - const auto moving = database.miis[index]; - const auto replacing = database.miis[new_index]; - if (replacing.uuid) { - database.miis[index] = replacing; - database.miis[new_index] = moving; - } else { - database.miis[index] = MiiStoreData{}; - database.miis[new_index] = moving; - } - - EnsureDatabasePartition(); - return true; -} - -bool MiiManager::AddOrReplace(const MiiStoreData& data) { - const auto index = IndexOf(data.uuid); - - updated_flag = true; - if (index == INVALID_INDEX) { - const auto size = Size(); - if (size == MAX_MIIS) - return false; - database.miis[size] = data; - } else { - database.miis[index] = data; - } - - return true; -} - -bool MiiManager::DestroyFile() { - database = DEFAULT_MII_DATABASE; - updated_flag = false; - return DeleteFile(); -} - -bool MiiManager::DeleteFile() { - const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + MII_SAVE_DATABASE_PATH; - return FileUtil::Exists(path) && FileUtil::Delete(path); -} - -void MiiManager::WriteToFile() { - const auto raw_path = - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000030"; - if (FileUtil::Exists(raw_path) && !FileUtil::IsDirectory(raw_path)) - FileUtil::Delete(raw_path); - - const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + MII_SAVE_DATABASE_PATH; - - if (!FileUtil::CreateFullPath(path)) { - LOG_WARNING(Service_Mii, - "Failed to create full path of MiiDatabase.dat. Create the directory " - "nand/system/save/8000000000000030 to mitigate this " - "issue."); - return; - } - - FileUtil::IOFile save(path, "wb"); - - if (!save.IsOpen()) { - LOG_WARNING(Service_Mii, "Failed to write save data to file... No changes to user data " - "made in current session will be saved."); - return; - } - - save.Resize(sizeof(MiiDatabase)); - if (save.WriteBytes(&database, sizeof(MiiDatabase)) != sizeof(MiiDatabase)) { - LOG_WARNING(Service_Mii, "Failed to write all data to save file... Data may be malformed " - "and/or regenerated on next run."); - save.Resize(0); - } -} - -void MiiManager::ReadFromFile() { - FileUtil::IOFile save( - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + MII_SAVE_DATABASE_PATH, "rb"); - - if (!save.IsOpen()) { - LOG_WARNING(Service_ACC, "Failed to load profile data from save data... Generating new " - "blank Mii database with no Miis."); - std::memcpy(&database, &DEFAULT_MII_DATABASE, sizeof(MiiDatabase)); - return; - } - - if (save.ReadBytes(&database, sizeof(MiiDatabase)) != sizeof(MiiDatabase)) { - LOG_WARNING(Service_ACC, "MiiDatabase.dat is smaller than expected... Generating new blank " - "Mii database with no Miis."); - std::memcpy(&database, &DEFAULT_MII_DATABASE, sizeof(MiiDatabase)); - return; - } - - EnsureDatabasePartition(); -} - -MiiStoreData MiiManager::CreateMiiWithUniqueUUID() const { - auto new_mii = DEFAULT_MII; - - do { - new_mii.uuid = Common::UUID::Generate(); - } while (IndexOf(new_mii.uuid) != INVALID_INDEX); - - return new_mii; -} - -void MiiManager::EnsureDatabasePartition() { - std::stable_partition(database.miis.begin(), database.miis.end(), - [](const MiiStoreData& elem) { return elem.uuid; }); -} - -} // namespace Service::Mii diff --git a/src/core/hle/service/mii/mii_manager.h b/src/core/hle/service/mii/mii_manager.h deleted file mode 100644 index fc742816a..000000000 --- a/src/core/hle/service/mii/mii_manager.h +++ /dev/null @@ -1,273 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "common/bit_field.h" -#include "common/common_funcs.h" -#include "common/uuid.h" - -namespace Service::Mii { - -constexpr std::size_t MAX_MIIS{100}; -constexpr u32 INVALID_INDEX{0xFFFFFFFF}; - -struct RandomParameters { - u32 unknown_1{}; - u32 unknown_2{}; - u32 unknown_3{}; -}; -static_assert(sizeof(RandomParameters) == 0xC, "RandomParameters has incorrect size."); - -enum class Source : u32 { - Database = 0, - Default = 1, - Account = 2, - Friend = 3, -}; - -std::ostream& operator<<(std::ostream& os, Source source); - -struct MiiInfo { - Common::UUID uuid{Common::INVALID_UUID}; - std::array<char16_t, 11> name{}; - u8 font_region{}; - u8 favorite_color{}; - u8 gender{}; - u8 height{}; - u8 weight{}; - u8 mii_type{}; - u8 mii_region{}; - u8 face_type{}; - u8 face_color{}; - u8 face_wrinkle{}; - u8 face_makeup{}; - u8 hair_type{}; - u8 hair_color{}; - bool hair_flip{}; - u8 eye_type{}; - u8 eye_color{}; - u8 eye_scale{}; - u8 eye_aspect_ratio{}; - u8 eye_rotate{}; - u8 eye_x{}; - u8 eye_y{}; - u8 eyebrow_type{}; - u8 eyebrow_color{}; - u8 eyebrow_scale{}; - u8 eyebrow_aspect_ratio{}; - u8 eyebrow_rotate{}; - u8 eyebrow_x{}; - u8 eyebrow_y{}; - u8 nose_type{}; - u8 nose_scale{}; - u8 nose_y{}; - u8 mouth_type{}; - u8 mouth_color{}; - u8 mouth_scale{}; - u8 mouth_aspect_ratio{}; - u8 mouth_y{}; - u8 facial_hair_color{}; - u8 beard_type{}; - u8 mustache_type{}; - u8 mustache_scale{}; - u8 mustache_y{}; - u8 glasses_type{}; - u8 glasses_color{}; - u8 glasses_scale{}; - u8 glasses_y{}; - u8 mole_type{}; - u8 mole_scale{}; - u8 mole_x{}; - u8 mole_y{}; - INSERT_PADDING_BYTES(1); - - std::u16string Name() const; -}; -static_assert(sizeof(MiiInfo) == 0x58, "MiiInfo has incorrect size."); -static_assert(std::has_unique_object_representations_v<MiiInfo>, - "All bits of MiiInfo must contribute to its value."); - -bool operator==(const MiiInfo& lhs, const MiiInfo& rhs); -bool operator!=(const MiiInfo& lhs, const MiiInfo& rhs); - -#pragma pack(push, 4) -struct MiiInfoElement { - MiiInfo info{}; - Source source{}; -}; -static_assert(sizeof(MiiInfoElement) == 0x5C, "MiiInfoElement has incorrect size."); - -struct MiiStoreBitFields { - union { - u32 word_0{}; - - BitField<24, 8, u32> hair_type; - BitField<23, 1, u32> mole_type; - BitField<16, 7, u32> height; - BitField<15, 1, u32> hair_flip; - BitField<8, 7, u32> weight; - BitField<0, 7, u32> hair_color; - }; - - union { - u32 word_1{}; - - BitField<31, 1, u32> gender; - BitField<24, 7, u32> eye_color; - BitField<16, 7, u32> eyebrow_color; - BitField<8, 7, u32> mouth_color; - BitField<0, 7, u32> facial_hair_color; - }; - - union { - u32 word_2{}; - - BitField<31, 1, u32> mii_type; - BitField<24, 7, u32> glasses_color; - BitField<22, 2, u32> font_region; - BitField<16, 6, u32> eye_type; - BitField<14, 2, u32> mii_region; - BitField<8, 6, u32> mouth_type; - BitField<5, 3, u32> glasses_scale; - BitField<0, 5, u32> eye_y; - }; - - union { - u32 word_3{}; - - BitField<29, 3, u32> mustache_type; - BitField<24, 5, u32> eyebrow_type; - BitField<21, 3, u32> beard_type; - BitField<16, 5, u32> nose_type; - BitField<13, 3, u32> mouth_aspect; - BitField<8, 5, u32> nose_y; - BitField<5, 3, u32> eyebrow_aspect; - BitField<0, 5, u32> mouth_y; - }; - - union { - u32 word_4{}; - - BitField<29, 3, u32> eye_rotate; - BitField<24, 5, u32> mustache_y; - BitField<21, 3, u32> eye_aspect; - BitField<16, 5, u32> glasses_y; - BitField<13, 3, u32> eye_scale; - BitField<8, 5, u32> mole_x; - BitField<0, 5, u32> mole_y; - }; - - union { - u32 word_5{}; - - BitField<24, 5, u32> glasses_type; - BitField<20, 4, u32> face_type; - BitField<16, 4, u32> favorite_color; - BitField<12, 4, u32> face_wrinkle; - BitField<8, 4, u32> face_color; - BitField<4, 4, u32> eye_x; - BitField<0, 4, u32> face_makeup; - }; - - union { - u32 word_6{}; - - BitField<28, 4, u32> eyebrow_rotate; - BitField<24, 4, u32> eyebrow_scale; - BitField<20, 4, u32> eyebrow_y; - BitField<16, 4, u32> eyebrow_x; - BitField<12, 4, u32> mouth_scale; - BitField<8, 4, u32> nose_scale; - BitField<4, 4, u32> mole_scale; - BitField<0, 4, u32> mustache_scale; - }; -}; -static_assert(sizeof(MiiStoreBitFields) == 0x1C, "MiiStoreBitFields has incorrect size."); -static_assert(std::is_trivially_copyable_v<MiiStoreBitFields>, - "MiiStoreBitFields is not trivially copyable."); - -struct MiiStoreData { - // This corresponds to the above structure MiiStoreBitFields. I did it like this because the - // BitField<> type makes this (and any thing that contains it) not trivially copyable, which is - // not suitable for our uses. - std::array<u8, 0x1C> data{}; - static_assert(sizeof(MiiStoreBitFields) == sizeof(data), "data field has incorrect size."); - - std::array<char16_t, 10> name{}; - Common::UUID uuid{Common::INVALID_UUID}; - u16 crc_1{}; - u16 crc_2{}; - - std::u16string Name() const; -}; -static_assert(sizeof(MiiStoreData) == 0x44, "MiiStoreData has incorrect size."); - -struct MiiStoreDataElement { - MiiStoreData data{}; - Source source{}; -}; -static_assert(sizeof(MiiStoreDataElement) == 0x48, "MiiStoreDataElement has incorrect size."); - -struct MiiDatabase { - u32 magic{}; // 'NFDB' - std::array<MiiStoreData, MAX_MIIS> miis{}; - INSERT_PADDING_BYTES(1); - u8 count{}; - u16 crc{}; -}; -static_assert(sizeof(MiiDatabase) == 0x1A98, "MiiDatabase has incorrect size."); -#pragma pack(pop) - -// The Mii manager is responsible for loading and storing the Miis to the database in NAND along -// with providing an easy interface for HLE emulation of the mii service. -class MiiManager { -public: - MiiManager(); - ~MiiManager(); - - MiiInfo CreateRandom(RandomParameters params); - MiiInfo CreateDefault(u32 index); - - bool CheckUpdatedFlag() const; - void ResetUpdatedFlag(); - - bool IsTestModeEnabled() const; - - bool Empty() const; - bool Full() const; - - void Clear(); - - u32 Size() const; - - MiiInfo GetInfo(u32 index) const; - MiiInfoElement GetInfoElement(u32 index) const; - MiiStoreData GetStoreData(u32 index) const; - MiiStoreDataElement GetStoreDataElement(u32 index) const; - - bool Remove(Common::UUID uuid); - u32 IndexOf(Common::UUID uuid) const; - u32 IndexOf(const MiiInfo& info) const; - - bool Move(Common::UUID uuid, u32 new_index); - bool AddOrReplace(const MiiStoreData& data); - - bool DestroyFile(); - bool DeleteFile(); - -private: - void WriteToFile(); - void ReadFromFile(); - - MiiStoreData CreateMiiWithUniqueUUID() const; - - void EnsureDatabasePartition(); - - MiiDatabase database; - bool updated_flag{}; - bool is_test_mode_enabled{}; -}; - -}; // namespace Service::Mii diff --git a/src/core/hle/service/mii/raw_data.cpp b/src/core/hle/service/mii/raw_data.cpp new file mode 100644 index 000000000..25d7bae0c --- /dev/null +++ b/src/core/hle/service/mii/raw_data.cpp @@ -0,0 +1,2261 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include "core/hle/service/mii/raw_data.h" + +namespace Service::Mii::RawData { + +const std::array<u8, 1728> DefaultMii{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x44, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x6e, 0x00, 0x6f, 0x00, 0x20, 0x00, 0x6e, 0x00, 0x61, 0x00, 0x6d, 0x00, + 0x65, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6e, 0x00, 0x6f, 0x00, + 0x20, 0x00, 0x6e, 0x00, 0x61, 0x00, 0x6d, 0x00, 0x65, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x6e, 0x00, 0x6f, 0x00, 0x20, 0x00, 0x6e, 0x00, 0x61, 0x00, 0x6d, 0x00, + 0x65, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6e, 0x00, 0x6f, 0x00, + 0x20, 0x00, 0x6e, 0x00, 0x61, 0x00, 0x6d, 0x00, 0x65, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x6e, 0x00, 0x6f, 0x00, 0x20, 0x00, 0x6e, 0x00, 0x61, 0x00, 0x6d, 0x00, + 0x65, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6e, 0x00, 0x6f, 0x00, + 0x20, 0x00, 0x6e, 0x00, 0x61, 0x00, 0x6d, 0x00, 0x65, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiFaceline{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 1200> RandomMiiFacelineColor{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiFacelineWrinkle{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiFacelineMakeup{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiHairType{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, + 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, + 0x34, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x4b, 0x00, 0x00, 0x00, 0x4c, 0x00, 0x00, 0x00, + 0x56, 0x00, 0x00, 0x00, 0x59, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, + 0x2c, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, + 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x34, 0x00, 0x00, 0x00, + 0x36, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x42, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, 0x4b, 0x00, 0x00, 0x00, 0x51, 0x00, 0x00, 0x00, + 0x56, 0x00, 0x00, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, + 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, + 0x32, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x34, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, + 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, + 0x49, 0x00, 0x00, 0x00, 0x4b, 0x00, 0x00, 0x00, 0x51, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, + 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, + 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, + 0x2a, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, + 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, + 0x33, 0x00, 0x00, 0x00, 0x34, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, + 0x37, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, + 0x3c, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x41, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, + 0x43, 0x00, 0x00, 0x00, 0x44, 0x00, 0x00, 0x00, 0x46, 0x00, 0x00, 0x00, 0x4b, 0x00, 0x00, 0x00, + 0x4c, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, 0x59, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, + 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, + 0x34, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, + 0x38, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x41, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00, 0x00, + 0x44, 0x00, 0x00, 0x00, 0x46, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, 0x4b, 0x00, 0x00, 0x00, + 0x51, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, + 0x2c, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, + 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x34, 0x00, 0x00, 0x00, + 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, + 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x41, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00, 0x00, 0x44, 0x00, 0x00, 0x00, + 0x46, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, 0x4b, 0x00, 0x00, 0x00, 0x51, 0x00, 0x00, 0x00, + 0x56, 0x00, 0x00, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x36, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, + 0x41, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, 0x58, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2d, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x36, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, + 0x41, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, 0x58, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x25, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, + 0x2f, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, + 0x37, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x41, 0x00, 0x00, 0x00, + 0x43, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, 0x58, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x1c, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x00, 0x00, + 0x3e, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, + 0x4c, 0x00, 0x00, 0x00, 0x4d, 0x00, 0x00, 0x00, 0x4f, 0x00, 0x00, 0x00, 0x50, 0x00, 0x00, 0x00, + 0x53, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x2e, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, + 0x3f, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, + 0x4a, 0x00, 0x00, 0x00, 0x4d, 0x00, 0x00, 0x00, 0x4e, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, + 0x53, 0x00, 0x00, 0x00, 0x54, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x57, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x2a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, + 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, + 0x32, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, + 0x4d, 0x00, 0x00, 0x00, 0x4e, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, + 0x54, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, + 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x2a, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, + 0x3e, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, + 0x47, 0x00, 0x00, 0x00, 0x4c, 0x00, 0x00, 0x00, 0x4f, 0x00, 0x00, 0x00, 0x50, 0x00, 0x00, 0x00, + 0x51, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x56, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, + 0x3a, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, 0x47, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, + 0x4a, 0x00, 0x00, 0x00, 0x4f, 0x00, 0x00, 0x00, 0x51, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, + 0x53, 0x00, 0x00, 0x00, 0x54, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, + 0x3c, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, + 0x45, 0x00, 0x00, 0x00, 0x47, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, + 0x4f, 0x00, 0x00, 0x00, 0x51, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, + 0x54, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x3a, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, 0x4c, 0x00, 0x00, 0x00, + 0x53, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, + 0x3e, 0x00, 0x00, 0x00, 0x45, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, 0x4c, 0x00, 0x00, 0x00, + 0x51, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, + 0x45, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, 0x4c, 0x00, 0x00, 0x00, 0x51, 0x00, 0x00, 0x00, + 0x53, 0x00, 0x00, 0x00, 0x55, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 1800> RandomMiiHairColor{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiEyeType{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2b, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, + 0x35, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, + 0x2f, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, + 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, + 0x2c, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, + 0x35, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, + 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, + 0x2f, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x37, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x1d, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0x29, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0x29, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, + 0x35, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2c, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x37, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x25, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, + 0x35, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, + 0x29, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, + 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, + 0x39, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, + 0x19, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, + 0x3b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x28, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, + 0x3b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, + 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, + 0x30, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, + 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x1d, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, + 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, + 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, + 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, + 0x3b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, + 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, + 0x36, 0x00, 0x00, 0x00, 0x39, 0x00, 0x00, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, + 0x27, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, + 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, + 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, + 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0x28, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, + 0x2e, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 588> RandomMiiEyeColor{ + 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiEyebrowType{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiNoseType{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 3672> RandomMiiMouthType{ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x1c, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, + 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x1d, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, + 0x12, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x17, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x05, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, + 0x19, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +const std::array<u8, 588> RandomMiiGlassType{ + 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x5a, 0x00, 0x00, 0x00, 0x5e, 0x00, 0x00, 0x00, + 0x60, 0x00, 0x00, 0x00, 0x64, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, + 0x56, 0x00, 0x00, 0x00, 0x5a, 0x00, 0x00, 0x00, 0x5d, 0x00, 0x00, 0x00, 0x5e, 0x00, 0x00, 0x00, + 0x60, 0x00, 0x00, 0x00, 0x62, 0x00, 0x00, 0x00, 0x64, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x4e, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x5d, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x62, 0x00, 0x00, 0x00, 0x64, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; + +} // namespace Service::Mii::RawData diff --git a/src/core/hle/service/mii/raw_data.h b/src/core/hle/service/mii/raw_data.h new file mode 100644 index 000000000..a02a5c0fd --- /dev/null +++ b/src/core/hle/service/mii/raw_data.h @@ -0,0 +1,27 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> + +#include "common/common_types.h" + +namespace Service::Mii::RawData { + +extern const std::array<u8, 1728> DefaultMii; +extern const std::array<u8, 3672> RandomMiiFaceline; +extern const std::array<u8, 1200> RandomMiiFacelineColor; +extern const std::array<u8, 3672> RandomMiiFacelineWrinkle; +extern const std::array<u8, 3672> RandomMiiFacelineMakeup; +extern const std::array<u8, 3672> RandomMiiHairType; +extern const std::array<u8, 1800> RandomMiiHairColor; +extern const std::array<u8, 3672> RandomMiiEyeType; +extern const std::array<u8, 588> RandomMiiEyeColor; +extern const std::array<u8, 3672> RandomMiiEyebrowType; +extern const std::array<u8, 3672> RandomMiiNoseType; +extern const std::array<u8, 3672> RandomMiiMouthType; +extern const std::array<u8, 588> RandomMiiGlassType; + +} // namespace Service::Mii::RawData diff --git a/src/core/hle/service/mii/types.h b/src/core/hle/service/mii/types.h new file mode 100644 index 000000000..d65a1055e --- /dev/null +++ b/src/core/hle/service/mii/types.h @@ -0,0 +1,67 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Service::Mii { + +enum class Age : u32 { + Young, + Normal, + Old, + All, +}; + +enum class BeardType : u32 { + None, + Beard1, + Beard2, + Beard3, + Beard4, + Beard5, +}; + +enum class BeardAndMustacheFlag : u32 { Beard = 1, Mustache, All = Beard | Mustache }; +DECLARE_ENUM_FLAG_OPERATORS(BeardAndMustacheFlag); + +enum class FontRegion : u32 { + Standard, + China, + Korea, + Taiwan, +}; + +enum class Gender : u32 { + Male, + Female, + All, + Maximum = Female, +}; + +enum class HairFlip : u32 { + Left, + Right, + Maximum = Right, +}; + +enum class MustacheType : u32 { + None, + Mustache1, + Mustache2, + Mustache3, + Mustache4, + Mustache5, +}; + +enum class Race : u32 { + Black, + White, + Asian, + All, +}; + +} // namespace Service::Mii diff --git a/src/core/hle/service/mm/mm_u.cpp b/src/core/hle/service/mm/mm_u.cpp index def63dc8a..25c24e537 100644 --- a/src/core/hle/service/mm/mm_u.cpp +++ b/src/core/hle/service/mm/mm_u.cpp @@ -14,14 +14,14 @@ public: explicit MM_U() : ServiceFramework{"mm:u"} { // clang-format off static const FunctionInfo functions[] = { - {0, &MM_U::Initialize, "Initialize"}, - {1, &MM_U::Finalize, "Finalize"}, - {2, &MM_U::SetAndWait, "SetAndWait"}, - {3, &MM_U::Get, "Get"}, - {4, &MM_U::InitializeWithId, "InitializeWithId"}, - {5, &MM_U::FinalizeWithId, "FinalizeWithId"}, - {6, &MM_U::SetAndWaitWithId, "SetAndWaitWithId"}, - {7, &MM_U::GetWithId, "GetWithId"}, + {0, &MM_U::InitializeOld, "InitializeOld"}, + {1, &MM_U::FinalizeOld, "FinalizeOld"}, + {2, &MM_U::SetAndWaitOld, "SetAndWaitOld"}, + {3, &MM_U::GetOld, "GetOld"}, + {4, &MM_U::Initialize, "Initialize"}, + {5, &MM_U::Finalize, "Finalize"}, + {6, &MM_U::SetAndWait, "SetAndWait"}, + {7, &MM_U::Get, "Get"}, }; // clang-format on @@ -29,21 +29,21 @@ public: } private: - void Initialize(Kernel::HLERequestContext& ctx) { + void InitializeOld(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_MM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } - void Finalize(Kernel::HLERequestContext& ctx) { + void FinalizeOld(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_MM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } - void SetAndWait(Kernel::HLERequestContext& ctx) { + void SetAndWaitOld(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; min = rp.Pop<u32>(); max = rp.Pop<u32>(); @@ -54,7 +54,7 @@ private: rb.Push(RESULT_SUCCESS); } - void Get(Kernel::HLERequestContext& ctx) { + void GetOld(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_MM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 3}; @@ -62,7 +62,7 @@ private: rb.Push(current); } - void InitializeWithId(Kernel::HLERequestContext& ctx) { + void Initialize(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_MM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 3}; @@ -70,14 +70,14 @@ private: rb.Push<u32>(id); // Any non zero value } - void FinalizeWithId(Kernel::HLERequestContext& ctx) { + void Finalize(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_MM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } - void SetAndWaitWithId(Kernel::HLERequestContext& ctx) { + void SetAndWait(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; u32 input_id = rp.Pop<u32>(); min = rp.Pop<u32>(); @@ -90,7 +90,7 @@ private: rb.Push(RESULT_SUCCESS); } - void GetWithId(Kernel::HLERequestContext& ctx) { + void Get(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_MM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 3}; diff --git a/src/core/hle/service/ncm/ncm.cpp b/src/core/hle/service/ncm/ncm.cpp index 89e283ca5..e38dea1f4 100644 --- a/src/core/hle/service/ncm/ncm.cpp +++ b/src/core/hle/service/ncm/ncm.cpp @@ -28,16 +28,16 @@ public: {7, nullptr, "ResolveApplicationLegalInformationPath"}, {8, nullptr, "RedirectApplicationLegalInformationPath"}, {9, nullptr, "Refresh"}, - {10, nullptr, "RedirectProgramPath2"}, - {11, nullptr, "Refresh2"}, - {12, nullptr, "DeleteProgramPath"}, - {13, nullptr, "DeleteApplicationControlPath"}, - {14, nullptr, "DeleteApplicationHtmlDocumentPath"}, - {15, nullptr, "DeleteApplicationLegalInformationPath"}, - {16, nullptr, ""}, - {17, nullptr, ""}, - {18, nullptr, ""}, - {19, nullptr, ""}, + {10, nullptr, "RedirectApplicationProgramPath"}, + {11, nullptr, "ClearApplicationRedirection"}, + {12, nullptr, "EraseProgramRedirection"}, + {13, nullptr, "EraseApplicationControlRedirection"}, + {14, nullptr, "EraseApplicationHtmlDocumentRedirection"}, + {15, nullptr, "EraseApplicationLegalInformationRedirection"}, + {16, nullptr, "ResolveProgramPathForDebug"}, + {17, nullptr, "RedirectProgramPathForDebug"}, + {18, nullptr, "RedirectApplicationProgramPathForDebug"}, + {19, nullptr, "EraseProgramRedirectionForDebug"}, }; // clang-format on @@ -122,6 +122,7 @@ public: {11, nullptr, "ActivateContentMetaDatabase"}, {12, nullptr, "InactivateContentMetaDatabase"}, {13, nullptr, "InvalidateRightsIdCache"}, + {14, nullptr, "GetMemoryReport"}, }; // clang-format on diff --git a/src/core/hle/service/nfc/nfc.cpp b/src/core/hle/service/nfc/nfc.cpp index b7b34ce7e..780ea30fe 100644 --- a/src/core/hle/service/nfc/nfc.cpp +++ b/src/core/hle/service/nfc/nfc.cpp @@ -198,9 +198,9 @@ public: static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, {1, nullptr, "Finalize"}, - {2, nullptr, "GetState"}, - {3, nullptr, "IsNfcEnabled"}, - {100, nullptr, "SetNfcEnabled"}, + {2, nullptr, "GetStateOld"}, + {3, nullptr, "IsNfcEnabledOld"}, + {100, nullptr, "SetNfcEnabledOld"}, {400, nullptr, "InitializeSystem"}, {401, nullptr, "FinalizeSystem"}, {402, nullptr, "GetState"}, diff --git a/src/core/hle/service/nfp/nfp.cpp b/src/core/hle/service/nfp/nfp.cpp index 4b79eb81d..a0469ffbd 100644 --- a/src/core/hle/service/nfp/nfp.cpp +++ b/src/core/hle/service/nfp/nfp.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> #include <atomic> #include "common/logging/log.h" @@ -72,10 +73,10 @@ private: std::array<u8, 10> uuid; u8 uuid_length; // TODO(ogniK): Figure out if this is actual the uuid length or does it // mean something else - INSERT_PADDING_BYTES(0x15); + std::array<u8, 0x15> padding_1; u32_le protocol; u32_le tag_type; - INSERT_PADDING_BYTES(0x2c); + std::array<u8, 0x2c> padding_2; }; static_assert(sizeof(TagInfo) == 0x54, "TagInfo is an invalid size"); @@ -127,7 +128,7 @@ private: const u32 array_size = rp.Pop<u32>(); LOG_DEBUG(Service_NFP, "called, array_size={}", array_size); - ctx.WriteBuffer(&device_handle, sizeof(device_handle)); + ctx.WriteBuffer(device_handle); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); @@ -213,14 +214,16 @@ private: LOG_DEBUG(Service_NFP, "called"); IPC::ResponseBuilder rb{ctx, 2}; - auto amiibo = nfp_interface.GetAmiiboBuffer(); - TagInfo tag_info{}; - tag_info.uuid = amiibo.uuid; - tag_info.uuid_length = static_cast<u8>(tag_info.uuid.size()); - - tag_info.protocol = 1; // TODO(ogniK): Figure out actual values - tag_info.tag_type = 2; - ctx.WriteBuffer(&tag_info, sizeof(TagInfo)); + const auto& amiibo = nfp_interface.GetAmiiboBuffer(); + const TagInfo tag_info{ + .uuid = amiibo.uuid, + .uuid_length = static_cast<u8>(tag_info.uuid.size()), + .padding_1 = {}, + .protocol = 1, // TODO(ogniK): Figure out actual values + .tag_type = 2, + .padding_2 = {}, + }; + ctx.WriteBuffer(tag_info); rb.Push(RESULT_SUCCESS); } @@ -236,8 +239,8 @@ private: LOG_DEBUG(Service_NFP, "called"); IPC::ResponseBuilder rb{ctx, 2}; - auto amiibo = nfp_interface.GetAmiiboBuffer(); - ctx.WriteBuffer(&amiibo.model_info, sizeof(amiibo.model_info)); + const auto& amiibo = nfp_interface.GetAmiiboBuffer(); + ctx.WriteBuffer(amiibo.model_info); rb.Push(RESULT_SUCCESS); } @@ -283,7 +286,7 @@ private: CommonInfo common_info{}; common_info.application_area_size = 0; - ctx.WriteBuffer(&common_info, sizeof(CommonInfo)); + ctx.WriteBuffer(common_info); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); diff --git a/src/core/hle/service/nifm/nifm.cpp b/src/core/hle/service/nifm/nifm.cpp index 767158444..db7ec6d0e 100644 --- a/src/core/hle/service/nifm/nifm.cpp +++ b/src/core/hle/service/nifm/nifm.cpp @@ -9,6 +9,7 @@ #include "core/hle/kernel/writable_event.h" #include "core/hle/service/nifm/nifm.h" #include "core/hle/service/service.h" +#include "core/network/network.h" #include "core/settings.h" namespace Service::NIFM { @@ -61,7 +62,7 @@ public: {18, nullptr, "SetRequirementByRevision"}, {19, nullptr, "GetRequirement"}, {20, nullptr, "GetRevision"}, - {21, nullptr, "GetAppletInfo"}, + {21, &IRequest::GetAppletInfo, "GetAppletInfo"}, {22, nullptr, "GetAdditionalInfo"}, {23, nullptr, "SetKeptInSleep"}, {24, nullptr, "RegisterSocketDescriptor"}, @@ -124,6 +125,16 @@ private: rb.Push(RESULT_SUCCESS); } + void GetAppletInfo(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIFM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 8}; + rb.Push(RESULT_SUCCESS); + rb.Push<u32>(0); + rb.Push<u32>(0); + rb.Push<u32>(0); + } + Kernel::EventPair event1, event2; }; @@ -174,10 +185,21 @@ private: IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } + void GetCurrentIpAddress(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIFM, "(STUBBED) called"); + + const auto [ipv4, error] = Network::GetHostIPv4Address(); + UNIMPLEMENTED_IF(error != Network::Errno::SUCCESS); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw(ipv4); + } void CreateTemporaryNetworkProfile(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NIFM, "called"); - ASSERT_MSG(ctx.GetReadBufferSize() == 0x17c, "NetworkProfileData is not the correct size"); + ASSERT_MSG(ctx.GetReadBufferSize() == 0x17c, + "SfNetworkProfileData is not the correct size"); u128 uuid{}; auto buffer = ctx.ReadBuffer(); std::memcpy(&uuid, buffer.data() + 8, sizeof(u128)); @@ -234,7 +256,7 @@ IGeneralService::IGeneralService(Core::System& system) {9, nullptr, "SetNetworkProfile"}, {10, &IGeneralService::RemoveNetworkProfile, "RemoveNetworkProfile"}, {11, nullptr, "GetScanDataOld"}, - {12, nullptr, "GetCurrentIpAddress"}, + {12, &IGeneralService::GetCurrentIpAddress, "GetCurrentIpAddress"}, {13, nullptr, "GetCurrentAccessPointOld"}, {14, &IGeneralService::CreateTemporaryNetworkProfile, "CreateTemporaryNetworkProfile"}, {15, nullptr, "GetCurrentIpConfigInfo"}, diff --git a/src/core/hle/service/nim/nim.cpp b/src/core/hle/service/nim/nim.cpp index e85f123e2..11aa74828 100644 --- a/src/core/hle/service/nim/nim.cpp +++ b/src/core/hle/service/nim/nim.cpp @@ -15,6 +15,66 @@ namespace Service::NIM { +class IShopServiceAsync final : public ServiceFramework<IShopServiceAsync> { +public: + IShopServiceAsync() : ServiceFramework("IShopServiceAsync") { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "Cancel"}, + {1, nullptr, "GetSize"}, + {2, nullptr, "Read"}, + {3, nullptr, "GetErrorCode"}, + {4, nullptr, "Request"}, + {5, nullptr, "Prepare"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IShopServiceAccessor final : public ServiceFramework<IShopServiceAccessor> { +public: + IShopServiceAccessor() : ServiceFramework("IShopServiceAccessor") { + // clang-format off + static const FunctionInfo functions[] = { + {0, &IShopServiceAccessor::CreateAsyncInterface, "CreateAsyncInterface"}, + }; + // clang-format on + + RegisterHandlers(functions); + } + +private: + void CreateAsyncInterface(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IShopServiceAsync>(); + } +}; + +class IShopServiceAccessServer final : public ServiceFramework<IShopServiceAccessServer> { +public: + IShopServiceAccessServer() : ServiceFramework("IShopServiceAccessServer") { + // clang-format off + static const FunctionInfo functions[] = { + {0, &IShopServiceAccessServer::CreateAccessorInterface, "CreateAccessorInterface"}, + }; + // clang-format on + + RegisterHandlers(functions); + } + +private: + void CreateAccessorInterface(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IShopServiceAccessor>(); + } +}; + class NIM final : public ServiceFramework<NIM> { public: explicit NIM() : ServiceFramework{"nim"} { @@ -61,11 +121,83 @@ public: {39, nullptr, "PrepareShutdown"}, {40, nullptr, "ListApplyDeltaTask"}, {41, nullptr, "ClearNotEnoughSpaceStateOfApplyDeltaTask"}, - {42, nullptr, "Unknown1"}, - {43, nullptr, "Unknown2"}, - {44, nullptr, "Unknown3"}, - {45, nullptr, "Unknown4"}, - {46, nullptr, "Unknown5"}, + {42, nullptr, "Unknown42"}, + {43, nullptr, "Unknown43"}, + {44, nullptr, "Unknown44"}, + {45, nullptr, "Unknown45"}, + {46, nullptr, "Unknown46"}, + {47, nullptr, "Unknown47"}, + {48, nullptr, "Unknown48"}, + {49, nullptr, "Unknown49"}, + {50, nullptr, "Unknown50"}, + {51, nullptr, "Unknown51"}, + {52, nullptr, "Unknown52"}, + {53, nullptr, "Unknown53"}, + {54, nullptr, "Unknown54"}, + {55, nullptr, "Unknown55"}, + {56, nullptr, "Unknown56"}, + {57, nullptr, "Unknown57"}, + {58, nullptr, "Unknown58"}, + {59, nullptr, "Unknown59"}, + {60, nullptr, "Unknown60"}, + {61, nullptr, "Unknown61"}, + {62, nullptr, "Unknown62"}, + {63, nullptr, "Unknown63"}, + {64, nullptr, "Unknown64"}, + {65, nullptr, "Unknown65"}, + {66, nullptr, "Unknown66"}, + {67, nullptr, "Unknown67"}, + {68, nullptr, "Unknown68"}, + {69, nullptr, "Unknown69"}, + {70, nullptr, "Unknown70"}, + {71, nullptr, "Unknown71"}, + {72, nullptr, "Unknown72"}, + {73, nullptr, "Unknown73"}, + {74, nullptr, "Unknown74"}, + {75, nullptr, "Unknown75"}, + {76, nullptr, "Unknown76"}, + {77, nullptr, "Unknown77"}, + {78, nullptr, "Unknown78"}, + {79, nullptr, "Unknown79"}, + {80, nullptr, "Unknown80"}, + {81, nullptr, "Unknown81"}, + {82, nullptr, "Unknown82"}, + {83, nullptr, "Unknown83"}, + {84, nullptr, "Unknown84"}, + {85, nullptr, "Unknown85"}, + {86, nullptr, "Unknown86"}, + {87, nullptr, "Unknown87"}, + {88, nullptr, "Unknown88"}, + {89, nullptr, "Unknown89"}, + {90, nullptr, "Unknown90"}, + {91, nullptr, "Unknown91"}, + {92, nullptr, "Unknown92"}, + {93, nullptr, "Unknown93"}, + {94, nullptr, "Unknown94"}, + {95, nullptr, "Unknown95"}, + {96, nullptr, "Unknown96"}, + {97, nullptr, "Unknown97"}, + {98, nullptr, "Unknown98"}, + {99, nullptr, "Unknown99"}, + {100, nullptr, "Unknown100"}, + {101, nullptr, "Unknown101"}, + {102, nullptr, "Unknown102"}, + {103, nullptr, "Unknown103"}, + {104, nullptr, "Unknown104"}, + {105, nullptr, "Unknown105"}, + {106, nullptr, "Unknown106"}, + {107, nullptr, "Unknown107"}, + {108, nullptr, "Unknown108"}, + {109, nullptr, "Unknown109"}, + {110, nullptr, "Unknown110"}, + {111, nullptr, "Unknown111"}, + {112, nullptr, "Unknown112"}, + {113, nullptr, "Unknown113"}, + {114, nullptr, "Unknown114"}, + {115, nullptr, "Unknown115"}, + {116, nullptr, "Unknown116"}, + {117, nullptr, "Unknown117"}, + {118, nullptr, "Unknown118"}, }; // clang-format on @@ -78,15 +210,24 @@ public: explicit NIM_ECA() : ServiceFramework{"nim:eca"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "CreateServerInterface"}, + {0, &NIM_ECA::CreateServerInterface, "CreateServerInterface"}, {1, nullptr, "RefreshDebugAvailability"}, {2, nullptr, "ClearDebugResponse"}, {3, nullptr, "RegisterDebugResponse"}, + {4, nullptr, "IsLargeResourceAvailable"}, }; // clang-format on RegisterHandlers(functions); } + +private: + void CreateServerInterface(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IShopServiceAccessServer>(); + } }; class NIM_SHP final : public ServiceFramework<NIM_SHP> { @@ -96,6 +237,8 @@ public: static const FunctionInfo functions[] = { {0, nullptr, "RequestDeviceAuthenticationToken"}, {1, nullptr, "RequestCachedDeviceAuthenticationToken"}, + {2, nullptr, "RequestEdgeToken"}, + {3, nullptr, "RequestCachedEdgeToken"}, {100, nullptr, "RequestRegisterDeviceAccount"}, {101, nullptr, "RequestUnregisterDeviceAccount"}, {102, nullptr, "RequestDeviceAccountStatus"}, @@ -113,7 +256,8 @@ public: {305, nullptr, "RequestCreateVirtualAccount"}, {306, nullptr, "RequestDeviceLinkStatus"}, {400, nullptr, "GetAccountByVirtualAccount"}, - {500, nullptr, "RequestSyncTicket"}, + {401, nullptr, "GetVirtualAccount"}, + {500, nullptr, "RequestSyncTicketLegacy"}, {501, nullptr, "RequestDownloadTicket"}, {502, nullptr, "RequestDownloadTicketForPrepurchasedContents"}, {503, nullptr, "RequestSyncTicket"}, diff --git a/src/core/hle/service/npns/npns.cpp b/src/core/hle/service/npns/npns.cpp index aa171473b..8fa16fb08 100644 --- a/src/core/hle/service/npns/npns.cpp +++ b/src/core/hle/service/npns/npns.cpp @@ -30,6 +30,7 @@ public: {23, nullptr, "DestroyToken"}, {24, nullptr, "DestroyTokenWithApplicationId"}, {25, nullptr, "QueryIsTokenValid"}, + {26, nullptr, "ListenToMyApplicationId"}, {31, nullptr, "UploadTokenToBaaS"}, {32, nullptr, "DestroyTokenForBaaS"}, {33, nullptr, "CreateTokenForBaaS"}, @@ -48,6 +49,8 @@ public: {151, nullptr, "GetStateWithHandover"}, {152, nullptr, "GetStateChangeEventWithHandover"}, {153, nullptr, "GetDropEventWithHandover"}, + {161, nullptr, "GetRequestChangeStateCancelEvent"}, + {162, nullptr, "RequestChangeStateForceTimedWithCancelEvent"}, {201, nullptr, "RequestChangeStateForceTimed"}, {202, nullptr, "RequestChangeStateForceAsync"}, }; diff --git a/src/core/hle/service/ns/ns.cpp b/src/core/hle/service/ns/ns.cpp index fdab3cf78..2594e6839 100644 --- a/src/core/hle/service/ns/ns.cpp +++ b/src/core/hle/service/ns/ns.cpp @@ -3,8 +3,10 @@ // Refer to the license.txt file included. #include "common/logging/log.h" +#include "core/core.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" +#include "core/file_sys/vfs.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/ns/errors.h" @@ -28,8 +30,8 @@ IAccountProxyInterface::IAccountProxyInterface() : ServiceFramework{"IAccountPro IAccountProxyInterface::~IAccountProxyInterface() = default; -IApplicationManagerInterface::IApplicationManagerInterface() - : ServiceFramework{"IApplicationManagerInterface"} { +IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_) + : ServiceFramework{"IApplicationManagerInterface"}, system{system_} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ListApplicationRecord"}, @@ -104,12 +106,16 @@ IApplicationManagerInterface::IApplicationManagerInterface() {94, nullptr, "LaunchApplication"}, {95, nullptr, "GetApplicationLaunchInfo"}, {96, nullptr, "AcquireApplicationLaunchInfo"}, - {97, nullptr, "GetMainApplicationProgramIndex2"}, + {97, nullptr, "GetMainApplicationProgramIndexByApplicationLaunchInfo"}, {98, nullptr, "EnableApplicationAllThreadDumpOnCrash"}, {99, nullptr, "LaunchDevMenu"}, {100, nullptr, "ResetToFactorySettings"}, {101, nullptr, "ResetToFactorySettingsWithoutUserSaveData"}, {102, nullptr, "ResetToFactorySettingsForRefurbishment"}, + {103, nullptr, "ResetToFactorySettingsWithPlatformRegion"}, + {104, nullptr, "ResetToFactorySettingsWithPlatformRegionAuthentication"}, + {105, nullptr, "RequestResetToFactorySettingsSecurely"}, + {106, nullptr, "RequestResetToFactorySettingsWithPlatformRegionAuthenticationSecurely"}, {200, nullptr, "CalculateUserSaveDataStatistics"}, {201, nullptr, "DeleteUserSaveDataAll"}, {210, nullptr, "DeleteUserSystemSaveData"}, @@ -191,6 +197,9 @@ IApplicationManagerInterface::IApplicationManagerInterface() {1307, nullptr, "TryDeleteRunningApplicationContentEntities"}, {1308, nullptr, "DeleteApplicationCompletelyForDebug"}, {1309, nullptr, "CleanupUnavailableAddOnContents"}, + {1310, nullptr, "RequestMoveApplicationEntity"}, + {1311, nullptr, "EstimateSizeToMove"}, + {1312, nullptr, "HasMovableEntity"}, {1400, nullptr, "PrepareShutdown"}, {1500, nullptr, "FormatSdCard"}, {1501, nullptr, "NeedsSystemUpdateToFormatSdCard"}, @@ -241,13 +250,13 @@ IApplicationManagerInterface::IApplicationManagerInterface() {2153, nullptr, "DeactivateRightsEnvironment"}, {2154, nullptr, "ForceActivateRightsContextForExit"}, {2155, nullptr, "UpdateRightsEnvironmentStatus"}, - {2156, nullptr, "CreateRightsEnvironmentForPreomia"}, + {2156, nullptr, "CreateRightsEnvironmentForMicroApplication"}, {2160, nullptr, "AddTargetApplicationToRightsEnvironment"}, {2161, nullptr, "SetUsersToRightsEnvironment"}, {2170, nullptr, "GetRightsEnvironmentStatus"}, {2171, nullptr, "GetRightsEnvironmentStatusChangedEvent"}, {2180, nullptr, "RequestExtendRightsInRightsEnvironment"}, - {2181, nullptr, "GetLastResultOfExtendRightsInRightsEnvironment"}, + {2181, nullptr, "GetResultOfExtendRightsInRightsEnvironment"}, {2182, nullptr, "SetActiveRightsContextUsingStateToRightsEnvironment"}, {2190, nullptr, "GetRightsEnvironmentHandleForApplication"}, {2199, nullptr, "GetRightsEnvironmentCountForDebug"}, @@ -258,6 +267,7 @@ IApplicationManagerInterface::IApplicationManagerInterface() {2350, nullptr, "PerformAutoUpdateByApplicationId"}, {2351, nullptr, "RequestNoDownloadRightsErrorResolution"}, {2352, nullptr, "RequestResolveNoDownloadRightsError"}, + {2353, nullptr, "GetApplicationDownloadTaskInfo"}, {2400, nullptr, "GetPromotionInfo"}, {2401, nullptr, "CountPromotionInfo"}, {2402, nullptr, "ListPromotionInfo"}, @@ -266,9 +276,12 @@ IApplicationManagerInterface::IApplicationManagerInterface() {2500, nullptr, "ConfirmAvailableTime"}, {2510, nullptr, "CreateApplicationResource"}, {2511, nullptr, "GetApplicationResource"}, - {2513, nullptr, "LaunchPreomia"}, + {2513, nullptr, "LaunchMicroApplication"}, {2514, nullptr, "ClearTaskOfAsyncTaskManager"}, + {2515, nullptr, "CleanupAllPlaceHolderAndFragmentsIfNoTask"}, + {2516, nullptr, "EnsureApplicationCertificate"}, {2800, nullptr, "GetApplicationIdOfPreomia"}, + {9999, nullptr, "GetApplicationCertificate"}, }; // clang-format on @@ -286,7 +299,8 @@ void IApplicationManagerInterface::GetApplicationControlData(Kernel::HLERequestC const auto size = ctx.GetWriteBufferSize(); - const FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto control = pm.GetControlMetadata(); std::vector<u8> out; @@ -355,15 +369,21 @@ ResultVal<u8> IApplicationManagerInterface::GetApplicationDesiredLanguage( LOG_DEBUG(Service_NS, "called with supported_languages={:08X}", supported_languages); // Get language code from settings - const auto language_code = Set::GetLanguageCodeFromIndex(Settings::values.language_index); + const auto language_code = + Set::GetLanguageCodeFromIndex(Settings::values.language_index.GetValue()); // Convert to application language, get priority list const auto application_language = ConvertToApplicationLanguage(language_code); if (application_language == std::nullopt) { + LOG_ERROR(Service_NS, "Could not convert application language! language_code={}", + language_code); return ERR_APPLICATION_LANGUAGE_NOT_FOUND; } const auto priority_list = GetApplicationLanguagePriorityList(*application_language); if (!priority_list) { + LOG_ERROR(Service_NS, + "Could not find application language priorities! application_language={}", + *application_language); return ERR_APPLICATION_LANGUAGE_NOT_FOUND; } @@ -375,6 +395,8 @@ ResultVal<u8> IApplicationManagerInterface::GetApplicationDesiredLanguage( } } + LOG_ERROR(Service_NS, "Could not find a valid language! supported_languages={:08X}", + supported_languages); return ERR_APPLICATION_LANGUAGE_NOT_FOUND; } @@ -399,6 +421,7 @@ ResultVal<u64> IApplicationManagerInterface::ConvertApplicationLanguageToLanguag const auto language_code = ConvertToLanguageCode(static_cast<ApplicationLanguage>(application_language)); if (language_code == std::nullopt) { + LOG_ERROR(Service_NS, "Language not found! application_language={}", application_language); return ERR_APPLICATION_LANGUAGE_NOT_FOUND; } @@ -426,8 +449,8 @@ IApplicationVersionInterface::IApplicationVersionInterface() IApplicationVersionInterface::~IApplicationVersionInterface() = default; -IContentManagerInterface::IContentManagerInterface() - : ServiceFramework{"IContentManagerInterface"} { +IContentManagementInterface::IContentManagementInterface() + : ServiceFramework{"IContentManagementInterface"} { // clang-format off static const FunctionInfo functions[] = { {11, nullptr, "CalculateApplicationOccupiedSize"}, @@ -444,7 +467,7 @@ IContentManagerInterface::IContentManagerInterface() RegisterHandlers(functions); } -IContentManagerInterface::~IContentManagerInterface() = default; +IContentManagementInterface::~IContentManagementInterface() = default; IDocumentInterface::IDocumentInterface() : ServiceFramework{"IDocumentInterface"} { // clang-format off @@ -505,6 +528,10 @@ IFactoryResetInterface::IFactoryResetInterface::IFactoryResetInterface() {100, nullptr, "ResetToFactorySettings"}, {101, nullptr, "ResetToFactorySettingsWithoutUserSaveData"}, {102, nullptr, "ResetToFactorySettingsForRefurbishment"}, + {103, nullptr, "ResetToFactorySettingsWithPlatformRegion"}, + {104, nullptr, "ResetToFactorySettingsWithPlatformRegionAuthentication"}, + {105, nullptr, "RequestResetToFactorySettingsSecurely"}, + {106, nullptr, "RequestResetToFactorySettingsWithPlatformRegionAuthenticationSecurely"}, }; // clang-format on @@ -513,16 +540,16 @@ IFactoryResetInterface::IFactoryResetInterface::IFactoryResetInterface() IFactoryResetInterface::~IFactoryResetInterface() = default; -NS::NS(const char* name) : ServiceFramework{name} { +NS::NS(const char* name, Core::System& system_) : ServiceFramework{name}, system{system_} { // clang-format off static const FunctionInfo functions[] = { {7992, &NS::PushInterface<IECommerceInterface>, "GetECommerceInterface"}, {7993, &NS::PushInterface<IApplicationVersionInterface>, "GetApplicationVersionInterface"}, {7994, &NS::PushInterface<IFactoryResetInterface>, "GetFactoryResetInterface"}, {7995, &NS::PushInterface<IAccountProxyInterface>, "GetAccountProxyInterface"}, - {7996, &NS::PushInterface<IApplicationManagerInterface>, "GetApplicationManagerInterface"}, + {7996, &NS::PushIApplicationManagerInterface, "GetApplicationManagerInterface"}, {7997, &NS::PushInterface<IDownloadTaskInterface>, "GetDownloadTaskInterface"}, - {7998, &NS::PushInterface<IContentManagerInterface>, "GetContentManagementInterface"}, + {7998, &NS::PushInterface<IContentManagementInterface>, "GetContentManagementInterface"}, {7999, &NS::PushInterface<IDocumentInterface>, "GetDocumentInterface"}, }; // clang-format on @@ -533,7 +560,7 @@ NS::NS(const char* name) : ServiceFramework{name} { NS::~NS() = default; std::shared_ptr<IApplicationManagerInterface> NS::GetApplicationManagerInterface() const { - return GetInterface<IApplicationManagerInterface>(); + return GetInterface<IApplicationManagerInterface>(system); } class NS_DEV final : public ServiceFramework<NS_DEV> { @@ -549,10 +576,13 @@ public: {6, nullptr, "TerminateApplication"}, {7, nullptr, "PrepareLaunchProgramFromHost"}, {8, nullptr, "LaunchApplication"}, - {9, nullptr, "LaunchApplicationWithStorageId"}, - {10, nullptr, "TerminateApplication2"}, - {11, nullptr, "GetRunningApplicationProcessId"}, + {9, nullptr, "LaunchApplicationWithStorageIdForDevelop"}, + {10, nullptr, "IsSystemMemoryResourceLimitBoosted"}, + {11, nullptr, "GetRunningApplicationProcessIdForDevelop"}, {12, nullptr, "SetCurrentApplicationRightsEnvironmentCanBeActive"}, + {13, nullptr, "CreateApplicationResourceForDevelop"}, + {14, nullptr, "IsPreomiaForDevelop"}, + {15, nullptr, "GetApplicationProgramIdFromHost"}, }; // clang-format on @@ -610,6 +640,10 @@ public: {9, nullptr, "GetSystemUpdateNotificationEventForContentDelivery"}, {10, nullptr, "NotifySystemUpdateForContentDelivery"}, {11, nullptr, "PrepareShutdown"}, + {12, nullptr, "Unknown12"}, + {13, nullptr, "Unknown13"}, + {14, nullptr, "Unknown14"}, + {15, nullptr, "Unknown15"}, {16, nullptr, "DestroySystemUpdateTask"}, {17, nullptr, "RequestSendSystemUpdate"}, {18, nullptr, "GetSendSystemUpdateProgress"}, @@ -646,11 +680,11 @@ public: void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { - std::make_shared<NS>("ns:am2")->InstallAsService(service_manager); - std::make_shared<NS>("ns:ec")->InstallAsService(service_manager); - std::make_shared<NS>("ns:rid")->InstallAsService(service_manager); - std::make_shared<NS>("ns:rt")->InstallAsService(service_manager); - std::make_shared<NS>("ns:web")->InstallAsService(service_manager); + std::make_shared<NS>("ns:am2", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:ec", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:rid", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:rt", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:web", system)->InstallAsService(service_manager); std::make_shared<NS_DEV>()->InstallAsService(service_manager); std::make_shared<NS_SU>()->InstallAsService(service_manager); diff --git a/src/core/hle/service/ns/ns.h b/src/core/hle/service/ns/ns.h index 13a64ad88..c90ccd755 100644 --- a/src/core/hle/service/ns/ns.h +++ b/src/core/hle/service/ns/ns.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service { namespace FileSystem { @@ -22,7 +26,7 @@ public: class IApplicationManagerInterface final : public ServiceFramework<IApplicationManagerInterface> { public: - explicit IApplicationManagerInterface(); + explicit IApplicationManagerInterface(Core::System& system_); ~IApplicationManagerInterface() override; ResultVal<u8> GetApplicationDesiredLanguage(u32 supported_languages); @@ -32,6 +36,8 @@ private: void GetApplicationControlData(Kernel::HLERequestContext& ctx); void GetApplicationDesiredLanguage(Kernel::HLERequestContext& ctx); void ConvertApplicationLanguageToLanguageCode(Kernel::HLERequestContext& ctx); + + Core::System& system; }; class IApplicationVersionInterface final : public ServiceFramework<IApplicationVersionInterface> { @@ -40,10 +46,10 @@ public: ~IApplicationVersionInterface() override; }; -class IContentManagerInterface final : public ServiceFramework<IContentManagerInterface> { +class IContentManagementInterface final : public ServiceFramework<IContentManagementInterface> { public: - explicit IContentManagerInterface(); - ~IContentManagerInterface() override; + explicit IContentManagementInterface(); + ~IContentManagementInterface() override; }; class IDocumentInterface final : public ServiceFramework<IDocumentInterface> { @@ -72,13 +78,13 @@ public: class NS final : public ServiceFramework<NS> { public: - explicit NS(const char* name); + explicit NS(const char* name, Core::System& system_); ~NS() override; std::shared_ptr<IApplicationManagerInterface> GetApplicationManagerInterface() const; private: - template <typename T> + template <typename T, typename... Args> void PushInterface(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NS, "called"); @@ -87,13 +93,23 @@ private: rb.PushIpcInterface<T>(); } - template <typename T> - std::shared_ptr<T> GetInterface() const { + void PushIApplicationManagerInterface(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_NS, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IApplicationManagerInterface>(system); + } + + template <typename T, typename... Args> + std::shared_ptr<T> GetInterface(Args&&... args) const { static_assert(std::is_base_of_v<Kernel::SessionRequestHandler, T>, "Not a base of ServiceFrameworkBase"); - return std::make_shared<T>(); + return std::make_shared<T>(std::forward<Args>(args)...); } + + Core::System& system; }; /// Registers all NS services with the specified service manager. diff --git a/src/core/hle/service/ns/pl_u.cpp b/src/core/hle/service/ns/pl_u.cpp index 8da4e52c5..5ccec2637 100644 --- a/src/core/hle/service/ns/pl_u.cpp +++ b/src/core/hle/service/ns/pl_u.cpp @@ -19,6 +19,7 @@ #include "core/file_sys/romfs.h" #include "core/file_sys/system_archive/system_archive.h" #include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/physical_memory.h" #include "core/hle/kernel/shared_memory.h" #include "core/hle/service/filesystem/filesystem.h" @@ -49,19 +50,9 @@ constexpr std::array<std::pair<FontArchives, const char*>, 7> SHARED_FONTS{ std::make_pair(FontArchives::Extension, "nintendo_ext2_003.bfttf"), }; -constexpr std::array<const char*, 7> SHARED_FONTS_TTF{ - "FontStandard.ttf", - "FontChineseSimplified.ttf", - "FontExtendedChineseSimplified.ttf", - "FontChineseTraditional.ttf", - "FontKorean.ttf", - "FontNintendoExtended.ttf", - "FontNintendoExtended2.ttf", -}; - // The below data is specific to shared font data dumped from Switch on f/w 2.2 // Virtual address and offsets/sizes likely will vary by dump -constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL}; +[[maybe_unused]] constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL}; constexpr u32 EXPECTED_RESULT{0x7f9a0218}; // What we expect the decrypted bfttf first 4 bytes to be constexpr u32 EXPECTED_MAGIC{0x36f81a1e}; // What we expect the encrypted bfttf first 4 bytes to be constexpr u64 SHARED_FONT_MEM_SIZE{0x1100000}; @@ -162,7 +153,8 @@ PL_U::PL_U(Core::System& system) {5, &PL_U::GetSharedFontInOrderOfPriority, "GetSharedFontInOrderOfPriority"}, {6, nullptr, "GetSharedFontInOrderOfPriorityForSystem"}, {100, nullptr, "RequestApplicationFunctionAuthorization"}, - {101, nullptr, "RequestApplicationFunctionAuthorizationForSystem"}, + {101, nullptr, "RequestApplicationFunctionAuthorizationByProcessId"}, + {102, nullptr, "RequestApplicationFunctionAuthorizationByApplicationId"}, {1000, nullptr, "LoadNgWordDataForPlatformRegionChina"}, {1001, nullptr, "GetNgWordDataSizeForPlatformRegionChina"}, }; @@ -265,16 +257,13 @@ void PL_U::GetSharedMemoryAddressOffset(Kernel::HLERequestContext& ctx) { void PL_U::GetSharedMemoryNativeHandle(Kernel::HLERequestContext& ctx) { // Map backing memory for the font data LOG_DEBUG(Service_NS, "called"); - system.CurrentProcess()->VMManager().MapMemoryBlock(SHARED_FONT_MEM_VADDR, impl->shared_font, 0, - SHARED_FONT_MEM_SIZE, - Kernel::MemoryState::Shared); // Create shared font memory object auto& kernel = system.Kernel(); - impl->shared_font_mem = Kernel::SharedMemory::Create( - kernel, system.CurrentProcess(), SHARED_FONT_MEM_SIZE, Kernel::MemoryPermission::ReadWrite, - Kernel::MemoryPermission::Read, SHARED_FONT_MEM_VADDR, Kernel::MemoryRegion::BASE, - "PL_U:shared_font_mem"); + impl->shared_font_mem = SharedFrom(&kernel.GetFontSharedMem()); + + std::memcpy(impl->shared_font_mem->GetPointer(), impl->shared_font->data(), + impl->shared_font->size()); IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); diff --git a/src/core/hle/service/nvdrv/devices/nvdevice.h b/src/core/hle/service/nvdrv/devices/nvdevice.h index 1b52511a5..5681599ba 100644 --- a/src/core/hle/service/nvdrv/devices/nvdevice.h +++ b/src/core/hle/service/nvdrv/devices/nvdevice.h @@ -21,27 +21,40 @@ namespace Service::Nvidia::Devices { /// implement the ioctl interface. class nvdevice { public: - explicit nvdevice(Core::System& system) : system{system} {}; + explicit nvdevice(Core::System& system) : system{system} {} virtual ~nvdevice() = default; - union Ioctl { - u32_le raw; - BitField<0, 8, u32> cmd; - BitField<8, 8, u32> group; - BitField<16, 14, u32> length; - BitField<30, 1, u32> is_in; - BitField<31, 1, u32> is_out; - }; /** - * Handles an ioctl request. + * Handles an ioctl1 request. * @param command The ioctl command id. * @param input A buffer containing the input data for the ioctl. * @param output A buffer where the output data will be written to. * @returns The result code of the ioctl. */ - virtual u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) = 0; + virtual NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) = 0; + + /** + * Handles an ioctl2 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param inline_input A buffer containing the input data for the ioctl which has been inlined. + * @param output A buffer where the output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) = 0; + + /** + * Handles an ioctl3 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param output A buffer where the output data will be written to. + * @param inline_output A buffer where the inlined output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) = 0; protected: Core::System& system; diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp index 3f7b8e670..ce615c758 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp @@ -18,11 +18,22 @@ nvdisp_disp0::nvdisp_disp0(Core::System& system, std::shared_ptr<nvmap> nvmap_de : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} nvdisp_disp0 ::~nvdisp_disp0() = default; -u32 nvdisp_disp0::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; +NvResult nvdisp_disp0::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvdisp_disp0::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvdisp_disp0::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h index 6fcdeee84..55a33b7e4 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h @@ -20,9 +20,11 @@ public: explicit nvdisp_disp0(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvdisp_disp0() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; /// Performs a screen flip, drawing the buffer pointed to by the handle. void flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, u32 stride, diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp index 195421cc0..6b062e10e 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp @@ -16,107 +16,152 @@ #include "video_core/renderer_base.h" namespace Service::Nvidia::Devices { -namespace NvErrCodes { -enum { - InvalidNmapHandle = -22, -}; -} nvhost_as_gpu::nvhost_as_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} nvhost_as_gpu::~nvhost_as_gpu() = default; -u32 nvhost_as_gpu::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocInitalizeExCommand: - return InitalizeEx(input, output); - case IoctlCommand::IocAllocateSpaceCommand: - return AllocateSpace(input, output); - case IoctlCommand::IocMapBufferExCommand: - return MapBufferEx(input, output); - case IoctlCommand::IocBindChannelCommand: - return BindChannel(input, output); - case IoctlCommand::IocGetVaRegionsCommand: - return GetVARegions(input, output); - case IoctlCommand::IocUnmapBufferCommand: - return UnmapBuffer(input, output); +NvResult nvhost_as_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 'A': + switch (command.cmd) { + case 0x1: + return BindChannel(input, output); + case 0x2: + return AllocateSpace(input, output); + case 0x3: + return FreeSpace(input, output); + case 0x5: + return UnmapBuffer(input, output); + case 0x6: + return MapBufferEx(input, output); + case 0x8: + return GetVARegions(input, output); + case 0x9: + return InitalizeEx(input, output); + case 0x14: + return Remap(input, output); + default: + break; + } + break; default: break; } - if (static_cast<IoctlCommand>(command.cmd.Value()) == IoctlCommand::IocRemapCommand) - return Remap(input, output); + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_as_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - UNIMPLEMENTED_MSG("Unimplemented ioctl command"); - return 0; +NvResult nvhost_as_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + switch (command.group) { + case 'A': + switch (command.cmd) { + case 0x8: + return GetVARegions(input, output, inline_output); + default: + break; + } + break; + default: + break; + } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_as_gpu::InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output) { IoctlInitalizeEx params{}; std::memcpy(¶ms, input.data(), input.size()); + LOG_WARNING(Service_NVDRV, "(STUBBED) called, big_page_size=0x{:X}", params.big_page_size); - return 0; + return NvResult::Success; } -u32 nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocSpace params{}; std::memcpy(¶ms, input.data(), input.size()); + LOG_DEBUG(Service_NVDRV, "called, pages={:X}, page_size={:X}, flags={:X}", params.pages, params.page_size, params.flags); - auto& gpu = system.GPU(); - const u64 size{static_cast<u64>(params.pages) * static_cast<u64>(params.page_size)}; - if (params.flags & 1) { - params.offset = gpu.MemoryManager().AllocateSpace(params.offset, size, 1); + const auto size{static_cast<u64>(params.pages) * static_cast<u64>(params.page_size)}; + if ((params.flags & AddressSpaceFlags::FixedOffset) != AddressSpaceFlags::None) { + params.offset = *system.GPU().MemoryManager().AllocateFixed(params.offset, size); } else { - params.offset = gpu.MemoryManager().AllocateSpace(size, params.align); + params.offset = system.GPU().MemoryManager().Allocate(size, params.align); + } + + auto result = NvResult::Success; + if (!params.offset) { + LOG_CRITICAL(Service_NVDRV, "allocation failed for size {}", size); + result = NvResult::InsufficientMemory; } std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return result; } -u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) { - std::size_t num_entries = input.size() / sizeof(IoctlRemapEntry); +NvResult nvhost_as_gpu::FreeSpace(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlFreeSpace params{}; + std::memcpy(¶ms, input.data(), input.size()); + + LOG_DEBUG(Service_NVDRV, "called, offset={:X}, pages={:X}, page_size={:X}", params.offset, + params.pages, params.page_size); + + system.GPU().MemoryManager().Unmap(params.offset, + static_cast<std::size_t>(params.pages) * params.page_size); + + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; +} - LOG_WARNING(Service_NVDRV, "(STUBBED) called, num_entries=0x{:X}", num_entries); +NvResult nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) { + const auto num_entries = input.size() / sizeof(IoctlRemapEntry); + LOG_DEBUG(Service_NVDRV, "called, num_entries=0x{:X}", num_entries); + + auto result = NvResult::Success; std::vector<IoctlRemapEntry> entries(num_entries); std::memcpy(entries.data(), input.data(), input.size()); - auto& gpu = system.GPU(); for (const auto& entry : entries) { - LOG_WARNING(Service_NVDRV, "remap entry, offset=0x{:X} handle=0x{:X} pages=0x{:X}", - entry.offset, entry.nvmap_handle, entry.pages); - GPUVAddr offset = static_cast<GPUVAddr>(entry.offset) << 0x10; - auto object = nvmap_dev->GetObject(entry.nvmap_handle); + LOG_DEBUG(Service_NVDRV, "remap entry, offset=0x{:X} handle=0x{:X} pages=0x{:X}", + entry.offset, entry.nvmap_handle, entry.pages); + + const auto object{nvmap_dev->GetObject(entry.nvmap_handle)}; if (!object) { - LOG_CRITICAL(Service_NVDRV, "nvmap {} is an invalid handle!", entry.nvmap_handle); - std::memcpy(output.data(), entries.data(), output.size()); - return static_cast<u32>(NvErrCodes::InvalidNmapHandle); + LOG_CRITICAL(Service_NVDRV, "invalid nvmap_handle={:X}", entry.nvmap_handle); + result = NvResult::InvalidState; + break; } - ASSERT(object->status == nvmap::Object::Status::Allocated); + const auto offset{static_cast<GPUVAddr>(entry.offset) << 0x10}; + const auto size{static_cast<u64>(entry.pages) << 0x10}; + const auto map_offset{static_cast<u64>(entry.map_offset) << 0x10}; + const auto addr{system.GPU().MemoryManager().Map(object->addr + map_offset, offset, size)}; - const u64 size = static_cast<u64>(entry.pages) << 0x10; - ASSERT(size <= object->size); - const u64 map_offset = static_cast<u64>(entry.map_offset) << 0x10; - - const GPUVAddr returned = - gpu.MemoryManager().MapBufferEx(object->addr + map_offset, offset, size); - ASSERT(returned == offset); + if (!addr) { + LOG_CRITICAL(Service_NVDRV, "map returned an invalid address!"); + result = NvResult::InvalidState; + break; + } } + std::memcpy(output.data(), entries.data(), output.size()); - return 0; + return result; } -u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) { IoctlMapBufferEx params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -126,79 +171,107 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou params.flags, params.nvmap_handle, params.buffer_offset, params.mapping_size, params.offset); - if (!params.nvmap_handle) { - return 0; + const auto object{nvmap_dev->GetObject(params.nvmap_handle)}; + if (!object) { + LOG_CRITICAL(Service_NVDRV, "invalid nvmap_handle={:X}", params.nvmap_handle); + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; } - auto object = nvmap_dev->GetObject(params.nvmap_handle); - ASSERT(object); - - // We can only map objects that have already been assigned a CPU address. - ASSERT(object->status == nvmap::Object::Status::Allocated); - - ASSERT(params.buffer_offset == 0); - // The real nvservices doesn't make a distinction between handles and ids, and // object can only have one handle and it will be the same as its id. Assert that this is the // case to prevent unexpected behavior. ASSERT(object->id == params.nvmap_handle); - auto& gpu = system.GPU(); - if (params.flags & 1) { - params.offset = gpu.MemoryManager().MapBufferEx(object->addr, params.offset, object->size); - } else { - params.offset = gpu.MemoryManager().MapBufferEx(object->addr, object->size); + u64 page_size{params.page_size}; + if (!page_size) { + page_size = object->align; + } + + if ((params.flags & AddressSpaceFlags::Remap) != AddressSpaceFlags::None) { + if (const auto buffer_map{FindBufferMap(params.offset)}; buffer_map) { + const auto cpu_addr{static_cast<VAddr>(buffer_map->CpuAddr() + params.buffer_offset)}; + const auto gpu_addr{static_cast<GPUVAddr>(params.offset + params.buffer_offset)}; + + if (!gpu.MemoryManager().Map(cpu_addr, gpu_addr, params.mapping_size)) { + LOG_CRITICAL(Service_NVDRV, + "remap failed, flags={:X}, nvmap_handle={:X}, buffer_offset={}, " + "mapping_size = {}, offset={}", + params.flags, params.nvmap_handle, params.buffer_offset, + params.mapping_size, params.offset); + + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; + } + + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; + } else { + LOG_CRITICAL(Service_NVDRV, "address not mapped offset={}", params.offset); + + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; + } } - // Create a new mapping entry for this operation. - ASSERT_MSG(buffer_mappings.find(params.offset) == buffer_mappings.end(), - "Offset is already mapped"); + // We can only map objects that have already been assigned a CPU address. + ASSERT(object->status == nvmap::Object::Status::Allocated); - BufferMapping mapping{}; - mapping.nvmap_handle = params.nvmap_handle; - mapping.offset = params.offset; - mapping.size = object->size; + const auto physical_address{object->addr + params.buffer_offset}; + u64 size{params.mapping_size}; + if (!size) { + size = object->size; + } + + const bool is_alloc{(params.flags & AddressSpaceFlags::FixedOffset) == AddressSpaceFlags::None}; + if (is_alloc) { + params.offset = gpu.MemoryManager().MapAllocate(physical_address, size, page_size); + } else { + params.offset = gpu.MemoryManager().Map(physical_address, params.offset, size); + } - buffer_mappings[params.offset] = mapping; + auto result = NvResult::Success; + if (!params.offset) { + LOG_CRITICAL(Service_NVDRV, "failed to map size={}", size); + result = NvResult::InvalidState; + } else { + AddBufferMap(params.offset, size, physical_address, is_alloc); + } std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return result; } -u32 nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { IoctlUnmapBuffer params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, offset=0x{:X}", params.offset); - const auto itr = buffer_mappings.find(params.offset); - if (itr == buffer_mappings.end()) { - LOG_WARNING(Service_NVDRV, "Tried to unmap an invalid offset 0x{:X}", params.offset); - // Hardware tests shows that unmapping an already unmapped buffer always returns successful - // and doesn't fail. - return 0; + if (const auto size{RemoveBufferMap(params.offset)}; size) { + system.GPU().MemoryManager().Unmap(params.offset, *size); + } else { + LOG_ERROR(Service_NVDRV, "invalid offset=0x{:X}", params.offset); } - params.offset = system.GPU().MemoryManager().UnmapBuffer(params.offset, itr->second.size); - buffer_mappings.erase(itr->second.offset); - std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_as_gpu::BindChannel(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::BindChannel(const std::vector<u8>& input, std::vector<u8>& output) { IoctlBindChannel params{}; std::memcpy(¶ms, input.data(), input.size()); - LOG_DEBUG(Service_NVDRV, "called, fd={:X}", params.fd); + LOG_WARNING(Service_NVDRV, "(STUBBED) called, fd={:X}", params.fd); channel = params.fd; - return 0; + return NvResult::Success; } -u32 nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output) { IoctlGetVaRegions params{}; std::memcpy(¶ms, input.data(), input.size()); + LOG_WARNING(Service_NVDRV, "(STUBBED) called, buf_addr={:X}, buf_size={:X}", params.buf_addr, params.buf_size); @@ -210,9 +283,67 @@ u32 nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& o params.regions[1].offset = 0x04000000; params.regions[1].page_size = 0x10000; params.regions[1].pages = 0x1bffff; + // TODO(ogniK): This probably can stay stubbed but should add support way way later + std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; +} + +NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + IoctlGetVaRegions params{}; + std::memcpy(¶ms, input.data(), input.size()); + + LOG_WARNING(Service_NVDRV, "(STUBBED) called, buf_addr={:X}, buf_size={:X}", params.buf_addr, + params.buf_size); + + params.buf_size = 0x30; + params.regions[0].offset = 0x04000000; + params.regions[0].page_size = 0x1000; + params.regions[0].pages = 0x3fbfff; + + params.regions[1].offset = 0x04000000; + params.regions[1].page_size = 0x10000; + params.regions[1].pages = 0x1bffff; + + // TODO(ogniK): This probably can stay stubbed but should add support way way later + + std::memcpy(output.data(), ¶ms, output.size()); + std::memcpy(inline_output.data(), ¶ms.regions, inline_output.size()); + return NvResult::Success; +} + +std::optional<nvhost_as_gpu::BufferMap> nvhost_as_gpu::FindBufferMap(GPUVAddr gpu_addr) const { + const auto end{buffer_mappings.upper_bound(gpu_addr)}; + for (auto iter{buffer_mappings.begin()}; iter != end; ++iter) { + if (gpu_addr >= iter->second.StartAddr() && gpu_addr < iter->second.EndAddr()) { + return iter->second; + } + } + + return std::nullopt; +} + +void nvhost_as_gpu::AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, + bool is_allocated) { + buffer_mappings[gpu_addr] = {gpu_addr, size, cpu_addr, is_allocated}; +} + +std::optional<std::size_t> nvhost_as_gpu::RemoveBufferMap(GPUVAddr gpu_addr) { + if (const auto iter{buffer_mappings.find(gpu_addr)}; iter != buffer_mappings.end()) { + std::size_t size{}; + + if (iter->second.IsAllocated()) { + size = iter->second.Size(); + } + + buffer_mappings.erase(iter); + + return size; + } + + return std::nullopt; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h index f79fcc065..08035fa0e 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h @@ -4,9 +4,12 @@ #pragma once +#include <map> #include <memory> -#include <unordered_map> +#include <optional> #include <vector> + +#include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" #include "core/hle/service/nvdrv/devices/nvdevice.h" @@ -15,41 +18,79 @@ namespace Service::Nvidia::Devices { class nvmap; +enum class AddressSpaceFlags : u32 { + None = 0x0, + FixedOffset = 0x1, + Remap = 0x100, +}; +DECLARE_ENUM_FLAG_OPERATORS(AddressSpaceFlags); + class nvhost_as_gpu final : public nvdevice { public: explicit nvhost_as_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvhost_as_gpu() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocInitalizeExCommand = 0x40284109, - IocAllocateSpaceCommand = 0xC0184102, - IocRemapCommand = 0x00000014, - IocMapBufferExCommand = 0xC0284106, - IocBindChannelCommand = 0x40044101, - IocGetVaRegionsCommand = 0xC0404108, - IocUnmapBufferCommand = 0xC0084105, + class BufferMap final { + public: + constexpr BufferMap() = default; + + constexpr BufferMap(GPUVAddr start_addr, std::size_t size) + : start_addr{start_addr}, end_addr{start_addr + size} {} + + constexpr BufferMap(GPUVAddr start_addr, std::size_t size, VAddr cpu_addr, + bool is_allocated) + : start_addr{start_addr}, end_addr{start_addr + size}, cpu_addr{cpu_addr}, + is_allocated{is_allocated} {} + + constexpr VAddr StartAddr() const { + return start_addr; + } + + constexpr VAddr EndAddr() const { + return end_addr; + } + + constexpr std::size_t Size() const { + return end_addr - start_addr; + } + + constexpr VAddr CpuAddr() const { + return cpu_addr; + } + + constexpr bool IsAllocated() const { + return is_allocated; + } + + private: + GPUVAddr start_addr{}; + GPUVAddr end_addr{}; + VAddr cpu_addr{}; + bool is_allocated{}; }; struct IoctlInitalizeEx { - u32_le big_page_size; // depends on GPU's available_big_page_sizes; 0=default - s32_le as_fd; // ignored; passes 0 - u32_le flags; // passes 0 - u32_le reserved; // ignored; passes 0 - u64_le unk0; - u64_le unk1; - u64_le unk2; + u32_le big_page_size{}; // depends on GPU's available_big_page_sizes; 0=default + s32_le as_fd{}; // ignored; passes 0 + u32_le flags{}; // passes 0 + u32_le reserved{}; // ignored; passes 0 + u64_le unk0{}; + u64_le unk1{}; + u64_le unk2{}; }; static_assert(sizeof(IoctlInitalizeEx) == 40, "IoctlInitalizeEx is incorrect size"); struct IoctlAllocSpace { - u32_le pages; - u32_le page_size; - u32_le flags; + u32_le pages{}; + u32_le page_size{}; + AddressSpaceFlags flags{}; INSERT_PADDING_WORDS(1); union { u64_le offset; @@ -58,74 +99,83 @@ private: }; static_assert(sizeof(IoctlAllocSpace) == 24, "IoctlInitalizeEx is incorrect size"); + struct IoctlFreeSpace { + u64_le offset{}; + u32_le pages{}; + u32_le page_size{}; + }; + static_assert(sizeof(IoctlFreeSpace) == 16, "IoctlFreeSpace is incorrect size"); + struct IoctlRemapEntry { - u16_le flags; - u16_le kind; - u32_le nvmap_handle; - u32_le map_offset; - u32_le offset; - u32_le pages; + u16_le flags{}; + u16_le kind{}; + u32_le nvmap_handle{}; + u32_le map_offset{}; + u32_le offset{}; + u32_le pages{}; }; static_assert(sizeof(IoctlRemapEntry) == 20, "IoctlRemapEntry is incorrect size"); struct IoctlMapBufferEx { - u32_le flags; // bit0: fixed_offset, bit2: cacheable - u32_le kind; // -1 is default - u32_le nvmap_handle; - u32_le page_size; // 0 means don't care - u64_le buffer_offset; - u64_le mapping_size; - u64_le offset; + AddressSpaceFlags flags{}; // bit0: fixed_offset, bit2: cacheable + u32_le kind{}; // -1 is default + u32_le nvmap_handle{}; + u32_le page_size{}; // 0 means don't care + s64_le buffer_offset{}; + u64_le mapping_size{}; + s64_le offset{}; }; static_assert(sizeof(IoctlMapBufferEx) == 40, "IoctlMapBufferEx is incorrect size"); struct IoctlUnmapBuffer { - u64_le offset; + s64_le offset{}; }; static_assert(sizeof(IoctlUnmapBuffer) == 8, "IoctlUnmapBuffer is incorrect size"); struct IoctlBindChannel { - u32_le fd; + s32_le fd{}; }; static_assert(sizeof(IoctlBindChannel) == 4, "IoctlBindChannel is incorrect size"); struct IoctlVaRegion { - u64_le offset; - u32_le page_size; + u64_le offset{}; + u32_le page_size{}; INSERT_PADDING_WORDS(1); - u64_le pages; + u64_le pages{}; }; static_assert(sizeof(IoctlVaRegion) == 24, "IoctlVaRegion is incorrect size"); struct IoctlGetVaRegions { - u64_le buf_addr; // (contained output user ptr on linux, ignored) - u32_le buf_size; // forced to 2*sizeof(struct va_region) - u32_le reserved; - IoctlVaRegion regions[2]; + u64_le buf_addr{}; // (contained output user ptr on linux, ignored) + u32_le buf_size{}; // forced to 2*sizeof(struct va_region) + u32_le reserved{}; + IoctlVaRegion regions[2]{}; }; static_assert(sizeof(IoctlGetVaRegions) == 16 + sizeof(IoctlVaRegion) * 2, "IoctlGetVaRegions is incorrect size"); - struct BufferMapping { - u64 offset; - u64 size; - u32 nvmap_handle; - }; + s32 channel{}; - /// Map containing the nvmap object mappings in GPU memory. - std::unordered_map<u64, BufferMapping> buffer_mappings; + NvResult InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output); + NvResult AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output); + NvResult Remap(const std::vector<u8>& input, std::vector<u8>& output); + NvResult MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output); + NvResult UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output); + NvResult FreeSpace(const std::vector<u8>& input, std::vector<u8>& output); + NvResult BindChannel(const std::vector<u8>& input, std::vector<u8>& output); - u32 channel{}; + NvResult GetVARegions(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetVARegions(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output); - u32 InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output); - u32 AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output); - u32 Remap(const std::vector<u8>& input, std::vector<u8>& output); - u32 MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output); - u32 UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output); - u32 BindChannel(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetVARegions(const std::vector<u8>& input, std::vector<u8>& output); + std::optional<BufferMap> FindBufferMap(GPUVAddr gpu_addr) const; + void AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, bool is_allocated); + std::optional<std::size_t> RemoveBufferMap(GPUVAddr gpu_addr); std::shared_ptr<nvmap> nvmap_dev; + + // This is expected to be ordered, therefore we must use a map, not unordered_map + std::map<GPUVAddr, BufferMap> buffer_mappings; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp index b27ee0502..d90cf90a8 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp @@ -15,45 +15,59 @@ namespace Service::Nvidia::Devices { -nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface) - : nvdevice(system), events_interface{events_interface} {} +nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface, + SyncpointManager& syncpoint_manager) + : nvdevice(system), events_interface{events_interface}, syncpoint_manager{syncpoint_manager} {} nvhost_ctrl::~nvhost_ctrl() = default; -u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocGetConfigCommand: - return NvOsGetConfigU32(input, output); - case IoctlCommand::IocCtrlEventWaitCommand: - return IocCtrlEventWait(input, output, false, ctrl); - case IoctlCommand::IocCtrlEventWaitAsyncCommand: - return IocCtrlEventWait(input, output, true, ctrl); - case IoctlCommand::IocCtrlEventRegisterCommand: - return IocCtrlEventRegister(input, output); - case IoctlCommand::IocCtrlEventUnregisterCommand: - return IocCtrlEventUnregister(input, output); - case IoctlCommand::IocCtrlEventSignalCommand: - return IocCtrlEventSignal(input, output); +NvResult nvhost_ctrl::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x1b: + return NvOsGetConfigU32(input, output); + case 0x1c: + return IocCtrlClearEventWait(input, output); + case 0x1d: + return IocCtrlEventWait(input, output, false); + case 0x1e: + return IocCtrlEventWait(input, output, true); + case 0x1f: + return IocCtrlEventRegister(input, output); + case 0x20: + return IocCtrlEventUnregister(input, output); + } + break; default: - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + break; } + + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output) { IocGetConfigParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_TRACE(Service_NVDRV, "called, setting={}!{}", params.domain_str.data(), params.param_str.data()); - return 0x30006; // Returns error on production mode + return NvResult::ConfigVarNotFound; // Returns error on production mode } -u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, - bool is_async, IoctlCtrl& ctrl) { +NvResult nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, + bool is_async) { IocCtrlEventWaitParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "syncpt_id={}, threshold={}, timeout={}, is_async={}", @@ -70,19 +84,33 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& return NvResult::BadParameter; } + if (syncpoint_manager.IsSyncpointExpired(params.syncpt_id, params.threshold)) { + params.value = syncpoint_manager.GetSyncpointMin(params.syncpt_id); + std::memcpy(output.data(), ¶ms, sizeof(params)); + return NvResult::Success; + } + + if (const auto new_value = syncpoint_manager.RefreshSyncpoint(params.syncpt_id); + syncpoint_manager.IsSyncpointExpired(params.syncpt_id, params.threshold)) { + params.value = new_value; + std::memcpy(output.data(), ¶ms, sizeof(params)); + return NvResult::Success; + } + auto event = events_interface.events[event_id]; auto& gpu = system.GPU(); + // This is mostly to take into account unimplemented features. As synced // gpu is always synced. if (!gpu.IsAsync()) { - event.writable->Signal(); + event.event.writable->Signal(); return NvResult::Success; } auto lock = gpu.LockSync(); - const u32 current_syncpoint_value = gpu.GetSyncpointValue(params.syncpt_id); + const u32 current_syncpoint_value = event.fence.value; const s32 diff = current_syncpoint_value - params.threshold; if (diff >= 0) { - event.writable->Signal(); + event.event.writable->Signal(); params.value = current_syncpoint_value; std::memcpy(output.data(), ¶ms, sizeof(params)); return NvResult::Success; @@ -109,12 +137,9 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& params.value = ((params.syncpt_id & 0xfff) << 16) | 0x10000000; } params.value |= event_id; - event.writable->Clear(); + event.event.writable->Clear(); gpu.RegisterSyncptInterrupt(params.syncpt_id, target_value); - if (!is_async && ctrl.fresh_call) { - ctrl.must_delay = true; - ctrl.timeout = params.timeout; - ctrl.event_id = event_id; + if (!is_async) { return NvResult::Timeout; } std::memcpy(output.data(), ¶ms, sizeof(params)); @@ -124,7 +149,7 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& return NvResult::BadParameter; } -u32 nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output) { IocCtrlEventRegisterParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); const u32 event_id = params.user_event_id & 0x00FF; @@ -139,7 +164,8 @@ u32 nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector< return NvResult::Success; } -u32 nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, + std::vector<u8>& output) { IocCtrlEventUnregisterParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); const u32 event_id = params.user_event_id & 0x00FF; @@ -154,24 +180,22 @@ u32 nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, std::vecto return NvResult::Success; } -u32 nvhost_ctrl::IocCtrlEventSignal(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output) { IocCtrlEventSignalParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); - // TODO(Blinkhawk): This is normally called when an NvEvents timeout on WaitSynchronization - // It is believed from RE to cancel the GPU Event. However, better research is required - u32 event_id = params.user_event_id & 0x00FF; - LOG_WARNING(Service_NVDRV, "(STUBBED) called, user_event_id: {:X}", event_id); + + u32 event_id = params.event_id & 0x00FF; + LOG_WARNING(Service_NVDRV, "cleared event wait on, event_id: {:X}", event_id); + if (event_id >= MaxNvEvents) { return NvResult::BadParameter; } if (events_interface.status[event_id] == EventState::Waiting) { - auto& gpu = system.GPU(); - if (gpu.CancelSyncptInterrupt(events_interface.assigned_syncpt[event_id], - events_interface.assigned_value[event_id])) { - events_interface.LiberateEvent(event_id); - events_interface.events[event_id].writable->Signal(); - } + events_interface.LiberateEvent(event_id); } + + syncpoint_manager.RefreshSyncpoint(events_interface.events[event_id].fence.id); + return NvResult::Success; } diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h index 9898623de..c5aa1362a 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h @@ -14,137 +14,120 @@ namespace Service::Nvidia::Devices { class nvhost_ctrl final : public nvdevice { public: - explicit nvhost_ctrl(Core::System& system, EventInterface& events_interface); + explicit nvhost_ctrl(Core::System& system, EventInterface& events_interface, + SyncpointManager& syncpoint_manager); ~nvhost_ctrl() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocSyncptReadCommand = 0xC0080014, - IocSyncptIncrCommand = 0x40040015, - IocSyncptWaitCommand = 0xC00C0016, - IocModuleMutexCommand = 0x40080017, - IocModuleRegRDWRCommand = 0xC0180018, - IocSyncptWaitexCommand = 0xC0100019, - IocSyncptReadMaxCommand = 0xC008001A, - IocGetConfigCommand = 0xC183001B, - IocCtrlEventSignalCommand = 0xC004001C, - IocCtrlEventWaitCommand = 0xC010001D, - IocCtrlEventWaitAsyncCommand = 0xC010001E, - IocCtrlEventRegisterCommand = 0xC004001F, - IocCtrlEventUnregisterCommand = 0xC0040020, - IocCtrlEventKillCommand = 0x40080021, - }; struct IocSyncptReadParams { - u32_le id; - u32_le value; + u32_le id{}; + u32_le value{}; }; static_assert(sizeof(IocSyncptReadParams) == 8, "IocSyncptReadParams is incorrect size"); struct IocSyncptIncrParams { - u32_le id; + u32_le id{}; }; static_assert(sizeof(IocSyncptIncrParams) == 4, "IocSyncptIncrParams is incorrect size"); struct IocSyncptWaitParams { - u32_le id; - u32_le thresh; - s32_le timeout; + u32_le id{}; + u32_le thresh{}; + s32_le timeout{}; }; static_assert(sizeof(IocSyncptWaitParams) == 12, "IocSyncptWaitParams is incorrect size"); struct IocModuleMutexParams { - u32_le id; - u32_le lock; // (0 = unlock and 1 = lock) + u32_le id{}; + u32_le lock{}; // (0 = unlock and 1 = lock) }; static_assert(sizeof(IocModuleMutexParams) == 8, "IocModuleMutexParams is incorrect size"); struct IocModuleRegRDWRParams { - u32_le id; - u32_le num_offsets; - u32_le block_size; - u32_le offsets; - u32_le values; - u32_le write; + u32_le id{}; + u32_le num_offsets{}; + u32_le block_size{}; + u32_le offsets{}; + u32_le values{}; + u32_le write{}; }; static_assert(sizeof(IocModuleRegRDWRParams) == 24, "IocModuleRegRDWRParams is incorrect size"); struct IocSyncptWaitexParams { - u32_le id; - u32_le thresh; - s32_le timeout; - u32_le value; + u32_le id{}; + u32_le thresh{}; + s32_le timeout{}; + u32_le value{}; }; static_assert(sizeof(IocSyncptWaitexParams) == 16, "IocSyncptWaitexParams is incorrect size"); struct IocSyncptReadMaxParams { - u32_le id; - u32_le value; + u32_le id{}; + u32_le value{}; }; static_assert(sizeof(IocSyncptReadMaxParams) == 8, "IocSyncptReadMaxParams is incorrect size"); struct IocGetConfigParams { - std::array<char, 0x41> domain_str; - std::array<char, 0x41> param_str; - std::array<char, 0x101> config_str; + std::array<char, 0x41> domain_str{}; + std::array<char, 0x41> param_str{}; + std::array<char, 0x101> config_str{}; }; static_assert(sizeof(IocGetConfigParams) == 387, "IocGetConfigParams is incorrect size"); struct IocCtrlEventSignalParams { - u32_le user_event_id; + u32_le event_id{}; }; static_assert(sizeof(IocCtrlEventSignalParams) == 4, "IocCtrlEventSignalParams is incorrect size"); struct IocCtrlEventWaitParams { - u32_le syncpt_id; - u32_le threshold; - s32_le timeout; - u32_le value; + u32_le syncpt_id{}; + u32_le threshold{}; + s32_le timeout{}; + u32_le value{}; }; static_assert(sizeof(IocCtrlEventWaitParams) == 16, "IocCtrlEventWaitParams is incorrect size"); struct IocCtrlEventWaitAsyncParams { - u32_le syncpt_id; - u32_le threshold; - u32_le timeout; - u32_le value; + u32_le syncpt_id{}; + u32_le threshold{}; + u32_le timeout{}; + u32_le value{}; }; static_assert(sizeof(IocCtrlEventWaitAsyncParams) == 16, "IocCtrlEventWaitAsyncParams is incorrect size"); struct IocCtrlEventRegisterParams { - u32_le user_event_id; + u32_le user_event_id{}; }; static_assert(sizeof(IocCtrlEventRegisterParams) == 4, "IocCtrlEventRegisterParams is incorrect size"); struct IocCtrlEventUnregisterParams { - u32_le user_event_id; + u32_le user_event_id{}; }; static_assert(sizeof(IocCtrlEventUnregisterParams) == 4, "IocCtrlEventUnregisterParams is incorrect size"); struct IocCtrlEventKill { - u64_le user_events; + u64_le user_events{}; }; static_assert(sizeof(IocCtrlEventKill) == 8, "IocCtrlEventKill is incorrect size"); - u32 NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output); - - u32 IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, bool is_async, - IoctlCtrl& ctrl); - - u32 IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output); - - u32 IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output); - - u32 IocCtrlEventSignal(const std::vector<u8>& input, std::vector<u8>& output); + NvResult NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, bool is_async); + NvResult IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output); EventInterface& events_interface; + SyncpointManager& syncpoint_manager; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp index cc2192e5c..2d7ea433c 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp @@ -15,39 +15,66 @@ namespace Service::Nvidia::Devices { nvhost_ctrl_gpu::nvhost_ctrl_gpu(Core::System& system) : nvdevice(system) {} nvhost_ctrl_gpu::~nvhost_ctrl_gpu() = default; -u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, - const std::vector<u8>& input2, std::vector<u8>& output, - std::vector<u8>& output2, IoctlCtrl& ctrl, IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocGetCharacteristicsCommand: - return GetCharacteristics(input, output, output2, version); - case IoctlCommand::IocGetTPCMasksCommand: - return GetTPCMasks(input, output); - case IoctlCommand::IocGetActiveSlotMaskCommand: - return GetActiveSlotMask(input, output); - case IoctlCommand::IocZcullGetCtxSizeCommand: - return ZCullGetCtxSize(input, output); - case IoctlCommand::IocZcullGetInfo: - return ZCullGetInfo(input, output); - case IoctlCommand::IocZbcSetTable: - return ZBCSetTable(input, output); - case IoctlCommand::IocZbcQueryTable: - return ZBCQueryTable(input, output); - case IoctlCommand::IocFlushL2: - return FlushL2(input, output); - case IoctlCommand::IocGetGpuTime: - return GetGpuTime(input, output); +NvResult nvhost_ctrl_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 'G': + switch (command.cmd) { + case 0x1: + return ZCullGetCtxSize(input, output); + case 0x2: + return ZCullGetInfo(input, output); + case 0x3: + return ZBCSetTable(input, output); + case 0x4: + return ZBCQueryTable(input, output); + case 0x5: + return GetCharacteristics(input, output); + case 0x6: + return GetTPCMasks(input, output); + case 0x7: + return FlushL2(input, output); + case 0x14: + return GetActiveSlotMask(input, output); + case 0x1c: + return GetGpuTime(input, output); + default: + break; + } + break; + } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { + switch (command.group) { + case 'G': + switch (command.cmd) { + case 0x5: + return GetCharacteristics(input, output, inline_output); + case 0x6: + return GetTPCMasks(input, output, inline_output); + default: + break; + } + break; default: - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + break; } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& output2, IoctlVersion version) { +NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, + std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlCharacteristics params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -88,31 +115,83 @@ u32 nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vecto params.gc.gr_compbit_store_base_hw = 0x0; params.gpu_characteristics_buf_size = 0xA0; params.gpu_characteristics_buf_addr = 0xdeadbeef; // Cannot be 0 (UNUSED) + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; +} - if (version == IoctlVersion::Version3) { - std::memcpy(output.data(), input.data(), output.size()); - std::memcpy(output2.data(), ¶ms.gc, output2.size()); - } else { - std::memcpy(output.data(), ¶ms, output.size()); +NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + LOG_DEBUG(Service_NVDRV, "called"); + IoctlCharacteristics params{}; + std::memcpy(¶ms, input.data(), input.size()); + params.gc.arch = 0x120; + params.gc.impl = 0xb; + params.gc.rev = 0xa1; + params.gc.num_gpc = 0x1; + params.gc.l2_cache_size = 0x40000; + params.gc.on_board_video_memory_size = 0x0; + params.gc.num_tpc_per_gpc = 0x2; + params.gc.bus_type = 0x20; + params.gc.big_page_size = 0x20000; + params.gc.compression_page_size = 0x20000; + params.gc.pde_coverage_bit_count = 0x1B; + params.gc.available_big_page_sizes = 0x30000; + params.gc.gpc_mask = 0x1; + params.gc.sm_arch_sm_version = 0x503; + params.gc.sm_arch_spa_version = 0x503; + params.gc.sm_arch_warp_count = 0x80; + params.gc.gpu_va_bit_count = 0x28; + params.gc.reserved = 0x0; + params.gc.flags = 0x55; + params.gc.twod_class = 0x902D; + params.gc.threed_class = 0xB197; + params.gc.compute_class = 0xB1C0; + params.gc.gpfifo_class = 0xB06F; + params.gc.inline_to_memory_class = 0xA140; + params.gc.dma_copy_class = 0xB0B5; + params.gc.max_fbps_count = 0x1; + params.gc.fbp_en_mask = 0x0; + params.gc.max_ltc_per_fbp = 0x2; + params.gc.max_lts_per_ltc = 0x1; + params.gc.max_tex_per_tpc = 0x0; + params.gc.max_gpc_count = 0x1; + params.gc.rop_l2_en_mask_0 = 0x21D70; + params.gc.rop_l2_en_mask_1 = 0x0; + params.gc.chipname = 0x6230326D67; + params.gc.gr_compbit_store_base_hw = 0x0; + params.gpu_characteristics_buf_size = 0xA0; + params.gpu_characteristics_buf_addr = 0xdeadbeef; // Cannot be 0 (UNUSED) + + std::memcpy(output.data(), input.data(), output.size()); + std::memcpy(inline_output.data(), ¶ms.gc, inline_output.size()); + return NvResult::Success; +} + +NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlGpuGetTpcMasksArgs params{}; + std::memcpy(¶ms, input.data(), input.size()); + LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); + if (params.mask_buffer_size != 0) { + params.tcp_mask = 3; } - return 0; + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; } -u32 nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { IoctlGpuGetTpcMasksArgs params{}; std::memcpy(¶ms, input.data(), input.size()); - LOG_INFO(Service_NVDRV, "called, mask=0x{:X}, mask_buf_addr=0x{:X}", params.mask_buf_size, - params.mask_buf_addr); - // TODO(ogniK): Confirm value on hardware - if (params.mask_buf_size) - params.tpc_mask_size = 4 * 1; // 4 * num_gpc - else - params.tpc_mask_size = 0; - std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); + if (params.mask_buffer_size != 0) { + params.tcp_mask = 3; + } + std::memcpy(output.data(), ¶ms, output.size()); + std::memcpy(inline_output.data(), ¶ms.tcp_mask, inline_output.size()); + return NvResult::Success; } -u32 nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlActiveSlotMask params{}; @@ -122,10 +201,10 @@ u32 nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector params.slot = 0x07; params.mask = 0x01; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlZcullGetCtxSize params{}; @@ -134,10 +213,10 @@ u32 nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u } params.size = 0x1; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlNvgpuGpuZcullGetInfoArgs params{}; @@ -157,48 +236,47 @@ u32 nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& params.subregion_height_align_pixels = 0x40; params.subregion_count = 0x10; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlZbcSetTable params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO(ogniK): What does this even actually do? std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlZbcQueryTable params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO : To implement properly std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlFlushL2 params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO : To implement properly std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlGetGpuTime params{}; std::memcpy(¶ms, input.data(), input.size()); - const auto ns = Core::Timing::CyclesToNs(system.CoreTiming().GetTicks()); - params.gpu_time = static_cast<u64_le>(ns.count()); + params.gpu_time = static_cast<u64_le>(system.CoreTiming().GetGlobalTimeNs().count()); std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h index 642b0a2cb..137b88238 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h @@ -16,32 +16,13 @@ public: explicit nvhost_ctrl_gpu(Core::System& system); ~nvhost_ctrl_gpu() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocGetCharacteristicsCommand = 0xC0B04705, - IocGetTPCMasksCommand = 0xC0184706, - IocGetActiveSlotMaskCommand = 0x80084714, - IocZcullGetCtxSizeCommand = 0x80044701, - IocZcullGetInfo = 0x80284702, - IocZbcSetTable = 0x402C4703, - IocZbcQueryTable = 0xC0344704, - IocFlushL2 = 0x40084707, - IocInvalICache = 0x4008470D, - IocSetMmudebugMode = 0x4008470E, - IocSetSmDebugMode = 0x4010470F, - IocWaitForPause = 0xC0084710, - IocGetTcpExceptionEnStatus = 0x80084711, - IocNumVsms = 0x80084712, - IocVsmsMapping = 0xC0044713, - IocGetErrorChannelUserData = 0xC008471B, - IocGetGpuTime = 0xC010471C, - IocGetCpuTimeCorrelationInfo = 0xC108471D, - }; - struct IoctlGpuCharacteristics { u32_le arch; // 0x120 (NVGPU_GPU_ARCH_GM200) u32_le impl; // 0xB (NVGPU_GPU_IMPL_GM20B) @@ -92,16 +73,11 @@ private: "IoctlCharacteristics is incorrect size"); struct IoctlGpuGetTpcMasksArgs { - /// [in] TPC mask buffer size reserved by userspace. Should be at least - /// sizeof(__u32) * fls(gpc_mask) to receive TPC mask for each GPC. - /// [out] full kernel buffer size - u32_le mask_buf_size; - u32_le reserved; - - /// [in] pointer to TPC mask buffer. It will receive one 32-bit TPC mask per GPC or 0 if - /// GPC is not enabled or not present. This parameter is ignored if mask_buf_size is 0. - u64_le mask_buf_addr; - u64_le tpc_mask_size; // Nintendo add this? + u32_le mask_buffer_size{}; + INSERT_PADDING_WORDS(1); + u64_le mask_buffer_address{}; + u32_le tcp_mask{}; + INSERT_PADDING_WORDS(1); }; static_assert(sizeof(IoctlGpuGetTpcMasksArgs) == 24, "IoctlGpuGetTpcMasksArgs is incorrect size"); @@ -159,20 +135,26 @@ private: static_assert(sizeof(IoctlFlushL2) == 8, "IoctlFlushL2 is incorrect size"); struct IoctlGetGpuTime { - u64_le gpu_time; + u64_le gpu_time{}; + INSERT_PADDING_WORDS(2); }; - static_assert(sizeof(IoctlGetGpuTime) == 8, "IoctlGetGpuTime is incorrect size"); - - u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& output2, IoctlVersion version); - u32 GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output); - u32 FlushL2(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output); + static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size"); + + NvResult GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output); + + NvResult GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output); + + NvResult GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output); + NvResult FlushL2(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp index f1966ac0e..af8b3d9f1 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp @@ -7,117 +7,148 @@ #include "common/logging/log.h" #include "core/core.h" #include "core/hle/service/nvdrv/devices/nvhost_gpu.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" #include "core/memory.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" namespace Service::Nvidia::Devices { -nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) - : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} +nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager) + : nvdevice(system), nvmap_dev(std::move(nvmap_dev)), syncpoint_manager{syncpoint_manager} { + channel_fence.id = syncpoint_manager.AllocateSyncpoint(); + channel_fence.value = system.GPU().GetSyncpointValue(channel_fence.id); +} + nvhost_gpu::~nvhost_gpu() = default; -u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); - case IoctlCommand::IocSetClientDataCommand: - return SetClientData(input, output); - case IoctlCommand::IocGetClientDataCommand: - return GetClientData(input, output); - case IoctlCommand::IocZCullBind: - return ZCullBind(input, output); - case IoctlCommand::IocSetErrorNotifierCommand: - return SetErrorNotifier(input, output); - case IoctlCommand::IocChannelSetPriorityCommand: - return SetChannelPriority(input, output); - case IoctlCommand::IocAllocGPFIFOEx2Command: - return AllocGPFIFOEx2(input, output); - case IoctlCommand::IocAllocObjCtxCommand: - return AllocateObjectContext(input, output); - case IoctlCommand::IocChannelGetWaitbaseCommand: - return GetWaitbase(input, output); - case IoctlCommand::IocChannelSetTimeoutCommand: - return ChannelSetTimeout(input, output); - case IoctlCommand::IocChannelSetTimeslice: - return ChannelSetTimeslice(input, output); - default: +NvResult nvhost_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x3: + return GetWaitbase(input, output); + default: + break; + } + break; + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input, output); + case 0x3: + return ChannelSetTimeout(input, output); + case 0x8: + return SubmitGPFIFOBase(input, output, false); + case 0x9: + return AllocateObjectContext(input, output); + case 0xb: + return ZCullBind(input, output); + case 0xc: + return SetErrorNotifier(input, output); + case 0xd: + return SetChannelPriority(input, output); + case 0x1a: + return AllocGPFIFOEx2(input, output); + case 0x1b: + return SubmitGPFIFOBase(input, output, true); + case 0x1d: + return ChannelSetTimeslice(input, output); + default: + break; + } + break; + case 'G': + switch (command.cmd) { + case 0x14: + return SetClientData(input, output); + case 0x15: + return GetClientData(input, output); + default: + break; + } break; } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +}; - if (command.group == NVGPU_IOCTL_MAGIC) { - if (command.cmd == NVGPU_IOCTL_CHANNEL_SUBMIT_GPFIFO) { - return SubmitGPFIFO(input, output); - } - if (command.cmd == NVGPU_IOCTL_CHANNEL_KICKOFF_PB) { - return KickoffPB(input, output, input2, version); +NvResult nvhost_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + switch (command.group) { + case 'H': + switch (command.cmd) { + case 0x1b: + return SubmitGPFIFOBase(input, inline_input, output); } + break; } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; -}; +NvResult nvhost_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} -u32 nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); nvmap_fd = params.nvmap_fd; - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SetClientData(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetClientData(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlClientData params{}; std::memcpy(¶ms, input.data(), input.size()); user_data = params.data; - return 0; + return NvResult::Success; } -u32 nvhost_gpu::GetClientData(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::GetClientData(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlClientData params{}; std::memcpy(¶ms, input.data(), input.size()); params.data = user_data; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::ZCullBind(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::ZCullBind(const std::vector<u8>& input, std::vector<u8>& output) { std::memcpy(&zcull_params, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, gpu_va={:X}, mode={:X}", zcull_params.gpu_va, zcull_params.mode); std::memcpy(output.data(), &zcull_params, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetErrorNotifier params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, offset={:X}, size={:X}, mem={:X}", params.offset, params.size, params.mem); std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output) { std::memcpy(&channel_priority, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "(STUBBED) called, priority={:X}", channel_priority); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocGpfifoEx2 params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, @@ -126,15 +157,15 @@ u32 nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& ou params.num_entries, params.flags, params.unk0, params.unk1, params.unk2, params.unk3); - auto& gpu = system.GPU(); - params.fence_out.id = assigned_syncpoints; - params.fence_out.value = gpu.GetSyncpointValue(assigned_syncpoints); - assigned_syncpoints++; + channel_fence.value = system.GPU().GetSyncpointValue(channel_fence.id); + + params.fence_out = channel_fence; + std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocObjCtx params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, class_num={:X}, flags={:X}", params.class_num, @@ -142,102 +173,149 @@ u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector< params.obj_id = 0x0; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output) { - if (input.size() < sizeof(IoctlSubmitGpfifo)) { - UNIMPLEMENTED(); +static std::vector<Tegra::CommandHeader> BuildWaitCommandList(Fence fence) { + return { + Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceValue, 1, + Tegra::SubmissionMode::Increasing), + {fence.value}, + Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceAction, 1, + Tegra::SubmissionMode::Increasing), + Tegra::GPU::FenceAction::Build(Tegra::GPU::FenceOperation::Acquire, fence.id), + }; +} + +static std::vector<Tegra::CommandHeader> BuildIncrementCommandList(Fence fence, u32 add_increment) { + std::vector<Tegra::CommandHeader> result{ + Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceValue, 1, + Tegra::SubmissionMode::Increasing), + {}}; + + for (u32 count = 0; count < add_increment; ++count) { + result.emplace_back(Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceAction, 1, + Tegra::SubmissionMode::Increasing)); + result.emplace_back( + Tegra::GPU::FenceAction::Build(Tegra::GPU::FenceOperation::Increment, fence.id)); } - IoctlSubmitGpfifo params{}; - std::memcpy(¶ms, input.data(), sizeof(IoctlSubmitGpfifo)); + + return result; +} + +static std::vector<Tegra::CommandHeader> BuildIncrementWithWfiCommandList(Fence fence, + u32 add_increment) { + std::vector<Tegra::CommandHeader> result{ + Tegra::BuildCommandHeader(Tegra::BufferMethods::WaitForInterrupt, 1, + Tegra::SubmissionMode::Increasing), + {}}; + const std::vector<Tegra::CommandHeader> increment{ + BuildIncrementCommandList(fence, add_increment)}; + + result.insert(result.end(), increment.begin(), increment.end()); + + return result; +} + +NvResult nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output, + Tegra::CommandList&& entries) { LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address, params.num_entries, params.flags.raw); - ASSERT_MSG(input.size() == sizeof(IoctlSubmitGpfifo) + - params.num_entries * sizeof(Tegra::CommandListHeader), - "Incorrect input size"); + auto& gpu = system.GPU(); - Tegra::CommandList entries(params.num_entries); - std::memcpy(entries.data(), &input[sizeof(IoctlSubmitGpfifo)], - params.num_entries * sizeof(Tegra::CommandListHeader)); + params.fence_out.id = channel_fence.id; - UNIMPLEMENTED_IF(params.flags.add_wait.Value() != 0); - UNIMPLEMENTED_IF(params.flags.add_increment.Value() != 0); + if (params.flags.add_wait.Value() && + !syncpoint_manager.IsSyncpointExpired(params.fence_out.id, params.fence_out.value)) { + gpu.PushGPUEntries(Tegra::CommandList{BuildWaitCommandList(params.fence_out)}); + } - auto& gpu = system.GPU(); - u32 current_syncpoint_value = gpu.GetSyncpointValue(params.fence_out.id); - if (params.flags.increment.Value()) { - params.fence_out.value += current_syncpoint_value; + if (params.flags.add_increment.Value() || params.flags.increment.Value()) { + const u32 increment_value = params.flags.increment.Value() ? params.fence_out.value : 0; + params.fence_out.value = syncpoint_manager.IncreaseSyncpoint( + params.fence_out.id, params.AddIncrementValue() + increment_value); } else { - params.fence_out.value = current_syncpoint_value; + params.fence_out.value = syncpoint_manager.GetSyncpointMax(params.fence_out.id); } + gpu.PushGPUEntries(std::move(entries)); + if (params.flags.add_increment.Value()) { + if (params.flags.suppress_wfi) { + gpu.PushGPUEntries(Tegra::CommandList{ + BuildIncrementCommandList(params.fence_out, params.AddIncrementValue())}); + } else { + gpu.PushGPUEntries(Tegra::CommandList{ + BuildIncrementWithWfiCommandList(params.fence_out, params.AddIncrementValue())}); + } + } + std::memcpy(output.data(), ¶ms, sizeof(IoctlSubmitGpfifo)); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::KickoffPB(const std::vector<u8>& input, std::vector<u8>& output, - const std::vector<u8>& input2, IoctlVersion version) { +NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<u8>& output, + bool kickoff) { if (input.size() < sizeof(IoctlSubmitGpfifo)) { UNIMPLEMENTED(); + return NvResult::InvalidSize; } IoctlSubmitGpfifo params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlSubmitGpfifo)); - LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address, - params.num_entries, params.flags.raw); - Tegra::CommandList entries(params.num_entries); - if (version == IoctlVersion::Version2) { - std::memcpy(entries.data(), input2.data(), - params.num_entries * sizeof(Tegra::CommandListHeader)); - } else { - system.Memory().ReadBlock(params.address, entries.data(), - params.num_entries * sizeof(Tegra::CommandListHeader)); - } - UNIMPLEMENTED_IF(params.flags.add_wait.Value() != 0); - UNIMPLEMENTED_IF(params.flags.add_increment.Value() != 0); - auto& gpu = system.GPU(); - u32 current_syncpoint_value = gpu.GetSyncpointValue(params.fence_out.id); - if (params.flags.increment.Value()) { - params.fence_out.value += current_syncpoint_value; + if (kickoff) { + system.Memory().ReadBlock(params.address, entries.command_lists.data(), + params.num_entries * sizeof(Tegra::CommandListHeader)); } else { - params.fence_out.value = current_syncpoint_value; + std::memcpy(entries.command_lists.data(), &input[sizeof(IoctlSubmitGpfifo)], + params.num_entries * sizeof(Tegra::CommandListHeader)); } - gpu.PushGPUEntries(std::move(entries)); - std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return SubmitGPFIFOImpl(params, output, std::move(entries)); +} + +NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input, + const std::vector<u8>& input_inline, + std::vector<u8>& output) { + if (input.size() < sizeof(IoctlSubmitGpfifo)) { + UNIMPLEMENTED(); + return NvResult::InvalidSize; + } + IoctlSubmitGpfifo params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlSubmitGpfifo)); + Tegra::CommandList entries(params.num_entries); + std::memcpy(entries.command_lists.data(), input_inline.data(), input_inline.size()); + return SubmitGPFIFOImpl(params, output, std::move(entries)); } -u32 nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) { IoctlGetWaitbase params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlGetWaitbase)); LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown); params.value = 0; // Seems to be hard coded at 0 std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output) { IoctlChannelSetTimeout params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlChannelSetTimeout)); LOG_INFO(Service_NVDRV, "called, timeout=0x{:X}", params.timeout); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetTimeslice params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlSetTimeslice)); LOG_INFO(Service_NVDRV, "called, timeslice=0x{:X}", params.timeslice); channel_timeslice = params.timeslice; - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h index 2ac74743f..e0298b4fe 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h @@ -11,46 +11,28 @@ #include "common/swap.h" #include "core/hle/service/nvdrv/devices/nvdevice.h" #include "core/hle/service/nvdrv/nvdata.h" +#include "video_core/dma_pusher.h" + +namespace Service::Nvidia { +class SyncpointManager; +} namespace Service::Nvidia::Devices { class nvmap; -constexpr u32 NVGPU_IOCTL_MAGIC('H'); -constexpr u32 NVGPU_IOCTL_CHANNEL_SUBMIT_GPFIFO(0x8); -constexpr u32 NVGPU_IOCTL_CHANNEL_KICKOFF_PB(0x1b); - class nvhost_gpu final : public nvdevice { public: - explicit nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); + explicit nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager); ~nvhost_gpu() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - IocAllocGPFIFOCommand = 0x40084805, - IocSetClientDataCommand = 0x40084714, - IocGetClientDataCommand = 0x80084715, - IocZCullBind = 0xc010480b, - IocSetErrorNotifierCommand = 0xC018480C, - IocChannelSetPriorityCommand = 0x4004480D, - IocEnableCommand = 0x0000480E, - IocDisableCommand = 0x0000480F, - IocPreemptCommand = 0x00004810, - IocForceResetCommand = 0x00004811, - IocEventIdControlCommand = 0x40084812, - IocGetErrorNotificationCommand = 0xC0104817, - IocAllocGPFIFOExCommand = 0x40204818, - IocAllocGPFIFOEx2Command = 0xC020481A, - IocAllocObjCtxCommand = 0xC0104809, - IocChannelGetWaitbaseCommand = 0xC0080003, - IocChannelSetTimeoutCommand = 0x40044803, - IocChannelSetTimeslice = 0xC004481D, - }; - enum class CtxObjects : u32_le { Ctx2D = 0x902D, Ctx3D = 0xB197, @@ -61,63 +43,63 @@ private: }; struct IoctlSetNvmapFD { - u32_le nvmap_fd; + s32_le nvmap_fd{}; }; static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); struct IoctlChannelSetTimeout { - u32_le timeout; + u32_le timeout{}; }; static_assert(sizeof(IoctlChannelSetTimeout) == 4, "IoctlChannelSetTimeout is incorrect size"); struct IoctlAllocGPFIFO { - u32_le num_entries; - u32_le flags; + u32_le num_entries{}; + u32_le flags{}; }; static_assert(sizeof(IoctlAllocGPFIFO) == 8, "IoctlAllocGPFIFO is incorrect size"); struct IoctlClientData { - u64_le data; + u64_le data{}; }; static_assert(sizeof(IoctlClientData) == 8, "IoctlClientData is incorrect size"); struct IoctlZCullBind { - u64_le gpu_va; - u32_le mode; // 0=global, 1=no_ctxsw, 2=separate_buffer, 3=part_of_regular_buf + u64_le gpu_va{}; + u32_le mode{}; // 0=global, 1=no_ctxsw, 2=separate_buffer, 3=part_of_regular_buf INSERT_PADDING_WORDS(1); }; static_assert(sizeof(IoctlZCullBind) == 16, "IoctlZCullBind is incorrect size"); struct IoctlSetErrorNotifier { - u64_le offset; - u64_le size; - u32_le mem; // nvmap object handle + u64_le offset{}; + u64_le size{}; + u32_le mem{}; // nvmap object handle INSERT_PADDING_WORDS(1); }; static_assert(sizeof(IoctlSetErrorNotifier) == 24, "IoctlSetErrorNotifier is incorrect size"); struct IoctlChannelSetPriority { - u32_le priority; + u32_le priority{}; }; static_assert(sizeof(IoctlChannelSetPriority) == 4, "IoctlChannelSetPriority is incorrect size"); struct IoctlSetTimeslice { - u32_le timeslice; + u32_le timeslice{}; }; static_assert(sizeof(IoctlSetTimeslice) == 4, "IoctlSetTimeslice is incorrect size"); struct IoctlEventIdControl { - u32_le cmd; // 0=disable, 1=enable, 2=clear - u32_le id; + u32_le cmd{}; // 0=disable, 1=enable, 2=clear + u32_le id{}; }; static_assert(sizeof(IoctlEventIdControl) == 8, "IoctlEventIdControl is incorrect size"); struct IoctlGetErrorNotification { - u64_le timestamp; - u32_le info32; - u16_le info16; - u16_le status; // always 0xFFFF + u64_le timestamp{}; + u32_le info32{}; + u16_le info16{}; + u16_le status{}; // always 0xFFFF }; static_assert(sizeof(IoctlGetErrorNotification) == 16, "IoctlGetErrorNotification is incorrect size"); @@ -125,80 +107,89 @@ private: static_assert(sizeof(Fence) == 8, "Fence is incorrect size"); struct IoctlAllocGpfifoEx { - u32_le num_entries; - u32_le flags; - u32_le unk0; - u32_le unk1; - u32_le unk2; - u32_le unk3; - u32_le unk4; - u32_le unk5; + u32_le num_entries{}; + u32_le flags{}; + u32_le unk0{}; + u32_le unk1{}; + u32_le unk2{}; + u32_le unk3{}; + u32_le unk4{}; + u32_le unk5{}; }; static_assert(sizeof(IoctlAllocGpfifoEx) == 32, "IoctlAllocGpfifoEx is incorrect size"); struct IoctlAllocGpfifoEx2 { - u32_le num_entries; // in - u32_le flags; // in - u32_le unk0; // in (1 works) - Fence fence_out; // out - u32_le unk1; // in - u32_le unk2; // in - u32_le unk3; // in + u32_le num_entries{}; // in + u32_le flags{}; // in + u32_le unk0{}; // in (1 works) + Fence fence_out{}; // out + u32_le unk1{}; // in + u32_le unk2{}; // in + u32_le unk3{}; // in }; static_assert(sizeof(IoctlAllocGpfifoEx2) == 32, "IoctlAllocGpfifoEx2 is incorrect size"); struct IoctlAllocObjCtx { - u32_le class_num; // 0x902D=2d, 0xB197=3d, 0xB1C0=compute, 0xA140=kepler, 0xB0B5=DMA, - // 0xB06F=channel_gpfifo - u32_le flags; - u64_le obj_id; // (ignored) used for FREE_OBJ_CTX ioctl, which is not supported + u32_le class_num{}; // 0x902D=2d, 0xB197=3d, 0xB1C0=compute, 0xA140=kepler, 0xB0B5=DMA, + // 0xB06F=channel_gpfifo + u32_le flags{}; + u64_le obj_id{}; // (ignored) used for FREE_OBJ_CTX ioctl, which is not supported }; static_assert(sizeof(IoctlAllocObjCtx) == 16, "IoctlAllocObjCtx is incorrect size"); struct IoctlSubmitGpfifo { - u64_le address; // pointer to gpfifo entry structs - u32_le num_entries; // number of fence objects being submitted + u64_le address{}; // pointer to gpfifo entry structs + u32_le num_entries{}; // number of fence objects being submitted union { u32_le raw; BitField<0, 1, u32_le> add_wait; // append a wait sync_point to the list BitField<1, 1, u32_le> add_increment; // append an increment to the list - BitField<2, 1, u32_le> new_hw_format; // Mostly ignored + BitField<2, 1, u32_le> new_hw_format; // mostly ignored + BitField<4, 1, u32_le> suppress_wfi; // suppress wait for interrupt BitField<8, 1, u32_le> increment; // increment the returned fence } flags; - Fence fence_out; // returned new fence object for others to wait on + Fence fence_out{}; // returned new fence object for others to wait on + + u32 AddIncrementValue() const { + return flags.add_increment.Value() << 1; + } }; static_assert(sizeof(IoctlSubmitGpfifo) == 16 + sizeof(Fence), "IoctlSubmitGpfifo is incorrect size"); struct IoctlGetWaitbase { - u32 unknown; // seems to be ignored? Nintendo added this - u32 value; + u32 unknown{}; // seems to be ignored? Nintendo added this + u32 value{}; }; static_assert(sizeof(IoctlGetWaitbase) == 8, "IoctlGetWaitbase is incorrect size"); - u32_le nvmap_fd{}; + s32_le nvmap_fd{}; u64_le user_data{}; IoctlZCullBind zcull_params{}; u32_le channel_priority{}; u32_le channel_timeslice{}; - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); - u32 SetClientData(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetClientData(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZCullBind(const std::vector<u8>& input, std::vector<u8>& output); - u32 SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output); - u32 SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output); - u32 AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output); - u32 AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output); - u32 SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output); - u32 KickoffPB(const std::vector<u8>& input, std::vector<u8>& output, - const std::vector<u8>& input2, IoctlVersion version); - u32 GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output); - u32 ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output); - u32 ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetClientData(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetClientData(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZCullBind(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output); + NvResult AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output); + NvResult AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output, + Tegra::CommandList&& entries); + NvResult SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<u8>& output, + bool kickoff = false); + NvResult SubmitGPFIFOBase(const std::vector<u8>& input, const std::vector<u8>& input_inline, + std::vector<u8>& output); + NvResult GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output); std::shared_ptr<nvmap> nvmap_dev; - u32 assigned_syncpoints{}; + SyncpointManager& syncpoint_manager; + Fence channel_fence; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp index bdae8b887..d8735491c 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp @@ -2,39 +2,71 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> - #include "common/assert.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/service/nvdrv/devices/nvhost_nvdec.h" +#include "video_core/memory_manager.h" +#include "video_core/renderer_base.h" namespace Service::Nvidia::Devices { -nvhost_nvdec::nvhost_nvdec(Core::System& system) : nvdevice(system) {} +nvhost_nvdec::nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) + : nvhost_nvdec_common(system, std::move(nvmap_dev)) {} nvhost_nvdec::~nvhost_nvdec() = default; -u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); +NvResult nvhost_nvdec::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x1: + return Submit(input, output); + case 0x2: + return GetSyncpoint(input, output); + case 0x3: + return GetWaitbase(input, output); + case 0x7: + return SetSubmitTimeout(input, output); + case 0x9: + return MapBuffer(input, output); + case 0xa: { + if (command.length == 0x1c) { + LOG_INFO(Service_NVDRV, "NVDEC video stream ended"); + Tegra::ChCommandHeaderList cmdlist(1); + cmdlist[0] = Tegra::ChCommandHeader{0xDEADB33F}; + system.GPU().PushCommandBuffer(cmdlist); + } + return UnmapBuffer(input, output); + } + default: + break; + } + break; + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input); + default: + break; + } + break; } - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_nvdec::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { - IoctlSetNvmapFD params{}; - std::memcpy(¶ms, input.data(), input.size()); - LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); +NvResult nvhost_nvdec::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - nvmap_fd = params.nvmap_fd; - return 0; +NvResult nvhost_nvdec::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h index cbdac8069..79b8b6de1 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h @@ -4,35 +4,21 @@ #pragma once -#include <vector> -#include "common/common_types.h" -#include "common/swap.h" -#include "core/hle/service/nvdrv/devices/nvdevice.h" +#include <memory> +#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h" namespace Service::Nvidia::Devices { -class nvhost_nvdec final : public nvdevice { +class nvhost_nvdec final : public nvhost_nvdec_common { public: - explicit nvhost_nvdec(Core::System& system); + explicit nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvhost_nvdec() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; - -private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - }; - - struct IoctlSetNvmapFD { - u32_le nvmap_fd; - }; - static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); - - u32_le nvmap_fd{}; - - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp new file mode 100644 index 000000000..b49cecb42 --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp @@ -0,0 +1,229 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <cstring> + +#include "common/assert.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "core/core.h" +#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h" +#include "core/hle/service/nvdrv/devices/nvmap.h" +#include "core/memory.h" +#include "video_core/memory_manager.h" +#include "video_core/renderer_base.h" + +namespace Service::Nvidia::Devices { + +namespace { +// Splice vectors will copy count amount of type T from the input vector into the dst vector. +template <typename T> +std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count, + std::size_t offset) { + std::memcpy(dst.data(), input.data() + offset, count * sizeof(T)); + offset += count * sizeof(T); + return offset; +} + +// Write vectors will write data to the output buffer +template <typename T> +std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) { + std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T)); + offset += src.size() * sizeof(T); + return offset; +} +} // Anonymous namespace + +nvhost_nvdec_common::nvhost_nvdec_common(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) + : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} +nvhost_nvdec_common::~nvhost_nvdec_common() = default; + +NvResult nvhost_nvdec_common::SetNVMAPfd(const std::vector<u8>& input) { + IoctlSetNvmapFD params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlSetNvmapFD)); + LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); + + nvmap_fd = params.nvmap_fd; + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlSubmit params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlSubmit)); + LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count); + + // Instantiate param buffers + std::size_t offset = sizeof(IoctlSubmit); + std::vector<CommandBuffer> command_buffers(params.cmd_buffer_count); + std::vector<Reloc> relocs(params.relocation_count); + std::vector<u32> reloc_shifts(params.relocation_count); + std::vector<SyncptIncr> syncpt_increments(params.syncpoint_count); + std::vector<SyncptIncr> wait_checks(params.syncpoint_count); + std::vector<Fence> fences(params.fence_count); + + // Splice input into their respective buffers + offset = SpliceVectors(input, command_buffers, params.cmd_buffer_count, offset); + offset = SpliceVectors(input, relocs, params.relocation_count, offset); + offset = SpliceVectors(input, reloc_shifts, params.relocation_count, offset); + offset = SpliceVectors(input, syncpt_increments, params.syncpoint_count, offset); + offset = SpliceVectors(input, wait_checks, params.syncpoint_count, offset); + offset = SpliceVectors(input, fences, params.fence_count, offset); + + // TODO(ameerj): For async gpu, utilize fences for syncpoint 'max' increment + + auto& gpu = system.GPU(); + + for (const auto& cmd_buffer : command_buffers) { + auto object = nvmap_dev->GetObject(cmd_buffer.memory_id); + ASSERT_OR_EXECUTE(object, return NvResult::InvalidState;); + const auto map = FindBufferMap(object->dma_map_addr); + if (!map) { + LOG_ERROR(Service_NVDRV, "Tried to submit an invalid offset 0x{:X} dma 0x{:X}", + object->addr, object->dma_map_addr); + return NvResult::Success; + } + Tegra::ChCommandHeaderList cmdlist(cmd_buffer.word_count); + gpu.MemoryManager().ReadBlock(map->StartAddr() + cmd_buffer.offset, cmdlist.data(), + cmdlist.size() * sizeof(u32)); + gpu.PushCommandBuffer(cmdlist); + } + + std::memcpy(output.data(), ¶ms, sizeof(IoctlSubmit)); + // Some games expect command_buffers to be written back + offset = sizeof(IoctlSubmit); + offset = WriteVectors(output, command_buffers, offset); + offset = WriteVectors(output, relocs, offset); + offset = WriteVectors(output, reloc_shifts, offset); + offset = WriteVectors(output, syncpt_increments, offset); + offset = WriteVectors(output, wait_checks, offset); + + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlGetSyncpoint params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlGetSyncpoint)); + LOG_DEBUG(Service_NVDRV, "called GetSyncpoint, id={}", params.param); + + // We found that implementing this causes deadlocks with async gpu, along with degraded + // performance. TODO: RE the nvdec async implementation + params.value = 0; + std::memcpy(output.data(), ¶ms, sizeof(IoctlGetSyncpoint)); + + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlGetWaitbase params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlGetWaitbase)); + params.value = 0; // Seems to be hard coded at 0 + std::memcpy(output.data(), ¶ms, sizeof(IoctlGetWaitbase)); + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlMapBuffer params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlMapBuffer)); + std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries); + + SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer)); + + auto& gpu = system.GPU(); + + for (auto& cmf_buff : cmd_buffer_handles) { + auto object{nvmap_dev->GetObject(cmf_buff.map_handle)}; + if (!object) { + LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle); + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; + } + if (object->dma_map_addr == 0) { + // NVDEC and VIC memory is in the 32-bit address space + // MapAllocate32 will attempt to map a lower 32-bit value in the shared gpu memory space + const GPUVAddr low_addr = gpu.MemoryManager().MapAllocate32(object->addr, object->size); + object->dma_map_addr = static_cast<u32>(low_addr); + // Ensure that the dma_map_addr is indeed in the lower 32-bit address space. + ASSERT(object->dma_map_addr == low_addr); + } + if (!object->dma_map_addr) { + LOG_ERROR(Service_NVDRV, "failed to map size={}", object->size); + } else { + cmf_buff.map_address = object->dma_map_addr; + AddBufferMap(object->dma_map_addr, object->size, object->addr, + object->status == nvmap::Object::Status::Allocated); + } + } + std::memcpy(output.data(), ¶ms, sizeof(IoctlMapBuffer)); + std::memcpy(output.data() + sizeof(IoctlMapBuffer), cmd_buffer_handles.data(), + cmd_buffer_handles.size() * sizeof(MapBufferEntry)); + + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlMapBuffer params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlMapBuffer)); + std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries); + SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer)); + + auto& gpu = system.GPU(); + + for (auto& cmf_buff : cmd_buffer_handles) { + const auto object{nvmap_dev->GetObject(cmf_buff.map_handle)}; + if (!object) { + LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle); + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; + } + if (const auto size{RemoveBufferMap(object->dma_map_addr)}; size) { + gpu.MemoryManager().Unmap(object->dma_map_addr, *size); + } else { + // This occurs quite frequently, however does not seem to impact functionality + LOG_DEBUG(Service_NVDRV, "invalid offset=0x{:X} dma=0x{:X}", object->addr, + object->dma_map_addr); + } + object->dma_map_addr = 0; + } + std::memset(output.data(), 0, output.size()); + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::SetSubmitTimeout(const std::vector<u8>& input, + std::vector<u8>& output) { + std::memcpy(&submit_timeout, input.data(), input.size()); + LOG_WARNING(Service_NVDRV, "(STUBBED) called"); + return NvResult::Success; +} + +std::optional<nvhost_nvdec_common::BufferMap> nvhost_nvdec_common::FindBufferMap( + GPUVAddr gpu_addr) const { + const auto it = std::find_if( + buffer_mappings.begin(), buffer_mappings.upper_bound(gpu_addr), [&](const auto& entry) { + return (gpu_addr >= entry.second.StartAddr() && gpu_addr < entry.second.EndAddr()); + }); + + ASSERT(it != buffer_mappings.end()); + return it->second; +} + +void nvhost_nvdec_common::AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, + bool is_allocated) { + buffer_mappings.insert_or_assign(gpu_addr, BufferMap{gpu_addr, size, cpu_addr, is_allocated}); +} + +std::optional<std::size_t> nvhost_nvdec_common::RemoveBufferMap(GPUVAddr gpu_addr) { + const auto iter{buffer_mappings.find(gpu_addr)}; + if (iter == buffer_mappings.end()) { + return std::nullopt; + } + std::size_t size = 0; + if (iter->second.IsAllocated()) { + size = iter->second.Size(); + } + buffer_mappings.erase(iter); + return size; +} + +} // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h new file mode 100644 index 000000000..86ba3a4d1 --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h @@ -0,0 +1,196 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <map> +#include <vector> +#include "common/common_types.h" +#include "common/swap.h" +#include "core/hle/service/nvdrv/devices/nvdevice.h" + +namespace Service::Nvidia::Devices { +class nvmap; + +class nvhost_nvdec_common : public nvdevice { +public: + explicit nvhost_nvdec_common(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); + ~nvhost_nvdec_common() override; + + /** + * Handles an ioctl1 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param output A buffer where the output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) = 0; + + /** + * Handles an ioctl2 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param inline_input A buffer containing the input data for the ioctl which has been inlined. + * @param output A buffer where the output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) = 0; + + /** + * Handles an ioctl3 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param output A buffer where the output data will be written to. + * @param inline_output A buffer where the inlined output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) = 0; + +protected: + class BufferMap final { + public: + constexpr BufferMap() = default; + + constexpr BufferMap(GPUVAddr start_addr, std::size_t size) + : start_addr{start_addr}, end_addr{start_addr + size} {} + + constexpr BufferMap(GPUVAddr start_addr, std::size_t size, VAddr cpu_addr, + bool is_allocated) + : start_addr{start_addr}, end_addr{start_addr + size}, cpu_addr{cpu_addr}, + is_allocated{is_allocated} {} + + constexpr VAddr StartAddr() const { + return start_addr; + } + + constexpr VAddr EndAddr() const { + return end_addr; + } + + constexpr std::size_t Size() const { + return end_addr - start_addr; + } + + constexpr VAddr CpuAddr() const { + return cpu_addr; + } + + constexpr bool IsAllocated() const { + return is_allocated; + } + + private: + GPUVAddr start_addr{}; + GPUVAddr end_addr{}; + VAddr cpu_addr{}; + bool is_allocated{}; + }; + + struct IoctlSetNvmapFD { + s32_le nvmap_fd{}; + }; + static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); + + struct IoctlSubmitCommandBuffer { + u32_le id{}; + u32_le offset{}; + u32_le count{}; + }; + static_assert(sizeof(IoctlSubmitCommandBuffer) == 0xC, + "IoctlSubmitCommandBuffer is incorrect size"); + struct IoctlSubmit { + u32_le cmd_buffer_count{}; + u32_le relocation_count{}; + u32_le syncpoint_count{}; + u32_le fence_count{}; + }; + static_assert(sizeof(IoctlSubmit) == 0x10, "IoctlSubmit has incorrect size"); + + struct CommandBuffer { + s32 memory_id{}; + u32 offset{}; + s32 word_count{}; + }; + static_assert(sizeof(CommandBuffer) == 0xC, "CommandBuffer has incorrect size"); + + struct Reloc { + s32 cmdbuffer_memory{}; + s32 cmdbuffer_offset{}; + s32 target{}; + s32 target_offset{}; + }; + static_assert(sizeof(Reloc) == 0x10, "CommandBuffer has incorrect size"); + + struct SyncptIncr { + u32 id{}; + u32 increments{}; + }; + static_assert(sizeof(SyncptIncr) == 0x8, "CommandBuffer has incorrect size"); + + struct Fence { + u32 id{}; + u32 value{}; + }; + static_assert(sizeof(Fence) == 0x8, "CommandBuffer has incorrect size"); + + struct IoctlGetSyncpoint { + // Input + u32_le param{}; + // Output + u32_le value{}; + }; + static_assert(sizeof(IoctlGetSyncpoint) == 8, "IocGetIdParams has wrong size"); + + struct IoctlGetWaitbase { + u32_le unknown{}; // seems to be ignored? Nintendo added this + u32_le value{}; + }; + static_assert(sizeof(IoctlGetWaitbase) == 0x8, "IoctlGetWaitbase is incorrect size"); + + struct IoctlMapBuffer { + u32_le num_entries{}; + u32_le data_address{}; // Ignored by the driver. + u32_le attach_host_ch_das{}; + }; + static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size"); + + struct IocGetIdParams { + // Input + u32_le param{}; + // Output + u32_le value{}; + }; + static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size"); + + // Used for mapping and unmapping command buffers + struct MapBufferEntry { + u32_le map_handle{}; + u32_le map_address{}; + }; + static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size"); + + /// Ioctl command implementations + NvResult SetNVMAPfd(const std::vector<u8>& input); + NvResult Submit(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output); + NvResult MapBuffer(const std::vector<u8>& input, std::vector<u8>& output); + NvResult UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetSubmitTimeout(const std::vector<u8>& input, std::vector<u8>& output); + + std::optional<BufferMap> FindBufferMap(GPUVAddr gpu_addr) const; + void AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, bool is_allocated); + std::optional<std::size_t> RemoveBufferMap(GPUVAddr gpu_addr); + + s32_le nvmap_fd{}; + u32_le submit_timeout{}; + std::shared_ptr<nvmap> nvmap_dev; + + // This is expected to be ordered, therefore we must use a map, not unordered_map + std::map<GPUVAddr, BufferMap> buffer_mappings; +}; +}; // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp index 96e7b7dab..2d06955c0 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp @@ -13,28 +13,44 @@ namespace Service::Nvidia::Devices { nvhost_nvjpg::nvhost_nvjpg(Core::System& system) : nvdevice(system) {} nvhost_nvjpg::~nvhost_nvjpg() = default; -u32 nvhost_nvjpg::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); +NvResult nvhost_nvjpg::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input, output); + default: + break; + } + break; + default: + break; } - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_nvjpg::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_nvjpg::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_nvjpg::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_nvjpg::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); nvmap_fd = params.nvmap_fd; - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h index 98dcac52f..43948d18d 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h @@ -16,23 +16,21 @@ public: explicit nvhost_nvjpg(Core::System& system); ~nvhost_nvjpg() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - }; - struct IoctlSetNvmapFD { - u32_le nvmap_fd; + s32_le nvmap_fd{}; }; static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); - u32_le nvmap_fd{}; + s32_le nvmap_fd{}; - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp index c695b8863..805fe86ae 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp @@ -2,39 +2,63 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> - #include "common/assert.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/service/nvdrv/devices/nvhost_vic.h" +#include "video_core/memory_manager.h" +#include "video_core/renderer_base.h" namespace Service::Nvidia::Devices { +nvhost_vic::nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) + : nvhost_nvdec_common(system, std::move(nvmap_dev)) {} -nvhost_vic::nvhost_vic(Core::System& system) : nvdevice(system) {} nvhost_vic::~nvhost_vic() = default; -u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); +NvResult nvhost_vic::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x1: + return Submit(input, output); + case 0x2: + return GetSyncpoint(input, output); + case 0x3: + return GetWaitbase(input, output); + case 0x9: + return MapBuffer(input, output); + case 0xa: + return UnmapBuffer(input, output); + default: + break; + } + break; + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input); + default: + break; + } + break; + default: + break; } - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_vic::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { - IoctlSetNvmapFD params{}; - std::memcpy(¶ms, input.data(), input.size()); - LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); +NvResult nvhost_vic::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - nvmap_fd = params.nvmap_fd; - return 0; +NvResult nvhost_vic::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.h b/src/core/hle/service/nvdrv/devices/nvhost_vic.h index bec32bea1..b2e11f4d4 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.h @@ -4,35 +4,20 @@ #pragma once -#include <vector> -#include "common/common_types.h" -#include "common/swap.h" -#include "core/hle/service/nvdrv/devices/nvdevice.h" +#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h" namespace Service::Nvidia::Devices { +class nvmap; -class nvhost_vic final : public nvdevice { +class nvhost_vic final : public nvhost_nvdec_common { public: - explicit nvhost_vic(Core::System& system); - ~nvhost_vic() override; - - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; - -private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - }; - - struct IoctlSetNvmapFD { - u32_le nvmap_fd; - }; - static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); - - u32_le nvmap_fd{}; - - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + explicit nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); + ~nvhost_vic(); + + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; }; - } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvmap.cpp b/src/core/hle/service/nvdrv/devices/nvmap.cpp index 8c742316c..4015a2740 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.cpp +++ b/src/core/hle/service/nvdrv/devices/nvmap.cpp @@ -11,16 +11,54 @@ namespace Service::Nvidia::Devices { -namespace NvErrCodes { -enum { - OperationNotPermitted = -1, - InvalidValue = -22, -}; +nvmap::nvmap(Core::System& system) : nvdevice(system) { + // Handle 0 appears to be used when remapping, so we create a placeholder empty nvmap object to + // represent this. + CreateObject(0); } -nvmap::nvmap(Core::System& system) : nvdevice(system) {} nvmap::~nvmap() = default; +NvResult nvmap::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x1: + switch (command.cmd) { + case 0x1: + return IocCreate(input, output); + case 0x3: + return IocFromId(input, output); + case 0x4: + return IocAlloc(input, output); + case 0x5: + return IocFree(input, output); + case 0x9: + return IocParam(input, output); + case 0xe: + return IocGetId(input, output); + default: + break; + } + break; + default: + break; + } + + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvmap::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvmap::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + VAddr nvmap::GetObjectAddress(u32 handle) const { auto object = GetObject(handle); ASSERT(object); @@ -28,66 +66,50 @@ VAddr nvmap::GetObjectAddress(u32 handle) const { return object->addr; } -u32 nvmap::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::Create: - return IocCreate(input, output); - case IoctlCommand::Alloc: - return IocAlloc(input, output); - case IoctlCommand::GetId: - return IocGetId(input, output); - case IoctlCommand::FromId: - return IocFromId(input, output); - case IoctlCommand::Param: - return IocParam(input, output); - case IoctlCommand::Free: - return IocFree(input, output); - } +u32 nvmap::CreateObject(u32 size) { + // Create a new nvmap object and obtain a handle to it. + auto object = std::make_shared<Object>(); + object->id = next_id++; + object->size = size; + object->status = Object::Status::Created; + object->refcount = 1; + + const u32 handle = next_handle++; + + handles.insert_or_assign(handle, std::move(object)); - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + return handle; } -u32 nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) { IocCreateParams params; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "size=0x{:08X}", params.size); if (!params.size) { LOG_ERROR(Service_NVDRV, "Size is 0"); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } - // Create a new nvmap object and obtain a handle to it. - auto object = std::make_shared<Object>(); - object->id = next_id++; - object->size = params.size; - object->status = Object::Status::Created; - object->refcount = 1; - - u32 handle = next_handle++; - handles[handle] = std::move(object); - params.handle = handle; + params.handle = CreateObject(params.size); std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { IocAllocParams params; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "called, addr={:X}", params.addr); if (!params.handle) { LOG_ERROR(Service_NVDRV, "Handle is 0"); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if ((params.align - 1) & params.align) { LOG_ERROR(Service_NVDRV, "Incorrect alignment used, alignment={:08X}", params.align); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } const u32 min_alignment = 0x1000; @@ -98,12 +120,12 @@ u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { auto object = GetObject(params.handle); if (!object) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if (object->status == Object::Status::Allocated) { LOG_ERROR(Service_NVDRV, "Object is already allocated, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::OperationNotPermitted); + return NvResult::InsufficientMemory; } object->flags = params.flags; @@ -113,10 +135,10 @@ u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { object->status = Object::Status::Allocated; std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) { IocGetIdParams params; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -124,22 +146,22 @@ u32 nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) { if (!params.handle) { LOG_ERROR(Service_NVDRV, "Handle is zero"); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } auto object = GetObject(params.handle); if (!object) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::OperationNotPermitted); + return NvResult::BadValue; } params.id = object->id; std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { IocFromIdParams params; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -149,13 +171,13 @@ u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { [&](const auto& entry) { return entry.second->id == params.id; }); if (itr == handles.end()) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } auto& object = itr->second; if (object->status != Object::Status::Allocated) { LOG_ERROR(Service_NVDRV, "Object is not allocated, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } itr->second->refcount++; @@ -164,10 +186,10 @@ u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { params.handle = itr->first; std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { enum class ParamTypes { Size = 1, Alignment = 2, Base = 3, Heap = 4, Kind = 5, Compr = 6 }; IocParamParams params; @@ -178,12 +200,12 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { auto object = GetObject(params.handle); if (!object) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if (object->status != Object::Status::Allocated) { LOG_ERROR(Service_NVDRV, "Object is not allocated, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::OperationNotPermitted); + return NvResult::BadValue; } switch (static_cast<ParamTypes>(params.param)) { @@ -205,10 +227,10 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { } std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { // TODO(Subv): These flags are unconfirmed. enum FreeFlags { Freed = 0, @@ -223,14 +245,14 @@ u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { auto itr = handles.find(params.handle); if (itr == handles.end()) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if (!itr->second->refcount) { LOG_ERROR( Service_NVDRV, "There is no references to this object. The object is already freed. handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } itr->second->refcount--; @@ -250,7 +272,7 @@ u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { handles.erase(params.handle); std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvmap.h b/src/core/hle/service/nvdrv/devices/nvmap.h index 73c2e8809..4484bd79f 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.h +++ b/src/core/hle/service/nvdrv/devices/nvmap.h @@ -19,13 +19,15 @@ public: explicit nvmap(Core::System& system); ~nvmap() override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; + /// Returns the allocated address of an nvmap object given its handle. VAddr GetObjectAddress(u32 handle) const; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; - /// Represents an nvmap object. struct Object { enum class Status { Created, Allocated }; @@ -37,6 +39,7 @@ public: VAddr addr; Status status; u32 refcount; + u32 dma_map_addr; }; std::shared_ptr<Object> GetObject(u32 handle) const { @@ -49,82 +52,76 @@ public: private: /// Id to use for the next handle that is created. - u32 next_handle = 1; + u32 next_handle = 0; /// Id to use for the next object that is created. - u32 next_id = 1; + u32 next_id = 0; /// Mapping of currently allocated handles to the objects they represent. std::unordered_map<u32, std::shared_ptr<Object>> handles; - enum class IoctlCommand : u32 { - Create = 0xC0080101, - FromId = 0xC0080103, - Alloc = 0xC0200104, - Free = 0xC0180105, - Param = 0xC00C0109, - GetId = 0xC008010E, - }; struct IocCreateParams { // Input - u32_le size; + u32_le size{}; // Output - u32_le handle; + u32_le handle{}; }; static_assert(sizeof(IocCreateParams) == 8, "IocCreateParams has wrong size"); struct IocFromIdParams { // Input - u32_le id; + u32_le id{}; // Output - u32_le handle; + u32_le handle{}; }; static_assert(sizeof(IocFromIdParams) == 8, "IocFromIdParams has wrong size"); struct IocAllocParams { // Input - u32_le handle; - u32_le heap_mask; - u32_le flags; - u32_le align; - u8 kind; + u32_le handle{}; + u32_le heap_mask{}; + u32_le flags{}; + u32_le align{}; + u8 kind{}; INSERT_PADDING_BYTES(7); - u64_le addr; + u64_le addr{}; }; static_assert(sizeof(IocAllocParams) == 32, "IocAllocParams has wrong size"); struct IocFreeParams { - u32_le handle; + u32_le handle{}; INSERT_PADDING_BYTES(4); - u64_le address; - u32_le size; - u32_le flags; + u64_le address{}; + u32_le size{}; + u32_le flags{}; }; static_assert(sizeof(IocFreeParams) == 24, "IocFreeParams has wrong size"); struct IocParamParams { // Input - u32_le handle; - u32_le param; + u32_le handle{}; + u32_le param{}; // Output - u32_le result; + u32_le result{}; }; static_assert(sizeof(IocParamParams) == 12, "IocParamParams has wrong size"); struct IocGetIdParams { // Output - u32_le id; + u32_le id{}; // Input - u32_le handle; + u32_le handle{}; }; static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size"); - u32 IocCreate(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocAlloc(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocGetId(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocFromId(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocParam(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocFree(const std::vector<u8>& input, std::vector<u8>& output); + u32 CreateObject(u32 size); + + NvResult IocCreate(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocAlloc(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocGetId(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocFromId(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocParam(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocFree(const std::vector<u8>& input, std::vector<u8>& output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/interface.cpp b/src/core/hle/service/nvdrv/interface.cpp index c8ea6c661..f6c38e853 100644 --- a/src/core/hle/service/nvdrv/interface.cpp +++ b/src/core/hle/service/nvdrv/interface.cpp @@ -23,138 +23,184 @@ void NVDRV::SignalGPUInterruptSyncpt(const u32 syncpoint_id, const u32 value) { void NVDRV::Open(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NVDRV, "called"); + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + const auto& buffer = ctx.ReadBuffer(); - std::string device_name(buffer.begin(), buffer.end()); + const std::string device_name(buffer.begin(), buffer.end()); + DeviceFD fd = nvdrv->Open(device_name); - u32 fd = nvdrv->Open(device_name); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(fd); - rb.Push<u32>(0); + rb.Push<DeviceFD>(fd); + rb.PushEnum(fd != INVALID_NVDRV_FD ? NvResult::Success : NvResult::FileOperationFailed); +} + +void NVDRV::ServiceError(Kernel::HLERequestContext& ctx, NvResult result) { + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(result); } -void NVDRV::IoctlBase(Kernel::HLERequestContext& ctx, IoctlVersion version) { +void NVDRV::Ioctl1(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - u32 fd = rp.Pop<u32>(); - u32 command = rp.Pop<u32>(); - - /// Ioctl 3 has 2 outputs, first in the input params, second is the result - std::vector<u8> output(ctx.GetWriteBufferSize(0)); - std::vector<u8> output2; - if (version == IoctlVersion::Version3) { - output2.resize((ctx.GetWriteBufferSize(1))); + const auto fd = rp.Pop<DeviceFD>(); + const auto command = rp.PopRaw<Ioctl>(); + LOG_DEBUG(Service_NVDRV, "called fd={}, ioctl=0x{:08X}", fd, command.raw); + + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; } - /// Ioctl2 has 2 inputs. It's used to pass data directly instead of providing a pointer. - /// KickOfPB uses this - auto input = ctx.ReadBuffer(0); + // Check device + std::vector<u8> output_buffer(ctx.GetWriteBufferSize(0)); + const auto input_buffer = ctx.ReadBuffer(0); - std::vector<u8> input2; - if (version == IoctlVersion::Version2) { - input2 = ctx.ReadBuffer(1); - } + const auto nv_result = nvdrv->Ioctl1(fd, command, input_buffer, output_buffer); - IoctlCtrl ctrl{}; - - u32 result = nvdrv->Ioctl(fd, command, input, input2, output, output2, ctrl, version); - - if (ctrl.must_delay) { - ctrl.fresh_call = false; - ctx.SleepClientThread("NVServices::DelayedResponse", ctrl.timeout, - [=](std::shared_ptr<Kernel::Thread> thread, - Kernel::HLERequestContext& ctx, - Kernel::ThreadWakeupReason reason) { - IoctlCtrl ctrl2{ctrl}; - std::vector<u8> tmp_output = output; - std::vector<u8> tmp_output2 = output2; - u32 result = nvdrv->Ioctl(fd, command, input, input2, tmp_output, - tmp_output2, ctrl2, version); - ctx.WriteBuffer(tmp_output, 0); - if (version == IoctlVersion::Version3) { - ctx.WriteBuffer(tmp_output2, 1); - } - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push(result); - }, - nvdrv->GetEventWriteable(ctrl.event_id)); - } else { - ctx.WriteBuffer(output); - if (version == IoctlVersion::Version3) { - ctx.WriteBuffer(output2, 1); - } + if (command.is_out != 0) { + ctx.WriteBuffer(output_buffer); } + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push(result); -} - -void NVDRV::Ioctl(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_NVDRV, "called"); - IoctlBase(ctx, IoctlVersion::Version1); + rb.PushEnum(nv_result); } void NVDRV::Ioctl2(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_NVDRV, "called"); - IoctlBase(ctx, IoctlVersion::Version2); + IPC::RequestParser rp{ctx}; + const auto fd = rp.Pop<DeviceFD>(); + const auto command = rp.PopRaw<Ioctl>(); + LOG_DEBUG(Service_NVDRV, "called fd={}, ioctl=0x{:08X}", fd, command.raw); + + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + + const auto input_buffer = ctx.ReadBuffer(0); + const auto input_inlined_buffer = ctx.ReadBuffer(1); + std::vector<u8> output_buffer(ctx.GetWriteBufferSize(0)); + + const auto nv_result = + nvdrv->Ioctl2(fd, command, input_buffer, input_inlined_buffer, output_buffer); + + if (command.is_out != 0) { + ctx.WriteBuffer(output_buffer); + } + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(nv_result); } void NVDRV::Ioctl3(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_NVDRV, "called"); - IoctlBase(ctx, IoctlVersion::Version3); + IPC::RequestParser rp{ctx}; + const auto fd = rp.Pop<DeviceFD>(); + const auto command = rp.PopRaw<Ioctl>(); + LOG_DEBUG(Service_NVDRV, "called fd={}, ioctl=0x{:08X}", fd, command.raw); + + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + + const auto input_buffer = ctx.ReadBuffer(0); + std::vector<u8> output_buffer(ctx.GetWriteBufferSize(0)); + std::vector<u8> output_buffer_inline(ctx.GetWriteBufferSize(1)); + + const auto nv_result = + nvdrv->Ioctl3(fd, command, input_buffer, output_buffer, output_buffer_inline); + + if (command.is_out != 0) { + ctx.WriteBuffer(output_buffer, 0); + ctx.WriteBuffer(output_buffer_inline, 1); + } + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(nv_result); } void NVDRV::Close(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NVDRV, "called"); - IPC::RequestParser rp{ctx}; - u32 fd = rp.Pop<u32>(); + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } - auto result = nvdrv->Close(fd); + IPC::RequestParser rp{ctx}; + const auto fd = rp.Pop<DeviceFD>(); + const auto result = nvdrv->Close(fd); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(result); + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(result); } void NVDRV::Initialize(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); + is_initialized = true; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); + rb.PushEnum(NvResult::Success); } void NVDRV::QueryEvent(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - u32 fd = rp.Pop<u32>(); - // TODO(Blinkhawk): Figure the meaning of the flag at bit 16 - u32 event_id = rp.Pop<u32>() & 0x000000FF; + const auto fd = rp.Pop<DeviceFD>(); + const auto event_id = rp.Pop<u32>() & 0x00FF; LOG_WARNING(Service_NVDRV, "(STUBBED) called, fd={:X}, event_id={:X}", fd, event_id); - IPC::ResponseBuilder rb{ctx, 3, 1}; - rb.Push(RESULT_SUCCESS); + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + + const auto nv_result = nvdrv->VerifyFD(fd); + if (nv_result != NvResult::Success) { + LOG_ERROR(Service_NVDRV, "Invalid FD specified DeviceFD={}!", fd); + ServiceError(ctx, nv_result); + return; + } + if (event_id < MaxNvEvents) { + IPC::ResponseBuilder rb{ctx, 3, 1}; + rb.Push(RESULT_SUCCESS); auto event = nvdrv->GetEvent(event_id); event->Clear(); rb.PushCopyObjects(event); - rb.Push<u32>(NvResult::Success); + rb.PushEnum(NvResult::Success); } else { - rb.Push<u32>(0); - rb.Push<u32>(NvResult::BadParameter); + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(NvResult::BadParameter); } } -void NVDRV::SetClientPID(Kernel::HLERequestContext& ctx) { +void NVDRV::SetAruid(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; pid = rp.Pop<u64>(); LOG_WARNING(Service_NVDRV, "(STUBBED) called, pid=0x{:X}", pid); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); + rb.PushEnum(NvResult::Success); } -void NVDRV::FinishInitialize(Kernel::HLERequestContext& ctx) { +void NVDRV::SetGraphicsFirmwareMemoryMarginEnabled(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2}; @@ -164,8 +210,9 @@ void NVDRV::FinishInitialize(Kernel::HLERequestContext& ctx) { void NVDRV::GetStatus(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); + rb.PushEnum(NvResult::Success); } void NVDRV::DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx) { @@ -181,19 +228,20 @@ NVDRV::NVDRV(std::shared_ptr<Module> nvdrv, const char* name) : ServiceFramework(name), nvdrv(std::move(nvdrv)) { static const FunctionInfo functions[] = { {0, &NVDRV::Open, "Open"}, - {1, &NVDRV::Ioctl, "Ioctl"}, + {1, &NVDRV::Ioctl1, "Ioctl"}, {2, &NVDRV::Close, "Close"}, {3, &NVDRV::Initialize, "Initialize"}, {4, &NVDRV::QueryEvent, "QueryEvent"}, {5, nullptr, "MapSharedMem"}, {6, &NVDRV::GetStatus, "GetStatus"}, - {7, nullptr, "ForceSetClientPID"}, - {8, &NVDRV::SetClientPID, "SetClientPID"}, + {7, nullptr, "SetAruidForTest"}, + {8, &NVDRV::SetAruid, "SetAruid"}, {9, &NVDRV::DumpGraphicsMemoryInfo, "DumpGraphicsMemoryInfo"}, {10, nullptr, "InitializeDevtools"}, {11, &NVDRV::Ioctl2, "Ioctl2"}, {12, &NVDRV::Ioctl3, "Ioctl3"}, - {13, &NVDRV::FinishInitialize, "FinishInitialize"}, + {13, &NVDRV::SetGraphicsFirmwareMemoryMarginEnabled, + "SetGraphicsFirmwareMemoryMarginEnabled"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/nvdrv/interface.h b/src/core/hle/service/nvdrv/interface.h index 9269ce00c..e05f905ae 100644 --- a/src/core/hle/service/nvdrv/interface.h +++ b/src/core/hle/service/nvdrv/interface.h @@ -23,21 +23,23 @@ public: private: void Open(Kernel::HLERequestContext& ctx); - void Ioctl(Kernel::HLERequestContext& ctx); + void Ioctl1(Kernel::HLERequestContext& ctx); void Ioctl2(Kernel::HLERequestContext& ctx); void Ioctl3(Kernel::HLERequestContext& ctx); void Close(Kernel::HLERequestContext& ctx); void Initialize(Kernel::HLERequestContext& ctx); void QueryEvent(Kernel::HLERequestContext& ctx); - void SetClientPID(Kernel::HLERequestContext& ctx); - void FinishInitialize(Kernel::HLERequestContext& ctx); + void SetAruid(Kernel::HLERequestContext& ctx); + void SetGraphicsFirmwareMemoryMarginEnabled(Kernel::HLERequestContext& ctx); void GetStatus(Kernel::HLERequestContext& ctx); void DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx); - void IoctlBase(Kernel::HLERequestContext& ctx, IoctlVersion version); + + void ServiceError(Kernel::HLERequestContext& ctx, NvResult result); std::shared_ptr<Module> nvdrv; u64 pid{}; + bool is_initialized{}; }; } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/nvdata.h b/src/core/hle/service/nvdrv/nvdata.h index 529b03471..3294bc0e7 100644 --- a/src/core/hle/service/nvdrv/nvdata.h +++ b/src/core/hle/service/nvdrv/nvdata.h @@ -1,12 +1,16 @@ #pragma once #include <array> +#include "common/bit_field.h" #include "common/common_types.h" namespace Service::Nvidia { constexpr u32 MaxSyncPoints = 192; constexpr u32 MaxNvEvents = 64; +using DeviceFD = s32; + +constexpr DeviceFD INVALID_NVDRV_FD = -1; struct Fence { s32 id; @@ -20,11 +24,61 @@ struct MultiFence { std::array<Fence, 4> fences; }; -enum NvResult : u32 { - Success = 0, - BadParameter = 4, - Timeout = 5, - ResourceError = 15, +enum class NvResult : u32 { + Success = 0x0, + NotImplemented = 0x1, + NotSupported = 0x2, + NotInitialized = 0x3, + BadParameter = 0x4, + Timeout = 0x5, + InsufficientMemory = 0x6, + ReadOnlyAttribute = 0x7, + InvalidState = 0x8, + InvalidAddress = 0x9, + InvalidSize = 0xA, + BadValue = 0xB, + AlreadyAllocated = 0xD, + Busy = 0xE, + ResourceError = 0xF, + CountMismatch = 0x10, + OverFlow = 0x11, + InsufficientTransferMemory = 0x1000, + InsufficientVideoMemory = 0x10000, + BadSurfaceColorScheme = 0x10001, + InvalidSurface = 0x10002, + SurfaceNotSupported = 0x10003, + DispInitFailed = 0x20000, + DispAlreadyAttached = 0x20001, + DispTooManyDisplays = 0x20002, + DispNoDisplaysAttached = 0x20003, + DispModeNotSupported = 0x20004, + DispNotFound = 0x20005, + DispAttachDissallowed = 0x20006, + DispTypeNotSupported = 0x20007, + DispAuthenticationFailed = 0x20008, + DispNotAttached = 0x20009, + DispSamePwrState = 0x2000A, + DispEdidFailure = 0x2000B, + DispDsiReadAckError = 0x2000C, + DispDsiReadInvalidResp = 0x2000D, + FileWriteFailed = 0x30000, + FileReadFailed = 0x30001, + EndOfFile = 0x30002, + FileOperationFailed = 0x30003, + DirOperationFailed = 0x30004, + EndOfDirList = 0x30005, + ConfigVarNotFound = 0x30006, + InvalidConfigVar = 0x30007, + LibraryNotFound = 0x30008, + SymbolNotFound = 0x30009, + MemoryMapFailed = 0x3000A, + IoctlFailed = 0x3000F, + AccessDenied = 0x30010, + DeviceNotFound = 0x30011, + KernelDriverNotFound = 0x30012, + FileNotFound = 0x30013, + PathAlreadyExists = 0x30014, + ModuleNotPresent = 0xA000E, }; enum class EventState { @@ -34,21 +88,13 @@ enum class EventState { Busy = 3, }; -enum class IoctlVersion : u32 { - Version1, - Version2, - Version3, -}; - -struct IoctlCtrl { - // First call done to the servioce for services that call itself again after a call. - bool fresh_call{true}; - // Tells the Ioctl Wrapper that it must delay the IPC response and send the thread to sleep - bool must_delay{}; - // Timeout for the delay - s64 timeout{}; - // NV Event Id - s32 event_id{-1}; +union Ioctl { + u32_le raw; + BitField<0, 8, u32> cmd; + BitField<8, 8, u32> group; + BitField<16, 14, u32> length; + BitField<30, 1, u32> is_in; + BitField<31, 1, u32> is_out; }; } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/nvdrv.cpp b/src/core/hle/service/nvdrv/nvdrv.cpp index 197c77db0..bdbbedd0d 100644 --- a/src/core/hle/service/nvdrv/nvdrv.cpp +++ b/src/core/hle/service/nvdrv/nvdrv.cpp @@ -5,6 +5,7 @@ #include <utility> #include <fmt/format.h> +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/writable_event.h" @@ -21,6 +22,7 @@ #include "core/hle/service/nvdrv/interface.h" #include "core/hle/service/nvdrv/nvdrv.h" #include "core/hle/service/nvdrv/nvmemp.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" #include "core/hle/service/nvflinger/nvflinger.h" namespace Service::Nvidia { @@ -36,58 +38,125 @@ void InstallInterfaces(SM::ServiceManager& service_manager, NVFlinger::NVFlinger nvflinger.SetNVDrvInstance(module_); } -Module::Module(Core::System& system) { +Module::Module(Core::System& system) : syncpoint_manager{system.GPU()} { auto& kernel = system.Kernel(); for (u32 i = 0; i < MaxNvEvents; i++) { std::string event_label = fmt::format("NVDRV::NvEvent_{}", i); - events_interface.events[i] = Kernel::WritableEvent::CreateEventPair(kernel, event_label); + events_interface.events[i] = {Kernel::WritableEvent::CreateEventPair(kernel, event_label)}; events_interface.status[i] = EventState::Free; events_interface.registered[i] = false; } auto nvmap_dev = std::make_shared<Devices::nvmap>(system); devices["/dev/nvhost-as-gpu"] = std::make_shared<Devices::nvhost_as_gpu>(system, nvmap_dev); - devices["/dev/nvhost-gpu"] = std::make_shared<Devices::nvhost_gpu>(system, nvmap_dev); + devices["/dev/nvhost-gpu"] = + std::make_shared<Devices::nvhost_gpu>(system, nvmap_dev, syncpoint_manager); devices["/dev/nvhost-ctrl-gpu"] = std::make_shared<Devices::nvhost_ctrl_gpu>(system); devices["/dev/nvmap"] = nvmap_dev; devices["/dev/nvdisp_disp0"] = std::make_shared<Devices::nvdisp_disp0>(system, nvmap_dev); - devices["/dev/nvhost-ctrl"] = std::make_shared<Devices::nvhost_ctrl>(system, events_interface); - devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>(system); + devices["/dev/nvhost-ctrl"] = + std::make_shared<Devices::nvhost_ctrl>(system, events_interface, syncpoint_manager); + devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>(system, nvmap_dev); devices["/dev/nvhost-nvjpg"] = std::make_shared<Devices::nvhost_nvjpg>(system); - devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>(system); + devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>(system, nvmap_dev); } Module::~Module() = default; -u32 Module::Open(const std::string& device_name) { - ASSERT_MSG(devices.find(device_name) != devices.end(), "Trying to open unknown device {}", - device_name); +NvResult Module::VerifyFD(DeviceFD fd) const { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + if (open_files.find(fd) == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return NvResult::Success; +} + +DeviceFD Module::Open(const std::string& device_name) { + if (devices.find(device_name) == devices.end()) { + LOG_ERROR(Service_NVDRV, "Trying to open unknown device {}", device_name); + return INVALID_NVDRV_FD; + } auto device = devices[device_name]; - const u32 fd = next_fd++; + const DeviceFD fd = next_fd++; open_files[fd] = std::move(device); return fd; } -u32 Module::Ioctl(u32 fd, u32 command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - auto itr = open_files.find(fd); - ASSERT_MSG(itr != open_files.end(), "Tried to talk to an invalid device"); +NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } - auto& device = itr->second; - return device->ioctl({command}, input, input2, output, output2, ctrl, version); + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return itr->second->Ioctl1(command, input, output); } -ResultCode Module::Close(u32 fd) { - auto itr = open_files.find(fd); - ASSERT_MSG(itr != open_files.end(), "Tried to talk to an invalid device"); +NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return itr->second->Ioctl2(command, input, inline_input, output); +} + +NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return itr->second->Ioctl3(command, input, output, inline_output); +} + +NvResult Module::Close(DeviceFD fd) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } open_files.erase(itr); - // TODO(flerovium): return correct result code if operation failed. - return RESULT_SUCCESS; + return NvResult::Success; } void Module::SignalSyncpt(const u32 syncpoint_id, const u32 value) { @@ -95,17 +164,17 @@ void Module::SignalSyncpt(const u32 syncpoint_id, const u32 value) { if (events_interface.assigned_syncpt[i] == syncpoint_id && events_interface.assigned_value[i] == value) { events_interface.LiberateEvent(i); - events_interface.events[i].writable->Signal(); + events_interface.events[i].event.writable->Signal(); } } } std::shared_ptr<Kernel::ReadableEvent> Module::GetEvent(const u32 event_id) const { - return events_interface.events[event_id].readable; + return events_interface.events[event_id].event.readable; } std::shared_ptr<Kernel::WritableEvent> Module::GetEventWriteable(const u32 event_id) const { - return events_interface.events[event_id].writable; + return events_interface.events[event_id].event.writable; } } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/nvdrv.h b/src/core/hle/service/nvdrv/nvdrv.h index d7a1bef91..7654bb026 100644 --- a/src/core/hle/service/nvdrv/nvdrv.h +++ b/src/core/hle/service/nvdrv/nvdrv.h @@ -10,6 +10,7 @@ #include "common/common_types.h" #include "core/hle/kernel/writable_event.h" #include "core/hle/service/nvdrv/nvdata.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" #include "core/hle/service/service.h" namespace Core { @@ -22,15 +23,23 @@ class NVFlinger; namespace Service::Nvidia { +class SyncpointManager; + namespace Devices { class nvdevice; } +/// Represents an Nvidia event +struct NvEvent { + Kernel::EventPair event; + Fence fence{}; +}; + struct EventInterface { // Mask representing currently busy events u64 events_mask{}; // Each kernel event associated to an NV event - std::array<Kernel::EventPair, MaxNvEvents> events; + std::array<NvEvent, MaxNvEvents> events; // The status of the current NVEvent std::array<EventState, MaxNvEvents> status{}; // Tells if an NVEvent is registered or not @@ -54,7 +63,7 @@ struct EventInterface { } mask = mask >> 1; } - return {}; + return std::nullopt; } void SetEventStatus(const u32 event_id, EventState new_status) { EventState old_status = status[event_id]; @@ -103,14 +112,23 @@ public: return std::static_pointer_cast<T>(itr->second); } + NvResult VerifyFD(DeviceFD fd) const; + /// Opens a device node and returns a file descriptor to it. - u32 Open(const std::string& device_name); + DeviceFD Open(const std::string& device_name); + /// Sends an ioctl command to the specified file descriptor. - u32 Ioctl(u32 fd, u32 command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version); + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output); + + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output); + + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output); + /// Closes a device file descriptor and returns operation success. - ResultCode Close(u32 fd); + NvResult Close(DeviceFD fd); void SignalSyncpt(const u32 syncpoint_id, const u32 value); @@ -119,11 +137,14 @@ public: std::shared_ptr<Kernel::WritableEvent> GetEventWriteable(u32 event_id) const; private: + /// Manages syncpoints on the host + SyncpointManager syncpoint_manager; + /// Id to use for the next open file descriptor. - u32 next_fd = 1; + DeviceFD next_fd = 1; /// Mapping of file descriptors to the devices they reference. - std::unordered_map<u32, std::shared_ptr<Devices::nvdevice>> open_files; + std::unordered_map<DeviceFD, std::shared_ptr<Devices::nvdevice>> open_files; /// Mapping of device node names to their implementation. std::unordered_map<std::string, std::shared_ptr<Devices::nvdevice>> devices; diff --git a/src/core/hle/service/nvdrv/nvmemp.cpp b/src/core/hle/service/nvdrv/nvmemp.cpp index b7b8b7a1b..73b37e805 100644 --- a/src/core/hle/service/nvdrv/nvmemp.cpp +++ b/src/core/hle/service/nvdrv/nvmemp.cpp @@ -10,19 +10,19 @@ namespace Service::Nvidia { NVMEMP::NVMEMP() : ServiceFramework("nvmemp") { static const FunctionInfo functions[] = { - {0, &NVMEMP::Cmd0, "Cmd0"}, - {1, &NVMEMP::Cmd1, "Cmd1"}, + {0, &NVMEMP::Open, "Open"}, + {1, &NVMEMP::GetAruid, "GetAruid"}, }; RegisterHandlers(functions); } NVMEMP::~NVMEMP() = default; -void NVMEMP::Cmd0(Kernel::HLERequestContext& ctx) { +void NVMEMP::Open(Kernel::HLERequestContext& ctx) { UNIMPLEMENTED(); } -void NVMEMP::Cmd1(Kernel::HLERequestContext& ctx) { +void NVMEMP::GetAruid(Kernel::HLERequestContext& ctx) { UNIMPLEMENTED(); } diff --git a/src/core/hle/service/nvdrv/nvmemp.h b/src/core/hle/service/nvdrv/nvmemp.h index 6eafb1346..c453ee4db 100644 --- a/src/core/hle/service/nvdrv/nvmemp.h +++ b/src/core/hle/service/nvdrv/nvmemp.h @@ -14,8 +14,8 @@ public: ~NVMEMP() override; private: - void Cmd0(Kernel::HLERequestContext& ctx); - void Cmd1(Kernel::HLERequestContext& ctx); + void Open(Kernel::HLERequestContext& ctx); + void GetAruid(Kernel::HLERequestContext& ctx); }; } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/syncpoint_manager.cpp b/src/core/hle/service/nvdrv/syncpoint_manager.cpp new file mode 100644 index 000000000..0151a03b7 --- /dev/null +++ b/src/core/hle/service/nvdrv/syncpoint_manager.cpp @@ -0,0 +1,39 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" +#include "video_core/gpu.h" + +namespace Service::Nvidia { + +SyncpointManager::SyncpointManager(Tegra::GPU& gpu) : gpu{gpu} {} + +SyncpointManager::~SyncpointManager() = default; + +u32 SyncpointManager::RefreshSyncpoint(u32 syncpoint_id) { + syncpoints[syncpoint_id].min = gpu.GetSyncpointValue(syncpoint_id); + return GetSyncpointMin(syncpoint_id); +} + +u32 SyncpointManager::AllocateSyncpoint() { + for (u32 syncpoint_id = 1; syncpoint_id < MaxSyncPoints; syncpoint_id++) { + if (!syncpoints[syncpoint_id].is_allocated) { + syncpoints[syncpoint_id].is_allocated = true; + return syncpoint_id; + } + } + UNREACHABLE_MSG("No more available syncpoints!"); + return {}; +} + +u32 SyncpointManager::IncreaseSyncpoint(u32 syncpoint_id, u32 value) { + for (u32 index = 0; index < value; ++index) { + syncpoints[syncpoint_id].max.fetch_add(1, std::memory_order_relaxed); + } + + return GetSyncpointMax(syncpoint_id); +} + +} // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/syncpoint_manager.h b/src/core/hle/service/nvdrv/syncpoint_manager.h new file mode 100644 index 000000000..4168b6c7e --- /dev/null +++ b/src/core/hle/service/nvdrv/syncpoint_manager.h @@ -0,0 +1,85 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <atomic> + +#include "common/common_types.h" +#include "core/hle/service/nvdrv/nvdata.h" + +namespace Tegra { +class GPU; +} + +namespace Service::Nvidia { + +class SyncpointManager final { +public: + explicit SyncpointManager(Tegra::GPU& gpu); + ~SyncpointManager(); + + /** + * Returns true if the specified syncpoint is expired for the given value. + * @param syncpoint_id Syncpoint ID to check. + * @param value Value to check against the specified syncpoint. + * @returns True if the specified syncpoint is expired for the given value, otherwise False. + */ + bool IsSyncpointExpired(u32 syncpoint_id, u32 value) const { + return (GetSyncpointMax(syncpoint_id) - value) >= (GetSyncpointMin(syncpoint_id) - value); + } + + /** + * Gets the lower bound for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to get the lower bound for. + * @returns The lower bound for the specified syncpoint. + */ + u32 GetSyncpointMin(u32 syncpoint_id) const { + return syncpoints[syncpoint_id].min.load(std::memory_order_relaxed); + } + + /** + * Gets the uper bound for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to get the upper bound for. + * @returns The upper bound for the specified syncpoint. + */ + u32 GetSyncpointMax(u32 syncpoint_id) const { + return syncpoints[syncpoint_id].max.load(std::memory_order_relaxed); + } + + /** + * Refreshes the minimum value for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to be refreshed. + * @returns The new syncpoint minimum value. + */ + u32 RefreshSyncpoint(u32 syncpoint_id); + + /** + * Allocates a new syncoint. + * @returns The syncpoint ID for the newly allocated syncpoint. + */ + u32 AllocateSyncpoint(); + + /** + * Increases the maximum value for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to be increased. + * @param value Value to increase the specified syncpoint by. + * @returns The new syncpoint maximum value. + */ + u32 IncreaseSyncpoint(u32 syncpoint_id, u32 value); + +private: + struct Syncpoint { + std::atomic<u32> min; + std::atomic<u32> max; + std::atomic<bool> is_allocated; + }; + + std::array<Syncpoint, MaxSyncPoints> syncpoints{}; + + Tegra::GPU& gpu; +}; + +} // namespace Service::Nvidia diff --git a/src/core/hle/service/nvflinger/buffer_queue.cpp b/src/core/hle/service/nvflinger/buffer_queue.cpp index f1e3d832a..b89a2d41b 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.cpp +++ b/src/core/hle/service/nvflinger/buffer_queue.cpp @@ -24,13 +24,17 @@ BufferQueue::~BufferQueue() = default; void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer) { LOG_WARNING(Service, "Adding graphics buffer {}", slot); - Buffer buffer{}; - buffer.slot = slot; - buffer.igbp_buffer = igbp_buffer; - buffer.status = Buffer::Status::Free; free_buffers.push_back(slot); + queue.push_back({ + .slot = slot, + .status = Buffer::Status::Free, + .igbp_buffer = igbp_buffer, + .transform = {}, + .crop_rect = {}, + .swap_interval = 0, + .multi_fence = {}, + }); - queue.emplace_back(buffer); buffer_wait_event.writable->Signal(); } @@ -38,7 +42,7 @@ std::optional<std::pair<u32, Service::Nvidia::MultiFence*>> BufferQueue::Dequeue u32 height) { if (free_buffers.empty()) { - return {}; + return std::nullopt; } auto f_itr = free_buffers.begin(); @@ -69,7 +73,7 @@ std::optional<std::pair<u32, Service::Nvidia::MultiFence*>> BufferQueue::Dequeue } if (itr == queue.end()) { - return {}; + return std::nullopt; } itr->status = Buffer::Status::Dequeued; @@ -99,18 +103,33 @@ void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform, queue_sequence.push_back(slot); } +void BufferQueue::CancelBuffer(u32 slot, const Service::Nvidia::MultiFence& multi_fence) { + const auto itr = std::find_if(queue.begin(), queue.end(), + [slot](const Buffer& buffer) { return buffer.slot == slot; }); + ASSERT(itr != queue.end()); + ASSERT(itr->status != Buffer::Status::Free); + itr->status = Buffer::Status::Free; + itr->multi_fence = multi_fence; + itr->swap_interval = 0; + + free_buffers.push_back(slot); + + buffer_wait_event.writable->Signal(); +} + std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() { auto itr = queue.end(); // Iterate to find a queued buffer matching the requested slot. while (itr == queue.end() && !queue_sequence.empty()) { - u32 slot = queue_sequence.front(); + const u32 slot = queue_sequence.front(); itr = std::find_if(queue.begin(), queue.end(), [&slot](const Buffer& buffer) { return buffer.status == Buffer::Status::Queued && buffer.slot == slot; }); queue_sequence.pop_front(); } - if (itr == queue.end()) - return {}; + if (itr == queue.end()) { + return std::nullopt; + } itr->status = Buffer::Status::Acquired; return *itr; } @@ -138,9 +157,7 @@ u32 BufferQueue::Query(QueryType type) { switch (type) { case QueryType::NativeWindowFormat: - // TODO(Subv): Use an enum for this - static constexpr u32 FormatABGR8 = 1; - return FormatABGR8; + return static_cast<u32>(PixelFormat::RGBA8888); } UNIMPLEMENTED(); diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h index d5f31e567..e7517c7e1 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.h +++ b/src/core/hle/service/nvflinger/buffer_queue.h @@ -66,6 +66,16 @@ public: Rotate270 = 0x07, }; + enum class PixelFormat : u32 { + RGBA8888 = 1, + RGBX8888 = 2, + RGB888 = 3, + RGB565 = 4, + BGRA8888 = 5, + RGBA5551 = 6, + RRGBA4444 = 7, + }; + struct Buffer { enum class Status { Free = 0, Queued = 1, Dequeued = 2, Acquired = 3 }; @@ -85,6 +95,7 @@ public: void QueueBuffer(u32 slot, BufferTransformFlags transform, const Common::Rectangle<int>& crop_rect, u32 swap_interval, Service::Nvidia::MultiFence& multi_fence); + void CancelBuffer(u32 slot, const Service::Nvidia::MultiFence& multi_fence); std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer(); void ReleaseBuffer(u32 slot); void Disconnect(); diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp index 437bc5dee..44aa2bdae 100644 --- a/src/core/hle/service/nvflinger/nvflinger.cpp +++ b/src/core/hle/service/nvflinger/nvflinger.cpp @@ -9,6 +9,7 @@ #include "common/logging/log.h" #include "common/microprofile.h" #include "common/scope_exit.h" +#include "common/thread.h" #include "core/core.h" #include "core/core_timing.h" #include "core/core_timing_util.h" @@ -27,8 +28,34 @@ namespace Service::NVFlinger { -constexpr s64 frame_ticks = static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 60); -constexpr s64 frame_ticks_30fps = static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 30); +constexpr auto frame_ns = std::chrono::nanoseconds{1000000000 / 60}; + +void NVFlinger::VSyncThread(NVFlinger& nv_flinger) { + nv_flinger.SplitVSync(); +} + +void NVFlinger::SplitVSync() { + system.RegisterHostThread(); + std::string name = "yuzu:VSyncThread"; + MicroProfileOnThreadCreate(name.c_str()); + Common::SetCurrentThreadName(name.c_str()); + Common::SetCurrentThreadPriority(Common::ThreadPriority::High); + s64 delay = 0; + while (is_running) { + guard->lock(); + const s64 time_start = system.CoreTiming().GetGlobalTimeNs().count(); + Compose(); + const auto ticks = GetNextTicks(); + const s64 time_end = system.CoreTiming().GetGlobalTimeNs().count(); + const s64 time_passed = time_end - time_start; + const s64 next_time = std::max<s64>(0, ticks - time_passed - delay); + guard->unlock(); + if (next_time > 0) { + wait_event->WaitFor(std::chrono::nanoseconds{next_time}); + } + delay = (system.CoreTiming().GetGlobalTimeNs().count() - time_end) - next_time; + } +} NVFlinger::NVFlinger(Core::System& system) : system(system) { displays.emplace_back(0, "Default", system); @@ -36,22 +63,40 @@ NVFlinger::NVFlinger(Core::System& system) : system(system) { displays.emplace_back(2, "Edid", system); displays.emplace_back(3, "Internal", system); displays.emplace_back(4, "Null", system); + guard = std::make_shared<std::mutex>(); // Schedule the screen composition events - composition_event = - Core::Timing::CreateEvent("ScreenComposition", [this](u64 userdata, s64 cycles_late) { + composition_event = Core::Timing::CreateEvent( + "ScreenComposition", [this](std::uintptr_t, std::chrono::nanoseconds ns_late) { + const auto guard = Lock(); Compose(); - const auto ticks = - Settings::values.force_30fps_mode ? frame_ticks_30fps : GetNextTicks(); - this->system.CoreTiming().ScheduleEvent(std::max<s64>(0LL, ticks - cycles_late), - composition_event); + + const auto ticks = std::chrono::nanoseconds{GetNextTicks()}; + const auto ticks_delta = ticks - ns_late; + const auto future_ns = std::max(std::chrono::nanoseconds::zero(), ticks_delta); + + this->system.CoreTiming().ScheduleEvent(future_ns, composition_event); }); - system.CoreTiming().ScheduleEvent(frame_ticks, composition_event); + if (system.IsMulticore()) { + is_running = true; + wait_event = std::make_unique<Common::Event>(); + vsync_thread = std::make_unique<std::thread>(VSyncThread, std::ref(*this)); + } else { + system.CoreTiming().ScheduleEvent(frame_ns, composition_event); + } } NVFlinger::~NVFlinger() { - system.CoreTiming().UnscheduleEvent(composition_event, 0); + if (system.IsMulticore()) { + is_running = false; + wait_event->Set(); + vsync_thread->join(); + vsync_thread.reset(); + wait_event.reset(); + } else { + system.CoreTiming().UnscheduleEvent(composition_event, 0); + } } void NVFlinger::SetNVDrvInstance(std::shared_ptr<Nvidia::Module> instance) { @@ -69,7 +114,7 @@ std::optional<u64> NVFlinger::OpenDisplay(std::string_view name) { [&](const VI::Display& display) { return display.GetName() == name; }); if (itr == displays.end()) { - return {}; + return std::nullopt; } return itr->GetID(); @@ -79,7 +124,7 @@ std::optional<u64> NVFlinger::CreateLayer(u64 display_id) { auto* const display = FindDisplay(display_id); if (display == nullptr) { - return {}; + return std::nullopt; } const u64 layer_id = next_layer_id++; @@ -99,7 +144,7 @@ std::optional<u32> NVFlinger::FindBufferQueueId(u64 display_id, u64 layer_id) co const auto* const layer = FindLayer(display_id, layer_id); if (layer == nullptr) { - return {}; + return std::nullopt; } return layer->GetBufferQueue().GetId(); @@ -197,12 +242,18 @@ void NVFlinger::Compose() { const auto& igbp_buffer = buffer->get().igbp_buffer; + if (!system.IsPoweredOn()) { + return; // We are likely shutting down + } + auto& gpu = system.GPU(); const auto& multi_fence = buffer->get().multi_fence; + guard->unlock(); for (u32 fence_id = 0; fence_id < multi_fence.num_fences; fence_id++) { const auto& fence = multi_fence.fences[fence_id]; gpu.WaitFence(fence.id, fence.value); } + guard->lock(); MicroProfileFlip(); @@ -223,7 +274,7 @@ void NVFlinger::Compose() { s64 NVFlinger::GetNextTicks() const { constexpr s64 max_hertz = 120LL; - return (Core::Hardware::BASE_CLOCK_RATE * (1LL << swap_interval)) / max_hertz; + return (1000000000 * (1LL << swap_interval)) / max_hertz; } } // namespace Service::NVFlinger diff --git a/src/core/hle/service/nvflinger/nvflinger.h b/src/core/hle/service/nvflinger/nvflinger.h index 57a21f33b..1ebe949c0 100644 --- a/src/core/hle/service/nvflinger/nvflinger.h +++ b/src/core/hle/service/nvflinger/nvflinger.h @@ -4,15 +4,22 @@ #pragma once +#include <atomic> #include <memory> +#include <mutex> #include <optional> #include <string> #include <string_view> +#include <thread> #include <vector> #include "common/common_types.h" #include "core/hle/kernel/object.h" +namespace Common { +class Event; +} // namespace Common + namespace Core::Timing { class CoreTiming; struct EventType; @@ -47,12 +54,12 @@ public: /// Opens the specified display and returns the ID. /// /// If an invalid display name is provided, then an empty optional is returned. - std::optional<u64> OpenDisplay(std::string_view name); + [[nodiscard]] std::optional<u64> OpenDisplay(std::string_view name); /// Creates a layer on the specified display and returns the layer ID. /// /// If an invalid display ID is specified, then an empty optional is returned. - std::optional<u64> CreateLayer(u64 display_id); + [[nodiscard]] std::optional<u64> CreateLayer(u64 display_id); /// Closes a layer on all displays for the given layer ID. void CloseLayer(u64 layer_id); @@ -60,37 +67,45 @@ public: /// Finds the buffer queue ID of the specified layer in the specified display. /// /// If an invalid display ID or layer ID is provided, then an empty optional is returned. - std::optional<u32> FindBufferQueueId(u64 display_id, u64 layer_id) const; + [[nodiscard]] std::optional<u32> FindBufferQueueId(u64 display_id, u64 layer_id) const; /// Gets the vsync event for the specified display. /// /// If an invalid display ID is provided, then nullptr is returned. - std::shared_ptr<Kernel::ReadableEvent> FindVsyncEvent(u64 display_id) const; + [[nodiscard]] std::shared_ptr<Kernel::ReadableEvent> FindVsyncEvent(u64 display_id) const; /// Obtains a buffer queue identified by the ID. - BufferQueue& FindBufferQueue(u32 id); + [[nodiscard]] BufferQueue& FindBufferQueue(u32 id); /// Obtains a buffer queue identified by the ID. - const BufferQueue& FindBufferQueue(u32 id) const; + [[nodiscard]] const BufferQueue& FindBufferQueue(u32 id) const; /// Performs a composition request to the emulated nvidia GPU and triggers the vsync events when /// finished. void Compose(); - s64 GetNextTicks() const; + [[nodiscard]] s64 GetNextTicks() const; + + [[nodiscard]] std::unique_lock<std::mutex> Lock() const { + return std::unique_lock{*guard}; + } private: /// Finds the display identified by the specified ID. - VI::Display* FindDisplay(u64 display_id); + [[nodiscard]] VI::Display* FindDisplay(u64 display_id); /// Finds the display identified by the specified ID. - const VI::Display* FindDisplay(u64 display_id) const; + [[nodiscard]] const VI::Display* FindDisplay(u64 display_id) const; /// Finds the layer identified by the specified ID in the desired display. - VI::Layer* FindLayer(u64 display_id, u64 layer_id); + [[nodiscard]] VI::Layer* FindLayer(u64 display_id, u64 layer_id); /// Finds the layer identified by the specified ID in the desired display. - const VI::Layer* FindLayer(u64 display_id, u64 layer_id) const; + [[nodiscard]] const VI::Layer* FindLayer(u64 display_id, u64 layer_id) const; + + static void VSyncThread(NVFlinger& nv_flinger); + + void SplitVSync(); std::shared_ptr<Nvidia::Module> nvdrv; @@ -108,7 +123,13 @@ private: /// Event that handles screen composition. std::shared_ptr<Core::Timing::EventType> composition_event; + std::shared_ptr<std::mutex> guard; + Core::System& system; + + std::unique_ptr<std::thread> vsync_thread; + std::unique_ptr<Common::Event> wait_event; + std::atomic<bool> is_running{}; }; } // namespace Service::NVFlinger diff --git a/src/core/hle/service/olsc/olsc.cpp b/src/core/hle/service/olsc/olsc.cpp new file mode 100644 index 000000000..aad4ca706 --- /dev/null +++ b/src/core/hle/service/olsc/olsc.cpp @@ -0,0 +1,69 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/hle_ipc.h" +#include "core/hle/service/olsc/olsc.h" +#include "core/hle/service/service.h" +#include "core/hle/service/sm/sm.h" + +namespace Service::OLSC { + +class OLSC final : public ServiceFramework<OLSC> { +public: + explicit OLSC() : ServiceFramework{"olsc:u"} { + // clang-format off + static const FunctionInfo functions[] = { + {0, &OLSC::Initialize, "Initialize"}, + {10, nullptr, "VerifySaveDataBackupLicenseAsync"}, + {13, nullptr, "GetSaveDataBackupSetting"}, + {14, &OLSC::SetSaveDataBackupSettingEnabled, "SetSaveDataBackupSettingEnabled"}, + {15, nullptr, "SetCustomData"}, + {16, nullptr, "DeleteSaveDataBackupSetting"}, + {18, nullptr, "GetSaveDataBackupInfoCache"}, + {19, nullptr, "UpdateSaveDataBackupInfoCacheAsync"}, + {22, nullptr, "DeleteSaveDataBackupAsync"}, + {25, nullptr, "ListDownloadableSaveDataBackupInfoAsync"}, + {26, nullptr, "DownloadSaveDataBackupAsync"}, + {9010, nullptr, "VerifySaveDataBackupLicenseAsyncForDebug"}, + {9013, nullptr, "GetSaveDataBackupSettingForDebug"}, + {9014, nullptr, "SetSaveDataBackupSettingEnabledForDebug"}, + {9015, nullptr, "SetCustomDataForDebug"}, + {9016, nullptr, "DeleteSaveDataBackupSettingForDebug"}, + {9018, nullptr, "GetSaveDataBackupInfoCacheForDebug"}, + {9019, nullptr, "UpdateSaveDataBackupInfoCacheAsyncForDebug"}, + {9022, nullptr, "DeleteSaveDataBackupAsyncForDebug"}, + {9025, nullptr, "ListDownloadableSaveDataBackupInfoAsyncForDebug"}, + {9026, nullptr, "DownloadSaveDataBackupAsyncForDebug"}, + }; + // clang-format on + + RegisterHandlers(functions); + } + +private: + void Initialize(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_OLSC, "(STUBBED) called"); + + initialized = true; + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void SetSaveDataBackupSettingEnabled(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_OLSC, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + bool initialized{}; +}; + +void InstallInterfaces(SM::ServiceManager& service_manager) { + std::make_shared<OLSC>()->InstallAsService(service_manager); +} + +} // namespace Service::OLSC diff --git a/src/core/hle/service/olsc/olsc.h b/src/core/hle/service/olsc/olsc.h new file mode 100644 index 000000000..edee4376b --- /dev/null +++ b/src/core/hle/service/olsc/olsc.h @@ -0,0 +1,16 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +namespace Service::SM { +class ServiceManager; +} + +namespace Service::OLSC { + +/// Registers all SSL services with the specified service manager. +void InstallInterfaces(SM::ServiceManager& service_manager); + +} // namespace Service::OLSC diff --git a/src/core/hle/service/pcie/pcie.cpp b/src/core/hle/service/pcie/pcie.cpp index 39cf05eba..c568a0adc 100644 --- a/src/core/hle/service/pcie/pcie.cpp +++ b/src/core/hle/service/pcie/pcie.cpp @@ -36,6 +36,9 @@ public: {18, nullptr, "ReleaseIrq"}, {19, nullptr, "SetIrqEnable"}, {20, nullptr, "SetAspmEnable"}, + {21, nullptr, "SetResetUponResumeEnable"}, + {22, nullptr, "Unknown22"}, + {23, nullptr, "Unknown23"}, }; // clang-format on diff --git a/src/core/hle/service/pctl/module.cpp b/src/core/hle/service/pctl/module.cpp index c75b4ee34..caf14ed61 100644 --- a/src/core/hle/service/pctl/module.cpp +++ b/src/core/hle/service/pctl/module.cpp @@ -31,6 +31,8 @@ public: {1014, nullptr, "ConfirmPlayableApplicationVideoOld"}, {1015, nullptr, "ConfirmPlayableApplicationVideo"}, {1016, nullptr, "ConfirmShowNewsPermission"}, + {1017, nullptr, "EndFreeCommunication"}, + {1018, nullptr, "IsFreeCommunicationAvailable"}, {1031, nullptr, "IsRestrictionEnabled"}, {1032, nullptr, "GetSafetyLevel"}, {1033, nullptr, "SetSafetyLevel"}, diff --git a/src/core/hle/service/pcv/pcv.cpp b/src/core/hle/service/pcv/pcv.cpp index d6891a659..8bfc0276e 100644 --- a/src/core/hle/service/pcv/pcv.cpp +++ b/src/core/hle/service/pcv/pcv.cpp @@ -42,6 +42,9 @@ public: {24, nullptr, "GetModuleStateTable"}, {25, nullptr, "GetPowerDomainStateTable"}, {26, nullptr, "GetFuseInfo"}, + {27, nullptr, "GetDramId"}, + {28, nullptr, "IsPoweredOn"}, + {29, nullptr, "GetVoltage"}, }; // clang-format on diff --git a/src/core/hle/service/pm/pm.cpp b/src/core/hle/service/pm/pm.cpp index 809eca0ab..a771a51b4 100644 --- a/src/core/hle/service/pm/pm.cpp +++ b/src/core/hle/service/pm/pm.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" @@ -78,13 +79,13 @@ public: : ServiceFramework{"pm:dmnt"}, kernel(kernel) { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "GetDebugProcesses"}, - {1, nullptr, "StartDebugProcess"}, - {2, &DebugMonitor::GetTitlePid, "GetTitlePid"}, - {3, nullptr, "EnableDebugForTitleId"}, - {4, &DebugMonitor::GetApplicationPid, "GetApplicationPid"}, - {5, nullptr, "EnableDebugForApplication"}, - {6, nullptr, "DisableDebug"}, + {0, nullptr, "GetJitDebugProcessIdList"}, + {1, nullptr, "StartProcess"}, + {2, &DebugMonitor::GetProcessId, "GetProcessId"}, + {3, nullptr, "HookToCreateProcess"}, + {4, &DebugMonitor::GetApplicationProcessId, "GetApplicationProcessId"}, + {5, nullptr, "HookToCreateApplicationProgress"}, + {6, nullptr, "ClearHook"}, }; // clang-format on @@ -92,7 +93,7 @@ public: } private: - void GetTitlePid(Kernel::HLERequestContext& ctx) { + void GetProcessId(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto title_id = rp.PopRaw<u64>(); @@ -114,7 +115,7 @@ private: rb.Push((*process)->GetProcessID()); } - void GetApplicationPid(Kernel::HLERequestContext& ctx) { + void GetApplicationProcessId(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_PM, "called"); GetApplicationPidGeneric(ctx, kernel.GetProcessList()); } @@ -163,15 +164,15 @@ public: : ServiceFramework{"pm:shell"}, kernel(kernel) { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "LaunchProcess"}, - {1, nullptr, "TerminateProcessByPid"}, - {2, nullptr, "TerminateProcessByTitleId"}, - {3, nullptr, "GetProcessEventWaiter"}, - {4, nullptr, "GetProcessEventType"}, + {0, nullptr, "LaunchProgram"}, + {1, nullptr, "TerminateProcess"}, + {2, nullptr, "TerminateProgram"}, + {3, nullptr, "GetProcessEventHandle"}, + {4, nullptr, "GetProcessEventInfo"}, {5, nullptr, "NotifyBootFinished"}, - {6, &Shell::GetApplicationPid, "GetApplicationPid"}, + {6, &Shell::GetApplicationProcessIdForShell, "GetApplicationProcessIdForShell"}, {7, nullptr, "BoostSystemMemoryResourceLimit"}, - {8, nullptr, "EnableAdditionalSystemThreads"}, + {8, nullptr, "BoostApplicationThreadResourceLimit"}, {9, nullptr, "GetBootFinishedEventHandle"}, }; // clang-format on @@ -180,7 +181,7 @@ public: } private: - void GetApplicationPid(Kernel::HLERequestContext& ctx) { + void GetApplicationProcessIdForShell(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_PM, "called"); GetApplicationPidGeneric(ctx, kernel.GetProcessList()); } diff --git a/src/core/hle/service/prepo/prepo.cpp b/src/core/hle/service/prepo/prepo.cpp index 8f1be0e48..b9ef86b72 100644 --- a/src/core/hle/service/prepo/prepo.cpp +++ b/src/core/hle/service/prepo/prepo.cpp @@ -4,6 +4,7 @@ #include "common/hex_util.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/process.h" #include "core/hle/service/acc/profile_manager.h" @@ -21,8 +22,10 @@ public: static const FunctionInfo functions[] = { {10100, &PlayReport::SaveReport<Core::Reporter::PlayReportType::Old>, "SaveReportOld"}, {10101, &PlayReport::SaveReportWithUser<Core::Reporter::PlayReportType::Old>, "SaveReportWithUserOld"}, - {10102, &PlayReport::SaveReport<Core::Reporter::PlayReportType::New>, "SaveReport"}, - {10103, &PlayReport::SaveReportWithUser<Core::Reporter::PlayReportType::New>, "SaveReportWithUser"}, + {10102, &PlayReport::SaveReport<Core::Reporter::PlayReportType::Old2>, "SaveReportOld2"}, + {10103, &PlayReport::SaveReportWithUser<Core::Reporter::PlayReportType::Old2>, "SaveReportWithUserOld2"}, + {10104, nullptr, "SaveReport"}, + {10105, nullptr, "SaveReportWithUser"}, {10200, nullptr, "RequestImmediateTransmission"}, {10300, nullptr, "GetTransmissionStatus"}, {10400, nullptr, "GetSystemSessionId"}, @@ -35,9 +38,16 @@ public: {30400, nullptr, "GetStatistics"}, {30401, nullptr, "GetThroughputHistory"}, {30500, nullptr, "GetLastUploadError"}, + {30600, nullptr, "GetApplicationUploadSummary"}, {40100, nullptr, "IsUserAgreementCheckEnabled"}, {40101, nullptr, "SetUserAgreementCheckEnabled"}, + {50100, nullptr, "ReadAllApplicationReportFiles"}, {90100, nullptr, "ReadAllReportFiles"}, + {90101, nullptr, "Unknown90101"}, + {90102, nullptr, "Unknown90102"}, + {90200, nullptr, "GetStatistics"}, + {90201, nullptr, "GetThroughputHistory"}, + {90300, nullptr, "GetLastUploadError"}, }; // clang-format on @@ -51,7 +61,7 @@ private: const auto process_id = rp.PopRaw<u64>(); std::vector<std::vector<u8>> data{ctx.ReadBuffer(0)}; - if (Type == Core::Reporter::PlayReportType::New) { + if constexpr (Type == Core::Reporter::PlayReportType::Old2) { data.emplace_back(ctx.ReadBuffer(1)); } @@ -71,8 +81,13 @@ private: const auto user_id = rp.PopRaw<u128>(); const auto process_id = rp.PopRaw<u64>(); std::vector<std::vector<u8>> data{ctx.ReadBuffer(0)}; - if (Type == Core::Reporter::PlayReportType::New) { - data.emplace_back(ctx.ReadBuffer(1)); + + if constexpr (Type == Core::Reporter::PlayReportType::Old2) { + const auto read_buffer_count = + ctx.BufferDescriptorX().size() + ctx.BufferDescriptorA().size(); + if (read_buffer_count > 1) { + data.emplace_back(ctx.ReadBuffer(1)); + } } LOG_DEBUG( diff --git a/src/core/hle/service/psc/psc.cpp b/src/core/hle/service/psc/psc.cpp index 53ec6b031..99e1c9042 100644 --- a/src/core/hle/service/psc/psc.cpp +++ b/src/core/hle/service/psc/psc.cpp @@ -24,6 +24,8 @@ public: {4, nullptr, "Cancel"}, {5, nullptr, "PrintModuleInformation"}, {6, nullptr, "GetModuleInformation"}, + {10, nullptr, "Unknown10"}, + {11, nullptr, "Unknown11"}, }; // clang-format on diff --git a/src/core/hle/service/ptm/psm.cpp b/src/core/hle/service/ptm/psm.cpp index c2d5fda94..6d9e6bd09 100644 --- a/src/core/hle/service/ptm/psm.cpp +++ b/src/core/hle/service/ptm/psm.cpp @@ -12,9 +12,6 @@ namespace Service::PSM { -constexpr u32 BATTERY_FULLY_CHARGED = 100; // 100% Full -constexpr u32 BATTERY_CURRENTLY_CHARGING = 1; // Plugged into an official dock - class PSM final : public ServiceFramework<PSM> { public: explicit PSM() : ServiceFramework{"psm"} { @@ -38,6 +35,7 @@ public: {15, nullptr, "GetBatteryAgePercentage"}, {16, nullptr, "GetBatteryChargeInfoEvent"}, {17, nullptr, "GetBatteryChargeInfoFields"}, + {18, nullptr, "GetBatteryChargeCalibratedEvent"}, }; // clang-format on @@ -48,20 +46,30 @@ public: private: void GetBatteryChargePercentage(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_PSM, "(STUBBED) called"); + LOG_DEBUG(Service_PSM, "called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(BATTERY_FULLY_CHARGED); + rb.Push<u32>(battery_charge_percentage); } void GetChargerType(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_PSM, "(STUBBED) called"); + LOG_DEBUG(Service_PSM, "called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(BATTERY_CURRENTLY_CHARGING); + rb.PushEnum(charger_type); } + + enum class ChargerType : u32 { + Unplugged = 0, + RegularCharger = 1, + LowPowerCharger = 2, + Unknown = 3, + }; + + u32 battery_charge_percentage{100}; // 100% + ChargerType charger_type{ChargerType::RegularCharger}; }; void InstallInterfaces(SM::ServiceManager& sm) { diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp index fa5347af9..fbfda2d5b 100644 --- a/src/core/hle/service/service.cpp +++ b/src/core/hle/service/service.cpp @@ -51,6 +51,7 @@ #include "core/hle/service/ns/ns.h" #include "core/hle/service/nvdrv/nvdrv.h" #include "core/hle/service/nvflinger/nvflinger.h" +#include "core/hle/service/olsc/olsc.h" #include "core/hle/service/pcie/pcie.h" #include "core/hle/service/pctl/module.h" #include "core/hle/service/pcv/pcv.h" @@ -89,8 +90,6 @@ namespace Service { return function_string; } -//////////////////////////////////////////////////////////////////////////////////////////////////// - ServiceFrameworkBase::ServiceFrameworkBase(const char* service_name, u32 max_sessions, InvokerFn* handler_invoker) : service_name(service_name), max_sessions(max_sessions), handler_invoker(handler_invoker) {} @@ -105,10 +104,9 @@ void ServiceFrameworkBase::InstallAsService(SM::ServiceManager& service_manager) port_installed = true; } -void ServiceFrameworkBase::InstallAsNamedPort() { +void ServiceFrameworkBase::InstallAsNamedPort(Kernel::KernelCore& kernel) { ASSERT(!port_installed); - auto& kernel = Core::System::GetInstance().Kernel(); auto [server_port, client_port] = Kernel::ServerPort::CreatePortPair(kernel, max_sessions, service_name); server_port->SetHleHandler(shared_from_this()); @@ -116,10 +114,9 @@ void ServiceFrameworkBase::InstallAsNamedPort() { port_installed = true; } -std::shared_ptr<Kernel::ClientPort> ServiceFrameworkBase::CreatePort() { +std::shared_ptr<Kernel::ClientPort> ServiceFrameworkBase::CreatePort(Kernel::KernelCore& kernel) { ASSERT(!port_installed); - auto& kernel = Core::System::GetInstance().Kernel(); auto [server_port, client_port] = Kernel::ServerPort::CreatePortPair(kernel, max_sessions, service_name); auto port = MakeResult(std::move(server_port)).Unwrap(); @@ -191,9 +188,6 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co return RESULT_SUCCESS; } -//////////////////////////////////////////////////////////////////////////////////////////////////// -// Module interface - /// Initialize ServiceManager void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) { // NVFlinger needs to be accessed by several services like Vi and AppletOE so we instantiate it @@ -238,6 +232,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) { NPNS::InstallInterfaces(*sm); NS::InstallInterfaces(*sm, system); Nvidia::InstallInterfaces(*sm, *nv_flinger, system); + OLSC::InstallInterfaces(*sm); PCIe::InstallInterfaces(*sm); PCTL::InstallInterfaces(*sm); PCV::InstallInterfaces(*sm); @@ -246,7 +241,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) { PSC::InstallInterfaces(*sm); PSM::InstallInterfaces(*sm); Set::InstallInterfaces(*sm); - Sockets::InstallInterfaces(*sm); + Sockets::InstallInterfaces(*sm, system); SPL::InstallInterfaces(*sm); SSL::InstallInterfaces(*sm); Time::InstallInterfaces(system); diff --git a/src/core/hle/service/service.h b/src/core/hle/service/service.h index 022d885b6..a01ef3353 100644 --- a/src/core/hle/service/service.h +++ b/src/core/hle/service/service.h @@ -63,9 +63,9 @@ public: /// Creates a port pair and registers this service with the given ServiceManager. void InstallAsService(SM::ServiceManager& service_manager); /// Creates a port pair and registers it on the kernel's global port registry. - void InstallAsNamedPort(); + void InstallAsNamedPort(Kernel::KernelCore& kernel); /// Creates and returns an unregistered port for the service. - std::shared_ptr<Kernel::ClientPort> CreatePort(); + std::shared_ptr<Kernel::ClientPort> CreatePort(Kernel::KernelCore& kernel); void InvokeRequest(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/set/set.cpp b/src/core/hle/service/set/set.cpp index 9e12c76fc..ffbf90b00 100644 --- a/src/core/hle/service/set/set.cpp +++ b/src/core/hle/service/set/set.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <algorithm> +#include <array> #include <chrono> #include "common/logging/log.h" #include "core/hle/ipc_helpers.h" @@ -31,6 +32,44 @@ constexpr std::array<LanguageCode, 17> available_language_codes = {{ LanguageCode::ZH_HANT, }}; +enum class KeyboardLayout : u64 { + Japanese = 0, + EnglishUs = 1, + EnglishUsInternational = 2, + EnglishUk = 3, + French = 4, + FrenchCa = 5, + Spanish = 6, + SpanishLatin = 7, + German = 8, + Italian = 9, + Portuguese = 10, + Russian = 11, + Korean = 12, + ChineseSimplified = 13, + ChineseTraditional = 14, +}; + +constexpr std::array<std::pair<LanguageCode, KeyboardLayout>, 17> language_to_layout{{ + {LanguageCode::JA, KeyboardLayout::Japanese}, + {LanguageCode::EN_US, KeyboardLayout::EnglishUs}, + {LanguageCode::FR, KeyboardLayout::French}, + {LanguageCode::DE, KeyboardLayout::German}, + {LanguageCode::IT, KeyboardLayout::Italian}, + {LanguageCode::ES, KeyboardLayout::Spanish}, + {LanguageCode::ZH_CN, KeyboardLayout::ChineseSimplified}, + {LanguageCode::KO, KeyboardLayout::Korean}, + {LanguageCode::NL, KeyboardLayout::EnglishUsInternational}, + {LanguageCode::PT, KeyboardLayout::Portuguese}, + {LanguageCode::RU, KeyboardLayout::Russian}, + {LanguageCode::ZH_TW, KeyboardLayout::ChineseTraditional}, + {LanguageCode::EN_GB, KeyboardLayout::EnglishUk}, + {LanguageCode::FR_CA, KeyboardLayout::FrenchCa}, + {LanguageCode::ES_419, KeyboardLayout::SpanishLatin}, + {LanguageCode::ZH_HANS, KeyboardLayout::ChineseSimplified}, + {LanguageCode::ZH_HANT, KeyboardLayout::ChineseTraditional}, +}}; + constexpr std::size_t pre4_0_0_max_entries = 15; constexpr std::size_t post4_0_0_max_entries = 17; @@ -50,6 +89,25 @@ void GetAvailableLanguageCodesImpl(Kernel::HLERequestContext& ctx, std::size_t m ctx.WriteBuffer(available_language_codes.data(), copy_size); PushResponseLanguageCode(ctx, copy_amount); } + +void GetKeyCodeMapImpl(Kernel::HLERequestContext& ctx) { + const auto language_code = available_language_codes[Settings::values.language_index.GetValue()]; + const auto key_code = + std::find_if(language_to_layout.cbegin(), language_to_layout.cend(), + [=](const auto& element) { return element.first == language_code; }); + KeyboardLayout layout = KeyboardLayout::EnglishUs; + if (key_code == language_to_layout.cend()) { + LOG_ERROR(Service_SET, + "Could not find keyboard layout for language index {}, defaulting to English us", + Settings::values.language_index.GetValue()); + } else { + layout = key_code->second; + } + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + ctx.WriteBuffer(layout); +} } // Anonymous namespace LanguageCode GetLanguageCodeFromIndex(std::size_t index) { @@ -67,6 +125,7 @@ void SET::MakeLanguageCode(Kernel::HLERequestContext& ctx) { const auto index = rp.Pop<u32>(); if (index >= available_language_codes.size()) { + LOG_ERROR(Service_SET, "Invalid language code index! index={}", index); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_INVALID_LANGUAGE); return; @@ -104,11 +163,11 @@ void SET::GetQuestFlag(Kernel::HLERequestContext& ctx) { } void SET::GetLanguageCode(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_SET, "called {}", Settings::values.language_index); + LOG_DEBUG(Service_SET, "called {}", Settings::values.language_index.GetValue()); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.PushEnum(available_language_codes[Settings::values.language_index]); + rb.PushEnum(available_language_codes[Settings::values.language_index.GetValue()]); } void SET::GetRegionCode(Kernel::HLERequestContext& ctx) { @@ -116,7 +175,17 @@ void SET::GetRegionCode(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push(Settings::values.region_index); + rb.Push(Settings::values.region_index.GetValue()); +} + +void SET::GetKeyCodeMap(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_SET, "Called {}", ctx.Description()); + GetKeyCodeMapImpl(ctx); +} + +void SET::GetKeyCodeMap2(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_SET, "Called {}", ctx.Description()); + GetKeyCodeMapImpl(ctx); } SET::SET() : ServiceFramework("set") { @@ -129,10 +198,11 @@ SET::SET() : ServiceFramework("set") { {4, &SET::GetRegionCode, "GetRegionCode"}, {5, &SET::GetAvailableLanguageCodes2, "GetAvailableLanguageCodes2"}, {6, &SET::GetAvailableLanguageCodeCount2, "GetAvailableLanguageCodeCount2"}, - {7, nullptr, "GetKeyCodeMap"}, + {7, &SET::GetKeyCodeMap, "GetKeyCodeMap"}, {8, &SET::GetQuestFlag, "GetQuestFlag"}, - {9, nullptr, "GetKeyCodeMap2"}, + {9, &SET::GetKeyCodeMap2, "GetKeyCodeMap2"}, {10, nullptr, "GetFirmwareVersionForDebug"}, + {11, nullptr, "GetDeviceNickName"}, }; // clang-format on diff --git a/src/core/hle/service/set/set.h b/src/core/hle/service/set/set.h index 6084b345d..8ac9c169d 100644 --- a/src/core/hle/service/set/set.h +++ b/src/core/hle/service/set/set.h @@ -44,6 +44,8 @@ private: void GetAvailableLanguageCodeCount2(Kernel::HLERequestContext& ctx); void GetQuestFlag(Kernel::HLERequestContext& ctx); void GetRegionCode(Kernel::HLERequestContext& ctx); + void GetKeyCodeMap(Kernel::HLERequestContext& ctx); + void GetKeyCodeMap2(Kernel::HLERequestContext& ctx); }; } // namespace Service::Set diff --git a/src/core/hle/service/set/set_cal.cpp b/src/core/hle/service/set/set_cal.cpp index 1398a4a48..3fbfecc9e 100644 --- a/src/core/hle/service/set/set_cal.cpp +++ b/src/core/hle/service/set/set_cal.cpp @@ -50,6 +50,8 @@ SET_CAL::SET_CAL() : ServiceFramework("set:cal") { {39, nullptr, "GetConsoleSixAxisSensorModuleType"}, {40, nullptr, "GetConsoleSixAxisSensorHorizontalOffset"}, {41, nullptr, "GetBatteryVersion"}, + {42, nullptr, "GetDeviceId"}, + {43, nullptr, "GetConsoleSixAxisSensorMountType"}, }; // clang-format on diff --git a/src/core/hle/service/set/set_sys.cpp b/src/core/hle/service/set/set_sys.cpp index b7c9ea74b..080b5743e 100644 --- a/src/core/hle/service/set/set_sys.cpp +++ b/src/core/hle/service/set/set_sys.cpp @@ -288,6 +288,20 @@ SET_SYS::SET_SYS() : ServiceFramework("set:sys") { {186, nullptr, "GetMemoryUsageRateFlag"}, {187, nullptr, "GetTouchScreenMode"}, {188, nullptr, "SetTouchScreenMode"}, + {189, nullptr, "GetButtonConfigSettingsFull"}, + {190, nullptr, "SetButtonConfigSettingsFull"}, + {191, nullptr, "GetButtonConfigSettingsEmbedded"}, + {192, nullptr, "SetButtonConfigSettingsEmbedded"}, + {193, nullptr, "GetButtonConfigSettingsLeft"}, + {194, nullptr, "SetButtonConfigSettingsLeft"}, + {195, nullptr, "GetButtonConfigSettingsRight"}, + {196, nullptr, "SetButtonConfigSettingsRight"}, + {197, nullptr, "GetButtonConfigRegisteredSettingsEmbedded"}, + {198, nullptr, "SetButtonConfigRegisteredSettingsEmbedded"}, + {199, nullptr, "GetButtonConfigRegisteredSettings"}, + {200, nullptr, "SetButtonConfigRegisteredSettings"}, + {201, nullptr, "GetFieldTestingFlag"}, + {202, nullptr, "SetFieldTestingFlag"}, }; // clang-format on diff --git a/src/core/hle/service/sm/controller.cpp b/src/core/hle/service/sm/controller.cpp index 9cca84b31..972aaa6d9 100644 --- a/src/core/hle/service/sm/controller.cpp +++ b/src/core/hle/service/sm/controller.cpp @@ -12,7 +12,7 @@ namespace Service::SM { -void Controller::ConvertSessionToDomain(Kernel::HLERequestContext& ctx) { +void Controller::ConvertCurrentObjectToDomain(Kernel::HLERequestContext& ctx) { ASSERT_MSG(ctx.Session()->IsSession(), "Session is already a domain"); LOG_DEBUG(Service, "called, server_session={}", ctx.Session()->GetObjectId()); ctx.Session()->ConvertToDomain(); @@ -22,7 +22,7 @@ void Controller::ConvertSessionToDomain(Kernel::HLERequestContext& ctx) { rb.Push<u32>(1); // Converted sessions start with 1 request handler } -void Controller::DuplicateSession(Kernel::HLERequestContext& ctx) { +void Controller::CloneCurrentObject(Kernel::HLERequestContext& ctx) { // TODO(bunnei): This is just creating a new handle to the same Session. I assume this is wrong // and that we probably want to actually make an entirely new Session, but we still need to // verify this on hardware. @@ -33,10 +33,10 @@ void Controller::DuplicateSession(Kernel::HLERequestContext& ctx) { rb.PushMoveObjects(ctx.Session()->GetParent()->Client()); } -void Controller::DuplicateSessionEx(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called, using DuplicateSession"); +void Controller::CloneCurrentObjectEx(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service, "(STUBBED) called, using CloneCurrentObject"); - DuplicateSession(ctx); + CloneCurrentObject(ctx); } void Controller::QueryPointerBufferSize(Kernel::HLERequestContext& ctx) { @@ -47,13 +47,14 @@ void Controller::QueryPointerBufferSize(Kernel::HLERequestContext& ctx) { rb.Push<u16>(0x1000); } +// https://switchbrew.org/wiki/IPC_Marshalling Controller::Controller() : ServiceFramework("IpcController") { static const FunctionInfo functions[] = { - {0x00000000, &Controller::ConvertSessionToDomain, "ConvertSessionToDomain"}, - {0x00000001, nullptr, "ConvertDomainToSession"}, - {0x00000002, &Controller::DuplicateSession, "DuplicateSession"}, - {0x00000003, &Controller::QueryPointerBufferSize, "QueryPointerBufferSize"}, - {0x00000004, &Controller::DuplicateSessionEx, "DuplicateSessionEx"}, + {0, &Controller::ConvertCurrentObjectToDomain, "ConvertCurrentObjectToDomain"}, + {1, nullptr, "CopyFromCurrentDomain"}, + {2, &Controller::CloneCurrentObject, "CloneCurrentObject"}, + {3, &Controller::QueryPointerBufferSize, "QueryPointerBufferSize"}, + {4, &Controller::CloneCurrentObjectEx, "CloneCurrentObjectEx"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/sm/controller.h b/src/core/hle/service/sm/controller.h index dc66c9e37..180c6da50 100644 --- a/src/core/hle/service/sm/controller.h +++ b/src/core/hle/service/sm/controller.h @@ -14,9 +14,9 @@ public: ~Controller() override; private: - void ConvertSessionToDomain(Kernel::HLERequestContext& ctx); - void DuplicateSession(Kernel::HLERequestContext& ctx); - void DuplicateSessionEx(Kernel::HLERequestContext& ctx); + void ConvertCurrentObjectToDomain(Kernel::HLERequestContext& ctx); + void CloneCurrentObject(Kernel::HLERequestContext& ctx); + void CloneCurrentObjectEx(Kernel::HLERequestContext& ctx); void QueryPointerBufferSize(Kernel::HLERequestContext& ctx); }; diff --git a/src/core/hle/service/sm/sm.cpp b/src/core/hle/service/sm/sm.cpp index 88909504d..9c1da361b 100644 --- a/src/core/hle/service/sm/sm.cpp +++ b/src/core/hle/service/sm/sm.cpp @@ -19,7 +19,7 @@ constexpr ResultCode ERR_ALREADY_REGISTERED(ErrorModule::SM, 4); constexpr ResultCode ERR_INVALID_NAME(ErrorModule::SM, 6); constexpr ResultCode ERR_SERVICE_NOT_REGISTERED(ErrorModule::SM, 7); -ServiceManager::ServiceManager() = default; +ServiceManager::ServiceManager(Kernel::KernelCore& kernel_) : kernel{kernel_} {} ServiceManager::~ServiceManager() = default; void ServiceManager::InvokeControlRequest(Kernel::HLERequestContext& context) { @@ -27,10 +27,12 @@ void ServiceManager::InvokeControlRequest(Kernel::HLERequestContext& context) { } static ResultCode ValidateServiceName(const std::string& name) { - if (name.size() <= 0 || name.size() > 8) { + if (name.empty() || name.size() > 8) { + LOG_ERROR(Service_SM, "Invalid service name! service={}", name); return ERR_INVALID_NAME; } - if (name.find('\0') != std::string::npos) { + if (name.rfind('\0') != std::string::npos) { + LOG_ERROR(Service_SM, "A non null terminated service was passed"); return ERR_INVALID_NAME; } return RESULT_SUCCESS; @@ -41,20 +43,21 @@ void ServiceManager::InstallInterfaces(std::shared_ptr<ServiceManager> self, ASSERT(self->sm_interface.expired()); auto sm = std::make_shared<SM>(self, kernel); - sm->InstallAsNamedPort(); + sm->InstallAsNamedPort(kernel); self->sm_interface = sm; self->controller_interface = std::make_unique<Controller>(); } -ResultVal<std::shared_ptr<Kernel::ServerPort>> ServiceManager::RegisterService( - std::string name, unsigned int max_sessions) { +ResultVal<std::shared_ptr<Kernel::ServerPort>> ServiceManager::RegisterService(std::string name, + u32 max_sessions) { CASCADE_CODE(ValidateServiceName(name)); - if (registered_services.find(name) != registered_services.end()) + if (registered_services.find(name) != registered_services.end()) { + LOG_ERROR(Service_SM, "Service is already registered! service={}", name); return ERR_ALREADY_REGISTERED; + } - auto& kernel = Core::System::GetInstance().Kernel(); auto [server_port, client_port] = Kernel::ServerPort::CreatePortPair(kernel, max_sessions, name); @@ -66,9 +69,10 @@ ResultCode ServiceManager::UnregisterService(const std::string& name) { CASCADE_CODE(ValidateServiceName(name)); const auto iter = registered_services.find(name); - if (iter == registered_services.end()) + if (iter == registered_services.end()) { + LOG_ERROR(Service_SM, "Server is not registered! service={}", name); return ERR_SERVICE_NOT_REGISTERED; - + } registered_services.erase(iter); return RESULT_SUCCESS; } @@ -79,6 +83,7 @@ ResultVal<std::shared_ptr<Kernel::ClientPort>> ServiceManager::GetServicePort( CASCADE_CODE(ValidateServiceName(name)); auto it = registered_services.find(name); if (it == registered_services.end()) { + LOG_ERROR(Service_SM, "Server is not registered! service={}", name); return ERR_SERVICE_NOT_REGISTERED; } @@ -136,7 +141,7 @@ void SM::GetService(Kernel::HLERequestContext& ctx) { } // Wake the threads waiting on the ServerPort - server_port->WakeupAllWaitingThreads(); + server_port->Signal(); LOG_DEBUG(Service_SM, "called service={} -> session={}", name, client->GetObjectId()); IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles}; diff --git a/src/core/hle/service/sm/sm.h b/src/core/hle/service/sm/sm.h index b06d2f103..6790c86f0 100644 --- a/src/core/hle/service/sm/sm.h +++ b/src/core/hle/service/sm/sm.h @@ -9,6 +9,7 @@ #include <type_traits> #include <unordered_map> +#include "common/concepts.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/server_port.h" @@ -47,19 +48,17 @@ class ServiceManager { public: static void InstallInterfaces(std::shared_ptr<ServiceManager> self, Kernel::KernelCore& kernel); - ServiceManager(); + explicit ServiceManager(Kernel::KernelCore& kernel_); ~ServiceManager(); ResultVal<std::shared_ptr<Kernel::ServerPort>> RegisterService(std::string name, - unsigned int max_sessions); + u32 max_sessions); ResultCode UnregisterService(const std::string& name); ResultVal<std::shared_ptr<Kernel::ClientPort>> GetServicePort(const std::string& name); ResultVal<std::shared_ptr<Kernel::ClientSession>> ConnectToService(const std::string& name); - template <typename T> + template <Common::DerivedFrom<Kernel::SessionRequestHandler> T> std::shared_ptr<T> GetService(const std::string& service_name) const { - static_assert(std::is_base_of_v<Kernel::SessionRequestHandler, T>, - "Not a base of ServiceFrameworkBase"); auto service = registered_services.find(service_name); if (service == registered_services.end()) { LOG_DEBUG(Service, "Can't find service: {}", service_name); @@ -80,6 +79,9 @@ private: /// Map of registered services, retrieved using GetServicePort or ConnectToService. std::unordered_map<std::string, std::shared_ptr<Kernel::ClientPort>> registered_services; + + /// Kernel context + Kernel::KernelCore& kernel; }; } // namespace Service::SM diff --git a/src/core/hle/service/sockets/blocking_worker.h b/src/core/hle/service/sockets/blocking_worker.h new file mode 100644 index 000000000..2d53e52b6 --- /dev/null +++ b/src/core/hle/service/sockets/blocking_worker.h @@ -0,0 +1,161 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <atomic> +#include <memory> +#include <string> +#include <string_view> +#include <thread> +#include <variant> +#include <vector> + +#include <fmt/format.h> + +#include "common/assert.h" +#include "common/microprofile.h" +#include "common/thread.h" +#include "core/core.h" +#include "core/hle/kernel/hle_ipc.h" +#include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/thread.h" +#include "core/hle/kernel/writable_event.h" + +namespace Service::Sockets { + +/** + * Worker abstraction to execute blocking calls on host without blocking the guest thread + * + * @tparam Service Service where the work is executed + * @tparam Types Types of work to execute + */ +template <class Service, class... Types> +class BlockingWorker { + using This = BlockingWorker<Service, Types...>; + using WorkVariant = std::variant<std::monostate, Types...>; + +public: + /// Create a new worker + static std::unique_ptr<This> Create(Core::System& system, Service* service, + std::string_view name) { + return std::unique_ptr<This>(new This(system, service, name)); + } + + ~BlockingWorker() { + while (!is_available.load(std::memory_order_relaxed)) { + // Busy wait until work is finished + std::this_thread::yield(); + } + // Monostate means to exit the thread + work = std::monostate{}; + work_event.Set(); + thread.join(); + } + + /** + * Try to capture the worker to send work after a success + * @returns True when the worker has been successfully captured + */ + bool TryCapture() { + bool expected = true; + return is_available.compare_exchange_weak(expected, false, std::memory_order_relaxed, + std::memory_order_relaxed); + } + + /** + * Send work to this worker abstraction + * @see TryCapture must be called before attempting to call this function + */ + template <class Work> + void SendWork(Work new_work) { + ASSERT_MSG(!is_available, "Trying to send work on a worker that's not captured"); + work = std::move(new_work); + work_event.Set(); + } + + /// Generate a callback for @see SleepClientThread + template <class Work> + auto Callback() { + return [this](std::shared_ptr<Kernel::Thread>, Kernel::HLERequestContext& ctx, + Kernel::ThreadWakeupReason reason) { + ASSERT(reason == Kernel::ThreadWakeupReason::Signal); + std::get<Work>(work).Response(ctx); + is_available.store(true); + }; + } + + /// Get kernel event that will be signalled by the worker when the host operation finishes + std::shared_ptr<Kernel::WritableEvent> KernelEvent() const { + return kernel_event; + } + +private: + explicit BlockingWorker(Core::System& system, Service* service, std::string_view name) { + auto pair = Kernel::WritableEvent::CreateEventPair(system.Kernel(), std::string(name)); + kernel_event = std::move(pair.writable); + thread = std::thread([this, &system, service, name] { Run(system, service, name); }); + } + + void Run(Core::System& system, Service* service, std::string_view name) { + system.RegisterHostThread(); + + const std::string thread_name = fmt::format("yuzu:{}", name); + MicroProfileOnThreadCreate(thread_name.c_str()); + Common::SetCurrentThreadName(thread_name.c_str()); + + bool keep_running = true; + while (keep_running) { + work_event.Wait(); + + const auto visit_fn = [service, &keep_running]<typename T>(T&& w) { + if constexpr (std::is_same_v<std::decay_t<T>, std::monostate>) { + keep_running = false; + } else { + w.Execute(service); + } + }; + std::visit(visit_fn, work); + + kernel_event->Signal(); + } + } + + std::thread thread; + WorkVariant work; + Common::Event work_event; + std::shared_ptr<Kernel::WritableEvent> kernel_event; + std::atomic_bool is_available{true}; +}; + +template <class Service, class... Types> +class BlockingWorkerPool { + using Worker = BlockingWorker<Service, Types...>; + +public: + explicit BlockingWorkerPool(Core::System& system_, Service* service_) + : system{system_}, service{service_} {} + + /// Returns a captured worker thread, creating new ones if necessary + Worker* CaptureWorker() { + for (auto& worker : workers) { + if (worker->TryCapture()) { + return worker.get(); + } + } + auto new_worker = Worker::Create(system, service, fmt::format("BSD:{}", workers.size())); + [[maybe_unused]] const bool success = new_worker->TryCapture(); + ASSERT(success); + + return workers.emplace_back(std::move(new_worker)).get(); + } + +private: + Core::System& system; + Service* const service; + + std::vector<std::unique_ptr<Worker>> workers; +}; + +} // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/bsd.cpp b/src/core/hle/service/sockets/bsd.cpp index f67fab2f9..a74be9370 100644 --- a/src/core/hle/service/sockets/bsd.cpp +++ b/src/core/hle/service/sockets/bsd.cpp @@ -2,18 +2,138 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> +#include <memory> +#include <string> +#include <utility> +#include <vector> + +#include <fmt/format.h> + +#include "common/microprofile.h" +#include "common/thread.h" #include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/thread.h" #include "core/hle/service/sockets/bsd.h" +#include "core/hle/service/sockets/sockets_translate.h" +#include "core/network/network.h" +#include "core/network/sockets.h" namespace Service::Sockets { +namespace { + +bool IsConnectionBased(Type type) { + switch (type) { + case Type::STREAM: + return true; + case Type::DGRAM: + return false; + default: + UNIMPLEMENTED_MSG("Unimplemented type={}", static_cast<int>(type)); + return false; + } +} + +} // Anonymous namespace + +void BSD::PollWork::Execute(BSD* bsd) { + std::tie(ret, bsd_errno) = bsd->PollImpl(write_buffer, read_buffer, nfds, timeout); +} + +void BSD::PollWork::Response(Kernel::HLERequestContext& ctx) { + ctx.WriteBuffer(write_buffer); + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); +} + +void BSD::AcceptWork::Execute(BSD* bsd) { + std::tie(ret, bsd_errno) = bsd->AcceptImpl(fd, write_buffer); +} + +void BSD::AcceptWork::Response(Kernel::HLERequestContext& ctx) { + ctx.WriteBuffer(write_buffer); + + IPC::ResponseBuilder rb{ctx, 5}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); + rb.Push<u32>(static_cast<u32>(write_buffer.size())); +} + +void BSD::ConnectWork::Execute(BSD* bsd) { + bsd_errno = bsd->ConnectImpl(fd, addr); +} + +void BSD::ConnectWork::Response(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(bsd_errno == Errno::SUCCESS ? 0 : -1); + rb.PushEnum(bsd_errno); +} + +void BSD::RecvWork::Execute(BSD* bsd) { + std::tie(ret, bsd_errno) = bsd->RecvImpl(fd, flags, message); +} + +void BSD::RecvWork::Response(Kernel::HLERequestContext& ctx) { + ctx.WriteBuffer(message); + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); +} + +void BSD::RecvFromWork::Execute(BSD* bsd) { + std::tie(ret, bsd_errno) = bsd->RecvFromImpl(fd, flags, message, addr); +} + +void BSD::RecvFromWork::Response(Kernel::HLERequestContext& ctx) { + ctx.WriteBuffer(message, 0); + if (!addr.empty()) { + ctx.WriteBuffer(addr, 1); + } + + IPC::ResponseBuilder rb{ctx, 5}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); + rb.Push<u32>(static_cast<u32>(addr.size())); +} + +void BSD::SendWork::Execute(BSD* bsd) { + std::tie(ret, bsd_errno) = bsd->SendImpl(fd, flags, message); +} + +void BSD::SendWork::Response(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); +} + +void BSD::SendToWork::Execute(BSD* bsd) { + std::tie(ret, bsd_errno) = bsd->SendToImpl(fd, flags, message, addr); +} + +void BSD::SendToWork::Response(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); +} + void BSD::RegisterClient(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // bsd errno + rb.Push<s32>(0); // bsd errno } void BSD::StartMonitoring(Kernel::HLERequestContext& ctx) { @@ -26,20 +146,19 @@ void BSD::StartMonitoring(Kernel::HLERequestContext& ctx) { void BSD::Socket(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; + const u32 domain = rp.Pop<u32>(); + const u32 type = rp.Pop<u32>(); + const u32 protocol = rp.Pop<u32>(); - u32 domain = rp.Pop<u32>(); - u32 type = rp.Pop<u32>(); - u32 protocol = rp.Pop<u32>(); - - LOG_WARNING(Service, "(STUBBED) called domain={} type={} protocol={}", domain, type, protocol); + LOG_DEBUG(Service, "called. domain={} type={} protocol={}", domain, type, protocol); - u32 fd = next_fd++; + const auto [fd, bsd_errno] = SocketImpl(static_cast<Domain>(domain), static_cast<Type>(type), + static_cast<Protocol>(protocol)); IPC::ResponseBuilder rb{ctx, 4}; - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(fd); - rb.Push<u32>(0); // bsd errno + rb.Push<s32>(fd); + rb.PushEnum(bsd_errno); } void BSD::Select(Kernel::HLERequestContext& ctx) { @@ -52,67 +171,664 @@ void BSD::Select(Kernel::HLERequestContext& ctx) { rb.Push<u32>(0); // bsd errno } +void BSD::Poll(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const s32 nfds = rp.Pop<s32>(); + const s32 timeout = rp.Pop<s32>(); + + LOG_DEBUG(Service, "called. nfds={} timeout={}", nfds, timeout); + + ExecuteWork(ctx, "BSD:Poll", timeout != 0, + PollWork{ + .nfds = nfds, + .timeout = timeout, + .read_buffer = ctx.ReadBuffer(), + .write_buffer = std::vector<u8>(ctx.GetWriteBufferSize()), + }); +} + +void BSD::Accept(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + + LOG_DEBUG(Service, "called. fd={}", fd); + + ExecuteWork(ctx, "BSD:Accept", IsBlockingSocket(fd), + AcceptWork{ + .fd = fd, + .write_buffer = std::vector<u8>(ctx.GetWriteBufferSize()), + }); +} + void BSD::Bind(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); - IPC::ResponseBuilder rb{ctx, 4}; + LOG_DEBUG(Service, "called. fd={} addrlen={}", fd, ctx.GetReadBufferSize()); - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // ret - rb.Push<u32>(0); // bsd errno + BuildErrnoResponse(ctx, BindImpl(fd, ctx.ReadBuffer())); } void BSD::Connect(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); - IPC::ResponseBuilder rb{ctx, 4}; + LOG_DEBUG(Service, "called. fd={} addrlen={}", fd, ctx.GetReadBufferSize()); + + ExecuteWork(ctx, "BSD:Connect", IsBlockingSocket(fd), + ConnectWork{ + .fd = fd, + .addr = ctx.ReadBuffer(), + }); +} + +void BSD::GetPeerName(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + + LOG_DEBUG(Service, "called. fd={}", fd); + std::vector<u8> write_buffer(ctx.GetWriteBufferSize()); + const Errno bsd_errno = GetPeerNameImpl(fd, write_buffer); + + ctx.WriteBuffer(write_buffer); + + IPC::ResponseBuilder rb{ctx, 5}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // ret - rb.Push<u32>(0); // bsd errno + rb.Push<s32>(bsd_errno != Errno::SUCCESS ? -1 : 0); + rb.PushEnum(bsd_errno); + rb.Push<u32>(static_cast<u32>(write_buffer.size())); +} + +void BSD::GetSockName(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + + LOG_DEBUG(Service, "called. fd={}", fd); + + std::vector<u8> write_buffer(ctx.GetWriteBufferSize()); + const Errno bsd_errno = GetSockNameImpl(fd, write_buffer); + + ctx.WriteBuffer(write_buffer); + + IPC::ResponseBuilder rb{ctx, 5}; + rb.Push(RESULT_SUCCESS); + rb.Push<s32>(bsd_errno != Errno::SUCCESS ? -1 : 0); + rb.PushEnum(bsd_errno); + rb.Push<u32>(static_cast<u32>(write_buffer.size())); } void BSD::Listen(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + const s32 backlog = rp.Pop<s32>(); - IPC::ResponseBuilder rb{ctx, 4}; + LOG_DEBUG(Service, "called. fd={} backlog={}", fd, backlog); + + BuildErrnoResponse(ctx, ListenImpl(fd, backlog)); +} + +void BSD::Fcntl(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + const s32 cmd = rp.Pop<s32>(); + const s32 arg = rp.Pop<s32>(); + LOG_DEBUG(Service, "called. fd={} cmd={} arg={}", fd, cmd, arg); + + const auto [ret, bsd_errno] = FcntlImpl(fd, static_cast<FcntlCmd>(cmd), arg); + + IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // ret - rb.Push<u32>(0); // bsd errno + rb.Push<s32>(ret); + rb.PushEnum(bsd_errno); } void BSD::SetSockOpt(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; - IPC::ResponseBuilder rb{ctx, 4}; + const s32 fd = rp.Pop<s32>(); + const u32 level = rp.Pop<u32>(); + const OptName optname = static_cast<OptName>(rp.Pop<u32>()); - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // ret - rb.Push<u32>(0); // bsd errno + const std::vector<u8> buffer = ctx.ReadBuffer(); + const u8* optval = buffer.empty() ? nullptr : buffer.data(); + size_t optlen = buffer.size(); + + std::array<u64, 2> values; + if ((optname == OptName::SNDTIMEO || optname == OptName::RCVTIMEO) && buffer.size() == 8) { + std::memcpy(values.data(), buffer.data(), sizeof(values)); + optlen = sizeof(values); + optval = reinterpret_cast<const u8*>(values.data()); + } + + LOG_DEBUG(Service, "called. fd={} level={} optname=0x{:x} optlen={}", fd, level, + static_cast<u32>(optname), optlen); + + BuildErrnoResponse(ctx, SetSockOptImpl(fd, level, optname, optlen, optval)); +} + +void BSD::Shutdown(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const s32 fd = rp.Pop<s32>(); + const s32 how = rp.Pop<s32>(); + + LOG_DEBUG(Service, "called. fd={} how={}", fd, how); + + BuildErrnoResponse(ctx, ShutdownImpl(fd, how)); +} + +void BSD::Recv(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const s32 fd = rp.Pop<s32>(); + const u32 flags = rp.Pop<u32>(); + + LOG_DEBUG(Service, "called. fd={} flags=0x{:x} len={}", fd, flags, ctx.GetWriteBufferSize()); + + ExecuteWork(ctx, "BSD:Recv", IsBlockingSocket(fd), + RecvWork{ + .fd = fd, + .flags = flags, + .message = std::vector<u8>(ctx.GetWriteBufferSize()), + }); +} + +void BSD::RecvFrom(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const s32 fd = rp.Pop<s32>(); + const u32 flags = rp.Pop<u32>(); + + LOG_DEBUG(Service, "called. fd={} flags=0x{:x} len={} addrlen={}", fd, flags, + ctx.GetWriteBufferSize(0), ctx.GetWriteBufferSize(1)); + + ExecuteWork(ctx, "BSD:RecvFrom", IsBlockingSocket(fd), + RecvFromWork{ + .fd = fd, + .flags = flags, + .message = std::vector<u8>(ctx.GetWriteBufferSize(0)), + .addr = std::vector<u8>(ctx.GetWriteBufferSize(1)), + }); +} + +void BSD::Send(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const s32 fd = rp.Pop<s32>(); + const u32 flags = rp.Pop<u32>(); + + LOG_DEBUG(Service, "called. fd={} flags=0x{:x} len={}", fd, flags, ctx.GetReadBufferSize()); + + ExecuteWork(ctx, "BSD:Send", IsBlockingSocket(fd), + SendWork{ + .fd = fd, + .flags = flags, + .message = ctx.ReadBuffer(), + }); } void BSD::SendTo(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + const u32 flags = rp.Pop<u32>(); + + LOG_DEBUG(Service, "called. fd={} flags=0x{} len={} addrlen={}", fd, flags, + ctx.GetReadBufferSize(0), ctx.GetReadBufferSize(1)); + + ExecuteWork(ctx, "BSD:SendTo", IsBlockingSocket(fd), + SendToWork{ + .fd = fd, + .flags = flags, + .message = ctx.ReadBuffer(0), + .addr = ctx.ReadBuffer(1), + }); +} - IPC::ResponseBuilder rb{ctx, 4}; +void BSD::Write(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); - rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // ret - rb.Push<u32>(0); // bsd errno + LOG_DEBUG(Service, "called. fd={} len={}", fd, ctx.GetReadBufferSize()); + + ExecuteWork(ctx, "BSD:Write", IsBlockingSocket(fd), + SendWork{ + .fd = fd, + .flags = 0, + .message = ctx.ReadBuffer(), + }); } void BSD::Close(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const s32 fd = rp.Pop<s32>(); + + LOG_DEBUG(Service, "called. fd={}", fd); + + BuildErrnoResponse(ctx, CloseImpl(fd)); +} + +template <typename Work> +void BSD::ExecuteWork(Kernel::HLERequestContext& ctx, std::string_view sleep_reason, + bool is_blocking, Work work) { + if (!is_blocking) { + work.Execute(this); + work.Response(ctx); + return; + } + + // Signal a dummy response to make IPC validation happy + // This will be overwritten by the SleepClientThread callback + work.Response(ctx); + + auto worker = worker_pool.CaptureWorker(); + + ctx.SleepClientThread(std::string(sleep_reason), std::numeric_limits<u64>::max(), + worker->Callback<Work>(), worker->KernelEvent()); + + worker->SendWork(std::move(work)); +} + +std::pair<s32, Errno> BSD::SocketImpl(Domain domain, Type type, Protocol protocol) { + if (type == Type::SEQPACKET) { + UNIMPLEMENTED_MSG("SOCK_SEQPACKET errno management"); + } else if (type == Type::RAW && (domain != Domain::INET || protocol != Protocol::ICMP)) { + UNIMPLEMENTED_MSG("SOCK_RAW errno management"); + } + + [[maybe_unused]] const bool unk_flag = (static_cast<u32>(type) & 0x20000000) != 0; + UNIMPLEMENTED_IF_MSG(unk_flag, "Unknown flag in type"); + type = static_cast<Type>(static_cast<u32>(type) & ~0x20000000); + + const s32 fd = FindFreeFileDescriptorHandle(); + if (fd < 0) { + LOG_ERROR(Service, "No more file descriptors available"); + return {-1, Errno::MFILE}; + } + + FileDescriptor& descriptor = file_descriptors[fd].emplace(); + // ENONMEM might be thrown here + + LOG_INFO(Service, "New socket fd={}", fd); + + descriptor.socket = std::make_unique<Network::Socket>(); + descriptor.socket->Initialize(Translate(domain), Translate(type), Translate(type, protocol)); + descriptor.is_connection_based = IsConnectionBased(type); + + return {fd, Errno::SUCCESS}; +} +std::pair<s32, Errno> BSD::PollImpl(std::vector<u8>& write_buffer, std::vector<u8> read_buffer, + s32 nfds, s32 timeout) { + if (write_buffer.size() < nfds * sizeof(PollFD)) { + return {-1, Errno::INVAL}; + } + + if (nfds == 0) { + // When no entries are provided, -1 is returned with errno zero + return {-1, Errno::SUCCESS}; + } + + const size_t length = std::min(read_buffer.size(), write_buffer.size()); + std::vector<PollFD> fds(nfds); + std::memcpy(fds.data(), read_buffer.data(), length); + + if (timeout >= 0) { + const s64 seconds = timeout / 1000; + const u64 nanoseconds = 1'000'000 * (static_cast<u64>(timeout) % 1000); + + if (seconds < 0) { + return {-1, Errno::INVAL}; + } + if (nanoseconds > 999'999'999) { + return {-1, Errno::INVAL}; + } + } else if (timeout != -1) { + return {-1, Errno::INVAL}; + } + + for (PollFD& pollfd : fds) { + ASSERT(pollfd.revents == 0); + + if (pollfd.fd > static_cast<s32>(MAX_FD) || pollfd.fd < 0) { + LOG_ERROR(Service, "File descriptor handle={} is invalid", pollfd.fd); + pollfd.revents = 0; + return {0, Errno::SUCCESS}; + } + + const std::optional<FileDescriptor>& descriptor = file_descriptors[pollfd.fd]; + if (!descriptor) { + LOG_ERROR(Service, "File descriptor handle={} is not allocated", pollfd.fd); + pollfd.revents = POLL_NVAL; + return {0, Errno::SUCCESS}; + } + } + + std::vector<Network::PollFD> host_pollfds(fds.size()); + std::transform(fds.begin(), fds.end(), host_pollfds.begin(), [this](PollFD pollfd) { + Network::PollFD result; + result.socket = file_descriptors[pollfd.fd]->socket.get(); + result.events = TranslatePollEventsToHost(pollfd.events); + result.revents = 0; + return result; + }); + + const auto result = Network::Poll(host_pollfds, timeout); + + const size_t num = host_pollfds.size(); + for (size_t i = 0; i < num; ++i) { + fds[i].revents = TranslatePollEventsToGuest(host_pollfds[i].revents); + } + std::memcpy(write_buffer.data(), fds.data(), length); + + return Translate(result); +} + +std::pair<s32, Errno> BSD::AcceptImpl(s32 fd, std::vector<u8>& write_buffer) { + if (!IsFileDescriptorValid(fd)) { + return {-1, Errno::BADF}; + } + + const s32 new_fd = FindFreeFileDescriptorHandle(); + if (new_fd < 0) { + LOG_ERROR(Service, "No more file descriptors available"); + return {-1, Errno::MFILE}; + } + + FileDescriptor& descriptor = *file_descriptors[fd]; + auto [result, bsd_errno] = descriptor.socket->Accept(); + if (bsd_errno != Network::Errno::SUCCESS) { + return {-1, Translate(bsd_errno)}; + } + + FileDescriptor& new_descriptor = file_descriptors[new_fd].emplace(); + new_descriptor.socket = std::move(result.socket); + new_descriptor.is_connection_based = descriptor.is_connection_based; + + ASSERT(write_buffer.size() == sizeof(SockAddrIn)); + const SockAddrIn guest_addr_in = Translate(result.sockaddr_in); + std::memcpy(write_buffer.data(), &guest_addr_in, sizeof(guest_addr_in)); + + return {new_fd, Errno::SUCCESS}; +} + +Errno BSD::BindImpl(s32 fd, const std::vector<u8>& addr) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + ASSERT(addr.size() == sizeof(SockAddrIn)); + SockAddrIn addr_in; + std::memcpy(&addr_in, addr.data(), sizeof(addr_in)); + + return Translate(file_descriptors[fd]->socket->Bind(Translate(addr_in))); +} + +Errno BSD::ConnectImpl(s32 fd, const std::vector<u8>& addr) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + + UNIMPLEMENTED_IF(addr.size() != sizeof(SockAddrIn)); + SockAddrIn addr_in; + std::memcpy(&addr_in, addr.data(), sizeof(addr_in)); + + return Translate(file_descriptors[fd]->socket->Connect(Translate(addr_in))); +} + +Errno BSD::GetPeerNameImpl(s32 fd, std::vector<u8>& write_buffer) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + + const auto [addr_in, bsd_errno] = file_descriptors[fd]->socket->GetPeerName(); + if (bsd_errno != Network::Errno::SUCCESS) { + return Translate(bsd_errno); + } + const SockAddrIn guest_addrin = Translate(addr_in); + + ASSERT(write_buffer.size() == sizeof(guest_addrin)); + std::memcpy(write_buffer.data(), &guest_addrin, sizeof(guest_addrin)); + return Translate(bsd_errno); +} + +Errno BSD::GetSockNameImpl(s32 fd, std::vector<u8>& write_buffer) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + + const auto [addr_in, bsd_errno] = file_descriptors[fd]->socket->GetSockName(); + if (bsd_errno != Network::Errno::SUCCESS) { + return Translate(bsd_errno); + } + const SockAddrIn guest_addrin = Translate(addr_in); + + ASSERT(write_buffer.size() == sizeof(guest_addrin)); + std::memcpy(write_buffer.data(), &guest_addrin, sizeof(guest_addrin)); + return Translate(bsd_errno); +} + +Errno BSD::ListenImpl(s32 fd, s32 backlog) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + return Translate(file_descriptors[fd]->socket->Listen(backlog)); +} + +std::pair<s32, Errno> BSD::FcntlImpl(s32 fd, FcntlCmd cmd, s32 arg) { + if (!IsFileDescriptorValid(fd)) { + return {-1, Errno::BADF}; + } + + FileDescriptor& descriptor = *file_descriptors[fd]; + + switch (cmd) { + case FcntlCmd::GETFL: + ASSERT(arg == 0); + return {descriptor.flags, Errno::SUCCESS}; + case FcntlCmd::SETFL: { + const bool enable = (arg & FLAG_O_NONBLOCK) != 0; + const Errno bsd_errno = Translate(descriptor.socket->SetNonBlock(enable)); + if (bsd_errno != Errno::SUCCESS) { + return {-1, bsd_errno}; + } + descriptor.flags = arg; + return {0, Errno::SUCCESS}; + } + default: + UNIMPLEMENTED_MSG("Unimplemented cmd={}", static_cast<int>(cmd)); + return {-1, Errno::SUCCESS}; + } +} + +Errno BSD::SetSockOptImpl(s32 fd, u32 level, OptName optname, size_t optlen, const void* optval) { + UNIMPLEMENTED_IF(level != 0xffff); // SOL_SOCKET + + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + + Network::Socket* const socket = file_descriptors[fd]->socket.get(); + + if (optname == OptName::LINGER) { + ASSERT(optlen == sizeof(Linger)); + Linger linger; + std::memcpy(&linger, optval, sizeof(linger)); + ASSERT(linger.onoff == 0 || linger.onoff == 1); + + return Translate(socket->SetLinger(linger.onoff != 0, linger.linger)); + } + + ASSERT(optlen == sizeof(u32)); + u32 value; + std::memcpy(&value, optval, sizeof(value)); + + switch (optname) { + case OptName::REUSEADDR: + ASSERT(value == 0 || value == 1); + return Translate(socket->SetReuseAddr(value != 0)); + case OptName::BROADCAST: + ASSERT(value == 0 || value == 1); + return Translate(socket->SetBroadcast(value != 0)); + case OptName::SNDBUF: + return Translate(socket->SetSndBuf(value)); + case OptName::RCVBUF: + return Translate(socket->SetRcvBuf(value)); + case OptName::SNDTIMEO: + return Translate(socket->SetSndTimeo(value)); + case OptName::RCVTIMEO: + return Translate(socket->SetRcvTimeo(value)); + default: + UNIMPLEMENTED_MSG("Unimplemented optname={}", static_cast<int>(optname)); + return Errno::SUCCESS; + } +} + +Errno BSD::ShutdownImpl(s32 fd, s32 how) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + const Network::ShutdownHow host_how = Translate(static_cast<ShutdownHow>(how)); + return Translate(file_descriptors[fd]->socket->Shutdown(host_how)); +} + +std::pair<s32, Errno> BSD::RecvImpl(s32 fd, u32 flags, std::vector<u8>& message) { + if (!IsFileDescriptorValid(fd)) { + return {-1, Errno::BADF}; + } + return Translate(file_descriptors[fd]->socket->Recv(flags, message)); +} + +std::pair<s32, Errno> BSD::RecvFromImpl(s32 fd, u32 flags, std::vector<u8>& message, + std::vector<u8>& addr) { + if (!IsFileDescriptorValid(fd)) { + return {-1, Errno::BADF}; + } + + FileDescriptor& descriptor = *file_descriptors[fd]; + + Network::SockAddrIn addr_in{}; + Network::SockAddrIn* p_addr_in = nullptr; + if (descriptor.is_connection_based) { + // Connection based file descriptors (e.g. TCP) zero addr + addr.clear(); + } else { + p_addr_in = &addr_in; + } + + // Apply flags + if ((flags & FLAG_MSG_DONTWAIT) != 0) { + flags &= ~FLAG_MSG_DONTWAIT; + if ((descriptor.flags & FLAG_O_NONBLOCK) == 0) { + descriptor.socket->SetNonBlock(true); + } + } + + const auto [ret, bsd_errno] = Translate(descriptor.socket->RecvFrom(flags, message, p_addr_in)); + + // Restore original state + if ((descriptor.flags & FLAG_O_NONBLOCK) == 0) { + descriptor.socket->SetNonBlock(false); + } + + if (p_addr_in) { + if (ret < 0) { + addr.clear(); + } else { + ASSERT(addr.size() == sizeof(SockAddrIn)); + const SockAddrIn result = Translate(addr_in); + std::memcpy(addr.data(), &result, sizeof(result)); + } + } + + return {ret, bsd_errno}; +} + +std::pair<s32, Errno> BSD::SendImpl(s32 fd, u32 flags, const std::vector<u8>& message) { + if (!IsFileDescriptorValid(fd)) { + return {-1, Errno::BADF}; + } + return Translate(file_descriptors[fd]->socket->Send(message, flags)); +} + +std::pair<s32, Errno> BSD::SendToImpl(s32 fd, u32 flags, const std::vector<u8>& message, + const std::vector<u8>& addr) { + if (!IsFileDescriptorValid(fd)) { + return {-1, Errno::BADF}; + } + + Network::SockAddrIn addr_in; + Network::SockAddrIn* p_addr_in = nullptr; + if (!addr.empty()) { + ASSERT(addr.size() == sizeof(SockAddrIn)); + SockAddrIn guest_addr_in; + std::memcpy(&guest_addr_in, addr.data(), sizeof(guest_addr_in)); + addr_in = Translate(guest_addr_in); + p_addr_in = &addr_in; + } + + return Translate(file_descriptors[fd]->socket->SendTo(flags, message, p_addr_in)); +} + +Errno BSD::CloseImpl(s32 fd) { + if (!IsFileDescriptorValid(fd)) { + return Errno::BADF; + } + + const Errno bsd_errno = Translate(file_descriptors[fd]->socket->Close()); + if (bsd_errno != Errno::SUCCESS) { + return bsd_errno; + } + + LOG_INFO(Service, "Close socket fd={}", fd); + + file_descriptors[fd].reset(); + return bsd_errno; +} + +s32 BSD::FindFreeFileDescriptorHandle() noexcept { + for (s32 fd = 0; fd < static_cast<s32>(file_descriptors.size()); ++fd) { + if (!file_descriptors[fd]) { + return fd; + } + } + return -1; +} + +bool BSD::IsFileDescriptorValid(s32 fd) const noexcept { + if (fd > static_cast<s32>(MAX_FD) || fd < 0) { + LOG_ERROR(Service, "Invalid file descriptor handle={}", fd); + return false; + } + if (!file_descriptors[fd]) { + LOG_ERROR(Service, "File descriptor handle={} is not allocated", fd); + return false; + } + return true; +} + +bool BSD::IsBlockingSocket(s32 fd) const noexcept { + // Inform invalid sockets as non-blocking + // This way we avoid using a worker thread as it will fail without blocking host + if (fd > static_cast<s32>(MAX_FD) || fd < 0) { + return false; + } + if (!file_descriptors[fd]) { + return false; + } + return (file_descriptors[fd]->flags & FLAG_O_NONBLOCK) != 0; +} + +void BSD::BuildErrnoResponse(Kernel::HLERequestContext& ctx, Errno bsd_errno) const noexcept { IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); // ret - rb.Push<u32>(0); // bsd errno + rb.Push<s32>(bsd_errno == Errno::SUCCESS ? 0 : -1); + rb.PushEnum(bsd_errno); } -BSD::BSD(const char* name) : ServiceFramework(name) { +BSD::BSD(Core::System& system, const char* name) + : ServiceFramework(name), worker_pool{system, this} { // clang-format off static const FunctionInfo functions[] = { {0, &BSD::RegisterClient, "RegisterClient"}, @@ -121,25 +837,25 @@ BSD::BSD(const char* name) : ServiceFramework(name) { {3, nullptr, "SocketExempt"}, {4, nullptr, "Open"}, {5, &BSD::Select, "Select"}, - {6, nullptr, "Poll"}, + {6, &BSD::Poll, "Poll"}, {7, nullptr, "Sysctl"}, - {8, nullptr, "Recv"}, - {9, nullptr, "RecvFrom"}, - {10, nullptr, "Send"}, + {8, &BSD::Recv, "Recv"}, + {9, &BSD::RecvFrom, "RecvFrom"}, + {10, &BSD::Send, "Send"}, {11, &BSD::SendTo, "SendTo"}, - {12, nullptr, "Accept"}, + {12, &BSD::Accept, "Accept"}, {13, &BSD::Bind, "Bind"}, {14, &BSD::Connect, "Connect"}, - {15, nullptr, "GetPeerName"}, - {16, nullptr, "GetSockName"}, + {15, &BSD::GetPeerName, "GetPeerName"}, + {16, &BSD::GetSockName, "GetSockName"}, {17, nullptr, "GetSockOpt"}, {18, &BSD::Listen, "Listen"}, {19, nullptr, "Ioctl"}, - {20, nullptr, "Fcntl"}, + {20, &BSD::Fcntl, "Fcntl"}, {21, &BSD::SetSockOpt, "SetSockOpt"}, - {22, nullptr, "Shutdown"}, + {22, &BSD::Shutdown, "Shutdown"}, {23, nullptr, "ShutdownAllSockets"}, - {24, nullptr, "Write"}, + {24, &BSD::Write, "Write"}, {25, nullptr, "Read"}, {26, &BSD::Close, "Close"}, {27, nullptr, "DuplicateSocket"}, @@ -148,6 +864,7 @@ BSD::BSD(const char* name) : ServiceFramework(name) { {30, nullptr, "SendMMsg"}, {31, nullptr, "EventFd"}, {32, nullptr, "RegisterResourceStatisticsName"}, + {33, nullptr, "Initialize2"}, }; // clang-format on diff --git a/src/core/hle/service/sockets/bsd.h b/src/core/hle/service/sockets/bsd.h index 3098e3baf..357531951 100644 --- a/src/core/hle/service/sockets/bsd.h +++ b/src/core/hle/service/sockets/bsd.h @@ -4,30 +4,174 @@ #pragma once +#include <memory> +#include <string_view> +#include <vector> + +#include "common/common_types.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/service.h" +#include "core/hle/service/sockets/blocking_worker.h" +#include "core/hle/service/sockets/sockets.h" + +namespace Core { +class System; +} + +namespace Network { +class Socket; +} namespace Service::Sockets { class BSD final : public ServiceFramework<BSD> { public: - explicit BSD(const char* name); + explicit BSD(Core::System& system, const char* name); ~BSD() override; private: + /// Maximum number of file descriptors + static constexpr size_t MAX_FD = 128; + + struct FileDescriptor { + std::unique_ptr<Network::Socket> socket; + s32 flags = 0; + bool is_connection_based = false; + }; + + struct PollWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 nfds; + s32 timeout; + std::vector<u8> read_buffer; + std::vector<u8> write_buffer; + s32 ret{}; + Errno bsd_errno{}; + }; + + struct AcceptWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 fd; + std::vector<u8> write_buffer; + s32 ret{}; + Errno bsd_errno{}; + }; + + struct ConnectWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 fd; + std::vector<u8> addr; + Errno bsd_errno{}; + }; + + struct RecvWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 fd; + u32 flags; + std::vector<u8> message; + s32 ret{}; + Errno bsd_errno{}; + }; + + struct RecvFromWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 fd; + u32 flags; + std::vector<u8> message; + std::vector<u8> addr; + s32 ret{}; + Errno bsd_errno{}; + }; + + struct SendWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 fd; + u32 flags; + std::vector<u8> message; + s32 ret{}; + Errno bsd_errno{}; + }; + + struct SendToWork { + void Execute(BSD* bsd); + void Response(Kernel::HLERequestContext& ctx); + + s32 fd; + u32 flags; + std::vector<u8> message; + std::vector<u8> addr; + s32 ret{}; + Errno bsd_errno{}; + }; + void RegisterClient(Kernel::HLERequestContext& ctx); void StartMonitoring(Kernel::HLERequestContext& ctx); void Socket(Kernel::HLERequestContext& ctx); void Select(Kernel::HLERequestContext& ctx); + void Poll(Kernel::HLERequestContext& ctx); + void Accept(Kernel::HLERequestContext& ctx); void Bind(Kernel::HLERequestContext& ctx); void Connect(Kernel::HLERequestContext& ctx); + void GetPeerName(Kernel::HLERequestContext& ctx); + void GetSockName(Kernel::HLERequestContext& ctx); void Listen(Kernel::HLERequestContext& ctx); + void Fcntl(Kernel::HLERequestContext& ctx); void SetSockOpt(Kernel::HLERequestContext& ctx); + void Shutdown(Kernel::HLERequestContext& ctx); + void Recv(Kernel::HLERequestContext& ctx); + void RecvFrom(Kernel::HLERequestContext& ctx); + void Send(Kernel::HLERequestContext& ctx); void SendTo(Kernel::HLERequestContext& ctx); + void Write(Kernel::HLERequestContext& ctx); void Close(Kernel::HLERequestContext& ctx); - /// Id to use for the next open file descriptor. - u32 next_fd = 1; + template <typename Work> + void ExecuteWork(Kernel::HLERequestContext& ctx, std::string_view sleep_reason, + bool is_blocking, Work work); + + std::pair<s32, Errno> SocketImpl(Domain domain, Type type, Protocol protocol); + std::pair<s32, Errno> PollImpl(std::vector<u8>& write_buffer, std::vector<u8> read_buffer, + s32 nfds, s32 timeout); + std::pair<s32, Errno> AcceptImpl(s32 fd, std::vector<u8>& write_buffer); + Errno BindImpl(s32 fd, const std::vector<u8>& addr); + Errno ConnectImpl(s32 fd, const std::vector<u8>& addr); + Errno GetPeerNameImpl(s32 fd, std::vector<u8>& write_buffer); + Errno GetSockNameImpl(s32 fd, std::vector<u8>& write_buffer); + Errno ListenImpl(s32 fd, s32 backlog); + std::pair<s32, Errno> FcntlImpl(s32 fd, FcntlCmd cmd, s32 arg); + Errno SetSockOptImpl(s32 fd, u32 level, OptName optname, size_t optlen, const void* optval); + Errno ShutdownImpl(s32 fd, s32 how); + std::pair<s32, Errno> RecvImpl(s32 fd, u32 flags, std::vector<u8>& message); + std::pair<s32, Errno> RecvFromImpl(s32 fd, u32 flags, std::vector<u8>& message, + std::vector<u8>& addr); + std::pair<s32, Errno> SendImpl(s32 fd, u32 flags, const std::vector<u8>& message); + std::pair<s32, Errno> SendToImpl(s32 fd, u32 flags, const std::vector<u8>& message, + const std::vector<u8>& addr); + Errno CloseImpl(s32 fd); + + s32 FindFreeFileDescriptorHandle() noexcept; + bool IsFileDescriptorValid(s32 fd) const noexcept; + bool IsBlockingSocket(s32 fd) const noexcept; + + void BuildErrnoResponse(Kernel::HLERequestContext& ctx, Errno bsd_errno) const noexcept; + + std::array<std::optional<FileDescriptor>, MAX_FD> file_descriptors; + + BlockingWorkerPool<BSD, PollWork, AcceptWork, ConnectWork, RecvWork, RecvFromWork, SendWork, + SendToWork> + worker_pool; }; class BSDCFG final : public ServiceFramework<BSDCFG> { diff --git a/src/core/hle/service/sockets/nsd.cpp b/src/core/hle/service/sockets/nsd.cpp index dc70fd6fe..40d781124 100644 --- a/src/core/hle/service/sockets/nsd.cpp +++ b/src/core/hle/service/sockets/nsd.cpp @@ -14,6 +14,7 @@ NSD::NSD(const char* name) : ServiceFramework(name) { {12, nullptr, "GetDeviceId"}, {13, nullptr, "DeleteSettings"}, {14, nullptr, "ImportSettings"}, + {15, nullptr, "SetChangeEnvironmentIdentifierDisabled"}, {20, nullptr, "Resolve"}, {21, nullptr, "ResolveEx"}, {30, nullptr, "GetNasServiceSetting"}, @@ -28,6 +29,11 @@ NSD::NSD(const char* name) : ServiceFramework(name) { {60, nullptr, "ReadSaveDataFromFsForTest"}, {61, nullptr, "WriteSaveDataToFsForTest"}, {62, nullptr, "DeleteSaveDataOfFsForTest"}, + {63, nullptr, "IsChangeEnvironmentIdentifierDisabled"}, + {64, nullptr, "SetWithoutDomainExchangeFqdns"}, + {100, nullptr, "GetApplicationServerEnvironmentType"}, + {101, nullptr, "SetApplicationServerEnvironmentType"}, + {102, nullptr, "DeleteApplicationServerEnvironmentType"}, }; // clang-format on diff --git a/src/core/hle/service/sockets/sfdnsres.cpp b/src/core/hle/service/sockets/sfdnsres.cpp index 852e71e4b..e3017451f 100644 --- a/src/core/hle/service/sockets/sfdnsres.cpp +++ b/src/core/hle/service/sockets/sfdnsres.cpp @@ -7,7 +7,7 @@ namespace Service::Sockets { -void SFDNSRES::GetAddrInfo(Kernel::HLERequestContext& ctx) { +void SFDNSRES::GetAddrInfoRequest(Kernel::HLERequestContext& ctx) { struct Parameters { u8 use_nsd_resolve; u32 unknown; @@ -29,15 +29,20 @@ SFDNSRES::SFDNSRES() : ServiceFramework("sfdnsres") { static const FunctionInfo functions[] = { {0, nullptr, "SetDnsAddressesPrivate"}, {1, nullptr, "GetDnsAddressPrivate"}, - {2, nullptr, "GetHostByName"}, - {3, nullptr, "GetHostByAddr"}, - {4, nullptr, "GetHostStringError"}, - {5, nullptr, "GetGaiStringError"}, - {6, &SFDNSRES::GetAddrInfo, "GetAddrInfo"}, - {7, nullptr, "GetNameInfo"}, - {8, nullptr, "RequestCancelHandle"}, - {9, nullptr, "CancelSocketCall"}, - {11, nullptr, "ClearDnsIpServerAddressArray"}, + {2, nullptr, "GetHostByNameRequest"}, + {3, nullptr, "GetHostByAddrRequest"}, + {4, nullptr, "GetHostStringErrorRequest"}, + {5, nullptr, "GetGaiStringErrorRequest"}, + {6, &SFDNSRES::GetAddrInfoRequest, "GetAddrInfoRequest"}, + {7, nullptr, "GetNameInfoRequest"}, + {8, nullptr, "RequestCancelHandleRequest"}, + {9, nullptr, "CancelRequest"}, + {10, nullptr, "GetHostByNameRequestWithOptions"}, + {11, nullptr, "GetHostByAddrRequestWithOptions"}, + {12, nullptr, "GetAddrInfoRequestWithOptions"}, + {13, nullptr, "GetNameInfoRequestWithOptions"}, + {14, nullptr, "ResolverSetOptionRequest"}, + {15, nullptr, "ResolverGetOptionRequest"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/sockets/sfdnsres.h b/src/core/hle/service/sockets/sfdnsres.h index eda432903..acd3647bb 100644 --- a/src/core/hle/service/sockets/sfdnsres.h +++ b/src/core/hle/service/sockets/sfdnsres.h @@ -15,7 +15,7 @@ public: ~SFDNSRES() override; private: - void GetAddrInfo(Kernel::HLERequestContext& ctx); + void GetAddrInfoRequest(Kernel::HLERequestContext& ctx); }; } // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/sockets.cpp b/src/core/hle/service/sockets/sockets.cpp index 08d2d306a..1d27f7906 100644 --- a/src/core/hle/service/sockets/sockets.cpp +++ b/src/core/hle/service/sockets/sockets.cpp @@ -10,9 +10,9 @@ namespace Service::Sockets { -void InstallInterfaces(SM::ServiceManager& service_manager) { - std::make_shared<BSD>("bsd:s")->InstallAsService(service_manager); - std::make_shared<BSD>("bsd:u")->InstallAsService(service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { + std::make_shared<BSD>(system, "bsd:s")->InstallAsService(service_manager); + std::make_shared<BSD>(system, "bsd:u")->InstallAsService(service_manager); std::make_shared<BSDCFG>()->InstallAsService(service_manager); std::make_shared<ETHC_C>()->InstallAsService(service_manager); diff --git a/src/core/hle/service/sockets/sockets.h b/src/core/hle/service/sockets/sockets.h index ca8a6a7e0..89a410076 100644 --- a/src/core/hle/service/sockets/sockets.h +++ b/src/core/hle/service/sockets/sockets.h @@ -4,11 +4,94 @@ #pragma once +#include "common/common_types.h" #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Sockets { +enum class Errno : u32 { + SUCCESS = 0, + BADF = 9, + AGAIN = 11, + INVAL = 22, + MFILE = 24, + NOTCONN = 107, +}; + +enum class Domain : u32 { + INET = 2, +}; + +enum class Type : u32 { + STREAM = 1, + DGRAM = 2, + RAW = 3, + SEQPACKET = 5, +}; + +enum class Protocol : u32 { + UNSPECIFIED = 0, + ICMP = 1, + TCP = 6, + UDP = 17, +}; + +enum class OptName : u32 { + REUSEADDR = 0x4, + BROADCAST = 0x20, + LINGER = 0x80, + SNDBUF = 0x1001, + RCVBUF = 0x1002, + SNDTIMEO = 0x1005, + RCVTIMEO = 0x1006, +}; + +enum class ShutdownHow : s32 { + RD = 0, + WR = 1, + RDWR = 2, +}; + +enum class FcntlCmd : s32 { + GETFL = 3, + SETFL = 4, +}; + +struct SockAddrIn { + u8 len; + u8 family; + u16 portno; + std::array<u8, 4> ip; + std::array<u8, 8> zeroes; +}; + +struct PollFD { + s32 fd; + u16 events; + u16 revents; +}; + +struct Linger { + u32 onoff; + u32 linger; +}; + +constexpr u16 POLL_IN = 0x01; +constexpr u16 POLL_PRI = 0x02; +constexpr u16 POLL_OUT = 0x04; +constexpr u16 POLL_ERR = 0x08; +constexpr u16 POLL_HUP = 0x10; +constexpr u16 POLL_NVAL = 0x20; + +constexpr u32 FLAG_MSG_DONTWAIT = 0x80; + +constexpr u32 FLAG_O_NONBLOCK = 0x800; + /// Registers all Sockets services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/sockets_translate.cpp b/src/core/hle/service/sockets/sockets_translate.cpp new file mode 100644 index 000000000..2e626fd86 --- /dev/null +++ b/src/core/hle/service/sockets/sockets_translate.cpp @@ -0,0 +1,165 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <utility> + +#include "common/assert.h" +#include "common/common_types.h" +#include "core/hle/service/sockets/sockets.h" +#include "core/hle/service/sockets/sockets_translate.h" +#include "core/network/network.h" + +namespace Service::Sockets { + +Errno Translate(Network::Errno value) { + switch (value) { + case Network::Errno::SUCCESS: + return Errno::SUCCESS; + case Network::Errno::BADF: + return Errno::BADF; + case Network::Errno::AGAIN: + return Errno::AGAIN; + case Network::Errno::INVAL: + return Errno::INVAL; + case Network::Errno::MFILE: + return Errno::MFILE; + case Network::Errno::NOTCONN: + return Errno::NOTCONN; + default: + UNIMPLEMENTED_MSG("Unimplemented errno={}", static_cast<int>(value)); + return Errno::SUCCESS; + } +} + +std::pair<s32, Errno> Translate(std::pair<s32, Network::Errno> value) { + return {value.first, Translate(value.second)}; +} + +Network::Domain Translate(Domain domain) { + switch (domain) { + case Domain::INET: + return Network::Domain::INET; + default: + UNIMPLEMENTED_MSG("Unimplemented domain={}", static_cast<int>(domain)); + return {}; + } +} + +Domain Translate(Network::Domain domain) { + switch (domain) { + case Network::Domain::INET: + return Domain::INET; + default: + UNIMPLEMENTED_MSG("Unimplemented domain={}", static_cast<int>(domain)); + return {}; + } +} + +Network::Type Translate(Type type) { + switch (type) { + case Type::STREAM: + return Network::Type::STREAM; + case Type::DGRAM: + return Network::Type::DGRAM; + default: + UNIMPLEMENTED_MSG("Unimplemented type={}", static_cast<int>(type)); + } +} + +Network::Protocol Translate(Type type, Protocol protocol) { + switch (protocol) { + case Protocol::UNSPECIFIED: + LOG_WARNING(Service, "Unspecified protocol, assuming protocol from type"); + switch (type) { + case Type::DGRAM: + return Network::Protocol::UDP; + case Type::STREAM: + return Network::Protocol::TCP; + default: + return Network::Protocol::TCP; + } + case Protocol::TCP: + return Network::Protocol::TCP; + case Protocol::UDP: + return Network::Protocol::UDP; + default: + UNIMPLEMENTED_MSG("Unimplemented protocol={}", static_cast<int>(protocol)); + return Network::Protocol::TCP; + } +} + +u16 TranslatePollEventsToHost(u32 flags) { + u32 result = 0; + const auto translate = [&result, &flags](u32 from, u32 to) { + if ((flags & from) != 0) { + flags &= ~from; + result |= to; + } + }; + translate(POLL_IN, Network::POLL_IN); + translate(POLL_PRI, Network::POLL_PRI); + translate(POLL_OUT, Network::POLL_OUT); + translate(POLL_ERR, Network::POLL_ERR); + translate(POLL_HUP, Network::POLL_HUP); + translate(POLL_NVAL, Network::POLL_NVAL); + + UNIMPLEMENTED_IF_MSG(flags != 0, "Unimplemented flags={}", flags); + return static_cast<u16>(result); +} + +u16 TranslatePollEventsToGuest(u32 flags) { + u32 result = 0; + const auto translate = [&result, &flags](u32 from, u32 to) { + if ((flags & from) != 0) { + flags &= ~from; + result |= to; + } + }; + + translate(Network::POLL_IN, POLL_IN); + translate(Network::POLL_PRI, POLL_PRI); + translate(Network::POLL_OUT, POLL_OUT); + translate(Network::POLL_ERR, POLL_ERR); + translate(Network::POLL_HUP, POLL_HUP); + translate(Network::POLL_NVAL, POLL_NVAL); + + UNIMPLEMENTED_IF_MSG(flags != 0, "Unimplemented flags={}", flags); + return static_cast<u16>(result); +} + +Network::SockAddrIn Translate(SockAddrIn value) { + ASSERT(value.len == 0 || value.len == sizeof(value)); + + return { + .family = Translate(static_cast<Domain>(value.family)), + .ip = value.ip, + .portno = static_cast<u16>(value.portno >> 8 | value.portno << 8), + }; +} + +SockAddrIn Translate(Network::SockAddrIn value) { + return { + .len = sizeof(SockAddrIn), + .family = static_cast<u8>(Translate(value.family)), + .portno = static_cast<u16>(value.portno >> 8 | value.portno << 8), + .ip = value.ip, + .zeroes = {}, + }; +} + +Network::ShutdownHow Translate(ShutdownHow how) { + switch (how) { + case ShutdownHow::RD: + return Network::ShutdownHow::RD; + case ShutdownHow::WR: + return Network::ShutdownHow::WR; + case ShutdownHow::RDWR: + return Network::ShutdownHow::RDWR; + default: + UNIMPLEMENTED_MSG("Unimplemented how={}", static_cast<int>(how)); + return {}; + } +} + +} // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/sockets_translate.h b/src/core/hle/service/sockets/sockets_translate.h new file mode 100644 index 000000000..e498913d4 --- /dev/null +++ b/src/core/hle/service/sockets/sockets_translate.h @@ -0,0 +1,48 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <utility> + +#include "common/common_types.h" +#include "core/hle/service/sockets/sockets.h" +#include "core/network/network.h" + +namespace Service::Sockets { + +/// Translate abstract errno to guest errno +Errno Translate(Network::Errno value); + +/// Translate abstract return value errno pair to guest return value errno pair +std::pair<s32, Errno> Translate(std::pair<s32, Network::Errno> value); + +/// Translate guest domain to abstract domain +Network::Domain Translate(Domain domain); + +/// Translate abstract domain to guest domain +Domain Translate(Network::Domain domain); + +/// Translate guest type to abstract type +Network::Type Translate(Type type); + +/// Translate guest protocol to abstract protocol +Network::Protocol Translate(Type type, Protocol protocol); + +/// Translate abstract poll event flags to guest poll event flags +u16 TranslatePollEventsToHost(u32 flags); + +/// Translate guest poll event flags to abstract poll event flags +u16 TranslatePollEventsToGuest(u32 flags); + +/// Translate guest socket address structure to abstract socket address structure +Network::SockAddrIn Translate(SockAddrIn value); + +/// Translate abstract socket address structure to guest socket address structure +SockAddrIn Translate(Network::SockAddrIn value); + +/// Translate guest shutdown mode to abstract shutdown mode +Network::ShutdownHow Translate(ShutdownHow how); + +} // namespace Service::Sockets diff --git a/src/core/hle/service/spl/module.cpp b/src/core/hle/service/spl/module.cpp index e724d4ab8..865ed3b91 100644 --- a/src/core/hle/service/spl/module.cpp +++ b/src/core/hle/service/spl/module.cpp @@ -19,7 +19,7 @@ namespace Service::SPL { Module::Interface::Interface(std::shared_ptr<Module> module, const char* name) : ServiceFramework(name), module(std::move(module)), - rng(Settings::values.rng_seed.value_or(std::time(nullptr))) {} + rng(Settings::values.rng_seed.GetValue().value_or(std::time(nullptr))) {} Module::Interface::~Interface() = default; diff --git a/src/core/hle/service/spl/spl.cpp b/src/core/hle/service/spl/spl.cpp index 70cb41905..773551464 100644 --- a/src/core/hle/service/spl/spl.cpp +++ b/src/core/hle/service/spl/spl.cpp @@ -9,35 +9,36 @@ namespace Service::SPL { SPL::SPL(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "spl:") { static const FunctionInfo functions[] = { {0, nullptr, "GetConfig"}, - {1, nullptr, "UserExpMod"}, + {1, nullptr, "ModularExponentiate"}, {2, nullptr, "GenerateAesKek"}, {3, nullptr, "LoadAesKey"}, {4, nullptr, "GenerateAesKey"}, {5, nullptr, "SetConfig"}, {7, &SPL::GetRandomBytes, "GetRandomBytes"}, - {9, nullptr, "LoadSecureExpModKey"}, - {10, nullptr, "SecureExpMod"}, + {9, nullptr, "ImportLotusKey"}, + {10, nullptr, "DecryptLotusMessage"}, {11, nullptr, "IsDevelopment"}, {12, nullptr, "GenerateSpecificAesKey"}, - {13, nullptr, "DecryptPrivk"}, + {13, nullptr, "DecryptDeviceUniqueData"}, {14, nullptr, "DecryptAesKey"}, - {15, nullptr, "DecryptAesCtr"}, + {15, nullptr, "CryptAesCtr"}, {16, nullptr, "ComputeCmac"}, - {17, nullptr, "LoadRsaOaepKey"}, - {18, nullptr, "UnwrapRsaOaepWrappedTitleKey"}, + {17, nullptr, "ImportEsKey"}, + {18, nullptr, "UnwrapTitleKey"}, {19, nullptr, "LoadTitleKey"}, - {20, nullptr, "UnwrapAesWrappedTitleKey"}, - {21, nullptr, "LockAesEngine"}, - {22, nullptr, "UnlockAesEngine"}, - {23, nullptr, "GetSplWaitEvent"}, - {24, nullptr, "SetSharedData"}, - {25, nullptr, "GetSharedData"}, - {26, nullptr, "ImportSslRsaKey"}, - {27, nullptr, "SecureExpModWithSslKey"}, - {28, nullptr, "ImportEsRsaKey"}, - {29, nullptr, "SecureExpModWithEsKey"}, - {30, nullptr, "EncryptManuRsaKeyForImport"}, - {31, nullptr, "GetPackage2Hash"}, + {20, nullptr, "PrepareEsCommonKey"}, + {21, nullptr, "AllocateAesKeyslot"}, + {22, nullptr, "DeallocateAesKeySlot"}, + {23, nullptr, "GetAesKeyslotAvailableEvent"}, + {24, nullptr, "SetBootReason"}, + {25, nullptr, "GetBootReason"}, + {26, nullptr, "DecryptAndStoreSslClientCertKey"}, + {27, nullptr, "ModularExponentiateWithSslClientCertKey"}, + {28, nullptr, "DecryptAndStoreDrmDeviceCertKey"}, + {29, nullptr, "ModularExponentiateWithDrmDeviceCertKey"}, + {30, nullptr, "ReencryptDeviceUniqueData "}, + {31, nullptr, "PrepareEsArchiveKey"}, // This is also GetPackage2Hash? + {32, nullptr, "LoadPreparedAesKey"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/time/interface.cpp b/src/core/hle/service/time/interface.cpp index f509653a3..ba8fd6152 100644 --- a/src/core/hle/service/time/interface.cpp +++ b/src/core/hle/service/time/interface.cpp @@ -29,7 +29,7 @@ Time::Time(std::shared_ptr<Module> module, Core::System& system, const char* nam {300, &Time::CalculateMonotonicSystemClockBaseTimePoint, "CalculateMonotonicSystemClockBaseTimePoint"}, {400, &Time::GetClockSnapshot, "GetClockSnapshot"}, {401, &Time::GetClockSnapshotFromSystemClockContext, "GetClockSnapshotFromSystemClockContext"}, - {500, nullptr, "CalculateStandardUserSystemClockDifferenceByUser"}, + {500, &Time::CalculateStandardUserSystemClockDifferenceByUser, "CalculateStandardUserSystemClockDifferenceByUser"}, {501, &Time::CalculateSpanBetween, "CalculateSpanBetween"}, }; // clang-format on diff --git a/src/core/hle/service/time/standard_network_system_clock_core.h b/src/core/hle/service/time/standard_network_system_clock_core.h index 3f505c37c..c993bdf79 100644 --- a/src/core/hle/service/time/standard_network_system_clock_core.h +++ b/src/core/hle/service/time/standard_network_system_clock_core.h @@ -23,7 +23,7 @@ public: standard_network_clock_sufficient_accuracy = value; } - bool IsStandardNetworkSystemClockAccuracySufficient(Core::System& system) { + bool IsStandardNetworkSystemClockAccuracySufficient(Core::System& system) const { SystemClockContext context{}; if (GetClockContext(system, context) != RESULT_SUCCESS) { return {}; diff --git a/src/core/hle/service/time/standard_steady_clock_core.cpp b/src/core/hle/service/time/standard_steady_clock_core.cpp index 1575f0b49..59a272f4a 100644 --- a/src/core/hle/service/time/standard_steady_clock_core.cpp +++ b/src/core/hle/service/time/standard_steady_clock_core.cpp @@ -11,9 +11,8 @@ namespace Service::Time::Clock { TimeSpanType StandardSteadyClockCore::GetCurrentRawTimePoint(Core::System& system) { - const TimeSpanType ticks_time_span{TimeSpanType::FromTicks( - Core::Timing::CpuCyclesToClockCycles(system.CoreTiming().GetTicks()), - Core::Hardware::CNTFREQ)}; + const TimeSpanType ticks_time_span{ + TimeSpanType::FromTicks(system.CoreTiming().GetClockTicks(), Core::Hardware::CNTFREQ)}; TimeSpanType raw_time_point{setup_value.nanoseconds + ticks_time_span.nanoseconds}; if (raw_time_point.nanoseconds < cached_raw_time_point.nanoseconds) { diff --git a/src/core/hle/service/time/steady_clock_core.h b/src/core/hle/service/time/steady_clock_core.h index 84af3d105..d80a2385f 100644 --- a/src/core/hle/service/time/steady_clock_core.h +++ b/src/core/hle/service/time/steady_clock_core.h @@ -16,6 +16,7 @@ namespace Service::Time::Clock { class SteadyClockCore { public: SteadyClockCore() = default; + virtual ~SteadyClockCore() = default; const Common::UUID& GetClockSourceId() const { return clock_source_id; diff --git a/src/core/hle/service/time/system_clock_context_update_callback.h b/src/core/hle/service/time/system_clock_context_update_callback.h index 6260de6c3..2b0fa7e75 100644 --- a/src/core/hle/service/time/system_clock_context_update_callback.h +++ b/src/core/hle/service/time/system_clock_context_update_callback.h @@ -20,7 +20,7 @@ namespace Service::Time::Clock { class SystemClockContextUpdateCallback { public: SystemClockContextUpdateCallback(); - ~SystemClockContextUpdateCallback(); + virtual ~SystemClockContextUpdateCallback(); bool NeedUpdate(const SystemClockContext& value) const; diff --git a/src/core/hle/service/time/system_clock_core.cpp b/src/core/hle/service/time/system_clock_core.cpp index 1a3ab8cfa..d31d4e2ca 100644 --- a/src/core/hle/service/time/system_clock_core.cpp +++ b/src/core/hle/service/time/system_clock_core.cpp @@ -9,7 +9,7 @@ namespace Service::Time::Clock { SystemClockCore::SystemClockCore(SteadyClockCore& steady_clock_core) - : steady_clock_core{steady_clock_core}, is_initialized{} { + : steady_clock_core{steady_clock_core} { context.steady_time_point.clock_source_id = steady_clock_core.GetClockSourceId(); } diff --git a/src/core/hle/service/time/system_clock_core.h b/src/core/hle/service/time/system_clock_core.h index 54407a6c5..608dd3b2e 100644 --- a/src/core/hle/service/time/system_clock_core.h +++ b/src/core/hle/service/time/system_clock_core.h @@ -22,7 +22,7 @@ class SystemClockContextUpdateCallback; class SystemClockCore { public: explicit SystemClockCore(SteadyClockCore& steady_clock_core); - ~SystemClockCore(); + virtual ~SystemClockCore(); SteadyClockCore& GetSteadyClockCore() const { return steady_clock_core; diff --git a/src/core/hle/service/time/tick_based_steady_clock_core.cpp b/src/core/hle/service/time/tick_based_steady_clock_core.cpp index 44d5bc651..8baaa2a6a 100644 --- a/src/core/hle/service/time/tick_based_steady_clock_core.cpp +++ b/src/core/hle/service/time/tick_based_steady_clock_core.cpp @@ -11,9 +11,8 @@ namespace Service::Time::Clock { SteadyClockTimePoint TickBasedSteadyClockCore::GetTimePoint(Core::System& system) { - const TimeSpanType ticks_time_span{TimeSpanType::FromTicks( - Core::Timing::CpuCyclesToClockCycles(system.CoreTiming().GetTicks()), - Core::Hardware::CNTFREQ)}; + const TimeSpanType ticks_time_span{ + TimeSpanType::FromTicks(system.CoreTiming().GetClockTicks(), Core::Hardware::CNTFREQ)}; return {ticks_time_span.ToSeconds(), GetClockSourceId()}; } diff --git a/src/core/hle/service/time/time.cpp b/src/core/hle/service/time/time.cpp index ce859f18d..7d0474e0b 100644 --- a/src/core/hle/service/time/time.cpp +++ b/src/core/hle/service/time/time.cpp @@ -10,6 +10,7 @@ #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/service/time/interface.h" #include "core/hle/service/time/time.h" @@ -20,8 +21,8 @@ namespace Service::Time { class ISystemClock final : public ServiceFramework<ISystemClock> { public: - ISystemClock(Clock::SystemClockCore& clock_core) - : ServiceFramework("ISystemClock"), clock_core{clock_core} { + explicit ISystemClock(Clock::SystemClockCore& clock_core, Core::System& system) + : ServiceFramework("ISystemClock"), clock_core{clock_core}, system{system} { // clang-format off static const FunctionInfo functions[] = { {0, &ISystemClock::GetCurrentTime, "GetCurrentTime"}, @@ -46,9 +47,8 @@ private: } s64 posix_time{}; - if (const ResultCode result{ - clock_core.GetCurrentTime(Core::System::GetInstance(), posix_time)}; - result != RESULT_SUCCESS) { + if (const ResultCode result{clock_core.GetCurrentTime(system, posix_time)}; + result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(result); return; @@ -69,9 +69,8 @@ private: } Clock::SystemClockContext system_clock_context{}; - if (const ResultCode result{ - clock_core.GetClockContext(Core::System::GetInstance(), system_clock_context)}; - result != RESULT_SUCCESS) { + if (const ResultCode result{clock_core.GetClockContext(system, system_clock_context)}; + result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(result); return; @@ -83,14 +82,22 @@ private: } Clock::SystemClockCore& clock_core; + Core::System& system; }; class ISteadyClock final : public ServiceFramework<ISteadyClock> { public: - ISteadyClock(Clock::SteadyClockCore& clock_core) - : ServiceFramework("ISteadyClock"), clock_core{clock_core} { + explicit ISteadyClock(Clock::SteadyClockCore& clock_core, Core::System& system) + : ServiceFramework("ISteadyClock"), clock_core{clock_core}, system{system} { static const FunctionInfo functions[] = { {0, &ISteadyClock::GetCurrentTimePoint, "GetCurrentTimePoint"}, + {2, nullptr, "GetTestOffset"}, + {3, nullptr, "SetTestOffset"}, + {100, nullptr, "GetRtcValue"}, + {101, nullptr, "IsRtcResetDetected"}, + {102, nullptr, "GetSetupResultValue"}, + {200, nullptr, "GetInternalOffset"}, + {201, nullptr, "SetInternalOffset"}, }; RegisterHandlers(functions); } @@ -105,21 +112,21 @@ private: return; } - const Clock::SteadyClockTimePoint time_point{ - clock_core.GetCurrentTimePoint(Core::System::GetInstance())}; + const Clock::SteadyClockTimePoint time_point{clock_core.GetCurrentTimePoint(system)}; IPC::ResponseBuilder rb{ctx, (sizeof(Clock::SteadyClockTimePoint) / 4) + 2}; rb.Push(RESULT_SUCCESS); rb.PushRaw(time_point); } Clock::SteadyClockCore& clock_core; + Core::System& system; }; ResultCode Module::Interface::GetClockSnapshotFromSystemClockContextInternal( Kernel::Thread* thread, Clock::SystemClockContext user_context, Clock::SystemClockContext network_context, u8 type, Clock::ClockSnapshot& clock_snapshot) { - auto& time_manager{module->GetTimeManager()}; + auto& time_manager{system.GetTimeManager()}; clock_snapshot.is_automatic_correction_enabled = time_manager.GetStandardUserSystemClockCore().IsAutomaticCorrectionEnabled(); @@ -134,7 +141,7 @@ ResultCode Module::Interface::GetClockSnapshotFromSystemClockContextInternal( } const auto current_time_point{ - time_manager.GetStandardSteadyClockCore().GetCurrentTimePoint(Core::System::GetInstance())}; + time_manager.GetStandardSteadyClockCore().GetCurrentTimePoint(system)}; if (const ResultCode result{Clock::ClockSnapshot::GetCurrentTime( clock_snapshot.user_time, current_time_point, clock_snapshot.user_context)}; result != RESULT_SUCCESS) { @@ -176,41 +183,44 @@ void Module::Interface::GetStandardUserSystemClock(Kernel::HLERequestContext& ct LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemClock>(module->GetTimeManager().GetStandardUserSystemClockCore()); + rb.PushIpcInterface<ISystemClock>(system.GetTimeManager().GetStandardUserSystemClockCore(), + system); } void Module::Interface::GetStandardNetworkSystemClock(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemClock>(module->GetTimeManager().GetStandardNetworkSystemClockCore()); + rb.PushIpcInterface<ISystemClock>(system.GetTimeManager().GetStandardNetworkSystemClockCore(), + system); } void Module::Interface::GetStandardSteadyClock(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISteadyClock>(module->GetTimeManager().GetStandardSteadyClockCore()); + rb.PushIpcInterface<ISteadyClock>(system.GetTimeManager().GetStandardSteadyClockCore(), system); } void Module::Interface::GetTimeZoneService(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ITimeZoneService>(module->GetTimeManager().GetTimeZoneContentManager()); + rb.PushIpcInterface<ITimeZoneService>(system.GetTimeManager().GetTimeZoneContentManager()); } void Module::Interface::GetStandardLocalSystemClock(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemClock>(module->GetTimeManager().GetStandardLocalSystemClockCore()); + rb.PushIpcInterface<ISystemClock>(system.GetTimeManager().GetStandardLocalSystemClockCore(), + system); } void Module::Interface::IsStandardNetworkSystemClockAccuracySufficient( Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); - auto& clock_core{module->GetTimeManager().GetStandardNetworkSystemClockCore()}; + auto& clock_core{system.GetTimeManager().GetStandardNetworkSystemClockCore()}; IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); rb.Push<u32>(clock_core.IsStandardNetworkSystemClockAccuracySufficient(system)); @@ -219,7 +229,7 @@ void Module::Interface::IsStandardNetworkSystemClockAccuracySufficient( void Module::Interface::CalculateMonotonicSystemClockBaseTimePoint(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); - auto& steady_clock_core{module->GetTimeManager().GetStandardSteadyClockCore()}; + auto& steady_clock_core{system.GetTimeManager().GetStandardSteadyClockCore()}; if (!steady_clock_core.IsInitialized()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_UNINITIALIZED_CLOCK); @@ -228,13 +238,11 @@ void Module::Interface::CalculateMonotonicSystemClockBaseTimePoint(Kernel::HLERe IPC::RequestParser rp{ctx}; const auto context{rp.PopRaw<Clock::SystemClockContext>()}; - const auto current_time_point{ - steady_clock_core.GetCurrentTimePoint(Core::System::GetInstance())}; + const auto current_time_point{steady_clock_core.GetCurrentTimePoint(system)}; if (current_time_point.clock_source_id == context.steady_time_point.clock_source_id) { - const auto ticks{Clock::TimeSpanType::FromTicks( - Core::Timing::CpuCyclesToClockCycles(system.CoreTiming().GetTicks()), - Core::Hardware::CNTFREQ)}; + const auto ticks{Clock::TimeSpanType::FromTicks(system.CoreTiming().GetClockTicks(), + Core::Hardware::CNTFREQ)}; const s64 base_time_point{context.offset + current_time_point.time_point - ticks.ToSeconds()}; IPC::ResponseBuilder rb{ctx, (sizeof(s64) / 4) + 2}; @@ -254,18 +262,18 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) { Clock::SystemClockContext user_context{}; if (const ResultCode result{ - module->GetTimeManager().GetStandardUserSystemClockCore().GetClockContext( - Core::System::GetInstance(), user_context)}; - result != RESULT_SUCCESS) { + system.GetTimeManager().GetStandardUserSystemClockCore().GetClockContext(system, + user_context)}; + result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(result); return; } Clock::SystemClockContext network_context{}; if (const ResultCode result{ - module->GetTimeManager().GetStandardNetworkSystemClockCore().GetClockContext( - Core::System::GetInstance(), network_context)}; - result != RESULT_SUCCESS) { + system.GetTimeManager().GetStandardNetworkSystemClockCore().GetClockContext( + system, network_context)}; + result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(result); return; @@ -274,7 +282,7 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) { Clock::ClockSnapshot clock_snapshot{}; if (const ResultCode result{GetClockSnapshotFromSystemClockContextInternal( &ctx.GetThread(), user_context, network_context, type, clock_snapshot)}; - result != RESULT_SUCCESS) { + result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(result); return; @@ -282,7 +290,7 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - ctx.WriteBuffer(&clock_snapshot, sizeof(Clock::ClockSnapshot)); + ctx.WriteBuffer(clock_snapshot); } void Module::Interface::GetClockSnapshotFromSystemClockContext(Kernel::HLERequestContext& ctx) { @@ -305,7 +313,30 @@ void Module::Interface::GetClockSnapshotFromSystemClockContext(Kernel::HLEReques IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - ctx.WriteBuffer(&clock_snapshot, sizeof(Clock::ClockSnapshot)); + ctx.WriteBuffer(clock_snapshot); +} + +void Module::Interface::CalculateStandardUserSystemClockDifferenceByUser( + Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_Time, "called"); + + IPC::RequestParser rp{ctx}; + const auto snapshot_a = rp.PopRaw<Clock::ClockSnapshot>(); + const auto snapshot_b = rp.PopRaw<Clock::ClockSnapshot>(); + + auto time_span_type{Clock::TimeSpanType::FromSeconds(snapshot_b.user_context.offset - + snapshot_a.user_context.offset)}; + + if ((snapshot_b.user_context.steady_time_point.clock_source_id != + snapshot_a.user_context.steady_time_point.clock_source_id) || + (snapshot_b.is_automatic_correction_enabled && + snapshot_a.is_automatic_correction_enabled)) { + time_span_type.nanoseconds = 0; + } + + IPC::ResponseBuilder rb{ctx, (sizeof(s64) / 4) + 2}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw(time_span_type.nanoseconds); } void Module::Interface::CalculateSpanBetween(Kernel::HLERequestContext& ctx) { @@ -341,7 +372,7 @@ void Module::Interface::GetSharedMemoryNativeHandle(Kernel::HLERequestContext& c LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); - rb.PushCopyObjects(module->GetTimeManager().GetSharedMemory().GetSharedMemoryHolder()); + rb.PushCopyObjects(SharedFrom(&system.Kernel().GetTimeSharedMem())); } Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name) @@ -350,7 +381,7 @@ Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& syste Module::Interface::~Interface() = default; void InstallInterfaces(Core::System& system) { - auto module{std::make_shared<Module>(system)}; + auto module{std::make_shared<Module>()}; std::make_shared<Time>(module, system, "time:a")->InstallAsService(system.ServiceManager()); std::make_shared<Time>(module, system, "time:s")->InstallAsService(system.ServiceManager()); std::make_shared<Time>(module, system, "time:u")->InstallAsService(system.ServiceManager()); diff --git a/src/core/hle/service/time/time.h b/src/core/hle/service/time/time.h index 351988468..49f4aac0a 100644 --- a/src/core/hle/service/time/time.h +++ b/src/core/hle/service/time/time.h @@ -16,7 +16,7 @@ namespace Service::Time { class Module final { public: - Module(Core::System& system) : time_manager{system} {} + Module() = default; class Interface : public ServiceFramework<Interface> { public: @@ -32,6 +32,7 @@ public: void CalculateMonotonicSystemClockBaseTimePoint(Kernel::HLERequestContext& ctx); void GetClockSnapshot(Kernel::HLERequestContext& ctx); void GetClockSnapshotFromSystemClockContext(Kernel::HLERequestContext& ctx); + void CalculateStandardUserSystemClockDifferenceByUser(Kernel::HLERequestContext& ctx); void CalculateSpanBetween(Kernel::HLERequestContext& ctx); void GetSharedMemoryNativeHandle(Kernel::HLERequestContext& ctx); @@ -45,13 +46,6 @@ public: std::shared_ptr<Module> module; Core::System& system; }; - - TimeManager& GetTimeManager() { - return time_manager; - } - -private: - TimeManager time_manager; }; /// Registers all Time services with the specified service manager. diff --git a/src/core/hle/service/time/time_manager.cpp b/src/core/hle/service/time/time_manager.cpp index 9d6c55865..858623e2b 100644 --- a/src/core/hle/service/time/time_manager.cpp +++ b/src/core/hle/service/time/time_manager.cpp @@ -5,6 +5,7 @@ #include <chrono> #include <ctime> +#include "common/time_zone.h" #include "core/hle/service/time/ephemeral_network_system_clock_context_writer.h" #include "core/hle/service/time/local_system_clock_context_writer.h" #include "core/hle/service/time/network_system_clock_context_writer.h" @@ -22,116 +23,281 @@ static std::chrono::seconds GetSecondsSinceEpoch() { } static s64 GetExternalRtcValue() { - return GetSecondsSinceEpoch().count(); -} - -TimeManager::TimeManager(Core::System& system) - : shared_memory{system}, standard_local_system_clock_core{standard_steady_clock_core}, - standard_network_system_clock_core{standard_steady_clock_core}, - standard_user_system_clock_core{standard_local_system_clock_core, - standard_network_system_clock_core, system}, - ephemeral_network_system_clock_core{tick_based_steady_clock_core}, - local_system_clock_context_writer{ - std::make_shared<Clock::LocalSystemClockContextWriter>(shared_memory)}, - network_system_clock_context_writer{ - std::make_shared<Clock::NetworkSystemClockContextWriter>(shared_memory)}, - ephemeral_network_system_clock_context_writer{ - std::make_shared<Clock::EphemeralNetworkSystemClockContextWriter>()}, - time_zone_content_manager{*this, system} { - - const auto system_time{Clock::TimeSpanType::FromSeconds(GetExternalRtcValue())}; - SetupStandardSteadyClock(system, Common::UUID::Generate(), system_time, {}, {}); - SetupStandardLocalSystemClock(system, {}, system_time.ToSeconds()); - SetupStandardNetworkSystemClock({}, standard_network_clock_accuracy); - SetupStandardUserSystemClock(system, {}, Clock::SteadyClockTimePoint::GetRandom()); - SetupEphemeralNetworkSystemClock(); + return GetSecondsSinceEpoch().count() + TimeManager::GetExternalTimeZoneOffset(); } -TimeManager::~TimeManager() = default; +struct TimeManager::Impl final { + explicit Impl(Core::System& system) + : shared_memory{system}, standard_local_system_clock_core{standard_steady_clock_core}, + standard_network_system_clock_core{standard_steady_clock_core}, + standard_user_system_clock_core{standard_local_system_clock_core, + standard_network_system_clock_core, system}, + ephemeral_network_system_clock_core{tick_based_steady_clock_core}, + local_system_clock_context_writer{ + std::make_shared<Clock::LocalSystemClockContextWriter>(shared_memory)}, + network_system_clock_context_writer{ + std::make_shared<Clock::NetworkSystemClockContextWriter>(shared_memory)}, + ephemeral_network_system_clock_context_writer{ + std::make_shared<Clock::EphemeralNetworkSystemClockContextWriter>()}, + time_zone_content_manager{system} { -void TimeManager::SetupTimeZoneManager(std::string location_name, - Clock::SteadyClockTimePoint time_zone_updated_time_point, - std::size_t total_location_name_count, - u128 time_zone_rule_version, - FileSys::VirtualFile& vfs_file) { - if (time_zone_content_manager.GetTimeZoneManager().SetDeviceLocationNameWithTimeZoneRule( - location_name, vfs_file) != RESULT_SUCCESS) { - UNREACHABLE(); - return; - } - - time_zone_content_manager.GetTimeZoneManager().SetUpdatedTime(time_zone_updated_time_point); - time_zone_content_manager.GetTimeZoneManager().SetTotalLocationNameCount( - total_location_name_count); - time_zone_content_manager.GetTimeZoneManager().SetTimeZoneRuleVersion(time_zone_rule_version); - time_zone_content_manager.GetTimeZoneManager().MarkAsInitialized(); -} - -void TimeManager::SetupStandardSteadyClock(Core::System& system, Common::UUID clock_source_id, - Clock::TimeSpanType setup_value, - Clock::TimeSpanType internal_offset, - bool is_rtc_reset_detected) { - standard_steady_clock_core.SetClockSourceId(clock_source_id); - standard_steady_clock_core.SetSetupValue(setup_value); - standard_steady_clock_core.SetInternalOffset(internal_offset); - standard_steady_clock_core.MarkAsInitialized(); - - const auto current_time_point{standard_steady_clock_core.GetCurrentRawTimePoint(system)}; - shared_memory.SetupStandardSteadyClock(system, clock_source_id, current_time_point); -} - -void TimeManager::SetupStandardLocalSystemClock(Core::System& system, - Clock::SystemClockContext clock_context, - s64 posix_time) { - standard_local_system_clock_core.SetUpdateCallbackInstance(local_system_clock_context_writer); - - const auto current_time_point{ - standard_local_system_clock_core.GetSteadyClockCore().GetCurrentTimePoint(system)}; - if (current_time_point.clock_source_id == clock_context.steady_time_point.clock_source_id) { - standard_local_system_clock_core.SetSystemClockContext(clock_context); - } else { - if (standard_local_system_clock_core.SetCurrentTime(system, posix_time) != RESULT_SUCCESS) { + const auto system_time{Clock::TimeSpanType::FromSeconds(GetExternalRtcValue())}; + SetupStandardSteadyClock(system, Common::UUID::Generate(), system_time, {}, {}); + SetupStandardLocalSystemClock(system, {}, system_time.ToSeconds()); + SetupStandardNetworkSystemClock({}, standard_network_clock_accuracy); + SetupStandardUserSystemClock(system, {}, Clock::SteadyClockTimePoint::GetRandom()); + SetupEphemeralNetworkSystemClock(); + } + + ~Impl() = default; + + Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() { + return standard_steady_clock_core; + } + + const Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() const { + return standard_steady_clock_core; + } + + Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() { + return standard_local_system_clock_core; + } + + const Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() const { + return standard_local_system_clock_core; + } + + Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() { + return standard_network_system_clock_core; + } + + const Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() const { + return standard_network_system_clock_core; + } + + Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() { + return standard_user_system_clock_core; + } + + const Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() const { + return standard_user_system_clock_core; + } + + TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() { + return time_zone_content_manager; + } + + const TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() const { + return time_zone_content_manager; + } + + SharedMemory& GetSharedMemory() { + return shared_memory; + } + + const SharedMemory& GetSharedMemory() const { + return shared_memory; + } + + void SetupTimeZoneManager(std::string location_name, + Clock::SteadyClockTimePoint time_zone_updated_time_point, + std::size_t total_location_name_count, u128 time_zone_rule_version, + FileSys::VirtualFile& vfs_file) { + if (time_zone_content_manager.GetTimeZoneManager().SetDeviceLocationNameWithTimeZoneRule( + location_name, vfs_file) != RESULT_SUCCESS) { UNREACHABLE(); return; } + + time_zone_content_manager.GetTimeZoneManager().SetUpdatedTime(time_zone_updated_time_point); + time_zone_content_manager.GetTimeZoneManager().SetTotalLocationNameCount( + total_location_name_count); + time_zone_content_manager.GetTimeZoneManager().SetTimeZoneRuleVersion( + time_zone_rule_version); + time_zone_content_manager.GetTimeZoneManager().MarkAsInitialized(); } - standard_local_system_clock_core.MarkAsInitialized(); -} + static s64 GetExternalTimeZoneOffset() { + // With "auto" timezone setting, we use the external system's timezone offset + if (Settings::GetTimeZoneString() == "auto") { + return Common::TimeZone::GetCurrentOffsetSeconds().count(); + } + return 0; + } -void TimeManager::SetupStandardNetworkSystemClock(Clock::SystemClockContext clock_context, - Clock::TimeSpanType sufficient_accuracy) { - standard_network_system_clock_core.SetUpdateCallbackInstance( - network_system_clock_context_writer); + void SetupStandardSteadyClock(Core::System& system, Common::UUID clock_source_id, + Clock::TimeSpanType setup_value, + Clock::TimeSpanType internal_offset, bool is_rtc_reset_detected) { + standard_steady_clock_core.SetClockSourceId(clock_source_id); + standard_steady_clock_core.SetSetupValue(setup_value); + standard_steady_clock_core.SetInternalOffset(internal_offset); + standard_steady_clock_core.MarkAsInitialized(); - if (standard_network_system_clock_core.SetSystemClockContext(clock_context) != RESULT_SUCCESS) { - UNREACHABLE(); - return; + const auto current_time_point{standard_steady_clock_core.GetCurrentRawTimePoint(system)}; + shared_memory.SetupStandardSteadyClock(system, clock_source_id, current_time_point); } - standard_network_system_clock_core.SetStandardNetworkClockSufficientAccuracy( - sufficient_accuracy); - standard_network_system_clock_core.MarkAsInitialized(); -} + void SetupStandardLocalSystemClock(Core::System& system, + Clock::SystemClockContext clock_context, s64 posix_time) { + standard_local_system_clock_core.SetUpdateCallbackInstance( + local_system_clock_context_writer); + + const auto current_time_point{ + standard_local_system_clock_core.GetSteadyClockCore().GetCurrentTimePoint(system)}; + if (current_time_point.clock_source_id == clock_context.steady_time_point.clock_source_id) { + standard_local_system_clock_core.SetSystemClockContext(clock_context); + } else { + if (standard_local_system_clock_core.SetCurrentTime(system, posix_time) != + RESULT_SUCCESS) { + UNREACHABLE(); + return; + } + } + + standard_local_system_clock_core.MarkAsInitialized(); + } + + void SetupStandardNetworkSystemClock(Clock::SystemClockContext clock_context, + Clock::TimeSpanType sufficient_accuracy) { + standard_network_system_clock_core.SetUpdateCallbackInstance( + network_system_clock_context_writer); + + if (standard_network_system_clock_core.SetSystemClockContext(clock_context) != + RESULT_SUCCESS) { + UNREACHABLE(); + return; + } + + standard_network_system_clock_core.SetStandardNetworkClockSufficientAccuracy( + sufficient_accuracy); + standard_network_system_clock_core.MarkAsInitialized(); + } + + void SetupStandardUserSystemClock(Core::System& system, bool is_automatic_correction_enabled, + Clock::SteadyClockTimePoint steady_clock_time_point) { + if (standard_user_system_clock_core.SetAutomaticCorrectionEnabled( + system, is_automatic_correction_enabled) != RESULT_SUCCESS) { + UNREACHABLE(); + return; + } + + standard_user_system_clock_core.SetAutomaticCorrectionUpdatedTime(steady_clock_time_point); + standard_user_system_clock_core.MarkAsInitialized(); + shared_memory.SetAutomaticCorrectionEnabled(is_automatic_correction_enabled); + } -void TimeManager::SetupStandardUserSystemClock( - Core::System& system, bool is_automatic_correction_enabled, - Clock::SteadyClockTimePoint steady_clock_time_point) { - if (standard_user_system_clock_core.SetAutomaticCorrectionEnabled( - system, is_automatic_correction_enabled) != RESULT_SUCCESS) { - UNREACHABLE(); - return; + void SetupEphemeralNetworkSystemClock() { + ephemeral_network_system_clock_core.SetUpdateCallbackInstance( + ephemeral_network_system_clock_context_writer); + ephemeral_network_system_clock_core.MarkAsInitialized(); } - standard_user_system_clock_core.SetAutomaticCorrectionUpdatedTime(steady_clock_time_point); - standard_user_system_clock_core.MarkAsInitialized(); - shared_memory.SetAutomaticCorrectionEnabled(is_automatic_correction_enabled); + void UpdateLocalSystemClockTime(Core::System& system, s64 posix_time) { + const auto timespan{Service::Time::Clock::TimeSpanType::FromSeconds(posix_time)}; + if (GetStandardLocalSystemClockCore() + .SetCurrentTime(system, timespan.ToSeconds()) + .IsError()) { + UNREACHABLE(); + return; + } + } + + SharedMemory shared_memory; + + Clock::StandardSteadyClockCore standard_steady_clock_core; + Clock::TickBasedSteadyClockCore tick_based_steady_clock_core; + Clock::StandardLocalSystemClockCore standard_local_system_clock_core; + Clock::StandardNetworkSystemClockCore standard_network_system_clock_core; + Clock::StandardUserSystemClockCore standard_user_system_clock_core; + Clock::EphemeralNetworkSystemClockCore ephemeral_network_system_clock_core; + + std::shared_ptr<Clock::LocalSystemClockContextWriter> local_system_clock_context_writer; + std::shared_ptr<Clock::NetworkSystemClockContextWriter> network_system_clock_context_writer; + std::shared_ptr<Clock::EphemeralNetworkSystemClockContextWriter> + ephemeral_network_system_clock_context_writer; + + TimeZone::TimeZoneContentManager time_zone_content_manager; +}; + +TimeManager::TimeManager(Core::System& system) : system{system} {} + +TimeManager::~TimeManager() = default; + +void TimeManager::Initialize() { + impl = std::make_unique<Impl>(system); + + // Time zones can only be initialized after impl is valid + impl->time_zone_content_manager.Initialize(*this); } -void TimeManager::SetupEphemeralNetworkSystemClock() { - ephemeral_network_system_clock_core.SetUpdateCallbackInstance( - ephemeral_network_system_clock_context_writer); - ephemeral_network_system_clock_core.MarkAsInitialized(); +Clock::StandardSteadyClockCore& TimeManager::GetStandardSteadyClockCore() { + return impl->standard_steady_clock_core; +} + +const Clock::StandardSteadyClockCore& TimeManager::GetStandardSteadyClockCore() const { + return impl->standard_steady_clock_core; +} + +Clock::StandardLocalSystemClockCore& TimeManager::GetStandardLocalSystemClockCore() { + return impl->standard_local_system_clock_core; +} + +const Clock::StandardLocalSystemClockCore& TimeManager::GetStandardLocalSystemClockCore() const { + return impl->standard_local_system_clock_core; +} + +Clock::StandardNetworkSystemClockCore& TimeManager::GetStandardNetworkSystemClockCore() { + return impl->standard_network_system_clock_core; +} + +const Clock::StandardNetworkSystemClockCore& TimeManager::GetStandardNetworkSystemClockCore() + const { + return impl->standard_network_system_clock_core; +} + +Clock::StandardUserSystemClockCore& TimeManager::GetStandardUserSystemClockCore() { + return impl->standard_user_system_clock_core; +} + +const Clock::StandardUserSystemClockCore& TimeManager::GetStandardUserSystemClockCore() const { + return impl->standard_user_system_clock_core; +} + +TimeZone::TimeZoneContentManager& TimeManager::GetTimeZoneContentManager() { + return impl->time_zone_content_manager; +} + +const TimeZone::TimeZoneContentManager& TimeManager::GetTimeZoneContentManager() const { + return impl->time_zone_content_manager; +} + +SharedMemory& TimeManager::GetSharedMemory() { + return impl->shared_memory; +} + +const SharedMemory& TimeManager::GetSharedMemory() const { + return impl->shared_memory; +} + +void TimeManager::UpdateLocalSystemClockTime(s64 posix_time) { + impl->UpdateLocalSystemClockTime(system, posix_time); +} + +void TimeManager::SetupTimeZoneManager(std::string location_name, + Clock::SteadyClockTimePoint time_zone_updated_time_point, + std::size_t total_location_name_count, + u128 time_zone_rule_version, + FileSys::VirtualFile& vfs_file) { + impl->SetupTimeZoneManager(location_name, time_zone_updated_time_point, + total_location_name_count, time_zone_rule_version, vfs_file); +} + +/*static*/ s64 TimeManager::GetExternalTimeZoneOffset() { + // With "auto" timezone setting, we use the external system's timezone offset + if (Settings::GetTimeZoneString() == "auto") { + return Common::TimeZone::GetCurrentOffsetSeconds().count(); + } + return 0; } } // namespace Service::Time diff --git a/src/core/hle/service/time/time_manager.h b/src/core/hle/service/time/time_manager.h index 8e65f0d22..993c7c288 100644 --- a/src/core/hle/service/time/time_manager.h +++ b/src/core/hle/service/time/time_manager.h @@ -5,6 +5,7 @@ #pragma once #include "common/common_types.h" +#include "common/time_zone.h" #include "core/file_sys/vfs_types.h" #include "core/hle/service/time/clock_types.h" #include "core/hle/service/time/ephemeral_network_system_clock_core.h" @@ -32,86 +33,46 @@ public: explicit TimeManager(Core::System& system); ~TimeManager(); - Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() { - return standard_steady_clock_core; - } + void Initialize(); - const Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() const { - return standard_steady_clock_core; - } + Clock::StandardSteadyClockCore& GetStandardSteadyClockCore(); - Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() { - return standard_local_system_clock_core; - } + const Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() const; - const Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() const { - return standard_local_system_clock_core; - } + Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore(); - Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() { - return standard_network_system_clock_core; - } + const Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() const; - const Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() const { - return standard_network_system_clock_core; - } + Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore(); - Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() { - return standard_user_system_clock_core; - } + const Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() const; - const Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() const { - return standard_user_system_clock_core; - } + Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore(); - TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() { - return time_zone_content_manager; - } + const Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() const; - const TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() const { - return time_zone_content_manager; - } + TimeZone::TimeZoneContentManager& GetTimeZoneContentManager(); - SharedMemory& GetSharedMemory() { - return shared_memory; - } + const TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() const; - const SharedMemory& GetSharedMemory() const { - return shared_memory; - } + void UpdateLocalSystemClockTime(s64 posix_time); + + SharedMemory& GetSharedMemory(); + + const SharedMemory& GetSharedMemory() const; void SetupTimeZoneManager(std::string location_name, Clock::SteadyClockTimePoint time_zone_updated_time_point, std::size_t total_location_name_count, u128 time_zone_rule_version, FileSys::VirtualFile& vfs_file); + static s64 GetExternalTimeZoneOffset(); + private: - void SetupStandardSteadyClock(Core::System& system, Common::UUID clock_source_id, - Clock::TimeSpanType setup_value, - Clock::TimeSpanType internal_offset, bool is_rtc_reset_detected); - void SetupStandardLocalSystemClock(Core::System& system, - Clock::SystemClockContext clock_context, s64 posix_time); - void SetupStandardNetworkSystemClock(Clock::SystemClockContext clock_context, - Clock::TimeSpanType sufficient_accuracy); - void SetupStandardUserSystemClock(Core::System& system, bool is_automatic_correction_enabled, - Clock::SteadyClockTimePoint steady_clock_time_point); - void SetupEphemeralNetworkSystemClock(); - - SharedMemory shared_memory; - - Clock::StandardSteadyClockCore standard_steady_clock_core; - Clock::TickBasedSteadyClockCore tick_based_steady_clock_core; - Clock::StandardLocalSystemClockCore standard_local_system_clock_core; - Clock::StandardNetworkSystemClockCore standard_network_system_clock_core; - Clock::StandardUserSystemClockCore standard_user_system_clock_core; - Clock::EphemeralNetworkSystemClockCore ephemeral_network_system_clock_core; - - std::shared_ptr<Clock::LocalSystemClockContextWriter> local_system_clock_context_writer; - std::shared_ptr<Clock::NetworkSystemClockContextWriter> network_system_clock_context_writer; - std::shared_ptr<Clock::EphemeralNetworkSystemClockContextWriter> - ephemeral_network_system_clock_context_writer; - - TimeZone::TimeZoneContentManager time_zone_content_manager; + Core::System& system; + + struct Impl; + std::unique_ptr<Impl> impl; }; } // namespace Service::Time diff --git a/src/core/hle/service/time/time_sharedmemory.cpp b/src/core/hle/service/time/time_sharedmemory.cpp index fdaef233f..e0ae9f874 100644 --- a/src/core/hle/service/time/time_sharedmemory.cpp +++ b/src/core/hle/service/time/time_sharedmemory.cpp @@ -6,6 +6,7 @@ #include "core/core_timing.h" #include "core/core_timing_util.h" #include "core/hardware_properties.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/service/time/clock_types.h" #include "core/hle/service/time/steady_clock_core.h" #include "core/hle/service/time/time_sharedmemory.h" @@ -15,9 +16,7 @@ namespace Service::Time { static constexpr std::size_t SHARED_MEMORY_SIZE{0x1000}; SharedMemory::SharedMemory(Core::System& system) : system(system) { - shared_memory_holder = Kernel::SharedMemory::Create( - system.Kernel(), nullptr, SHARED_MEMORY_SIZE, Kernel::MemoryPermission::ReadWrite, - Kernel::MemoryPermission::Read, 0, Kernel::MemoryRegion::BASE, "Time:SharedMemory"); + shared_memory_holder = SharedFrom(&system.Kernel().GetTimeSharedMem()); std::memset(shared_memory_holder->GetPointer(), 0, SHARED_MEMORY_SIZE); } @@ -31,8 +30,7 @@ void SharedMemory::SetupStandardSteadyClock(Core::System& system, const Common::UUID& clock_source_id, Clock::TimeSpanType current_time_point) { const Clock::TimeSpanType ticks_time_span{Clock::TimeSpanType::FromTicks( - Core::Timing::CpuCyclesToClockCycles(system.CoreTiming().GetTicks()), - Core::Hardware::CNTFREQ)}; + system.CoreTiming().GetClockTicks(), Core::Hardware::CNTFREQ)}; const Clock::SteadyClockContext context{ static_cast<u64>(current_time_point.nanoseconds - ticks_time_span.nanoseconds), clock_source_id}; diff --git a/src/core/hle/service/time/time_zone_content_manager.cpp b/src/core/hle/service/time/time_zone_content_manager.cpp index 78d4acd95..4177d0a41 100644 --- a/src/core/hle/service/time/time_zone_content_manager.cpp +++ b/src/core/hle/service/time/time_zone_content_manager.cpp @@ -5,6 +5,7 @@ #include <sstream> #include "common/logging/log.h" +#include "common/time_zone.h" #include "core/core.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/nca_metadata.h" @@ -14,6 +15,7 @@ #include "core/hle/service/filesystem/filesystem.h" #include "core/hle/service/time/time_manager.h" #include "core/hle/service/time/time_zone_content_manager.h" +#include "core/settings.h" namespace Service::Time::TimeZone { @@ -66,12 +68,23 @@ static std::vector<std::string> BuildLocationNameCache(Core::System& system) { return location_name_cache; } -TimeZoneContentManager::TimeZoneContentManager(TimeManager& time_manager, Core::System& system) - : system{system}, location_name_cache{BuildLocationNameCache(system)} { - if (FileSys::VirtualFile vfs_file; GetTimeZoneInfoFile("GMT", vfs_file) == RESULT_SUCCESS) { +TimeZoneContentManager::TimeZoneContentManager(Core::System& system) + : system{system}, location_name_cache{BuildLocationNameCache(system)} {} + +void TimeZoneContentManager::Initialize(TimeManager& time_manager) { + std::string location_name; + const auto timezone_setting = Settings::GetTimeZoneString(); + if (timezone_setting == "auto" || timezone_setting == "default") { + location_name = Common::TimeZone::GetDefaultTimeZone(); + } else { + location_name = timezone_setting; + } + + if (FileSys::VirtualFile vfs_file; + GetTimeZoneInfoFile(location_name, vfs_file) == RESULT_SUCCESS) { const auto time_point{ time_manager.GetStandardSteadyClockCore().GetCurrentTimePoint(system)}; - time_manager.SetupTimeZoneManager("GMT", time_point, location_name_cache.size(), {}, + time_manager.SetupTimeZoneManager(location_name, time_point, location_name_cache.size(), {}, vfs_file); } else { time_zone_manager.MarkAsInitialized(); @@ -114,6 +127,12 @@ ResultCode TimeZoneContentManager::GetTimeZoneInfoFile(const std::string& locati vfs_file = zoneinfo_dir->GetFile(location_name); if (!vfs_file) { + LOG_ERROR(Service_Time, "{:016X} has no file \"{}\"! Using default timezone.", + time_zone_binary_titleid, location_name); + vfs_file = zoneinfo_dir->GetFile(Common::TimeZone::GetDefaultTimeZone()); + } + + if (!vfs_file) { LOG_ERROR(Service_Time, "{:016X} has no file \"{}\"!", time_zone_binary_titleid, location_name); return ERROR_TIME_NOT_FOUND; diff --git a/src/core/hle/service/time/time_zone_content_manager.h b/src/core/hle/service/time/time_zone_content_manager.h index 4f302c3b9..52dd1a020 100644 --- a/src/core/hle/service/time/time_zone_content_manager.h +++ b/src/core/hle/service/time/time_zone_content_manager.h @@ -21,7 +21,9 @@ namespace Service::Time::TimeZone { class TimeZoneContentManager final { public: - TimeZoneContentManager(TimeManager& time_manager, Core::System& system); + explicit TimeZoneContentManager(Core::System& system); + + void Initialize(TimeManager& time_manager); TimeZoneManager& GetTimeZoneManager() { return time_zone_manager; diff --git a/src/core/hle/service/time/time_zone_manager.cpp b/src/core/hle/service/time/time_zone_manager.cpp index 07b553a43..bdf0439f2 100644 --- a/src/core/hle/service/time/time_zone_manager.cpp +++ b/src/core/hle/service/time/time_zone_manager.cpp @@ -309,7 +309,7 @@ static bool ParsePosixName(const char* name, TimeZoneRule& rule) { offset = GetTZName(name, offset); std_len = offset; } - if (!std_len) { + if (std_len == 0) { return {}; } if (!GetOffset(name, offset, std_offset)) { @@ -320,7 +320,7 @@ static bool ParsePosixName(const char* name, TimeZoneRule& rule) { int dest_len{}; int dest_offset{}; const char* dest_name{name + offset}; - if (rule.chars.size() < char_count) { + if (rule.chars.size() < std::size_t(char_count)) { return {}; } @@ -343,7 +343,7 @@ static bool ParsePosixName(const char* name, TimeZoneRule& rule) { return {}; } char_count += dest_len + 1; - if (rule.chars.size() < char_count) { + if (rule.chars.size() < std::size_t(char_count)) { return {}; } if (name[offset] != '\0' && name[offset] != ',' && name[offset] != ';') { @@ -414,7 +414,7 @@ static bool ParsePosixName(const char* name, TimeZoneRule& rule) { if (is_reversed || (start_time < end_time && (end_time - start_time < (year_seconds + (std_offset - dest_offset))))) { - if (rule.ats.size() - 2 < time_count) { + if (rule.ats.size() - 2 < std::size_t(time_count)) { break; } @@ -518,8 +518,8 @@ static bool ParseTimeZoneBinary(TimeZoneRule& time_zone_rule, FileSys::VirtualFi constexpr s32 time_zone_max_leaps{50}; constexpr s32 time_zone_max_chars{50}; if (!(0 <= header.leap_count && header.leap_count < time_zone_max_leaps && - 0 < header.type_count && header.type_count < time_zone_rule.ttis.size() && - 0 <= header.time_count && header.time_count < time_zone_rule.ats.size() && + 0 < header.type_count && header.type_count < s32(time_zone_rule.ttis.size()) && + 0 <= header.time_count && header.time_count < s32(time_zone_rule.ats.size()) && 0 <= header.char_count && header.char_count < time_zone_max_chars && (header.ttis_std_count == header.type_count || header.ttis_std_count == 0) && (header.ttis_gmt_count == header.type_count || header.ttis_gmt_count == 0))) { @@ -609,7 +609,7 @@ static bool ParseTimeZoneBinary(TimeZoneRule& time_zone_rule, FileSys::VirtualFi } const u64 position{(read_offset - sizeof(TzifHeader))}; - const std::size_t bytes_read{vfs_file->GetSize() - sizeof(TzifHeader) - position}; + const s64 bytes_read = s64(vfs_file->GetSize() - sizeof(TzifHeader) - position); if (bytes_read < 0) { return {}; } @@ -621,11 +621,11 @@ static bool ParseTimeZoneBinary(TimeZoneRule& time_zone_rule, FileSys::VirtualFi std::array<char, time_zone_name_max + 1> temp_name{}; vfs_file->ReadArray(temp_name.data(), bytes_read, read_offset); if (bytes_read > 2 && temp_name[0] == '\n' && temp_name[bytes_read - 1] == '\n' && - time_zone_rule.type_count + 2 <= time_zone_rule.ttis.size()) { + std::size_t(time_zone_rule.type_count) + 2 <= time_zone_rule.ttis.size()) { temp_name[bytes_read - 1] = '\0'; std::array<char, time_zone_name_max> name{}; - std::memcpy(name.data(), temp_name.data() + 1, bytes_read - 1); + std::memcpy(name.data(), temp_name.data() + 1, std::size_t(bytes_read - 1)); TimeZoneRule temp_rule; if (ParsePosixName(name.data(), temp_rule)) { @@ -820,7 +820,10 @@ static ResultCode ToCalendarTimeImpl(const TimeZoneRule& rules, s64 time, Calend const ResultCode result{ ToCalendarTimeInternal(rules, time, calendar_time, calendar.additiona_info)}; calendar.time.year = static_cast<s16>(calendar_time.year); - calendar.time.month = calendar_time.month + 1; // Internal impl. uses 0-indexed month + + // Internal impl. uses 0-indexed month + calendar.time.month = static_cast<s8>(calendar_time.month + 1); + calendar.time.day = calendar_time.day; calendar.time.hour = calendar_time.hour; calendar.time.minute = calendar_time.minute; @@ -872,13 +875,15 @@ ResultCode TimeZoneManager::ToPosixTime(const TimeZoneRule& rules, const CalendarTime& calendar_time, s64& posix_time) const { posix_time = 0; - CalendarTimeInternal internal_time{}; - internal_time.year = calendar_time.year; - internal_time.month = calendar_time.month - 1; // Internal impl. uses 0-indexed month - internal_time.day = calendar_time.day; - internal_time.hour = calendar_time.hour; - internal_time.minute = calendar_time.minute; - internal_time.second = calendar_time.second; + CalendarTimeInternal internal_time{ + .year = calendar_time.year, + // Internal impl. uses 0-indexed month + .month = static_cast<s8>(calendar_time.month - 1), + .day = calendar_time.day, + .hour = calendar_time.hour, + .minute = calendar_time.minute, + .second = calendar_time.second, + }; s32 hour{internal_time.hour}; s32 minute{internal_time.minute}; diff --git a/src/core/hle/service/time/time_zone_service.cpp b/src/core/hle/service/time/time_zone_service.cpp index db57ae069..ff3a10b3e 100644 --- a/src/core/hle/service/time/time_zone_service.cpp +++ b/src/core/hle/service/time/time_zone_service.cpp @@ -142,7 +142,7 @@ void ITimeZoneService::ToPosixTime(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); rb.PushRaw<u32>(1); // Number of times we're returning - ctx.WriteBuffer(&posix_time, sizeof(s64)); + ctx.WriteBuffer(posix_time); } void ITimeZoneService::ToPosixTimeWithMyRule(Kernel::HLERequestContext& ctx) { @@ -164,7 +164,7 @@ void ITimeZoneService::ToPosixTimeWithMyRule(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); rb.PushRaw<u32>(1); // Number of times we're returning - ctx.WriteBuffer(&posix_time, sizeof(s64)); + ctx.WriteBuffer(posix_time); } } // namespace Service::Time diff --git a/src/core/hle/service/usb/usb.cpp b/src/core/hle/service/usb/usb.cpp index 58a9845fc..d033f8603 100644 --- a/src/core/hle/service/usb/usb.cpp +++ b/src/core/hle/service/usb/usb.cpp @@ -20,7 +20,7 @@ public: static const FunctionInfo functions[] = { {0, nullptr, "GetDsEndpoint"}, {1, nullptr, "GetSetupEvent"}, - {2, nullptr, "Unknown"}, + {2, nullptr, "Unknown2"}, {3, nullptr, "EnableInterface"}, {4, nullptr, "DisableInterface"}, {5, nullptr, "CtrlInPostBufferAsync"}, @@ -55,6 +55,7 @@ public: {9, nullptr, "SetBinaryObjectStore"}, {10, nullptr, "Enable"}, {11, nullptr, "Disable"}, + {12, nullptr, "Unknown12"}, }; // clang-format on @@ -69,13 +70,13 @@ public: static const FunctionInfo functions[] = { {0, nullptr, "Open"}, {1, nullptr, "Close"}, - {2, nullptr, "Unknown1"}, + {2, nullptr, "Unknown2"}, {3, nullptr, "Populate"}, {4, nullptr, "PostBufferAsync"}, {5, nullptr, "GetXferReport"}, {6, nullptr, "PostBufferMultiAsync"}, - {7, nullptr, "Unknown3"}, - {8, nullptr, "Unknown4"}, + {7, nullptr, "Unknown7"}, + {8, nullptr, "Unknown8"}, }; // clang-format on @@ -88,13 +89,13 @@ public: explicit IClientIfSession() : ServiceFramework{"IClientIfSession"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, + {0, nullptr, "Unknown0"}, {1, nullptr, "SetInterface"}, {2, nullptr, "GetInterface"}, {3, nullptr, "GetAlternateInterface"}, {4, nullptr, "GetCurrentFrame"}, {5, nullptr, "CtrlXferAsync"}, - {6, nullptr, "Unknown2"}, + {6, nullptr, "Unknown6"}, {7, nullptr, "GetCtrlXferReport"}, {8, nullptr, "ResetDevice"}, {9, nullptr, "OpenUsbEp"}, @@ -118,7 +119,7 @@ public: {5, nullptr, "DestroyInterfaceAvailableEvent"}, {6, nullptr, "GetInterfaceStateChangeEvent"}, {7, nullptr, "AcquireUsbIf"}, - {8, nullptr, "Unknown1"}, + {8, nullptr, "Unknown8"}, }; // clang-format on @@ -179,8 +180,8 @@ public: {4, nullptr, "GetFwRevision"}, {5, nullptr, "GetManufacturerId"}, {6, nullptr, "GetDeviceId"}, - {7, nullptr, "Unknown1"}, - {8, nullptr, "Unknown2"}, + {7, nullptr, "Unknown7"}, + {8, nullptr, "Unknown8"}, }; // clang-format on @@ -215,12 +216,12 @@ public: explicit USB_PM() : ServiceFramework{"usb:pm"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, - {1, nullptr, "Unknown2"}, - {2, nullptr, "Unknown3"}, - {3, nullptr, "Unknown4"}, - {4, nullptr, "Unknown5"}, - {5, nullptr, "Unknown6"}, + {0, nullptr, "Unknown0"}, + {1, nullptr, "Unknown1"}, + {2, nullptr, "Unknown2"}, + {3, nullptr, "Unknown3"}, + {4, nullptr, "Unknown4"}, + {5, nullptr, "Unknown5"}, }; // clang-format on diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp index fdc62d05b..55e00dd93 100644 --- a/src/core/hle/service/vi/vi.cpp +++ b/src/core/hle/service/vi/vi.cpp @@ -101,8 +101,8 @@ public: } std::u16string ReadInterfaceToken() { - u32 unknown = Read<u32_le>(); - u32 length = Read<u32_le>(); + [[maybe_unused]] const u32 unknown = Read<u32_le>(); + const u32 length = Read<u32_le>(); std::u16string token{}; @@ -159,7 +159,7 @@ public: header.data_size = static_cast<u32_le>(write_index - sizeof(Header)); header.data_offset = sizeof(Header); header.objects_size = 4; - header.objects_offset = sizeof(Header) + header.data_size; + header.objects_offset = static_cast<u32>(sizeof(Header) + header.data_size); std::memcpy(buffer.data(), &header, sizeof(Header)); return buffer; @@ -215,10 +215,9 @@ public: explicit IGBPConnectRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPConnectRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); } @@ -267,7 +266,7 @@ protected: private: struct Data { - u32_le unk_0; + u32_le unk_0{}; }; Data data{}; @@ -279,10 +278,9 @@ public: : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPSetPreallocatedBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); buffer = Read<NVFlinger::IGBPBuffer>(); } @@ -306,15 +304,40 @@ protected: } }; +class IGBPCancelBufferRequestParcel : public Parcel { +public: + explicit IGBPCancelBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { + Deserialize(); + } + + void DeserializeData() override { + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); + data = Read<Data>(); + } + + struct Data { + u32_le slot; + Service::Nvidia::MultiFence multi_fence; + }; + + Data data; +}; + +class IGBPCancelBufferResponseParcel : public Parcel { +protected: + void SerializeData() override { + Write<u32>(0); // Success + } +}; + class IGBPDequeueBufferRequestParcel : public Parcel { public: explicit IGBPDequeueBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPDequeueBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); } @@ -333,7 +356,6 @@ class IGBPDequeueBufferResponseParcel : public Parcel { public: explicit IGBPDequeueBufferResponseParcel(u32 slot, Service::Nvidia::MultiFence& multi_fence) : slot(slot), multi_fence(multi_fence) {} - ~IGBPDequeueBufferResponseParcel() override = default; protected: void SerializeData() override { @@ -352,10 +374,9 @@ public: explicit IGBPRequestBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPRequestBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); slot = Read<u32_le>(); } @@ -384,10 +405,9 @@ public: explicit IGBPQueueBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPQueueBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); } @@ -447,10 +467,9 @@ public: explicit IGBPQueryRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPQueryRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); type = Read<u32_le>(); } @@ -511,6 +530,7 @@ private: LOG_DEBUG(Service_VI, "called. id=0x{:08X} transaction={:X}, flags=0x{:08X}", id, static_cast<u32>(transaction), flags); + const auto guard = nv_flinger->Lock(); auto& buffer_queue = nv_flinger->FindBufferQueue(id); switch (transaction) { @@ -518,9 +538,9 @@ private: IGBPConnectRequestParcel request{ctx.ReadBuffer()}; IGBPConnectResponseParcel response{ static_cast<u32>(static_cast<u32>(DisplayResolution::UndockedWidth) * - Settings::values.resolution_factor), + Settings::values.resolution_factor.GetValue()), static_cast<u32>(static_cast<u32>(DisplayResolution::UndockedHeight) * - Settings::values.resolution_factor)}; + Settings::values.resolution_factor.GetValue())}; ctx.WriteBuffer(response.Serialize()); break; } @@ -547,9 +567,10 @@ private: // Wait the current thread until a buffer becomes available ctx.SleepClientThread( "IHOSBinderDriver::DequeueBuffer", UINT64_MAX, - [=](std::shared_ptr<Kernel::Thread> thread, Kernel::HLERequestContext& ctx, - Kernel::ThreadWakeupReason reason) { + [=, this](std::shared_ptr<Kernel::Thread> thread, + Kernel::HLERequestContext& ctx, Kernel::ThreadWakeupReason reason) { // Repeat TransactParcel DequeueBuffer when a buffer is available + const auto guard = nv_flinger->Lock(); auto& buffer_queue = nv_flinger->FindBufferQueue(id); auto result = buffer_queue.DequeueBuffer(width, height); ASSERT_MSG(result != std::nullopt, "Could not dequeue buffer."); @@ -594,7 +615,12 @@ private: break; } case TransactionId::CancelBuffer: { - LOG_CRITICAL(Service_VI, "(STUBBED) called, transaction=CancelBuffer"); + IGBPCancelBufferRequestParcel request{ctx.ReadBuffer()}; + + buffer_queue.CancelBuffer(request.data.slot, request.data.multi_fence); + + IGBPCancelBufferResponseParcel response{}; + ctx.WriteBuffer(response.Serialize()); break; } case TransactionId::Disconnect: { @@ -614,6 +640,14 @@ private: ctx.WriteBuffer(response.Serialize()); break; } + case TransactionId::SetBufferCount: { + LOG_WARNING(Service_VI, "(STUBBED) called, transaction=SetBufferCount"); + [[maybe_unused]] const auto buffer = ctx.ReadBuffer(); + + IGBPEmptyResponseParcel response{}; + ctx.WriteBuffer(response.Serialize()); + break; + } default: ASSERT_MSG(false, "Unimplemented"); } @@ -690,6 +724,7 @@ public: {3215, nullptr, "SetDisplayGamma"}, {3216, nullptr, "GetDisplayCmuLuma"}, {3217, nullptr, "SetDisplayCmuLuma"}, + {6013, nullptr, "GetLayerPresentationSubmissionTimestamps"}, {8225, nullptr, "GetSharedBufferMemoryHandleId"}, {8250, nullptr, "OpenSharedLayer"}, {8251, nullptr, "CloseSharedLayer"}, @@ -736,16 +771,16 @@ private: IPC::ResponseBuilder rb{ctx, 6}; rb.Push(RESULT_SUCCESS); - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); } else { rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedWidth) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u32>(Service::VI::DisplayResolution::UndockedHeight) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); } rb.PushRaw<float>(60.0f); // This wouldn't seem to be correct for 30 fps games. @@ -775,6 +810,7 @@ public: {2300, nullptr, "AcquireLayerTexturePresentingEvent"}, {2301, nullptr, "ReleaseLayerTexturePresentingEvent"}, {2302, nullptr, "GetDisplayHotplugEvent"}, + {2303, nullptr, "GetDisplayModeChangedEvent"}, {2402, nullptr, "GetDisplayHotplugState"}, {2501, nullptr, "GetCompositorErrorInfo"}, {2601, nullptr, "GetDisplayErrorEvent"}, @@ -859,6 +895,7 @@ private: const auto layer_id = nv_flinger->CreateLayer(display); if (!layer_id) { + LOG_ERROR(Service_VI, "Layer not found! display=0x{:016X}", display); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -975,6 +1012,7 @@ private: const auto display_id = nv_flinger->OpenDisplay(name); if (!display_id) { + LOG_ERROR(Service_VI, "Display not found! display_name={}", name); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -1017,9 +1055,9 @@ private: // between docked and undocked dimensions. We take the liberty of applying // the resolution scaling factor here. rb.Push(static_cast<u64>(DisplayResolution::UndockedWidth) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u64>(DisplayResolution::UndockedHeight) * - static_cast<u32>(Settings::values.resolution_factor)); + static_cast<u32>(Settings::values.resolution_factor.GetValue())); } void SetLayerScalingMode(Kernel::HLERequestContext& ctx) { @@ -1052,8 +1090,8 @@ private: LOG_WARNING(Service_VI, "(STUBBED) called"); DisplayInfo display_info; - display_info.width *= static_cast<u64>(Settings::values.resolution_factor); - display_info.height *= static_cast<u64>(Settings::values.resolution_factor); + display_info.width *= static_cast<u64>(Settings::values.resolution_factor.GetValue()); + display_info.height *= static_cast<u64>(Settings::values.resolution_factor.GetValue()); ctx.WriteBuffer(&display_info, sizeof(DisplayInfo)); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); @@ -1074,6 +1112,7 @@ private: const auto display_id = nv_flinger->OpenDisplay(display_name); if (!display_id) { + LOG_ERROR(Service_VI, "Layer not found! layer_id={}", layer_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -1081,6 +1120,7 @@ private: const auto buffer_queue_id = nv_flinger->FindBufferQueueId(*display_id, layer_id); if (!buffer_queue_id) { + LOG_ERROR(Service_VI, "Buffer queue id not found! display_id={}", *display_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -1116,6 +1156,7 @@ private: const auto layer_id = nv_flinger->CreateLayer(display_id); if (!layer_id) { + LOG_ERROR(Service_VI, "Layer not found! layer_id={}", *layer_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -1123,6 +1164,7 @@ private: const auto buffer_queue_id = nv_flinger->FindBufferQueueId(display_id, *layer_id); if (!buffer_queue_id) { + LOG_ERROR(Service_VI, "Buffer queue id not found! display_id={}", display_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -1153,6 +1195,7 @@ private: const auto vsync_event = nv_flinger->FindVsyncEvent(display_id); if (!vsync_event) { + LOG_ERROR(Service_VI, "Vsync event was not found for display_id={}", display_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_NOT_FOUND); return; @@ -1180,6 +1223,23 @@ private: } } + void GetIndirectLayerImageRequiredMemoryInfo(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto width = rp.Pop<u64>(); + const auto height = rp.Pop<u64>(); + LOG_DEBUG(Service_VI, "called width={}, height={}", width, height); + + constexpr std::size_t base_size = 0x20000; + constexpr std::size_t alignment = 0x1000; + const auto texture_size = width * height * 4; + const auto out_size = (texture_size + base_size - 1) / base_size * base_size; + + IPC::ResponseBuilder rb{ctx, 6}; + rb.Push(RESULT_SUCCESS); + rb.Push(out_size); + rb.Push(alignment); + } + static ResultVal<ConvertedScaleMode> ConvertScalingModeImpl(NintendoScaleMode mode) { switch (mode) { case NintendoScaleMode::None: @@ -1193,6 +1253,7 @@ private: case NintendoScaleMode::PreserveAspectRatio: return MakeResult(ConvertedScaleMode::PreserveAspectRatio); default: + LOG_ERROR(Service_VI, "Invalid scaling mode specified, mode={}", mode); return ERR_OPERATION_FAILED; } } @@ -1223,7 +1284,8 @@ IApplicationDisplayService::IApplicationDisplayService( {2102, &IApplicationDisplayService::ConvertScalingMode, "ConvertScalingMode"}, {2450, nullptr, "GetIndirectLayerImageMap"}, {2451, nullptr, "GetIndirectLayerImageCropMap"}, - {2460, nullptr, "GetIndirectLayerImageRequiredMemoryInfo"}, + {2460, &IApplicationDisplayService::GetIndirectLayerImageRequiredMemoryInfo, + "GetIndirectLayerImageRequiredMemoryInfo"}, {5202, &IApplicationDisplayService::GetDisplayVsyncEvent, "GetDisplayVsyncEvent"}, {5203, nullptr, "GetDisplayVsyncEventForDebug"}, }; @@ -1249,6 +1311,7 @@ void detail::GetDisplayServiceImpl(Kernel::HLERequestContext& ctx, const auto policy = rp.PopEnum<Policy>(); if (!IsValidServiceAccess(permission, policy)) { + LOG_ERROR(Service_VI, "Permission denied for policy {}", static_cast<u32>(policy)); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_PERMISSION_DENIED); return; diff --git a/src/core/hle/service/vi/vi_u.cpp b/src/core/hle/service/vi/vi_u.cpp index 9d5ceb608..6b7329345 100644 --- a/src/core/hle/service/vi/vi_u.cpp +++ b/src/core/hle/service/vi/vi_u.cpp @@ -12,6 +12,7 @@ VI_U::VI_U(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) : ServiceFramework{"vi:u"}, nv_flinger{std::move(nv_flinger)} { static const FunctionInfo functions[] = { {0, &VI_U::GetDisplayService, "GetDisplayService"}, + {1, nullptr, "GetDisplayServiceWithProxyNameExchange"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/wlan/wlan.cpp b/src/core/hle/service/wlan/wlan.cpp index 2654594c1..0260d7dcf 100644 --- a/src/core/hle/service/wlan/wlan.cpp +++ b/src/core/hle/service/wlan/wlan.cpp @@ -15,34 +15,37 @@ public: explicit WLANInfra() : ServiceFramework{"wlan:inf"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, - {1, nullptr, "Unknown2"}, + {0, nullptr, "OpenMode"}, + {1, nullptr, "CloseMode"}, {2, nullptr, "GetMacAddress"}, {3, nullptr, "StartScan"}, {4, nullptr, "StopScan"}, {5, nullptr, "Connect"}, {6, nullptr, "CancelConnect"}, {7, nullptr, "Disconnect"}, - {8, nullptr, "Unknown3"}, - {9, nullptr, "Unknown4"}, + {8, nullptr, "GetConnectionEvent"}, + {9, nullptr, "GetConnectionStatus"}, {10, nullptr, "GetState"}, {11, nullptr, "GetScanResult"}, {12, nullptr, "GetRssi"}, {13, nullptr, "ChangeRxAntenna"}, - {14, nullptr, "Unknown5"}, - {15, nullptr, "Unknown6"}, + {14, nullptr, "GetFwVersion"}, + {15, nullptr, "RequestSleep"}, {16, nullptr, "RequestWakeUp"}, {17, nullptr, "RequestIfUpDown"}, - {18, nullptr, "Unknown7"}, - {19, nullptr, "Unknown8"}, - {20, nullptr, "Unknown9"}, - {21, nullptr, "Unknown10"}, - {22, nullptr, "Unknown11"}, - {23, nullptr, "Unknown12"}, - {24, nullptr, "Unknown13"}, - {25, nullptr, "Unknown14"}, - {26, nullptr, "Unknown15"}, - {27, nullptr, "Unknown16"}, + {18, nullptr, "Unknown18"}, + {19, nullptr, "Unknown19"}, + {20, nullptr, "Unknown20"}, + {21, nullptr, "Unknown21"}, + {22, nullptr, "Unknown22"}, + {23, nullptr, "Unknown23"}, + {24, nullptr, "Unknown24"}, + {25, nullptr, "Unknown25"}, + {26, nullptr, "Unknown26"}, + {27, nullptr, "Unknown27"}, + {28, nullptr, "Unknown28"}, + {29, nullptr, "Unknown29"}, + {30, nullptr, "Unknown30"}, }; // clang-format on @@ -55,12 +58,12 @@ public: explicit WLANLocal() : ServiceFramework{"wlan:lcl"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, - {1, nullptr, "Unknown2"}, - {2, nullptr, "Unknown3"}, - {3, nullptr, "Unknown4"}, - {4, nullptr, "Unknown5"}, - {5, nullptr, "Unknown6"}, + {0, nullptr, "Unknown0"}, + {1, nullptr, "Unknown1"}, + {2, nullptr, "Unknown2"}, + {3, nullptr, "Unknown3"}, + {4, nullptr, "Unknown4"}, + {5, nullptr, "Unknown5"}, {6, nullptr, "GetMacAddress"}, {7, nullptr, "CreateBss"}, {8, nullptr, "DestroyBss"}, @@ -72,38 +75,42 @@ public: {14, nullptr, "CancelJoin"}, {15, nullptr, "Disconnect"}, {16, nullptr, "SetBeaconLostCount"}, - {17, nullptr, "Unknown7"}, - {18, nullptr, "Unknown8"}, - {19, nullptr, "Unknown9"}, + {17, nullptr, "Unknown17"}, + {18, nullptr, "Unknown18"}, + {19, nullptr, "Unknown19"}, {20, nullptr, "GetBssIndicationEvent"}, {21, nullptr, "GetBssIndicationInfo"}, {22, nullptr, "GetState"}, {23, nullptr, "GetAllowedChannels"}, {24, nullptr, "AddIe"}, {25, nullptr, "DeleteIe"}, - {26, nullptr, "Unknown10"}, - {27, nullptr, "Unknown11"}, + {26, nullptr, "Unknown26"}, + {27, nullptr, "Unknown27"}, {28, nullptr, "CreateRxEntry"}, {29, nullptr, "DeleteRxEntry"}, - {30, nullptr, "Unknown12"}, - {31, nullptr, "Unknown13"}, + {30, nullptr, "Unknown30"}, + {31, nullptr, "Unknown31"}, {32, nullptr, "AddMatchingDataToRxEntry"}, {33, nullptr, "RemoveMatchingDataFromRxEntry"}, {34, nullptr, "GetScanResult"}, - {35, nullptr, "Unknown14"}, + {35, nullptr, "Unknown35"}, {36, nullptr, "SetActionFrameWithBeacon"}, {37, nullptr, "CancelActionFrameWithBeacon"}, {38, nullptr, "CreateRxEntryForActionFrame"}, {39, nullptr, "DeleteRxEntryForActionFrame"}, - {40, nullptr, "Unknown15"}, - {41, nullptr, "Unknown16"}, + {40, nullptr, "Unknown40"}, + {41, nullptr, "Unknown41"}, {42, nullptr, "CancelGetActionFrame"}, {43, nullptr, "GetRssi"}, - {44, nullptr, "Unknown17"}, - {45, nullptr, "Unknown18"}, - {46, nullptr, "Unknown19"}, - {47, nullptr, "Unknown20"}, - {48, nullptr, "Unknown21"}, + {44, nullptr, "Unknown44"}, + {45, nullptr, "Unknown45"}, + {46, nullptr, "Unknown46"}, + {47, nullptr, "Unknown47"}, + {48, nullptr, "Unknown48"}, + {49, nullptr, "Unknown49"}, + {50, nullptr, "Unknown50"}, + {51, nullptr, "Unknown51"}, + {52, nullptr, "Unknown52"}, }; // clang-format on @@ -142,18 +149,19 @@ public: explicit WLANSocketManager() : ServiceFramework{"wlan:soc"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, - {1, nullptr, "Unknown2"}, - {2, nullptr, "Unknown3"}, - {3, nullptr, "Unknown4"}, - {4, nullptr, "Unknown5"}, - {5, nullptr, "Unknown6"}, + {0, nullptr, "Unknown0"}, + {1, nullptr, "Unknown1"}, + {2, nullptr, "Unknown2"}, + {3, nullptr, "Unknown3"}, + {4, nullptr, "Unknown4"}, + {5, nullptr, "Unknown5"}, {6, nullptr, "GetMacAddress"}, {7, nullptr, "SwitchTsfTimerFunction"}, - {8, nullptr, "Unknown7"}, - {9, nullptr, "Unknown8"}, - {10, nullptr, "Unknown9"}, - {11, nullptr, "Unknown10"}, + {8, nullptr, "Unknown8"}, + {9, nullptr, "Unknown9"}, + {10, nullptr, "Unknown10"}, + {11, nullptr, "Unknown11"}, + {12, nullptr, "Unknown12"}, }; // clang-format on diff --git a/src/core/loader/deconstructed_rom_directory.cpp b/src/core/loader/deconstructed_rom_directory.cpp index 53559e8b1..2002dc4f2 100644 --- a/src/core/loader/deconstructed_rom_directory.cpp +++ b/src/core/loader/deconstructed_rom_directory.cpp @@ -14,6 +14,7 @@ #include "core/file_sys/romfs_factory.h" #include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/deconstructed_rom_directory.h" @@ -88,7 +89,7 @@ FileType AppLoader_DeconstructedRomDirectory::IdentifyType(const FileSys::Virtua } AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirectory::Load( - Kernel::Process& process) { + Kernel::Process& process, Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -113,7 +114,8 @@ AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirect } if (override_update) { - const FileSys::PatchManager patch_manager(metadata.GetTitleID()); + const FileSys::PatchManager patch_manager( + metadata.GetTitleID(), system.GetFileSystemController(), system.GetContentProvider()); dir = patch_manager.PatchExeFS(dir); } @@ -129,27 +131,48 @@ AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirect } metadata.Print(); - if (process.LoadFromMetadata(metadata).IsError()) { - return {ResultStatus::ErrorUnableToParseKernelMetadata, {}}; + const auto static_modules = {"rtld", "main", "subsdk0", "subsdk1", "subsdk2", "subsdk3", + "subsdk4", "subsdk5", "subsdk6", "subsdk7", "sdk"}; + + // Use the NSO module loader to figure out the code layout + std::size_t code_size{}; + for (const auto& module : static_modules) { + const FileSys::VirtualFile module_file{dir->GetFile(module)}; + if (!module_file) { + continue; + } + + const bool should_pass_arguments = std::strcmp(module, "rtld") == 0; + const auto tentative_next_load_addr = AppLoader_NSO::LoadModule( + process, system, *module_file, code_size, should_pass_arguments, false); + if (!tentative_next_load_addr) { + return {ResultStatus::ErrorLoadingNSO, {}}; + } + + code_size = *tentative_next_load_addr; } - const FileSys::PatchManager pm(metadata.GetTitleID()); + // Setup the process code layout + if (process.LoadFromMetadata(metadata, code_size).IsError()) { + return {ResultStatus::ErrorUnableToParseKernelMetadata, {}}; + } // Load NSO modules modules.clear(); - const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress(); - VAddr next_load_addr = base_address; - for (const auto& module : {"rtld", "main", "subsdk0", "subsdk1", "subsdk2", "subsdk3", - "subsdk4", "subsdk5", "subsdk6", "subsdk7", "sdk"}) { - const FileSys::VirtualFile module_file = dir->GetFile(module); - if (module_file == nullptr) { + const VAddr base_address{process.PageTable().GetCodeRegionStart()}; + VAddr next_load_addr{base_address}; + const FileSys::PatchManager pm{metadata.GetTitleID(), system.GetFileSystemController(), + system.GetContentProvider()}; + for (const auto& module : static_modules) { + const FileSys::VirtualFile module_file{dir->GetFile(module)}; + if (!module_file) { continue; } - const VAddr load_addr = next_load_addr; + const VAddr load_addr{next_load_addr}; const bool should_pass_arguments = std::strcmp(module, "rtld") == 0; - const auto tentative_next_load_addr = - AppLoader_NSO::LoadModule(process, *module_file, load_addr, should_pass_arguments, pm); + const auto tentative_next_load_addr = AppLoader_NSO::LoadModule( + process, system, *module_file, load_addr, should_pass_arguments, true, pm); if (!tentative_next_load_addr) { return {ResultStatus::ErrorLoadingNSO, {}}; } @@ -171,8 +194,8 @@ AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirect // Register the RomFS if a ".romfs" file was found if (romfs_iter != files.end() && *romfs_iter != nullptr) { romfs = *romfs_iter; - Core::System::GetInstance().GetFileSystemController().RegisterRomFS( - std::make_unique<FileSys::RomFSFactory>(*this)); + system.GetFileSystemController().RegisterRomFS(std::make_unique<FileSys::RomFSFactory>( + *this, system.GetContentProvider(), system.GetFileSystemController())); } is_loaded = true; diff --git a/src/core/loader/deconstructed_rom_directory.h b/src/core/loader/deconstructed_rom_directory.h index 1c0a354a4..35d340317 100644 --- a/src/core/loader/deconstructed_rom_directory.h +++ b/src/core/loader/deconstructed_rom_directory.h @@ -9,6 +9,10 @@ #include "core/file_sys/program_metadata.h" #include "core/loader/loader.h" +namespace Core { +class System; +} + namespace Loader { /** @@ -37,7 +41,7 @@ public: return IdentifyType(file); } - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override; ResultStatus ReadIcon(std::vector<u8>& buffer) override; diff --git a/src/core/loader/elf.cpp b/src/core/loader/elf.cpp index 8908e5328..dca1fcb18 100644 --- a/src/core/loader/elf.cpp +++ b/src/core/loader/elf.cpp @@ -10,8 +10,8 @@ #include "common/file_util.h" #include "common/logging/log.h" #include "core/hle/kernel/code_set.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/vm_manager.h" #include "core/loader/elf.h" #include "core/memory.h" @@ -383,7 +383,8 @@ FileType AppLoader_ELF::IdentifyType(const FileSys::VirtualFile& file) { return FileType::Error; } -AppLoader_ELF::LoadResult AppLoader_ELF::Load(Kernel::Process& process) { +AppLoader_ELF::LoadResult AppLoader_ELF::Load(Kernel::Process& process, + [[maybe_unused]] Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -393,15 +394,20 @@ AppLoader_ELF::LoadResult AppLoader_ELF::Load(Kernel::Process& process) { return {ResultStatus::ErrorIncorrectELFFileSize, {}}; } - const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress(); + const VAddr base_address = process.PageTable().GetCodeRegionStart(); ElfReader elf_reader(&buffer[0]); Kernel::CodeSet codeset = elf_reader.LoadInto(base_address); const VAddr entry_point = codeset.entrypoint; + // Setup the process code layout + if (process.LoadFromMetadata(FileSys::ProgramMetadata::GetDefault(), buffer.size()).IsError()) { + return {ResultStatus::ErrorNotInitialized, {}}; + } + process.LoadModule(std::move(codeset), entry_point); is_loaded = true; - return {ResultStatus::Success, LoadParameters{48, Memory::DEFAULT_STACK_SIZE}}; + return {ResultStatus::Success, LoadParameters{48, Core::Memory::DEFAULT_STACK_SIZE}}; } } // namespace Loader diff --git a/src/core/loader/elf.h b/src/core/loader/elf.h index 7ef7770a6..3527933ad 100644 --- a/src/core/loader/elf.h +++ b/src/core/loader/elf.h @@ -8,6 +8,10 @@ #include "common/common_types.h" #include "core/loader/loader.h" +namespace Core { +class System; +} + namespace Loader { /// Loads an ELF/AXF file @@ -26,7 +30,7 @@ public: return IdentifyType(file); } - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; }; } // namespace Loader diff --git a/src/core/loader/kip.cpp b/src/core/loader/kip.cpp index 092103abe..2a905d3e4 100644 --- a/src/core/loader/kip.cpp +++ b/src/core/loader/kip.cpp @@ -7,14 +7,16 @@ #include "core/file_sys/program_metadata.h" #include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/code_set.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" #include "core/loader/kip.h" +#include "core/memory.h" namespace Loader { namespace { constexpr u32 PageAlignSize(u32 size) { - return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK; + return static_cast<u32>((size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK); } } // Anonymous namespace @@ -41,7 +43,8 @@ FileType AppLoader_KIP::GetFileType() const { : FileType::Error; } -AppLoader::LoadResult AppLoader_KIP::Load(Kernel::Process& process) { +AppLoader::LoadResult AppLoader_KIP::Load(Kernel::Process& process, + [[maybe_unused]] Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -68,7 +71,7 @@ AppLoader::LoadResult AppLoader_KIP::Load(Kernel::Process& process) { kip->GetMainThreadCpuCore(), kip->GetMainThreadStackSize(), kip->GetTitleID(), 0xFFFFFFFFFFFFFFFF, kip->GetKernelCapabilities()); - const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress(); + const VAddr base_address = process.PageTable().GetCodeRegionStart(); Kernel::CodeSet codeset; Kernel::PhysicalMemory program_image; diff --git a/src/core/loader/kip.h b/src/core/loader/kip.h index 12ca40269..dee05a7b5 100644 --- a/src/core/loader/kip.h +++ b/src/core/loader/kip.h @@ -6,6 +6,10 @@ #include "core/loader/loader.h" +namespace Core { +class System; +} + namespace FileSys { class KIP; } @@ -26,7 +30,7 @@ public: FileType GetFileType() const override; - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; private: std::unique_ptr<FileSys::KIP> kip; diff --git a/src/core/loader/loader.cpp b/src/core/loader/loader.cpp index 59ca7091a..deffe7379 100644 --- a/src/core/loader/loader.cpp +++ b/src/core/loader/loader.cpp @@ -3,11 +3,14 @@ // Refer to the license.txt file included. #include <memory> +#include <optional> #include <ostream> #include <string> +#include "common/concepts.h" #include "common/file_util.h" #include "common/logging/log.h" #include "common/string_util.h" +#include "core/core.h" #include "core/hle/kernel/process.h" #include "core/loader/deconstructed_rom_directory.h" #include "core/loader/elf.h" @@ -21,27 +24,41 @@ namespace Loader { -FileType IdentifyFile(FileSys::VirtualFile file) { - FileType type; - -#define CHECK_TYPE(loader) \ - type = AppLoader_##loader::IdentifyType(file); \ - if (FileType::Error != type) \ - return type; +namespace { - CHECK_TYPE(DeconstructedRomDirectory) - CHECK_TYPE(ELF) - CHECK_TYPE(NSO) - CHECK_TYPE(NRO) - CHECK_TYPE(NCA) - CHECK_TYPE(XCI) - CHECK_TYPE(NAX) - CHECK_TYPE(NSP) - CHECK_TYPE(KIP) +template <Common::DerivedFrom<AppLoader> T> +std::optional<FileType> IdentifyFileLoader(FileSys::VirtualFile file) { + const auto file_type = T::IdentifyType(file); + if (file_type != FileType::Error) { + return file_type; + } + return std::nullopt; +} -#undef CHECK_TYPE +} // namespace - return FileType::Unknown; +FileType IdentifyFile(FileSys::VirtualFile file) { + if (const auto romdir_type = IdentifyFileLoader<AppLoader_DeconstructedRomDirectory>(file)) { + return *romdir_type; + } else if (const auto elf_type = IdentifyFileLoader<AppLoader_ELF>(file)) { + return *elf_type; + } else if (const auto nso_type = IdentifyFileLoader<AppLoader_NSO>(file)) { + return *nso_type; + } else if (const auto nro_type = IdentifyFileLoader<AppLoader_NRO>(file)) { + return *nro_type; + } else if (const auto nca_type = IdentifyFileLoader<AppLoader_NCA>(file)) { + return *nca_type; + } else if (const auto xci_type = IdentifyFileLoader<AppLoader_XCI>(file)) { + return *xci_type; + } else if (const auto nax_type = IdentifyFileLoader<AppLoader_NAX>(file)) { + return *nax_type; + } else if (const auto nsp_type = IdentifyFileLoader<AppLoader_NSP>(file)) { + return *nsp_type; + } else if (const auto kip_type = IdentifyFileLoader<AppLoader_KIP>(file)) { + return *kip_type; + } else { + return FileType::Unknown; + } } FileType GuessFromFilename(const std::string& name) { @@ -51,7 +68,7 @@ FileType GuessFromFilename(const std::string& name) { return FileType::NCA; const std::string extension = - Common::ToLower(std::string(FileUtil::GetExtensionFromFilename(name))); + Common::ToLower(std::string(Common::FS::GetExtensionFromFilename(name))); if (extension == "elf") return FileType::ELF; @@ -178,15 +195,14 @@ AppLoader::~AppLoader() = default; /** * Get a loader for a file with a specific type - * @param file The file to load - * @param type The type of the file - * @param file the file to retrieve the loader for - * @param type the file type + * @param system The system context to use. + * @param file The file to retrieve the loader for + * @param type The file type * @return std::unique_ptr<AppLoader> a pointer to a loader object; nullptr for unsupported type */ -static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileType type) { +static std::unique_ptr<AppLoader> GetFileLoader(Core::System& system, FileSys::VirtualFile file, + FileType type) { switch (type) { - // Standard ELF file format. case FileType::ELF: return std::make_unique<AppLoader_ELF>(std::move(file)); @@ -205,7 +221,8 @@ static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileT // NX XCI (nX Card Image) file format. case FileType::XCI: - return std::make_unique<AppLoader_XCI>(std::move(file)); + return std::make_unique<AppLoader_XCI>(std::move(file), system.GetFileSystemController(), + system.GetContentProvider()); // NX NAX (NintendoAesXts) file format. case FileType::NAX: @@ -213,7 +230,8 @@ static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileT // NX NSP (Nintendo Submission Package) file format case FileType::NSP: - return std::make_unique<AppLoader_NSP>(std::move(file)); + return std::make_unique<AppLoader_NSP>(std::move(file), system.GetFileSystemController(), + system.GetContentProvider()); // NX KIP (Kernel Internal Process) file format case FileType::KIP: @@ -228,20 +246,21 @@ static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileT } } -std::unique_ptr<AppLoader> GetLoader(FileSys::VirtualFile file) { +std::unique_ptr<AppLoader> GetLoader(Core::System& system, FileSys::VirtualFile file) { FileType type = IdentifyFile(file); - FileType filename_type = GuessFromFilename(file->GetName()); + const FileType filename_type = GuessFromFilename(file->GetName()); // Special case: 00 is either a NCA or NAX. if (type != filename_type && !(file->GetName() == "00" && type == FileType::NAX)) { LOG_WARNING(Loader, "File {} has a different type than its extension.", file->GetName()); - if (FileType::Unknown == type) + if (FileType::Unknown == type) { type = filename_type; + } } LOG_DEBUG(Loader, "Loading file {} as {}...", file->GetName(), GetFileTypeString(type)); - return GetFileLoader(std::move(file), type); + return GetFileLoader(system, std::move(file), type); } } // namespace Loader diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h index 227ecc704..8dc2d7615 100644 --- a/src/core/loader/loader.h +++ b/src/core/loader/loader.h @@ -15,6 +15,10 @@ #include "core/file_sys/control_metadata.h" #include "core/file_sys/vfs.h" +namespace Core { +class System; +} + namespace FileSys { class NACP; } // namespace FileSys @@ -154,9 +158,10 @@ public: /** * Load the application and return the created Process instance * @param process The newly created process. + * @param system The system that this process is being loaded under. * @return The status result of the operation. */ - virtual LoadResult Load(Kernel::Process& process) = 0; + virtual LoadResult Load(Kernel::Process& process, Core::System& system) = 0; /** * Get the code (typically .code section) of the application @@ -285,9 +290,12 @@ protected: /** * Identifies a bootable file and return a suitable loader - * @param file The bootable file - * @return the best loader for this file + * + * @param system The system context. + * @param file The bootable file. + * + * @return the best loader for this file. */ -std::unique_ptr<AppLoader> GetLoader(FileSys::VirtualFile file); +std::unique_ptr<AppLoader> GetLoader(Core::System& system, FileSys::VirtualFile file); } // namespace Loader diff --git a/src/core/loader/nax.cpp b/src/core/loader/nax.cpp index a152981a0..49028177b 100644 --- a/src/core/loader/nax.cpp +++ b/src/core/loader/nax.cpp @@ -41,7 +41,8 @@ FileType AppLoader_NAX::GetFileType() const { return IdentifyTypeImpl(*nax); } -AppLoader_NAX::LoadResult AppLoader_NAX::Load(Kernel::Process& process) { +AppLoader_NAX::LoadResult AppLoader_NAX::Load(Kernel::Process& process, + [[maybe_unused]] Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -65,7 +66,7 @@ AppLoader_NAX::LoadResult AppLoader_NAX::Load(Kernel::Process& process) { return {nca_status, {}}; } - const auto result = nca_loader->Load(process); + const auto result = nca_loader->Load(process, system); if (result.first != ResultStatus::Success) { return result; } diff --git a/src/core/loader/nax.h b/src/core/loader/nax.h index eaec9bf58..c2b7722b5 100644 --- a/src/core/loader/nax.h +++ b/src/core/loader/nax.h @@ -8,10 +8,12 @@ #include "common/common_types.h" #include "core/loader/loader.h" -namespace FileSys { +namespace Core { +class System; +} +namespace FileSys { class NAX; - } // namespace FileSys namespace Loader { @@ -33,7 +35,7 @@ public: FileType GetFileType() const override; - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override; u64 ReadRomFSIVFCOffset() const override; diff --git a/src/core/loader/nca.cpp b/src/core/loader/nca.cpp index 5a0469978..fa694de37 100644 --- a/src/core/loader/nca.cpp +++ b/src/core/loader/nca.cpp @@ -31,7 +31,7 @@ FileType AppLoader_NCA::IdentifyType(const FileSys::VirtualFile& file) { return FileType::Error; } -AppLoader_NCA::LoadResult AppLoader_NCA::Load(Kernel::Process& process) { +AppLoader_NCA::LoadResult AppLoader_NCA::Load(Kernel::Process& process, Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -52,14 +52,14 @@ AppLoader_NCA::LoadResult AppLoader_NCA::Load(Kernel::Process& process) { directory_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(exefs, true); - const auto load_result = directory_loader->Load(process); + const auto load_result = directory_loader->Load(process, system); if (load_result.first != ResultStatus::Success) { return load_result; } if (nca->GetRomFS() != nullptr && nca->GetRomFS()->GetSize() > 0) { - Core::System::GetInstance().GetFileSystemController().RegisterRomFS( - std::make_unique<FileSys::RomFSFactory>(*this)); + system.GetFileSystemController().RegisterRomFS(std::make_unique<FileSys::RomFSFactory>( + *this, system.GetContentProvider(), system.GetFileSystemController())); } is_loaded = true; diff --git a/src/core/loader/nca.h b/src/core/loader/nca.h index e47dc0e47..711070294 100644 --- a/src/core/loader/nca.h +++ b/src/core/loader/nca.h @@ -8,6 +8,10 @@ #include "core/file_sys/vfs.h" #include "core/loader/loader.h" +namespace Core { +class System; +} + namespace FileSys { class NCA; } @@ -33,7 +37,7 @@ public: return IdentifyType(file); } - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override; u64 ReadRomFSIVFCOffset() const override; diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp index 175898b91..5f4b3104b 100644 --- a/src/core/loader/nro.cpp +++ b/src/core/loader/nro.cpp @@ -16,8 +16,8 @@ #include "core/file_sys/vfs_offset.h" #include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/code_set.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/vm_manager.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/nro.h" #include "core/loader/nso.h" @@ -127,11 +127,11 @@ FileType AppLoader_NRO::IdentifyType(const FileSys::VirtualFile& file) { } static constexpr u32 PageAlignSize(u32 size) { - return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK; + return static_cast<u32>((size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK); } static bool LoadNroImpl(Kernel::Process& process, const std::vector<u8>& data, - const std::string& name, VAddr load_base) { + const std::string& name) { if (data.size() < sizeof(NroHeader)) { return {}; } @@ -187,41 +187,44 @@ static bool LoadNroImpl(Kernel::Process& process, const std::vector<u8>& data, codeset.DataSegment().size += bss_size; program_image.resize(static_cast<u32>(program_image.size()) + bss_size); + // Setup the process code layout + if (process.LoadFromMetadata(FileSys::ProgramMetadata::GetDefault(), program_image.size()) + .IsError()) { + return false; + } + // Load codeset for current process codeset.memory = std::move(program_image); - process.LoadModule(std::move(codeset), load_base); + process.LoadModule(std::move(codeset), process.PageTable().GetCodeRegionStart()); // Register module with GDBStub - GDBStub::RegisterModule(name, load_base, load_base); + GDBStub::RegisterModule(name, process.PageTable().GetCodeRegionStart(), + process.PageTable().GetCodeRegionEnd()); return true; } -bool AppLoader_NRO::LoadNro(Kernel::Process& process, const FileSys::VfsFile& file, - VAddr load_base) { - return LoadNroImpl(process, file.ReadAllBytes(), file.GetName(), load_base); +bool AppLoader_NRO::LoadNro(Kernel::Process& process, const FileSys::VfsFile& file) { + return LoadNroImpl(process, file.ReadAllBytes(), file.GetName()); } -AppLoader_NRO::LoadResult AppLoader_NRO::Load(Kernel::Process& process) { +AppLoader_NRO::LoadResult AppLoader_NRO::Load(Kernel::Process& process, Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } - // Load NRO - const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress(); - - if (!LoadNro(process, *file, base_address)) { + if (!LoadNro(process, *file)) { return {ResultStatus::ErrorLoadingNRO, {}}; } if (romfs != nullptr) { - Core::System::GetInstance().GetFileSystemController().RegisterRomFS( - std::make_unique<FileSys::RomFSFactory>(*this)); + system.GetFileSystemController().RegisterRomFS(std::make_unique<FileSys::RomFSFactory>( + *this, system.GetContentProvider(), system.GetFileSystemController())); } is_loaded = true; return {ResultStatus::Success, - LoadParameters{Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE}}; + LoadParameters{Kernel::THREADPRIO_DEFAULT, Core::Memory::DEFAULT_STACK_SIZE}}; } ResultStatus AppLoader_NRO::ReadIcon(std::vector<u8>& buffer) { diff --git a/src/core/loader/nro.h b/src/core/loader/nro.h index 71811bc29..a2aab2ecc 100644 --- a/src/core/loader/nro.h +++ b/src/core/loader/nro.h @@ -10,6 +10,10 @@ #include "common/common_types.h" #include "core/loader/loader.h" +namespace Core { +class System; +} + namespace FileSys { class NACP; } @@ -37,7 +41,7 @@ public: return IdentifyType(file); } - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadIcon(std::vector<u8>& buffer) override; ResultStatus ReadProgramId(u64& out_program_id) override; @@ -47,7 +51,7 @@ public: bool IsRomFSUpdatable() const override; private: - bool LoadNro(Kernel::Process& process, const FileSys::VfsFile& file, VAddr load_base); + bool LoadNro(Kernel::Process& process, const FileSys::VfsFile& file); std::vector<u8> icon_data; std::unique_ptr<FileSys::NACP> nacp; diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp index 044067a5b..aa85c1a29 100644 --- a/src/core/loader/nso.cpp +++ b/src/core/loader/nso.cpp @@ -16,8 +16,8 @@ #include "core/file_sys/patch_manager.h" #include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/code_set.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/vm_manager.h" #include "core/loader/nso.h" #include "core/memory.h" #include "core/settings.h" @@ -37,7 +37,7 @@ static_assert(sizeof(MODHeader) == 0x1c, "MODHeader has incorrect size."); std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data, const NSOSegmentHeader& header) { - const std::vector<u8> uncompressed_data = + std::vector<u8> uncompressed_data = Common::Compression::DecompressDataLZ4(compressed_data, header.size); ASSERT_MSG(uncompressed_data.size() == header.size, "{} != {}", header.size, @@ -47,7 +47,7 @@ std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data, } constexpr u32 PageAlignSize(u32 size) { - return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK; + return static_cast<u32>((size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK); } } // Anonymous namespace @@ -71,21 +71,21 @@ FileType AppLoader_NSO::IdentifyType(const FileSys::VirtualFile& file) { return FileType::NSO; } -std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, +std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, Core::System& system, const FileSys::VfsFile& file, VAddr load_base, - bool should_pass_arguments, + bool should_pass_arguments, bool load_into_process, std::optional<FileSys::PatchManager> pm) { if (file.GetSize() < sizeof(NSOHeader)) { - return {}; + return std::nullopt; } NSOHeader nso_header{}; if (sizeof(NSOHeader) != file.ReadObject(&nso_header)) { - return {}; + return std::nullopt; } if (nso_header.magic != Common::MakeMagic('N', 'S', 'O', '0')) { - return {}; + return std::nullopt; } // Build program image @@ -97,21 +97,17 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, if (nso_header.IsSegmentCompressed(i)) { data = DecompressSegment(data, nso_header.segments[i]); } - program_image.resize(nso_header.segments[i].location + - PageAlignSize(static_cast<u32>(data.size()))); + program_image.resize(nso_header.segments[i].location + static_cast<u32>(data.size())); std::memcpy(program_image.data() + nso_header.segments[i].location, data.data(), data.size()); codeset.segments[i].addr = nso_header.segments[i].location; codeset.segments[i].offset = nso_header.segments[i].location; - codeset.segments[i].size = PageAlignSize(static_cast<u32>(data.size())); + codeset.segments[i].size = nso_header.segments[i].size; } - if (should_pass_arguments) { - std::vector<u8> arg_data{Settings::values.program_args.begin(), - Settings::values.program_args.end()}; - if (arg_data.empty()) { - arg_data.resize(NSO_ARGUMENT_DEFAULT_SIZE); - } + if (should_pass_arguments && !Settings::values.program_args.empty()) { + const auto arg_data{Settings::values.program_args}; + codeset.DataSegment().size += NSO_ARGUMENT_DATA_ALLOCATION_SIZE; NSOArgumentHeader args_header{ NSO_ARGUMENT_DATA_ALLOCATION_SIZE, static_cast<u32_le>(arg_data.size()), {}}; @@ -123,24 +119,15 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, arg_data.size()); } - // MOD header pointer is at .text offset + 4 - u32 module_offset; - std::memcpy(&module_offset, program_image.data() + 4, sizeof(u32)); - - // Read MOD header - MODHeader mod_header{}; - // Default .bss to size in segment header if MOD0 section doesn't exist - u32 bss_size{PageAlignSize(nso_header.segments[2].bss_size)}; - std::memcpy(&mod_header, program_image.data() + module_offset, sizeof(MODHeader)); - const bool has_mod_header{mod_header.magic == Common::MakeMagic('M', 'O', 'D', '0')}; - if (has_mod_header) { - // Resize program image to include .bss section and page align each section - bss_size = PageAlignSize(mod_header.bss_end_offset - mod_header.bss_start_offset); - } - codeset.DataSegment().size += bss_size; - const u32 image_size{PageAlignSize(static_cast<u32>(program_image.size()) + bss_size)}; + codeset.DataSegment().size += nso_header.segments[2].bss_size; + const u32 image_size{ + PageAlignSize(static_cast<u32>(program_image.size()) + nso_header.segments[2].bss_size)}; program_image.resize(image_size); + for (std::size_t i = 0; i < nso_header.segments.size(); ++i) { + codeset.segments[i].size = PageAlignSize(codeset.segments[i].size); + } + // Apply patches if necessary if (pm && (pm->HasNSOPatch(nso_header.build_id) || Settings::values.dump_nso)) { std::vector<u8> pi_header; @@ -154,11 +141,15 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, std::copy(pi_header.begin() + sizeof(NSOHeader), pi_header.end(), program_image.data()); } + // If we aren't actually loading (i.e. just computing the process code layout), we are done + if (!load_into_process) { + return load_base + image_size; + } + // Apply cheats if they exist and the program has a valid title ID if (pm) { - auto& system = Core::System::GetInstance(); system.SetCurrentProcessBuildID(nso_header.build_id); - const auto cheats = pm->CreateCheatList(system, nso_header.build_id); + const auto cheats = pm->CreateCheatList(nso_header.build_id); if (!cheats.empty()) { system.RegisterCheatList(cheats, nso_header.build_id, load_base, image_size); } @@ -174,7 +165,7 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, return load_base + image_size; } -AppLoader_NSO::LoadResult AppLoader_NSO::Load(Kernel::Process& process) { +AppLoader_NSO::LoadResult AppLoader_NSO::Load(Kernel::Process& process, Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -182,8 +173,8 @@ AppLoader_NSO::LoadResult AppLoader_NSO::Load(Kernel::Process& process) { modules.clear(); // Load module - const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress(); - if (!LoadModule(process, *file, base_address, true)) { + const VAddr base_address = process.PageTable().GetCodeRegionStart(); + if (!LoadModule(process, system, *file, base_address, true, true)) { return {ResultStatus::ErrorLoadingNSO, {}}; } @@ -192,7 +183,7 @@ AppLoader_NSO::LoadResult AppLoader_NSO::Load(Kernel::Process& process) { is_loaded = true; return {ResultStatus::Success, - LoadParameters{Kernel::THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE}}; + LoadParameters{Kernel::THREADPRIO_DEFAULT, Core::Memory::DEFAULT_STACK_SIZE}}; } ResultStatus AppLoader_NSO::ReadNSOModules(Modules& modules) { diff --git a/src/core/loader/nso.h b/src/core/loader/nso.h index d2d600cd9..d331096ae 100644 --- a/src/core/loader/nso.h +++ b/src/core/loader/nso.h @@ -12,6 +12,10 @@ #include "core/file_sys/patch_manager.h" #include "core/loader/loader.h" +namespace Core { +class System; +} + namespace Kernel { class Process; } @@ -55,9 +59,7 @@ struct NSOHeader { static_assert(sizeof(NSOHeader) == 0x100, "NSOHeader has incorrect size."); static_assert(std::is_trivially_copyable_v<NSOHeader>, "NSOHeader must be trivially copyable."); -constexpr u64 NSO_ARGUMENT_DATA_ALLOCATION_SIZE = 0x9000; -// NOTE: Official software default argument state is unverified. -constexpr u64 NSO_ARGUMENT_DEFAULT_SIZE = 1; +constexpr u32 NSO_ARGUMENT_DATA_ALLOCATION_SIZE = 0x9000; struct NSOArgumentHeader { u32_le allocated_size; @@ -82,11 +84,12 @@ public: return IdentifyType(file); } - static std::optional<VAddr> LoadModule(Kernel::Process& process, const FileSys::VfsFile& file, - VAddr load_base, bool should_pass_arguments, + static std::optional<VAddr> LoadModule(Kernel::Process& process, Core::System& system, + const FileSys::VfsFile& file, VAddr load_base, + bool should_pass_arguments, bool load_into_process, std::optional<FileSys::PatchManager> pm = {}); - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadNSOModules(Modules& modules) override; diff --git a/src/core/loader/nsp.cpp b/src/core/loader/nsp.cpp index 13950fc08..e821937fd 100644 --- a/src/core/loader/nsp.cpp +++ b/src/core/loader/nsp.cpp @@ -21,26 +21,33 @@ namespace Loader { -AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file) +AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider) : AppLoader(file), nsp(std::make_unique<FileSys::NSP>(file)), title_id(nsp->GetProgramTitleID()) { - if (nsp->GetStatus() != ResultStatus::Success) + if (nsp->GetStatus() != ResultStatus::Success) { return; + } if (nsp->IsExtractedType()) { secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS()); } else { const auto control_nca = nsp->GetNCA(nsp->GetProgramTitleID(), FileSys::ContentRecordType::Control); - if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) + if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) { return; + } - std::tie(nacp_file, icon_file) = - FileSys::PatchManager(nsp->GetProgramTitleID()).ParseControlNCA(*control_nca); + std::tie(nacp_file, icon_file) = [this, &content_provider, &control_nca, &fsc] { + const FileSys::PatchManager pm{nsp->GetProgramTitleID(), fsc, content_provider}; + return pm.ParseControlNCA(*control_nca); + }(); - if (title_id == 0) + if (title_id == 0) { return; + } secondary_loader = std::make_unique<AppLoader_NCA>( nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program)); @@ -71,7 +78,7 @@ FileType AppLoader_NSP::IdentifyType(const FileSys::VirtualFile& file) { return FileType::Error; } -AppLoader_NSP::LoadResult AppLoader_NSP::Load(Kernel::Process& process) { +AppLoader_NSP::LoadResult AppLoader_NSP::Load(Kernel::Process& process, Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -99,15 +106,14 @@ AppLoader_NSP::LoadResult AppLoader_NSP::Load(Kernel::Process& process) { return {ResultStatus::ErrorNSPMissingProgramNCA, {}}; } - const auto result = secondary_loader->Load(process); + const auto result = secondary_loader->Load(process, system); if (result.first != ResultStatus::Success) { return result; } FileSys::VirtualFile update_raw; if (ReadUpdateRaw(update_raw) == ResultStatus::Success && update_raw != nullptr) { - Core::System::GetInstance().GetFileSystemController().SetPackedUpdate( - std::move(update_raw)); + system.GetFileSystemController().SetPackedUpdate(std::move(update_raw)); } is_loaded = true; diff --git a/src/core/loader/nsp.h b/src/core/loader/nsp.h index 868b028d3..36e8e3533 100644 --- a/src/core/loader/nsp.h +++ b/src/core/loader/nsp.h @@ -10,10 +10,15 @@ #include "core/loader/loader.h" namespace FileSys { +class ContentProvider; class NACP; class NSP; } // namespace FileSys +namespace Service::FileSystem { +class FileSystemController; +} + namespace Loader { class AppLoader_NCA; @@ -21,7 +26,9 @@ class AppLoader_NCA; /// Loads an XCI file class AppLoader_NSP final : public AppLoader { public: - explicit AppLoader_NSP(FileSys::VirtualFile file); + explicit AppLoader_NSP(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider); ~AppLoader_NSP() override; /** @@ -35,7 +42,7 @@ public: return IdentifyType(file); } - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadRomFS(FileSys::VirtualFile& file) override; u64 ReadRomFSIVFCOffset() const override; diff --git a/src/core/loader/xci.cpp b/src/core/loader/xci.cpp index 7186ad1ff..536e721fc 100644 --- a/src/core/loader/xci.cpp +++ b/src/core/loader/xci.cpp @@ -20,18 +20,24 @@ namespace Loader { -AppLoader_XCI::AppLoader_XCI(FileSys::VirtualFile file) +AppLoader_XCI::AppLoader_XCI(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider) : AppLoader(file), xci(std::make_unique<FileSys::XCI>(file)), nca_loader(std::make_unique<AppLoader_NCA>(xci->GetProgramNCAFile())) { - if (xci->GetStatus() != ResultStatus::Success) + if (xci->GetStatus() != ResultStatus::Success) { return; + } const auto control_nca = xci->GetNCAByType(FileSys::NCAContentType::Control); - if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) + if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) { return; + } - std::tie(nacp_file, icon_file) = - FileSys::PatchManager(xci->GetProgramTitleID()).ParseControlNCA(*control_nca); + std::tie(nacp_file, icon_file) = [this, &content_provider, &control_nca, &fsc] { + const FileSys::PatchManager pm{xci->GetProgramTitleID(), fsc, content_provider}; + return pm.ParseControlNCA(*control_nca); + }(); } AppLoader_XCI::~AppLoader_XCI() = default; @@ -49,7 +55,7 @@ FileType AppLoader_XCI::IdentifyType(const FileSys::VirtualFile& file) { return FileType::Error; } -AppLoader_XCI::LoadResult AppLoader_XCI::Load(Kernel::Process& process) { +AppLoader_XCI::LoadResult AppLoader_XCI::Load(Kernel::Process& process, Core::System& system) { if (is_loaded) { return {ResultStatus::ErrorAlreadyLoaded, {}}; } @@ -66,15 +72,14 @@ AppLoader_XCI::LoadResult AppLoader_XCI::Load(Kernel::Process& process) { return {ResultStatus::ErrorMissingProductionKeyFile, {}}; } - const auto result = nca_loader->Load(process); + const auto result = nca_loader->Load(process, system); if (result.first != ResultStatus::Success) { return result; } FileSys::VirtualFile update_raw; if (ReadUpdateRaw(update_raw) == ResultStatus::Success && update_raw != nullptr) { - Core::System::GetInstance().GetFileSystemController().SetPackedUpdate( - std::move(update_raw)); + system.GetFileSystemController().SetPackedUpdate(std::move(update_raw)); } is_loaded = true; diff --git a/src/core/loader/xci.h b/src/core/loader/xci.h index 618ae2f47..6dc1f9243 100644 --- a/src/core/loader/xci.h +++ b/src/core/loader/xci.h @@ -10,10 +10,15 @@ #include "core/loader/loader.h" namespace FileSys { +class ContentProvider; class NACP; class XCI; } // namespace FileSys +namespace Service::FileSystem { +class FileSystemController; +} + namespace Loader { class AppLoader_NCA; @@ -21,7 +26,9 @@ class AppLoader_NCA; /// Loads an XCI file class AppLoader_XCI final : public AppLoader { public: - explicit AppLoader_XCI(FileSys::VirtualFile file); + explicit AppLoader_XCI(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider); ~AppLoader_XCI() override; /** @@ -35,7 +42,7 @@ public: return IdentifyType(file); } - LoadResult Load(Kernel::Process& process) override; + LoadResult Load(Kernel::Process& process, Core::System& system) override; ResultStatus ReadRomFS(FileSys::VirtualFile& file) override; u64 ReadRomFSIVFCOffset() const override; diff --git a/src/core/memory.cpp b/src/core/memory.cpp index 6061d37ae..b88aa5c40 100644 --- a/src/core/memory.cpp +++ b/src/core/memory.cpp @@ -8,19 +8,21 @@ #include <utility> #include "common/assert.h" +#include "common/atomic_ops.h" #include "common/common_types.h" #include "common/logging/log.h" #include "common/page_table.h" #include "common/swap.h" #include "core/arm/arm_interface.h" #include "core/core.h" +#include "core/device_memory.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/physical_memory.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/vm_manager.h" #include "core/memory.h" #include "video_core/gpu.h" -namespace Memory { +namespace Core::Memory { // Implementation class used to keep the specifics of the memory subsystem hidden // from outside classes. This also allows modification to the internals of the memory @@ -28,23 +30,15 @@ namespace Memory { struct Memory::Impl { explicit Impl(Core::System& system_) : system{system_} {} - void SetCurrentPageTable(Kernel::Process& process) { - current_page_table = &process.VMManager().page_table; + void SetCurrentPageTable(Kernel::Process& process, u32 core_id) { + current_page_table = &process.PageTable().PageTableImpl(); - const std::size_t address_space_width = process.VMManager().GetAddressSpaceWidth(); + const std::size_t address_space_width = process.PageTable().GetAddressSpaceWidth(); - system.ArmInterface(0).PageTableChanged(*current_page_table, address_space_width); - system.ArmInterface(1).PageTableChanged(*current_page_table, address_space_width); - system.ArmInterface(2).PageTableChanged(*current_page_table, address_space_width); - system.ArmInterface(3).PageTableChanged(*current_page_table, address_space_width); + system.ArmInterface(core_id).PageTableChanged(*current_page_table, address_space_width); } - void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, - Kernel::PhysicalMemory& memory, VAddr offset) { - MapMemoryRegion(page_table, base, size, memory.data() + offset); - } - - void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, u8* target) { + void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, PAddr target) { ASSERT_MSG((size & PAGE_MASK) == 0, "non-page aligned size: {:016X}", size); ASSERT_MSG((base & PAGE_MASK) == 0, "non-page aligned base: {:016X}", base); MapPages(page_table, base / PAGE_SIZE, size / PAGE_SIZE, target, Common::PageType::Memory); @@ -52,46 +46,27 @@ struct Memory::Impl { void MapIoRegion(Common::PageTable& page_table, VAddr base, u64 size, Common::MemoryHookPointer mmio_handler) { - ASSERT_MSG((size & PAGE_MASK) == 0, "non-page aligned size: {:016X}", size); - ASSERT_MSG((base & PAGE_MASK) == 0, "non-page aligned base: {:016X}", base); - MapPages(page_table, base / PAGE_SIZE, size / PAGE_SIZE, nullptr, - Common::PageType::Special); - - const auto interval = boost::icl::discrete_interval<VAddr>::closed(base, base + size - 1); - const Common::SpecialRegion region{Common::SpecialRegion::Type::IODevice, - std::move(mmio_handler)}; - page_table.special_regions.add( - std::make_pair(interval, std::set<Common::SpecialRegion>{region})); + UNIMPLEMENTED(); } void UnmapRegion(Common::PageTable& page_table, VAddr base, u64 size) { ASSERT_MSG((size & PAGE_MASK) == 0, "non-page aligned size: {:016X}", size); ASSERT_MSG((base & PAGE_MASK) == 0, "non-page aligned base: {:016X}", base); - MapPages(page_table, base / PAGE_SIZE, size / PAGE_SIZE, nullptr, - Common::PageType::Unmapped); - - const auto interval = boost::icl::discrete_interval<VAddr>::closed(base, base + size - 1); - page_table.special_regions.erase(interval); + MapPages(page_table, base / PAGE_SIZE, size / PAGE_SIZE, 0, Common::PageType::Unmapped); } void AddDebugHook(Common::PageTable& page_table, VAddr base, u64 size, Common::MemoryHookPointer hook) { - const auto interval = boost::icl::discrete_interval<VAddr>::closed(base, base + size - 1); - const Common::SpecialRegion region{Common::SpecialRegion::Type::DebugHook, std::move(hook)}; - page_table.special_regions.add( - std::make_pair(interval, std::set<Common::SpecialRegion>{region})); + UNIMPLEMENTED(); } void RemoveDebugHook(Common::PageTable& page_table, VAddr base, u64 size, Common::MemoryHookPointer hook) { - const auto interval = boost::icl::discrete_interval<VAddr>::closed(base, base + size - 1); - const Common::SpecialRegion region{Common::SpecialRegion::Type::DebugHook, std::move(hook)}; - page_table.special_regions.subtract( - std::make_pair(interval, std::set<Common::SpecialRegion>{region})); + UNIMPLEMENTED(); } bool IsValidVirtualAddress(const Kernel::Process& process, const VAddr vaddr) const { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); const u8* const page_pointer = page_table.pointers[vaddr >> PAGE_BITS]; if (page_pointer != nullptr) { @@ -113,55 +88,28 @@ struct Memory::Impl { return IsValidVirtualAddress(*system.CurrentProcess(), vaddr); } - /** - * Gets a pointer to the exact memory at the virtual address (i.e. not page aligned) - * using a VMA from the current process - */ - u8* GetPointerFromVMA(const Kernel::Process& process, VAddr vaddr) { - const auto& vm_manager = process.VMManager(); - - const auto it = vm_manager.FindVMA(vaddr); - DEBUG_ASSERT(vm_manager.IsValidHandle(it)); + u8* GetPointerFromRasterizerCachedMemory(VAddr vaddr) const { + const PAddr paddr{current_page_table->backing_addr[vaddr >> PAGE_BITS]}; - u8* direct_pointer = nullptr; - const auto& vma = it->second; - switch (vma.type) { - case Kernel::VMAType::AllocatedMemoryBlock: - direct_pointer = vma.backing_block->data() + vma.offset; - break; - case Kernel::VMAType::BackingMemory: - direct_pointer = vma.backing_memory; - break; - case Kernel::VMAType::Free: - return nullptr; - default: - UNREACHABLE(); + if (!paddr) { + return {}; } - return direct_pointer + (vaddr - vma.base); + return system.DeviceMemory().GetPointer(paddr) + vaddr; } - /** - * Gets a pointer to the exact memory at the virtual address (i.e. not page aligned) - * using a VMA from the current process. - */ - u8* GetPointerFromVMA(VAddr vaddr) { - return GetPointerFromVMA(*system.CurrentProcess(), vaddr); - } - - u8* GetPointer(const VAddr vaddr) { - u8* const page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; - if (page_pointer != nullptr) { + u8* GetPointer(const VAddr vaddr) const { + u8* const page_pointer{current_page_table->pointers[vaddr >> PAGE_BITS]}; + if (page_pointer) { return page_pointer + vaddr; } if (current_page_table->attributes[vaddr >> PAGE_BITS] == Common::PageType::RasterizerCachedMemory) { - return GetPointerFromVMA(vaddr); + return GetPointerFromRasterizerCachedMemory(vaddr); } - LOG_ERROR(HW_Memory, "Unknown GetPointer @ 0x{:016X}", vaddr); - return nullptr; + return {}; } u8 Read8(const VAddr addr) { @@ -169,15 +117,33 @@ struct Memory::Impl { } u16 Read16(const VAddr addr) { - return Read<u16_le>(addr); + if ((addr & 1) == 0) { + return Read<u16_le>(addr); + } else { + const u32 a{Read<u8>(addr)}; + const u32 b{Read<u8>(addr + sizeof(u8))}; + return static_cast<u16>((b << 8) | a); + } } u32 Read32(const VAddr addr) { - return Read<u32_le>(addr); + if ((addr & 3) == 0) { + return Read<u32_le>(addr); + } else { + const u32 a{Read16(addr)}; + const u32 b{Read16(addr + sizeof(u16))}; + return (b << 16) | a; + } } u64 Read64(const VAddr addr) { - return Read<u64_le>(addr); + if ((addr & 7) == 0) { + return Read<u64_le>(addr); + } else { + const u32 a{Read32(addr)}; + const u32 b{Read32(addr + sizeof(u32))}; + return (static_cast<u64>(b) << 32) | a; + } } void Write8(const VAddr addr, const u8 data) { @@ -185,15 +151,46 @@ struct Memory::Impl { } void Write16(const VAddr addr, const u16 data) { - Write<u16_le>(addr, data); + if ((addr & 1) == 0) { + Write<u16_le>(addr, data); + } else { + Write<u8>(addr, static_cast<u8>(data)); + Write<u8>(addr + sizeof(u8), static_cast<u8>(data >> 8)); + } } void Write32(const VAddr addr, const u32 data) { - Write<u32_le>(addr, data); + if ((addr & 3) == 0) { + Write<u32_le>(addr, data); + } else { + Write16(addr, static_cast<u16>(data)); + Write16(addr + sizeof(u16), static_cast<u16>(data >> 16)); + } } void Write64(const VAddr addr, const u64 data) { - Write<u64_le>(addr, data); + if ((addr & 7) == 0) { + Write<u64_le>(addr, data); + } else { + Write32(addr, static_cast<u32>(data)); + Write32(addr + sizeof(u32), static_cast<u32>(data >> 32)); + } + } + + bool WriteExclusive8(const VAddr addr, const u8 data, const u8 expected) { + return WriteExclusive<u8>(addr, data, expected); + } + + bool WriteExclusive16(const VAddr addr, const u16 data, const u16 expected) { + return WriteExclusive<u16_le>(addr, data, expected); + } + + bool WriteExclusive32(const VAddr addr, const u32 data, const u32 expected) { + return WriteExclusive<u32_le>(addr, data, expected); + } + + bool WriteExclusive64(const VAddr addr, const u64 data, const u64 expected) { + return WriteExclusive<u64_le>(addr, data, expected); } std::string ReadCString(VAddr vaddr, std::size_t max_length) { @@ -213,7 +210,7 @@ struct Memory::Impl { void ReadBlock(const Kernel::Process& process, const VAddr src_addr, void* dest_buffer, const std::size_t size) { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); std::size_t remaining_size = size; std::size_t page_index = src_addr >> PAGE_BITS; @@ -241,7 +238,7 @@ struct Memory::Impl { break; } case Common::PageType::RasterizerCachedMemory: { - const u8* const host_ptr = GetPointerFromVMA(process, current_vaddr); + const u8* const host_ptr{GetPointerFromRasterizerCachedMemory(current_vaddr)}; system.GPU().FlushRegion(current_vaddr, copy_amount); std::memcpy(dest_buffer, host_ptr, copy_amount); break; @@ -259,7 +256,7 @@ struct Memory::Impl { void ReadBlockUnsafe(const Kernel::Process& process, const VAddr src_addr, void* dest_buffer, const std::size_t size) { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); std::size_t remaining_size = size; std::size_t page_index = src_addr >> PAGE_BITS; @@ -287,7 +284,7 @@ struct Memory::Impl { break; } case Common::PageType::RasterizerCachedMemory: { - const u8* const host_ptr = GetPointerFromVMA(process, current_vaddr); + const u8* const host_ptr{GetPointerFromRasterizerCachedMemory(current_vaddr)}; std::memcpy(dest_buffer, host_ptr, copy_amount); break; } @@ -312,7 +309,7 @@ struct Memory::Impl { void WriteBlock(const Kernel::Process& process, const VAddr dest_addr, const void* src_buffer, const std::size_t size) { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); std::size_t remaining_size = size; std::size_t page_index = dest_addr >> PAGE_BITS; std::size_t page_offset = dest_addr & PAGE_MASK; @@ -338,7 +335,7 @@ struct Memory::Impl { break; } case Common::PageType::RasterizerCachedMemory: { - u8* const host_ptr = GetPointerFromVMA(process, current_vaddr); + u8* const host_ptr{GetPointerFromRasterizerCachedMemory(current_vaddr)}; system.GPU().InvalidateRegion(current_vaddr, copy_amount); std::memcpy(host_ptr, src_buffer, copy_amount); break; @@ -356,7 +353,7 @@ struct Memory::Impl { void WriteBlockUnsafe(const Kernel::Process& process, const VAddr dest_addr, const void* src_buffer, const std::size_t size) { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); std::size_t remaining_size = size; std::size_t page_index = dest_addr >> PAGE_BITS; std::size_t page_offset = dest_addr & PAGE_MASK; @@ -382,7 +379,7 @@ struct Memory::Impl { break; } case Common::PageType::RasterizerCachedMemory: { - u8* const host_ptr = GetPointerFromVMA(process, current_vaddr); + u8* const host_ptr{GetPointerFromRasterizerCachedMemory(current_vaddr)}; std::memcpy(host_ptr, src_buffer, copy_amount); break; } @@ -406,7 +403,7 @@ struct Memory::Impl { } void ZeroBlock(const Kernel::Process& process, const VAddr dest_addr, const std::size_t size) { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); std::size_t remaining_size = size; std::size_t page_index = dest_addr >> PAGE_BITS; std::size_t page_offset = dest_addr & PAGE_MASK; @@ -432,7 +429,7 @@ struct Memory::Impl { break; } case Common::PageType::RasterizerCachedMemory: { - u8* const host_ptr = GetPointerFromVMA(process, current_vaddr); + u8* const host_ptr{GetPointerFromRasterizerCachedMemory(current_vaddr)}; system.GPU().InvalidateRegion(current_vaddr, copy_amount); std::memset(host_ptr, 0, copy_amount); break; @@ -453,7 +450,7 @@ struct Memory::Impl { void CopyBlock(const Kernel::Process& process, VAddr dest_addr, VAddr src_addr, const std::size_t size) { - const auto& page_table = process.VMManager().page_table; + const auto& page_table = process.PageTable().PageTableImpl(); std::size_t remaining_size = size; std::size_t page_index = src_addr >> PAGE_BITS; std::size_t page_offset = src_addr & PAGE_MASK; @@ -479,7 +476,7 @@ struct Memory::Impl { break; } case Common::PageType::RasterizerCachedMemory: { - const u8* const host_ptr = GetPointerFromVMA(process, current_vaddr); + const u8* const host_ptr{GetPointerFromRasterizerCachedMemory(current_vaddr)}; system.GPU().FlushRegion(current_vaddr, copy_amount); WriteBlock(process, dest_addr, host_ptr, copy_amount); break; @@ -512,7 +509,7 @@ struct Memory::Impl { u64 num_pages = ((vaddr + size - 1) >> PAGE_BITS) - (vaddr >> PAGE_BITS) + 1; for (unsigned i = 0; i < num_pages; ++i, vaddr += PAGE_SIZE) { - Common::PageType& page_type = current_page_table->attributes[vaddr >> PAGE_BITS]; + Common::PageType& page_type{current_page_table->attributes[vaddr >> PAGE_BITS]}; if (cached) { // Switch page type to cached if now cached @@ -544,16 +541,16 @@ struct Memory::Impl { // that this area is already unmarked as cached. break; case Common::PageType::RasterizerCachedMemory: { - u8* pointer = GetPointerFromVMA(vaddr & ~PAGE_MASK); + u8* pointer{GetPointerFromRasterizerCachedMemory(vaddr & ~PAGE_MASK)}; if (pointer == nullptr) { // It's possible that this function has been called while updating the // pagetable after unmapping a VMA. In that case the underlying VMA will no // longer exist, and we should just leave the pagetable entry blank. page_type = Common::PageType::Unmapped; } else { - page_type = Common::PageType::Memory; current_page_table->pointers[vaddr >> PAGE_BITS] = pointer - (vaddr & ~PAGE_MASK); + page_type = Common::PageType::Memory; } break; } @@ -570,12 +567,12 @@ struct Memory::Impl { * @param page_table The page table to use to perform the mapping. * @param base The base address to begin mapping at. * @param size The total size of the range in bytes. - * @param memory The memory to map. + * @param target The target address to begin mapping from. * @param type The page type to map the memory as. */ - void MapPages(Common::PageTable& page_table, VAddr base, u64 size, u8* memory, + void MapPages(Common::PageTable& page_table, VAddr base, u64 size, PAddr target, Common::PageType type) { - LOG_DEBUG(HW_Memory, "Mapping {} onto {:016X}-{:016X}", fmt::ptr(memory), base * PAGE_SIZE, + LOG_DEBUG(HW_Memory, "Mapping {:016X} onto {:016X}-{:016X}", target, base * PAGE_SIZE, (base + size) * PAGE_SIZE); // During boot, current_page_table might not be set yet, in which case we need not flush @@ -593,19 +590,29 @@ struct Memory::Impl { ASSERT_MSG(end <= page_table.pointers.size(), "out of range mapping at {:016X}", base + page_table.pointers.size()); - std::fill(page_table.attributes.begin() + base, page_table.attributes.begin() + end, type); + if (!target) { + ASSERT_MSG(type != Common::PageType::Memory, + "Mapping memory page without a pointer @ {:016x}", base * PAGE_SIZE); + + while (base != end) { + page_table.attributes[base] = type; + page_table.pointers[base] = nullptr; + page_table.backing_addr[base] = 0; - if (memory == nullptr) { - std::fill(page_table.pointers.begin() + base, page_table.pointers.begin() + end, - memory); + base += 1; + } } else { while (base != end) { - page_table.pointers[base] = memory - (base << PAGE_BITS); + page_table.pointers[base] = + system.DeviceMemory().GetPointer(target) - (base << PAGE_BITS); + page_table.attributes[base] = type; + page_table.backing_addr[base] = target - (base << PAGE_BITS); + ASSERT_MSG(page_table.pointers[base], "memory mapping base yield a nullptr within the table"); base += 1; - memory += PAGE_SIZE; + target += PAGE_SIZE; } } } @@ -640,7 +647,7 @@ struct Memory::Impl { ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", vaddr); break; case Common::PageType::RasterizerCachedMemory: { - const u8* const host_ptr = GetPointerFromVMA(vaddr); + const u8* const host_ptr{GetPointerFromRasterizerCachedMemory(vaddr)}; system.GPU().FlushRegion(vaddr, sizeof(T)); T value; std::memcpy(&value, host_ptr, sizeof(T)); @@ -682,7 +689,7 @@ struct Memory::Impl { ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", vaddr); break; case Common::PageType::RasterizerCachedMemory: { - u8* const host_ptr{GetPointerFromVMA(vaddr)}; + u8* const host_ptr{GetPointerFromRasterizerCachedMemory(vaddr)}; system.GPU().InvalidateRegion(vaddr, sizeof(T)); std::memcpy(host_ptr, &data, sizeof(T)); break; @@ -692,6 +699,65 @@ struct Memory::Impl { } } + template <typename T> + bool WriteExclusive(const VAddr vaddr, const T data, const T expected) { + u8* page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; + if (page_pointer != nullptr) { + // NOTE: Avoid adding any extra logic to this fast-path block + auto* pointer = reinterpret_cast<volatile T*>(&page_pointer[vaddr]); + return Common::AtomicCompareAndSwap(pointer, data, expected); + } + + const Common::PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; + switch (type) { + case Common::PageType::Unmapped: + LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X}", sizeof(data) * 8, + static_cast<u32>(data), vaddr); + return true; + case Common::PageType::Memory: + ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", vaddr); + break; + case Common::PageType::RasterizerCachedMemory: { + u8* host_ptr{GetPointerFromRasterizerCachedMemory(vaddr)}; + system.GPU().InvalidateRegion(vaddr, sizeof(T)); + auto* pointer = reinterpret_cast<volatile T*>(&host_ptr); + return Common::AtomicCompareAndSwap(pointer, data, expected); + } + default: + UNREACHABLE(); + } + return true; + } + + bool WriteExclusive128(const VAddr vaddr, const u128 data, const u128 expected) { + u8* const page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; + if (page_pointer != nullptr) { + // NOTE: Avoid adding any extra logic to this fast-path block + auto* pointer = reinterpret_cast<volatile u64*>(&page_pointer[vaddr]); + return Common::AtomicCompareAndSwap(pointer, data, expected); + } + + const Common::PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; + switch (type) { + case Common::PageType::Unmapped: + LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X}{:016X}", sizeof(data) * 8, + static_cast<u64>(data[1]), static_cast<u64>(data[0]), vaddr); + return true; + case Common::PageType::Memory: + ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", vaddr); + break; + case Common::PageType::RasterizerCachedMemory: { + u8* host_ptr{GetPointerFromRasterizerCachedMemory(vaddr)}; + system.GPU().InvalidateRegion(vaddr, sizeof(u128)); + auto* pointer = reinterpret_cast<volatile u64*>(&host_ptr); + return Common::AtomicCompareAndSwap(pointer, data, expected); + } + default: + UNREACHABLE(); + } + return true; + } + Common::PageTable* current_page_table = nullptr; Core::System& system; }; @@ -699,16 +765,11 @@ struct Memory::Impl { Memory::Memory(Core::System& system) : impl{std::make_unique<Impl>(system)} {} Memory::~Memory() = default; -void Memory::SetCurrentPageTable(Kernel::Process& process) { - impl->SetCurrentPageTable(process); +void Memory::SetCurrentPageTable(Kernel::Process& process, u32 core_id) { + impl->SetCurrentPageTable(process, core_id); } -void Memory::MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, - Kernel::PhysicalMemory& memory, VAddr offset) { - impl->MapMemoryRegion(page_table, base, size, memory, offset); -} - -void Memory::MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, u8* target) { +void Memory::MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, PAddr target) { impl->MapMemoryRegion(page_table, base, size, target); } @@ -779,6 +840,26 @@ void Memory::Write64(VAddr addr, u64 data) { impl->Write64(addr, data); } +bool Memory::WriteExclusive8(VAddr addr, u8 data, u8 expected) { + return impl->WriteExclusive8(addr, data, expected); +} + +bool Memory::WriteExclusive16(VAddr addr, u16 data, u16 expected) { + return impl->WriteExclusive16(addr, data, expected); +} + +bool Memory::WriteExclusive32(VAddr addr, u32 data, u32 expected) { + return impl->WriteExclusive32(addr, data, expected); +} + +bool Memory::WriteExclusive64(VAddr addr, u64 data, u64 expected) { + return impl->WriteExclusive64(addr, data, expected); +} + +bool Memory::WriteExclusive128(VAddr addr, u128 data, u128 expected) { + return impl->WriteExclusive128(addr, data, expected); +} + std::string Memory::ReadCString(VAddr vaddr, std::size_t max_length) { return impl->ReadCString(vaddr, max_length); } @@ -845,4 +926,4 @@ bool IsKernelVirtualAddress(const VAddr vaddr) { return KERNEL_REGION_VADDR <= vaddr && vaddr < KERNEL_REGION_END; } -} // namespace Memory +} // namespace Core::Memory diff --git a/src/core/memory.h b/src/core/memory.h index b92d678a4..4a1cc63f4 100644 --- a/src/core/memory.h +++ b/src/core/memory.h @@ -23,7 +23,7 @@ class PhysicalMemory; class Process; } // namespace Kernel -namespace Memory { +namespace Core::Memory { /** * Page size used by the ARM architecture. This is the smallest granularity with which memory can @@ -64,20 +64,7 @@ public: * * @param process The process to use the page table of. */ - void SetCurrentPageTable(Kernel::Process& process); - - /** - * Maps an physical buffer onto a region of the emulated process address space. - * - * @param page_table The page table of the emulated process. - * @param base The address to start mapping at. Must be page-aligned. - * @param size The amount of bytes to map. Must be page-aligned. - * @param memory Physical buffer with the memory backing the mapping. Must be of length - * at least `size + offset`. - * @param offset The offset within the physical memory. Must be page-aligned. - */ - void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, - Kernel::PhysicalMemory& memory, VAddr offset); + void SetCurrentPageTable(Kernel::Process& process, u32 core_id); /** * Maps an allocated buffer onto a region of the emulated process address space. @@ -88,7 +75,7 @@ public: * @param target Buffer with the memory backing the mapping. Must be of length at least * `size`. */ - void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, u8* target); + void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, PAddr target); /** * Maps a region of the emulated process address space as a IO region. @@ -258,6 +245,71 @@ public: void Write64(VAddr addr, u64 data); /** + * Writes a 8-bit unsigned integer to the given virtual address in + * the current process' address space if and only if the address contains + * the expected value. This operation is atomic. + * + * @param addr The virtual address to write the 8-bit unsigned integer to. + * @param data The 8-bit unsigned integer to write to the given virtual address. + * @param expected The 8-bit unsigned integer to check against the given virtual address. + * + * @post The memory range [addr, sizeof(data)) contains the given data value. + */ + bool WriteExclusive8(VAddr addr, u8 data, u8 expected); + + /** + * Writes a 16-bit unsigned integer to the given virtual address in + * the current process' address space if and only if the address contains + * the expected value. This operation is atomic. + * + * @param addr The virtual address to write the 16-bit unsigned integer to. + * @param data The 16-bit unsigned integer to write to the given virtual address. + * @param expected The 16-bit unsigned integer to check against the given virtual address. + * + * @post The memory range [addr, sizeof(data)) contains the given data value. + */ + bool WriteExclusive16(VAddr addr, u16 data, u16 expected); + + /** + * Writes a 32-bit unsigned integer to the given virtual address in + * the current process' address space if and only if the address contains + * the expected value. This operation is atomic. + * + * @param addr The virtual address to write the 32-bit unsigned integer to. + * @param data The 32-bit unsigned integer to write to the given virtual address. + * @param expected The 32-bit unsigned integer to check against the given virtual address. + * + * @post The memory range [addr, sizeof(data)) contains the given data value. + */ + bool WriteExclusive32(VAddr addr, u32 data, u32 expected); + + /** + * Writes a 64-bit unsigned integer to the given virtual address in + * the current process' address space if and only if the address contains + * the expected value. This operation is atomic. + * + * @param addr The virtual address to write the 64-bit unsigned integer to. + * @param data The 64-bit unsigned integer to write to the given virtual address. + * @param expected The 64-bit unsigned integer to check against the given virtual address. + * + * @post The memory range [addr, sizeof(data)) contains the given data value. + */ + bool WriteExclusive64(VAddr addr, u64 data, u64 expected); + + /** + * Writes a 128-bit unsigned integer to the given virtual address in + * the current process' address space if and only if the address contains + * the expected value. This operation is atomic. + * + * @param addr The virtual address to write the 128-bit unsigned integer to. + * @param data The 128-bit unsigned integer to write to the given virtual address. + * @param expected The 128-bit unsigned integer to check against the given virtual address. + * + * @post The memory range [addr, sizeof(data)) contains the given data value. + */ + bool WriteExclusive128(VAddr addr, u128 data, u128 expected); + + /** * Reads a null-terminated string from the given virtual address. * This function will continually read characters until either: * @@ -503,4 +555,4 @@ private: /// Determines if the given VAddr is a kernel address bool IsKernelVirtualAddress(VAddr vaddr); -} // namespace Memory +} // namespace Core::Memory diff --git a/src/core/memory/cheat_engine.cpp b/src/core/memory/cheat_engine.cpp index 4472500d2..2dd0eb0f8 100644 --- a/src/core/memory/cheat_engine.cpp +++ b/src/core/memory/cheat_engine.cpp @@ -10,17 +10,33 @@ #include "core/core_timing.h" #include "core/core_timing_util.h" #include "core/hardware_properties.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" #include "core/hle/service/hid/controllers/npad.h" #include "core/hle/service/hid/hid.h" #include "core/hle/service/sm/sm.h" +#include "core/memory.h" #include "core/memory/cheat_engine.h" -namespace Memory { - -constexpr s64 CHEAT_ENGINE_TICKS = static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 12); +namespace Core::Memory { +namespace { +constexpr auto CHEAT_ENGINE_NS = std::chrono::nanoseconds{1000000000 / 12}; constexpr u32 KEYPAD_BITMASK = 0x3FFFFFF; +std::string_view ExtractName(std::string_view data, std::size_t start_index, char match) { + auto end_index = start_index; + while (data[end_index] != match) { + ++end_index; + if (end_index > data.size() || + (end_index - start_index - 1) > sizeof(CheatDefinition::readable_name)) { + return {}; + } + } + + return data.substr(start_index, end_index - start_index); +} +} // Anonymous namespace + StandardVmCallbacks::StandardVmCallbacks(Core::System& system, const CheatProcessMetadata& metadata) : metadata(metadata), system(system) {} @@ -40,7 +56,7 @@ u64 StandardVmCallbacks::HidKeysDown() { if (applet_resource == nullptr) { LOG_WARNING(CheatEngine, "Attempted to read input state, but applet resource is not initialized!"); - return false; + return 0; } const auto press_state = @@ -80,26 +96,9 @@ CheatParser::~CheatParser() = default; TextCheatParser::~TextCheatParser() = default; -namespace { -template <char match> -std::string_view ExtractName(std::string_view data, std::size_t start_index) { - auto end_index = start_index; - while (data[end_index] != match) { - ++end_index; - if (end_index > data.size() || - (end_index - start_index - 1) > sizeof(CheatDefinition::readable_name)) { - return {}; - } - } - - return data.substr(start_index, end_index - start_index); -} -} // Anonymous namespace - -std::vector<CheatEntry> TextCheatParser::Parse(const Core::System& system, - std::string_view data) const { +std::vector<CheatEntry> TextCheatParser::Parse(std::string_view data) const { std::vector<CheatEntry> out(1); - std::optional<u64> current_entry = std::nullopt; + std::optional<u64> current_entry; for (std::size_t i = 0; i < data.size(); ++i) { if (::isspace(data[i])) { @@ -113,7 +112,7 @@ std::vector<CheatEntry> TextCheatParser::Parse(const Core::System& system, return {}; } - const auto name = ExtractName<'}'>(data, i + 1); + const auto name = ExtractName(data, i + 1, '}'); if (name.empty()) { return {}; } @@ -130,7 +129,7 @@ std::vector<CheatEntry> TextCheatParser::Parse(const Core::System& system, current_entry = out.size(); out.emplace_back(); - const auto name = ExtractName<']'>(data, i + 1); + const auto name = ExtractName(data, i + 1, ']'); if (name.empty()) { return {}; } @@ -154,8 +153,9 @@ std::vector<CheatEntry> TextCheatParser::Parse(const Core::System& system, return {}; } + const auto value = static_cast<u32>(std::stoul(hex, nullptr, 0x10)); out[*current_entry].definition.opcodes[out[*current_entry].definition.num_opcodes++] = - std::stoul(hex, nullptr, 0x10); + value; i += 8; } else { @@ -188,23 +188,38 @@ CheatEngine::~CheatEngine() { void CheatEngine::Initialize() { event = Core::Timing::CreateEvent( "CheatEngine::FrameCallback::" + Common::HexToString(metadata.main_nso_build_id), - [this](u64 userdata, s64 cycles_late) { FrameCallback(userdata, cycles_late); }); - core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS, event); + [this](std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + FrameCallback(user_data, ns_late); + }); + core_timing.ScheduleEvent(CHEAT_ENGINE_NS, event); metadata.process_id = system.CurrentProcess()->GetProcessID(); metadata.title_id = system.CurrentProcess()->GetTitleID(); - const auto& vm_manager = system.CurrentProcess()->VMManager(); - metadata.heap_extents = {vm_manager.GetHeapRegionBaseAddress(), vm_manager.GetHeapRegionSize()}; - metadata.address_space_extents = {vm_manager.GetAddressSpaceBaseAddress(), - vm_manager.GetAddressSpaceSize()}; - metadata.alias_extents = {vm_manager.GetMapRegionBaseAddress(), vm_manager.GetMapRegionSize()}; + const auto& page_table = system.CurrentProcess()->PageTable(); + metadata.heap_extents = { + .base = page_table.GetHeapRegionStart(), + .size = page_table.GetHeapRegionSize(), + }; + + metadata.address_space_extents = { + .base = page_table.GetAddressSpaceStart(), + .size = page_table.GetAddressSpaceSize(), + }; + + metadata.alias_extents = { + .base = page_table.GetAliasCodeRegionStart(), + .size = page_table.GetAliasCodeRegionSize(), + }; is_pending_reload.exchange(true); } void CheatEngine::SetMainMemoryParameters(VAddr main_region_begin, u64 main_region_size) { - metadata.main_nso_extents = {main_region_begin, main_region_size}; + metadata.main_nso_extents = { + .base = main_region_begin, + .size = main_region_size, + }; } void CheatEngine::Reload(std::vector<CheatEntry> cheats) { @@ -214,7 +229,7 @@ void CheatEngine::Reload(std::vector<CheatEntry> cheats) { MICROPROFILE_DEFINE(Cheat_Engine, "Add-Ons", "Cheat Engine", MP_RGB(70, 200, 70)); -void CheatEngine::FrameCallback(u64 userdata, s64 cycles_late) { +void CheatEngine::FrameCallback(std::uintptr_t, std::chrono::nanoseconds ns_late) { if (is_pending_reload.exchange(false)) { vm.LoadProgram(cheats); } @@ -227,7 +242,7 @@ void CheatEngine::FrameCallback(u64 userdata, s64 cycles_late) { vm.Execute(metadata); - core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS - cycles_late, event); + core_timing.ScheduleEvent(CHEAT_ENGINE_NS - ns_late, event); } -} // namespace Memory +} // namespace Core::Memory diff --git a/src/core/memory/cheat_engine.h b/src/core/memory/cheat_engine.h index 3d6b2298a..a31002346 100644 --- a/src/core/memory/cheat_engine.h +++ b/src/core/memory/cheat_engine.h @@ -5,6 +5,7 @@ #pragma once #include <atomic> +#include <chrono> #include <memory> #include <vector> #include "common/common_types.h" @@ -20,7 +21,7 @@ class CoreTiming; struct EventType; } // namespace Core::Timing -namespace Memory { +namespace Core::Memory { class StandardVmCallbacks : public DmntCheatVm::Callbacks { public: @@ -46,8 +47,7 @@ class CheatParser { public: virtual ~CheatParser(); - virtual std::vector<CheatEntry> Parse(const Core::System& system, - std::string_view data) const = 0; + [[nodiscard]] virtual std::vector<CheatEntry> Parse(std::string_view data) const = 0; }; // CheatParser implementation that parses text files @@ -55,7 +55,7 @@ class TextCheatParser final : public CheatParser { public: ~TextCheatParser() override; - std::vector<CheatEntry> Parse(const Core::System& system, std::string_view data) const override; + [[nodiscard]] std::vector<CheatEntry> Parse(std::string_view data) const override; }; // Class that encapsulates a CheatList and manages its interaction with memory and CoreTiming @@ -71,7 +71,7 @@ public: void Reload(std::vector<CheatEntry> cheats); private: - void FrameCallback(u64 userdata, s64 cycles_late); + void FrameCallback(std::uintptr_t user_data, std::chrono::nanoseconds ns_late); DmntCheatVm vm; CheatProcessMetadata metadata; @@ -84,4 +84,4 @@ private: Core::System& system; }; -} // namespace Memory +} // namespace Core::Memory diff --git a/src/core/memory/dmnt_cheat_types.h b/src/core/memory/dmnt_cheat_types.h index bf68fa0fe..5e60733dc 100644 --- a/src/core/memory/dmnt_cheat_types.h +++ b/src/core/memory/dmnt_cheat_types.h @@ -26,7 +26,7 @@ #include "common/common_types.h" -namespace Memory { +namespace Core::Memory { struct MemoryRegionExtents { u64 base{}; @@ -55,4 +55,4 @@ struct CheatEntry { CheatDefinition definition{}; }; -} // namespace Memory +} // namespace Core::Memory diff --git a/src/core/memory/dmnt_cheat_vm.cpp b/src/core/memory/dmnt_cheat_vm.cpp index 4f4fa5099..48be80c12 100644 --- a/src/core/memory/dmnt_cheat_vm.cpp +++ b/src/core/memory/dmnt_cheat_vm.cpp @@ -27,7 +27,7 @@ #include "core/memory/dmnt_cheat_types.h" #include "core/memory/dmnt_cheat_vm.h" -namespace Memory { +namespace Core::Memory { DmntCheatVm::DmntCheatVm(std::unique_ptr<Callbacks> callbacks) : callbacks(std::move(callbacks)) {} @@ -55,7 +55,7 @@ void DmntCheatVm::LogOpcode(const CheatVmOpcode& opcode) { fmt::format("Cond Type: {:X}", static_cast<u32>(begin_cond->cond_type))); callbacks->CommandLog(fmt::format("Rel Addr: {:X}", begin_cond->rel_address)); callbacks->CommandLog(fmt::format("Value: {:X}", begin_cond->value.bit64)); - } else if (auto end_cond = std::get_if<EndConditionalOpcode>(&opcode.opcode)) { + } else if (std::holds_alternative<EndConditionalOpcode>(opcode.opcode)) { callbacks->CommandLog("Opcode: End Conditional"); } else if (auto ctrl_loop = std::get_if<ControlLoopOpcode>(&opcode.opcode)) { if (ctrl_loop->start_loop) { @@ -190,6 +190,15 @@ void DmntCheatVm::LogOpcode(const CheatVmOpcode& opcode) { callbacks->CommandLog( fmt::format("Act[{:02X}]: {:d}", i, save_restore_regmask->should_operate[i])); } + } else if (auto rw_static_reg = std::get_if<ReadWriteStaticRegisterOpcode>(&opcode.opcode)) { + callbacks->CommandLog("Opcode: Read/Write Static Register"); + if (rw_static_reg->static_idx < NumReadableStaticRegisters) { + callbacks->CommandLog("Op Type: ReadStaticRegister"); + } else { + callbacks->CommandLog("Op Type: WriteStaticRegister"); + } + callbacks->CommandLog(fmt::format("Reg Idx {:X}", rw_static_reg->idx)); + callbacks->CommandLog(fmt::format("Stc Idx {:X}", rw_static_reg->static_idx)); } else if (auto debug_log = std::get_if<DebugLogOpcode>(&opcode.opcode)) { callbacks->CommandLog("Opcode: Debug Log"); callbacks->CommandLog(fmt::format("Bit Width: {:X}", debug_log->bit_width)); @@ -304,30 +313,32 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { switch (opcode_type) { case CheatVmOpcodeType::StoreStatic: { - StoreStaticOpcode store_static{}; // 0TMR00AA AAAAAAAA YYYYYYYY (YYYYYYYY) // Read additional words. const u32 second_dword = GetNextDword(); - store_static.bit_width = (first_dword >> 24) & 0xF; - store_static.mem_type = static_cast<MemoryAccessType>((first_dword >> 20) & 0xF); - store_static.offset_register = ((first_dword >> 16) & 0xF); - store_static.rel_address = - (static_cast<u64>(first_dword & 0xFF) << 32ul) | static_cast<u64>(second_dword); - store_static.value = GetNextVmInt(store_static.bit_width); - opcode.opcode = store_static; + const u32 bit_width = (first_dword >> 24) & 0xF; + + opcode.opcode = StoreStaticOpcode{ + .bit_width = bit_width, + .mem_type = static_cast<MemoryAccessType>((first_dword >> 20) & 0xF), + .offset_register = (first_dword >> 16) & 0xF, + .rel_address = (static_cast<u64>(first_dword & 0xFF) << 32) | second_dword, + .value = GetNextVmInt(bit_width), + }; } break; case CheatVmOpcodeType::BeginConditionalBlock: { - BeginConditionalOpcode begin_cond{}; // 1TMC00AA AAAAAAAA YYYYYYYY (YYYYYYYY) // Read additional words. const u32 second_dword = GetNextDword(); - begin_cond.bit_width = (first_dword >> 24) & 0xF; - begin_cond.mem_type = static_cast<MemoryAccessType>((first_dword >> 20) & 0xF); - begin_cond.cond_type = static_cast<ConditionalComparisonType>((first_dword >> 16) & 0xF); - begin_cond.rel_address = - (static_cast<u64>(first_dword & 0xFF) << 32ul) | static_cast<u64>(second_dword); - begin_cond.value = GetNextVmInt(begin_cond.bit_width); - opcode.opcode = begin_cond; + const u32 bit_width = (first_dword >> 24) & 0xF; + + opcode.opcode = BeginConditionalOpcode{ + .bit_width = bit_width, + .mem_type = static_cast<MemoryAccessType>((first_dword >> 20) & 0xF), + .cond_type = static_cast<ConditionalComparisonType>((first_dword >> 16) & 0xF), + .rel_address = (static_cast<u64>(first_dword & 0xFF) << 32) | second_dword, + .value = GetNextVmInt(bit_width), + }; } break; case CheatVmOpcodeType::EndConditionalBlock: { // 20000000 @@ -335,12 +346,14 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { opcode.opcode = EndConditionalOpcode{}; } break; case CheatVmOpcodeType::ControlLoop: { - ControlLoopOpcode ctrl_loop{}; // 300R0000 VVVVVVVV // 310R0000 // Parse register, whether loop start or loop end. - ctrl_loop.start_loop = ((first_dword >> 24) & 0xF) == 0; - ctrl_loop.reg_index = ((first_dword >> 20) & 0xF); + ControlLoopOpcode ctrl_loop{ + .start_loop = ((first_dword >> 24) & 0xF) == 0, + .reg_index = (first_dword >> 20) & 0xF, + .num_iters = 0, + }; // Read number of iters if loop start. if (ctrl_loop.start_loop) { @@ -349,65 +362,65 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { opcode.opcode = ctrl_loop; } break; case CheatVmOpcodeType::LoadRegisterStatic: { - LoadRegisterStaticOpcode ldr_static{}; // 400R0000 VVVVVVVV VVVVVVVV // Read additional words. - ldr_static.reg_index = ((first_dword >> 16) & 0xF); - ldr_static.value = - (static_cast<u64>(GetNextDword()) << 32ul) | static_cast<u64>(GetNextDword()); - opcode.opcode = ldr_static; + opcode.opcode = LoadRegisterStaticOpcode{ + .reg_index = (first_dword >> 16) & 0xF, + .value = (static_cast<u64>(GetNextDword()) << 32) | GetNextDword(), + }; } break; case CheatVmOpcodeType::LoadRegisterMemory: { - LoadRegisterMemoryOpcode ldr_memory{}; // 5TMRI0AA AAAAAAAA // Read additional words. const u32 second_dword = GetNextDword(); - ldr_memory.bit_width = (first_dword >> 24) & 0xF; - ldr_memory.mem_type = static_cast<MemoryAccessType>((first_dword >> 20) & 0xF); - ldr_memory.reg_index = ((first_dword >> 16) & 0xF); - ldr_memory.load_from_reg = ((first_dword >> 12) & 0xF) != 0; - ldr_memory.rel_address = - (static_cast<u64>(first_dword & 0xFF) << 32ul) | static_cast<u64>(second_dword); - opcode.opcode = ldr_memory; + opcode.opcode = LoadRegisterMemoryOpcode{ + .bit_width = (first_dword >> 24) & 0xF, + .mem_type = static_cast<MemoryAccessType>((first_dword >> 20) & 0xF), + .reg_index = ((first_dword >> 16) & 0xF), + .load_from_reg = ((first_dword >> 12) & 0xF) != 0, + .rel_address = (static_cast<u64>(first_dword & 0xFF) << 32) | second_dword, + }; } break; case CheatVmOpcodeType::StoreStaticToAddress: { - StoreStaticToAddressOpcode str_static{}; // 6T0RIor0 VVVVVVVV VVVVVVVV // Read additional words. - str_static.bit_width = (first_dword >> 24) & 0xF; - str_static.reg_index = ((first_dword >> 16) & 0xF); - str_static.increment_reg = ((first_dword >> 12) & 0xF) != 0; - str_static.add_offset_reg = ((first_dword >> 8) & 0xF) != 0; - str_static.offset_reg_index = ((first_dword >> 4) & 0xF); - str_static.value = - (static_cast<u64>(GetNextDword()) << 32ul) | static_cast<u64>(GetNextDword()); - opcode.opcode = str_static; + opcode.opcode = StoreStaticToAddressOpcode{ + .bit_width = (first_dword >> 24) & 0xF, + .reg_index = (first_dword >> 16) & 0xF, + .increment_reg = ((first_dword >> 12) & 0xF) != 0, + .add_offset_reg = ((first_dword >> 8) & 0xF) != 0, + .offset_reg_index = (first_dword >> 4) & 0xF, + .value = (static_cast<u64>(GetNextDword()) << 32) | GetNextDword(), + }; } break; case CheatVmOpcodeType::PerformArithmeticStatic: { - PerformArithmeticStaticOpcode perform_math_static{}; // 7T0RC000 VVVVVVVV // Read additional words. - perform_math_static.bit_width = (first_dword >> 24) & 0xF; - perform_math_static.reg_index = ((first_dword >> 16) & 0xF); - perform_math_static.math_type = - static_cast<RegisterArithmeticType>((first_dword >> 12) & 0xF); - perform_math_static.value = GetNextDword(); - opcode.opcode = perform_math_static; + opcode.opcode = PerformArithmeticStaticOpcode{ + .bit_width = (first_dword >> 24) & 0xF, + .reg_index = ((first_dword >> 16) & 0xF), + .math_type = static_cast<RegisterArithmeticType>((first_dword >> 12) & 0xF), + .value = GetNextDword(), + }; } break; case CheatVmOpcodeType::BeginKeypressConditionalBlock: { - BeginKeypressConditionalOpcode begin_keypress_cond{}; // 8kkkkkkk // Just parse the mask. - begin_keypress_cond.key_mask = first_dword & 0x0FFFFFFF; + opcode.opcode = BeginKeypressConditionalOpcode{ + .key_mask = first_dword & 0x0FFFFFFF, + }; } break; case CheatVmOpcodeType::PerformArithmeticRegister: { - PerformArithmeticRegisterOpcode perform_math_reg{}; // 9TCRSIs0 (VVVVVVVV (VVVVVVVV)) - perform_math_reg.bit_width = (first_dword >> 24) & 0xF; - perform_math_reg.math_type = static_cast<RegisterArithmeticType>((first_dword >> 20) & 0xF); - perform_math_reg.dst_reg_index = ((first_dword >> 16) & 0xF); - perform_math_reg.src_reg_1_index = ((first_dword >> 12) & 0xF); - perform_math_reg.has_immediate = ((first_dword >> 8) & 0xF) != 0; + PerformArithmeticRegisterOpcode perform_math_reg{ + .bit_width = (first_dword >> 24) & 0xF, + .math_type = static_cast<RegisterArithmeticType>((first_dword >> 20) & 0xF), + .dst_reg_index = (first_dword >> 16) & 0xF, + .src_reg_1_index = (first_dword >> 12) & 0xF, + .src_reg_2_index = 0, + .has_immediate = ((first_dword >> 8) & 0xF) != 0, + .value = {}, + }; if (perform_math_reg.has_immediate) { perform_math_reg.src_reg_2_index = 0; perform_math_reg.value = GetNextVmInt(perform_math_reg.bit_width); @@ -417,7 +430,6 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { opcode.opcode = perform_math_reg; } break; case CheatVmOpcodeType::StoreRegisterToAddress: { - StoreRegisterToAddressOpcode str_register{}; // ATSRIOxa (aaaaaaaa) // A = opcode 10 // T = bit width @@ -429,20 +441,23 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { // Relative Address // x = offset register (for offset type 1), memory type (for offset type 3) // a = relative address (for offset type 2+3) - str_register.bit_width = (first_dword >> 24) & 0xF; - str_register.str_reg_index = ((first_dword >> 20) & 0xF); - str_register.addr_reg_index = ((first_dword >> 16) & 0xF); - str_register.increment_reg = ((first_dword >> 12) & 0xF) != 0; - str_register.ofs_type = static_cast<StoreRegisterOffsetType>(((first_dword >> 8) & 0xF)); - str_register.ofs_reg_index = ((first_dword >> 4) & 0xF); + StoreRegisterToAddressOpcode str_register{ + .bit_width = (first_dword >> 24) & 0xF, + .str_reg_index = (first_dword >> 20) & 0xF, + .addr_reg_index = (first_dword >> 16) & 0xF, + .increment_reg = ((first_dword >> 12) & 0xF) != 0, + .ofs_type = static_cast<StoreRegisterOffsetType>(((first_dword >> 8) & 0xF)), + .mem_type = MemoryAccessType::MainNso, + .ofs_reg_index = (first_dword >> 4) & 0xF, + .rel_address = 0, + }; switch (str_register.ofs_type) { case StoreRegisterOffsetType::None: case StoreRegisterOffsetType::Reg: // Nothing more to do break; case StoreRegisterOffsetType::Imm: - str_register.rel_address = - ((static_cast<u64>(first_dword & 0xF) << 32ul) | static_cast<u64>(GetNextDword())); + str_register.rel_address = (static_cast<u64>(first_dword & 0xF) << 32) | GetNextDword(); break; case StoreRegisterOffsetType::MemReg: str_register.mem_type = static_cast<MemoryAccessType>((first_dword >> 4) & 0xF); @@ -450,8 +465,7 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { case StoreRegisterOffsetType::MemImm: case StoreRegisterOffsetType::MemImmReg: str_register.mem_type = static_cast<MemoryAccessType>((first_dword >> 4) & 0xF); - str_register.rel_address = - ((static_cast<u64>(first_dword & 0xF) << 32ul) | static_cast<u64>(GetNextDword())); + str_register.rel_address = (static_cast<u64>(first_dword & 0xF) << 32) | GetNextDword(); break; default: str_register.ofs_type = StoreRegisterOffsetType::None; @@ -460,7 +474,6 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { opcode.opcode = str_register; } break; case CheatVmOpcodeType::BeginRegisterConditionalBlock: { - BeginRegisterConditionalOpcode begin_reg_cond{}; // C0TcSX## // C0TcS0Ma aaaaaaaa // C0TcS1Mr @@ -482,11 +495,19 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { // r = offset register. // X = other register. // V = value. - begin_reg_cond.bit_width = (first_dword >> 20) & 0xF; - begin_reg_cond.cond_type = - static_cast<ConditionalComparisonType>((first_dword >> 16) & 0xF); - begin_reg_cond.val_reg_index = ((first_dword >> 12) & 0xF); - begin_reg_cond.comp_type = static_cast<CompareRegisterValueType>((first_dword >> 8) & 0xF); + + BeginRegisterConditionalOpcode begin_reg_cond{ + .bit_width = (first_dword >> 20) & 0xF, + .cond_type = static_cast<ConditionalComparisonType>((first_dword >> 16) & 0xF), + .val_reg_index = (first_dword >> 12) & 0xF, + .comp_type = static_cast<CompareRegisterValueType>((first_dword >> 8) & 0xF), + .mem_type = MemoryAccessType::MainNso, + .addr_reg_index = 0, + .other_reg_index = 0, + .ofs_reg_index = 0, + .rel_address = 0, + .value = {}, + }; switch (begin_reg_cond.comp_type) { case CompareRegisterValueType::StaticValue: @@ -498,26 +519,25 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { case CompareRegisterValueType::MemoryRelAddr: begin_reg_cond.mem_type = static_cast<MemoryAccessType>((first_dword >> 4) & 0xF); begin_reg_cond.rel_address = - ((static_cast<u64>(first_dword & 0xF) << 32ul) | static_cast<u64>(GetNextDword())); + (static_cast<u64>(first_dword & 0xF) << 32) | GetNextDword(); break; case CompareRegisterValueType::MemoryOfsReg: begin_reg_cond.mem_type = static_cast<MemoryAccessType>((first_dword >> 4) & 0xF); begin_reg_cond.ofs_reg_index = (first_dword & 0xF); break; case CompareRegisterValueType::RegisterRelAddr: - begin_reg_cond.addr_reg_index = ((first_dword >> 4) & 0xF); + begin_reg_cond.addr_reg_index = (first_dword >> 4) & 0xF; begin_reg_cond.rel_address = - ((static_cast<u64>(first_dword & 0xF) << 32ul) | static_cast<u64>(GetNextDword())); + (static_cast<u64>(first_dword & 0xF) << 32) | GetNextDword(); break; case CompareRegisterValueType::RegisterOfsReg: - begin_reg_cond.addr_reg_index = ((first_dword >> 4) & 0xF); - begin_reg_cond.ofs_reg_index = (first_dword & 0xF); + begin_reg_cond.addr_reg_index = (first_dword >> 4) & 0xF; + begin_reg_cond.ofs_reg_index = first_dword & 0xF; break; } opcode.opcode = begin_reg_cond; } break; case CheatVmOpcodeType::SaveRestoreRegister: { - SaveRestoreRegisterOpcode save_restore_reg{}; // C10D0Sx0 // C1 = opcode 0xC1 // D = destination index. @@ -525,26 +545,37 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { // x = 3 if clearing reg, 2 if clearing saved value, 1 if saving a register, 0 if restoring // a register. // NOTE: If we add more save slots later, current encoding is backwards compatible. - save_restore_reg.dst_index = (first_dword >> 16) & 0xF; - save_restore_reg.src_index = (first_dword >> 8) & 0xF; - save_restore_reg.op_type = static_cast<SaveRestoreRegisterOpType>((first_dword >> 4) & 0xF); - opcode.opcode = save_restore_reg; + opcode.opcode = SaveRestoreRegisterOpcode{ + .dst_index = (first_dword >> 16) & 0xF, + .src_index = (first_dword >> 8) & 0xF, + .op_type = static_cast<SaveRestoreRegisterOpType>((first_dword >> 4) & 0xF), + }; } break; case CheatVmOpcodeType::SaveRestoreRegisterMask: { - SaveRestoreRegisterMaskOpcode save_restore_regmask{}; // C2x0XXXX // C2 = opcode 0xC2 // x = 3 if clearing reg, 2 if clearing saved value, 1 if saving, 0 if restoring. // X = 16-bit bitmask, bit i --> save or restore register i. - save_restore_regmask.op_type = - static_cast<SaveRestoreRegisterOpType>((first_dword >> 20) & 0xF); + SaveRestoreRegisterMaskOpcode save_restore_regmask{ + .op_type = static_cast<SaveRestoreRegisterOpType>((first_dword >> 20) & 0xF), + .should_operate = {}, + }; for (std::size_t i = 0; i < NumRegisters; i++) { - save_restore_regmask.should_operate[i] = (first_dword & (1u << i)) != 0; + save_restore_regmask.should_operate[i] = (first_dword & (1U << i)) != 0; } opcode.opcode = save_restore_regmask; } break; + case CheatVmOpcodeType::ReadWriteStaticRegister: { + // C3000XXx + // C3 = opcode 0xC3. + // XX = static register index. + // x = register index. + opcode.opcode = ReadWriteStaticRegisterOpcode{ + .static_idx = (first_dword >> 4) & 0xFF, + .idx = first_dword & 0xF, + }; + } break; case CheatVmOpcodeType::DebugLog: { - DebugLogOpcode debug_log{}; // FFFTIX## // FFFTI0Ma aaaaaaaa // FFFTI1Mr @@ -563,31 +594,36 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) { // a = relative address. // r = offset register. // X = value register. - debug_log.bit_width = (first_dword >> 16) & 0xF; - debug_log.log_id = ((first_dword >> 12) & 0xF); - debug_log.val_type = static_cast<DebugLogValueType>((first_dword >> 8) & 0xF); + DebugLogOpcode debug_log{ + .bit_width = (first_dword >> 16) & 0xF, + .log_id = (first_dword >> 12) & 0xF, + .val_type = static_cast<DebugLogValueType>((first_dword >> 8) & 0xF), + .mem_type = MemoryAccessType::MainNso, + .addr_reg_index = 0, + .val_reg_index = 0, + .ofs_reg_index = 0, + .rel_address = 0, + }; switch (debug_log.val_type) { case DebugLogValueType::RegisterValue: - debug_log.val_reg_index = ((first_dword >> 4) & 0xF); + debug_log.val_reg_index = (first_dword >> 4) & 0xF; break; case DebugLogValueType::MemoryRelAddr: debug_log.mem_type = static_cast<MemoryAccessType>((first_dword >> 4) & 0xF); - debug_log.rel_address = - ((static_cast<u64>(first_dword & 0xF) << 32ul) | static_cast<u64>(GetNextDword())); + debug_log.rel_address = (static_cast<u64>(first_dword & 0xF) << 32) | GetNextDword(); break; case DebugLogValueType::MemoryOfsReg: debug_log.mem_type = static_cast<MemoryAccessType>((first_dword >> 4) & 0xF); - debug_log.ofs_reg_index = (first_dword & 0xF); + debug_log.ofs_reg_index = first_dword & 0xF; break; case DebugLogValueType::RegisterRelAddr: - debug_log.addr_reg_index = ((first_dword >> 4) & 0xF); - debug_log.rel_address = - ((static_cast<u64>(first_dword & 0xF) << 32ul) | static_cast<u64>(GetNextDword())); + debug_log.addr_reg_index = (first_dword >> 4) & 0xF; + debug_log.rel_address = (static_cast<u64>(first_dword & 0xF) << 32) | GetNextDword(); break; case DebugLogValueType::RegisterOfsReg: - debug_log.addr_reg_index = ((first_dword >> 4) & 0xF); - debug_log.ofs_reg_index = (first_dword & 0xF); + debug_log.addr_reg_index = (first_dword >> 4) & 0xF; + debug_log.ofs_reg_index = first_dword & 0xF; break; } opcode.opcode = debug_log; @@ -666,6 +702,7 @@ void DmntCheatVm::ResetState() { registers.fill(0); saved_values.fill(0); loop_tops.fill(0); + static_registers.fill(0); instruction_ptr = 0; condition_depth = 0; decode_success = true; @@ -779,7 +816,7 @@ void DmntCheatVm::Execute(const CheatProcessMetadata& metadata) { if (!cond_met) { SkipConditionalBlock(); } - } else if (auto end_cond = std::get_if<EndConditionalOpcode>(&cur_opcode.opcode)) { + } else if (std::holds_alternative<EndConditionalOpcode>(cur_opcode.opcode)) { // Decrement the condition depth. // We will assume, graciously, that mismatched conditional block ends are a nop. if (condition_depth > 0) { @@ -1152,6 +1189,15 @@ void DmntCheatVm::Execute(const CheatProcessMetadata& metadata) { } } } + } else if (auto rw_static_reg = + std::get_if<ReadWriteStaticRegisterOpcode>(&cur_opcode.opcode)) { + if (rw_static_reg->static_idx < NumReadableStaticRegisters) { + // Load a register with a static register. + registers[rw_static_reg->idx] = static_registers[rw_static_reg->static_idx]; + } else { + // Store a register to a static register. + static_registers[rw_static_reg->static_idx] = registers[rw_static_reg->idx]; + } } else if (auto debug_log = std::get_if<DebugLogOpcode>(&cur_opcode.opcode)) { // Read value from memory. u64 log_value = 0; @@ -1209,4 +1255,4 @@ void DmntCheatVm::Execute(const CheatProcessMetadata& metadata) { } } -} // namespace Memory +} // namespace Core::Memory diff --git a/src/core/memory/dmnt_cheat_vm.h b/src/core/memory/dmnt_cheat_vm.h index c36212cf1..21b86b72c 100644 --- a/src/core/memory/dmnt_cheat_vm.h +++ b/src/core/memory/dmnt_cheat_vm.h @@ -30,7 +30,7 @@ #include "common/common_types.h" #include "core/memory/dmnt_cheat_types.h" -namespace Memory { +namespace Core::Memory { enum class CheatVmOpcodeType : u32 { StoreStatic = 0, @@ -56,6 +56,7 @@ enum class CheatVmOpcodeType : u32 { BeginRegisterConditionalBlock = 0xC0, SaveRestoreRegister = 0xC1, SaveRestoreRegisterMask = 0xC2, + ReadWriteStaticRegister = 0xC3, // This is a meta entry, and not a real opcode. // This is to facilitate multi-nybble instruction decoding. @@ -237,6 +238,11 @@ struct SaveRestoreRegisterMaskOpcode { std::array<bool, 0x10> should_operate{}; }; +struct ReadWriteStaticRegisterOpcode { + u32 static_idx{}; + u32 idx{}; +}; + struct DebugLogOpcode { u32 bit_width{}; u32 log_id{}; @@ -259,7 +265,8 @@ struct CheatVmOpcode { PerformArithmeticStaticOpcode, BeginKeypressConditionalOpcode, PerformArithmeticRegisterOpcode, StoreRegisterToAddressOpcode, BeginRegisterConditionalOpcode, SaveRestoreRegisterOpcode, - SaveRestoreRegisterMaskOpcode, DebugLogOpcode, UnrecognizedInstruction> + SaveRestoreRegisterMaskOpcode, ReadWriteStaticRegisterOpcode, DebugLogOpcode, + UnrecognizedInstruction> opcode{}; }; @@ -281,6 +288,10 @@ public: static constexpr std::size_t MaximumProgramOpcodeCount = 0x400; static constexpr std::size_t NumRegisters = 0x10; + static constexpr std::size_t NumReadableStaticRegisters = 0x80; + static constexpr std::size_t NumWritableStaticRegisters = 0x80; + static constexpr std::size_t NumStaticRegisters = + NumReadableStaticRegisters + NumWritableStaticRegisters; explicit DmntCheatVm(std::unique_ptr<Callbacks> callbacks); ~DmntCheatVm(); @@ -302,6 +313,7 @@ private: std::array<u32, MaximumProgramOpcodeCount> program{}; std::array<u64, NumRegisters> registers{}; std::array<u64, NumRegisters> saved_values{}; + std::array<u64, NumStaticRegisters> static_registers{}; std::array<std::size_t, NumRegisters> loop_tops{}; bool DecodeNextOpcode(CheatVmOpcode& out); @@ -318,4 +330,4 @@ private: MemoryAccessType mem_type, u64 rel_address); }; -}; // namespace Memory +}; // namespace Core::Memory diff --git a/src/core/network/network.cpp b/src/core/network/network.cpp new file mode 100644 index 000000000..5ef2e8511 --- /dev/null +++ b/src/core/network/network.cpp @@ -0,0 +1,654 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <cstring> +#include <limits> +#include <utility> +#include <vector> + +#ifdef _WIN32 +#define _WINSOCK_DEPRECATED_NO_WARNINGS // gethostname +#include <winsock2.h> +#elif __unix__ +#include <errno.h> +#include <fcntl.h> +#include <netdb.h> +#include <netinet/in.h> +#include <poll.h> +#include <sys/socket.h> +#include <unistd.h> +#else +#error "Unimplemented platform" +#endif + +#include "common/assert.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "core/network/network.h" +#include "core/network/sockets.h" + +namespace Network { + +namespace { + +#ifdef _WIN32 + +using socklen_t = int; + +void Initialize() { + WSADATA wsa_data; + (void)WSAStartup(MAKEWORD(2, 2), &wsa_data); +} + +void Finalize() { + WSACleanup(); +} + +constexpr IPv4Address TranslateIPv4(in_addr addr) { + auto& bytes = addr.S_un.S_un_b; + return IPv4Address{bytes.s_b1, bytes.s_b2, bytes.s_b3, bytes.s_b4}; +} + +sockaddr TranslateFromSockAddrIn(SockAddrIn input) { + sockaddr_in result; + +#ifdef __unix__ + result.sin_len = sizeof(result); +#endif + + switch (static_cast<Domain>(input.family)) { + case Domain::INET: + result.sin_family = AF_INET; + break; + default: + UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", static_cast<int>(input.family)); + result.sin_family = AF_INET; + break; + } + + result.sin_port = htons(input.portno); + + auto& ip = result.sin_addr.S_un.S_un_b; + ip.s_b1 = input.ip[0]; + ip.s_b2 = input.ip[1]; + ip.s_b3 = input.ip[2]; + ip.s_b4 = input.ip[3]; + + sockaddr addr; + std::memcpy(&addr, &result, sizeof(addr)); + return addr; +} + +LINGER MakeLinger(bool enable, u32 linger_value) { + ASSERT(linger_value <= std::numeric_limits<u_short>::max()); + + LINGER value; + value.l_onoff = enable ? 1 : 0; + value.l_linger = static_cast<u_short>(linger_value); + return value; +} + +int LastError() { + return WSAGetLastError(); +} + +bool EnableNonBlock(SOCKET fd, bool enable) { + u_long value = enable ? 1 : 0; + return ioctlsocket(fd, FIONBIO, &value) != SOCKET_ERROR; +} + +#elif __unix__ // ^ _WIN32 v __unix__ + +using SOCKET = int; +using WSAPOLLFD = pollfd; +using ULONG = u64; + +constexpr SOCKET INVALID_SOCKET = -1; +constexpr SOCKET SOCKET_ERROR = -1; + +constexpr int WSAEWOULDBLOCK = EAGAIN; +constexpr int WSAENOTCONN = ENOTCONN; + +constexpr int SD_RECEIVE = SHUT_RD; +constexpr int SD_SEND = SHUT_WR; +constexpr int SD_BOTH = SHUT_RDWR; + +void Initialize() {} + +void Finalize() {} + +constexpr IPv4Address TranslateIPv4(in_addr addr) { + const u32 bytes = addr.s_addr; + return IPv4Address{static_cast<u8>(bytes), static_cast<u8>(bytes >> 8), + static_cast<u8>(bytes >> 16), static_cast<u8>(bytes >> 24)}; +} + +sockaddr TranslateFromSockAddrIn(SockAddrIn input) { + sockaddr_in result; + + switch (static_cast<Domain>(input.family)) { + case Domain::INET: + result.sin_family = AF_INET; + break; + default: + UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", static_cast<int>(input.family)); + result.sin_family = AF_INET; + break; + } + + result.sin_port = htons(input.portno); + + result.sin_addr.s_addr = input.ip[0] | input.ip[1] << 8 | input.ip[2] << 16 | input.ip[3] << 24; + + sockaddr addr; + std::memcpy(&addr, &result, sizeof(addr)); + return addr; +} + +int WSAPoll(WSAPOLLFD* fds, ULONG nfds, int timeout) { + return poll(fds, static_cast<nfds_t>(nfds), timeout); +} + +int closesocket(SOCKET fd) { + return close(fd); +} + +linger MakeLinger(bool enable, u32 linger_value) { + linger value; + value.l_onoff = enable ? 1 : 0; + value.l_linger = linger_value; + return value; +} + +int LastError() { + return errno; +} + +bool EnableNonBlock(int fd, bool enable) { + int flags = fcntl(fd, F_GETFD); + if (flags == -1) { + return false; + } + if (enable) { + flags |= O_NONBLOCK; + } else { + flags &= ~O_NONBLOCK; + } + return fcntl(fd, F_SETFD, flags) == 0; +} + +#endif + +int TranslateDomain(Domain domain) { + switch (domain) { + case Domain::INET: + return AF_INET; + default: + UNIMPLEMENTED_MSG("Unimplemented domain={}", static_cast<int>(domain)); + return 0; + } +} + +int TranslateType(Type type) { + switch (type) { + case Type::STREAM: + return SOCK_STREAM; + case Type::DGRAM: + return SOCK_DGRAM; + default: + UNIMPLEMENTED_MSG("Unimplemented type={}", static_cast<int>(type)); + return 0; + } +} + +int TranslateProtocol(Protocol protocol) { + switch (protocol) { + case Protocol::TCP: + return IPPROTO_TCP; + case Protocol::UDP: + return IPPROTO_UDP; + default: + UNIMPLEMENTED_MSG("Unimplemented protocol={}", static_cast<int>(protocol)); + return 0; + } +} + +SockAddrIn TranslateToSockAddrIn(sockaddr input_) { + sockaddr_in input; + std::memcpy(&input, &input_, sizeof(input)); + + SockAddrIn result; + + switch (input.sin_family) { + case AF_INET: + result.family = Domain::INET; + break; + default: + UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", input.sin_family); + result.family = Domain::INET; + break; + } + + result.portno = ntohs(input.sin_port); + + result.ip = TranslateIPv4(input.sin_addr); + + return result; +} + +u16 TranslatePollEvents(u32 events) { + u32 result = 0; + + if ((events & POLL_IN) != 0) { + events &= ~POLL_IN; + result |= POLLIN; + } + if ((events & POLL_PRI) != 0) { + events &= ~POLL_PRI; +#ifdef _WIN32 + LOG_WARNING(Service, "Winsock doesn't support POLLPRI"); +#else + result |= POLL_PRI; +#endif + } + if ((events & POLL_OUT) != 0) { + events &= ~POLL_OUT; + result |= POLLOUT; + } + + UNIMPLEMENTED_IF_MSG(events != 0, "Unhandled guest events=0x{:x}", events); + + return static_cast<u16>(result); +} + +u16 TranslatePollRevents(u32 revents) { + u32 result = 0; + const auto translate = [&result, &revents](u32 host, u32 guest) { + if ((revents & host) != 0) { + revents &= ~host; + result |= guest; + } + }; + + translate(POLLIN, POLL_IN); + translate(POLLPRI, POLL_PRI); + translate(POLLOUT, POLL_OUT); + translate(POLLERR, POLL_ERR); + translate(POLLHUP, POLL_HUP); + + UNIMPLEMENTED_IF_MSG(revents != 0, "Unhandled host revents=0x{:x}", revents); + + return static_cast<u16>(result); +} + +template <typename T> +Errno SetSockOpt(SOCKET fd, int option, T value) { + const int result = + setsockopt(fd, SOL_SOCKET, option, reinterpret_cast<const char*>(&value), sizeof(value)); + if (result != SOCKET_ERROR) { + return Errno::SUCCESS; + } + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; +} + +} // Anonymous namespace + +NetworkInstance::NetworkInstance() { + Initialize(); +} + +NetworkInstance::~NetworkInstance() { + Finalize(); +} + +std::pair<IPv4Address, Errno> GetHostIPv4Address() { + std::array<char, 256> name{}; + if (gethostname(name.data(), static_cast<int>(name.size()) - 1) == SOCKET_ERROR) { + UNIMPLEMENTED_MSG("Unhandled gethostname error"); + return {IPv4Address{}, Errno::SUCCESS}; + } + + hostent* const ent = gethostbyname(name.data()); + if (!ent) { + UNIMPLEMENTED_MSG("Unhandled gethostbyname error"); + return {IPv4Address{}, Errno::SUCCESS}; + } + if (ent->h_addr_list == nullptr) { + UNIMPLEMENTED_MSG("No addr provided in hostent->h_addr_list"); + return {IPv4Address{}, Errno::SUCCESS}; + } + if (ent->h_length != sizeof(in_addr)) { + UNIMPLEMENTED_MSG("Unexpected size={} in hostent->h_length", ent->h_length); + } + + in_addr addr; + std::memcpy(&addr, ent->h_addr_list[0], sizeof(addr)); + return {TranslateIPv4(addr), Errno::SUCCESS}; +} + +std::pair<s32, Errno> Poll(std::vector<PollFD>& pollfds, s32 timeout) { + const size_t num = pollfds.size(); + + std::vector<WSAPOLLFD> host_pollfds(pollfds.size()); + std::transform(pollfds.begin(), pollfds.end(), host_pollfds.begin(), [](PollFD fd) { + WSAPOLLFD result; + result.fd = fd.socket->fd; + result.events = TranslatePollEvents(fd.events); + result.revents = 0; + return result; + }); + + const int result = WSAPoll(host_pollfds.data(), static_cast<ULONG>(num), timeout); + if (result == 0) { + ASSERT(std::all_of(host_pollfds.begin(), host_pollfds.end(), + [](WSAPOLLFD fd) { return fd.revents == 0; })); + return {0, Errno::SUCCESS}; + } + + for (size_t i = 0; i < num; ++i) { + pollfds[i].revents = TranslatePollRevents(static_cast<u32>(host_pollfds[i].revents)); + } + + if (result > 0) { + return {result, Errno::SUCCESS}; + } + + ASSERT(result == SOCKET_ERROR); + + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {-1, Errno::SUCCESS}; +} + +Socket::~Socket() { + if (fd == INVALID_SOCKET) { + return; + } + (void)closesocket(fd); + fd = INVALID_SOCKET; +} + +Socket::Socket(Socket&& rhs) noexcept : fd{std::exchange(rhs.fd, INVALID_SOCKET)} {} + +Errno Socket::Initialize(Domain domain, Type type, Protocol protocol) { + fd = socket(TranslateDomain(domain), TranslateType(type), TranslateProtocol(protocol)); + if (fd != INVALID_SOCKET) { + return Errno::SUCCESS; + } + + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; +} + +std::pair<Socket::AcceptResult, Errno> Socket::Accept() { + sockaddr addr; + socklen_t addrlen = sizeof(addr); + const SOCKET new_socket = accept(fd, &addr, &addrlen); + + if (new_socket == INVALID_SOCKET) { + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {AcceptResult{}, Errno::SUCCESS}; + } + + AcceptResult result; + result.socket = std::make_unique<Socket>(); + result.socket->fd = new_socket; + + ASSERT(addrlen == sizeof(sockaddr_in)); + result.sockaddr_in = TranslateToSockAddrIn(addr); + + return {std::move(result), Errno::SUCCESS}; +} + +Errno Socket::Connect(SockAddrIn addr_in) { + const sockaddr host_addr_in = TranslateFromSockAddrIn(addr_in); + if (connect(fd, &host_addr_in, sizeof(host_addr_in)) != SOCKET_ERROR) { + return Errno::SUCCESS; + } + + switch (const int ec = LastError()) { + case WSAEWOULDBLOCK: + LOG_DEBUG(Service, "EAGAIN generated"); + return Errno::AGAIN; + default: + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; + } +} + +std::pair<SockAddrIn, Errno> Socket::GetPeerName() { + sockaddr addr; + socklen_t addrlen = sizeof(addr); + if (getpeername(fd, &addr, &addrlen) == SOCKET_ERROR) { + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {SockAddrIn{}, Errno::SUCCESS}; + } + + ASSERT(addrlen == sizeof(sockaddr_in)); + return {TranslateToSockAddrIn(addr), Errno::SUCCESS}; +} + +std::pair<SockAddrIn, Errno> Socket::GetSockName() { + sockaddr addr; + socklen_t addrlen = sizeof(addr); + if (getsockname(fd, &addr, &addrlen) == SOCKET_ERROR) { + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {SockAddrIn{}, Errno::SUCCESS}; + } + + ASSERT(addrlen == sizeof(sockaddr_in)); + return {TranslateToSockAddrIn(addr), Errno::SUCCESS}; +} + +Errno Socket::Bind(SockAddrIn addr) { + const sockaddr addr_in = TranslateFromSockAddrIn(addr); + if (bind(fd, &addr_in, sizeof(addr_in)) != SOCKET_ERROR) { + return Errno::SUCCESS; + } + + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; +} + +Errno Socket::Listen(s32 backlog) { + if (listen(fd, backlog) != SOCKET_ERROR) { + return Errno::SUCCESS; + } + + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; +} + +Errno Socket::Shutdown(ShutdownHow how) { + int host_how = 0; + switch (how) { + case ShutdownHow::RD: + host_how = SD_RECEIVE; + break; + case ShutdownHow::WR: + host_how = SD_SEND; + break; + case ShutdownHow::RDWR: + host_how = SD_BOTH; + break; + default: + UNIMPLEMENTED_MSG("Unimplemented flag how={}", static_cast<int>(how)); + return Errno::SUCCESS; + } + if (shutdown(fd, host_how) != SOCKET_ERROR) { + return Errno::SUCCESS; + } + + switch (const int ec = LastError()) { + case WSAENOTCONN: + LOG_ERROR(Service, "ENOTCONN generated"); + return Errno::NOTCONN; + default: + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; + } +} + +std::pair<s32, Errno> Socket::Recv(int flags, std::vector<u8>& message) { + ASSERT(flags == 0); + ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max())); + + const auto result = + recv(fd, reinterpret_cast<char*>(message.data()), static_cast<int>(message.size()), 0); + if (result != SOCKET_ERROR) { + return {static_cast<s32>(result), Errno::SUCCESS}; + } + + switch (const int ec = LastError()) { + case WSAEWOULDBLOCK: + LOG_DEBUG(Service, "EAGAIN generated"); + return {-1, Errno::AGAIN}; + case WSAENOTCONN: + LOG_ERROR(Service, "ENOTCONN generated"); + return {-1, Errno::NOTCONN}; + default: + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {0, Errno::SUCCESS}; + } +} + +std::pair<s32, Errno> Socket::RecvFrom(int flags, std::vector<u8>& message, SockAddrIn* addr) { + ASSERT(flags == 0); + ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max())); + + sockaddr addr_in{}; + socklen_t addrlen = sizeof(addr_in); + socklen_t* const p_addrlen = addr ? &addrlen : nullptr; + sockaddr* const p_addr_in = addr ? &addr_in : nullptr; + + const auto result = recvfrom(fd, reinterpret_cast<char*>(message.data()), + static_cast<int>(message.size()), 0, p_addr_in, p_addrlen); + if (result != SOCKET_ERROR) { + if (addr) { + ASSERT(addrlen == sizeof(addr_in)); + *addr = TranslateToSockAddrIn(addr_in); + } + return {static_cast<s32>(result), Errno::SUCCESS}; + } + + switch (const int ec = LastError()) { + case WSAEWOULDBLOCK: + LOG_DEBUG(Service, "EAGAIN generated"); + return {-1, Errno::AGAIN}; + case WSAENOTCONN: + LOG_ERROR(Service, "ENOTCONN generated"); + return {-1, Errno::NOTCONN}; + default: + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {-1, Errno::SUCCESS}; + } +} + +std::pair<s32, Errno> Socket::Send(const std::vector<u8>& message, int flags) { + ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max())); + ASSERT(flags == 0); + + const auto result = send(fd, reinterpret_cast<const char*>(message.data()), + static_cast<int>(message.size()), 0); + if (result != SOCKET_ERROR) { + return {static_cast<s32>(result), Errno::SUCCESS}; + } + + const int ec = LastError(); + switch (ec) { + case WSAEWOULDBLOCK: + LOG_DEBUG(Service, "EAGAIN generated"); + return {-1, Errno::AGAIN}; + case WSAENOTCONN: + LOG_ERROR(Service, "ENOTCONN generated"); + return {-1, Errno::NOTCONN}; + default: + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {-1, Errno::SUCCESS}; + } +} + +std::pair<s32, Errno> Socket::SendTo(u32 flags, const std::vector<u8>& message, + const SockAddrIn* addr) { + ASSERT(flags == 0); + + const sockaddr* to = nullptr; + const int tolen = addr ? 0 : sizeof(sockaddr); + sockaddr host_addr_in; + + if (addr) { + host_addr_in = TranslateFromSockAddrIn(*addr); + to = &host_addr_in; + } + + const auto result = sendto(fd, reinterpret_cast<const char*>(message.data()), + static_cast<int>(message.size()), 0, to, tolen); + if (result != SOCKET_ERROR) { + return {static_cast<s32>(result), Errno::SUCCESS}; + } + + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return {-1, Errno::SUCCESS}; +} + +Errno Socket::Close() { + [[maybe_unused]] const int result = closesocket(fd); + ASSERT(result == 0); + fd = INVALID_SOCKET; + + return Errno::SUCCESS; +} + +Errno Socket::SetLinger(bool enable, u32 linger) { + return SetSockOpt(fd, SO_LINGER, MakeLinger(enable, linger)); +} + +Errno Socket::SetReuseAddr(bool enable) { + return SetSockOpt<u32>(fd, SO_REUSEADDR, enable ? 1 : 0); +} + +Errno Socket::SetBroadcast(bool enable) { + return SetSockOpt<u32>(fd, SO_BROADCAST, enable ? 1 : 0); +} + +Errno Socket::SetSndBuf(u32 value) { + return SetSockOpt(fd, SO_SNDBUF, value); +} + +Errno Socket::SetRcvBuf(u32 value) { + return SetSockOpt(fd, SO_RCVBUF, value); +} + +Errno Socket::SetSndTimeo(u32 value) { + return SetSockOpt(fd, SO_SNDTIMEO, value); +} + +Errno Socket::SetRcvTimeo(u32 value) { + return SetSockOpt(fd, SO_RCVTIMEO, value); +} + +Errno Socket::SetNonBlock(bool enable) { + if (EnableNonBlock(fd, enable)) { + return Errno::SUCCESS; + } + const int ec = LastError(); + UNREACHABLE_MSG("Unhandled host socket error={}", ec); + return Errno::SUCCESS; +} + +bool Socket::IsOpened() const { + return fd != INVALID_SOCKET; +} + +} // namespace Network diff --git a/src/core/network/network.h b/src/core/network/network.h new file mode 100644 index 000000000..0622e4593 --- /dev/null +++ b/src/core/network/network.h @@ -0,0 +1,87 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <utility> + +#include "common/common_types.h" + +namespace Network { + +class Socket; + +/// Error code for network functions +enum class Errno { + SUCCESS, + BADF, + INVAL, + MFILE, + NOTCONN, + AGAIN, +}; + +/// Address families +enum class Domain { + INET, ///< Address family for IPv4 +}; + +/// Socket types +enum class Type { + STREAM, + DGRAM, + RAW, + SEQPACKET, +}; + +/// Protocol values for sockets +enum class Protocol { + ICMP, + TCP, + UDP, +}; + +/// Shutdown mode +enum class ShutdownHow { + RD, + WR, + RDWR, +}; + +/// Array of IPv4 address +using IPv4Address = std::array<u8, 4>; + +/// Cross-platform sockaddr structure +struct SockAddrIn { + Domain family; + IPv4Address ip; + u16 portno; +}; + +/// Cross-platform poll fd structure +struct PollFD { + Socket* socket; + u16 events; + u16 revents; +}; + +constexpr u16 POLL_IN = 1 << 0; +constexpr u16 POLL_PRI = 1 << 1; +constexpr u16 POLL_OUT = 1 << 2; +constexpr u16 POLL_ERR = 1 << 3; +constexpr u16 POLL_HUP = 1 << 4; +constexpr u16 POLL_NVAL = 1 << 5; + +class NetworkInstance { +public: + explicit NetworkInstance(); + ~NetworkInstance(); +}; + +/// @brief Returns host's IPv4 address +/// @return Pair of an array of human ordered IPv4 address (e.g. 192.168.0.1) and an error code +std::pair<IPv4Address, Errno> GetHostIPv4Address(); + +} // namespace Network diff --git a/src/core/network/sockets.h b/src/core/network/sockets.h new file mode 100644 index 000000000..7bdff0fe4 --- /dev/null +++ b/src/core/network/sockets.h @@ -0,0 +1,94 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <utility> + +#if defined(_WIN32) +#include <winsock.h> +#elif !defined(__unix__) +#error "Platform not implemented" +#endif + +#include "common/common_types.h" +#include "core/network/network.h" + +// TODO: C++20 Replace std::vector usages with std::span + +namespace Network { + +class Socket { +public: + struct AcceptResult { + std::unique_ptr<Socket> socket; + SockAddrIn sockaddr_in; + }; + + explicit Socket() = default; + ~Socket(); + + Socket(const Socket&) = delete; + Socket& operator=(const Socket&) = delete; + + Socket(Socket&& rhs) noexcept; + + // Avoid closing sockets implicitly + Socket& operator=(Socket&&) noexcept = delete; + + Errno Initialize(Domain domain, Type type, Protocol protocol); + + Errno Close(); + + std::pair<AcceptResult, Errno> Accept(); + + Errno Connect(SockAddrIn addr_in); + + std::pair<SockAddrIn, Errno> GetPeerName(); + + std::pair<SockAddrIn, Errno> GetSockName(); + + Errno Bind(SockAddrIn addr); + + Errno Listen(s32 backlog); + + Errno Shutdown(ShutdownHow how); + + std::pair<s32, Errno> Recv(int flags, std::vector<u8>& message); + + std::pair<s32, Errno> RecvFrom(int flags, std::vector<u8>& message, SockAddrIn* addr); + + std::pair<s32, Errno> Send(const std::vector<u8>& message, int flags); + + std::pair<s32, Errno> SendTo(u32 flags, const std::vector<u8>& message, const SockAddrIn* addr); + + Errno SetLinger(bool enable, u32 linger); + + Errno SetReuseAddr(bool enable); + + Errno SetBroadcast(bool enable); + + Errno SetSndBuf(u32 value); + + Errno SetRcvBuf(u32 value); + + Errno SetSndTimeo(u32 value); + + Errno SetRcvTimeo(u32 value); + + Errno SetNonBlock(bool enable); + + bool IsOpened() const; + +#if defined(_WIN32) + SOCKET fd = INVALID_SOCKET; +#elif defined(__unix__) + int fd = -1; +#endif +}; + +std::pair<s32, Errno> Poll(std::vector<PollFD>& poll_fds, s32 timeout); + +} // namespace Network diff --git a/src/core/perf_stats.cpp b/src/core/perf_stats.cpp index f1ae9d4df..b93396a80 100644 --- a/src/core/perf_stats.cpp +++ b/src/core/perf_stats.cpp @@ -38,11 +38,11 @@ PerfStats::~PerfStats() { std::ostringstream stream; std::copy(perf_history.begin() + IgnoreFrames, perf_history.begin() + current_index, std::ostream_iterator<double>(stream, "\n")); - const std::string& path = FileUtil::GetUserPath(FileUtil::UserPath::LogDir); + const std::string& path = Common::FS::GetUserPath(Common::FS::UserPath::LogDir); // %F Date format expanded is "%Y-%m-%d" const std::string filename = fmt::format("{}/{:%F-%H-%M}_{:016X}.csv", path, *std::localtime(&t), title_id); - FileUtil::IOFile file(filename, "w"); + Common::FS::IOFile file(filename, "w"); file.WriteString(stream.str()); } @@ -74,15 +74,16 @@ void PerfStats::EndGameFrame() { game_frames += 1; } -double PerfStats::GetMeanFrametime() { +double PerfStats::GetMeanFrametime() const { std::lock_guard lock{object_mutex}; if (current_index <= IgnoreFrames) { return 0; } + const double sum = std::accumulate(perf_history.begin() + IgnoreFrames, perf_history.begin() + current_index, 0.0); - return sum / (current_index - IgnoreFrames); + return sum / static_cast<double>(current_index - IgnoreFrames); } PerfStatsResults PerfStats::GetAndResetStats(microseconds current_system_time_us) { @@ -94,12 +95,13 @@ PerfStatsResults PerfStats::GetAndResetStats(microseconds current_system_time_us const auto system_us_per_second = (current_system_time_us - reset_point_system_us) / interval; - PerfStatsResults results{}; - results.system_fps = static_cast<double>(system_frames) / interval; - results.game_fps = static_cast<double>(game_frames) / interval; - results.frametime = duration_cast<DoubleSecs>(accumulated_frametime).count() / - static_cast<double>(system_frames); - results.emulation_speed = system_us_per_second.count() / 1'000'000.0; + const PerfStatsResults results{ + .system_fps = static_cast<double>(system_frames) / interval, + .game_fps = static_cast<double>(game_frames) / interval, + .frametime = duration_cast<DoubleSecs>(accumulated_frametime).count() / + static_cast<double>(system_frames), + .emulation_speed = system_us_per_second.count() / 1'000'000.0, + }; // Reset counters reset_point = now; @@ -111,7 +113,7 @@ PerfStatsResults PerfStats::GetAndResetStats(microseconds current_system_time_us return results; } -double PerfStats::GetLastFrameTimeScale() { +double PerfStats::GetLastFrameTimeScale() const { std::lock_guard lock{object_mutex}; constexpr double FRAME_LENGTH = 1.0 / 60; @@ -119,13 +121,14 @@ double PerfStats::GetLastFrameTimeScale() { } void FrameLimiter::DoFrameLimiting(microseconds current_system_time_us) { - if (!Settings::values.use_frame_limit) { + if (!Settings::values.use_frame_limit.GetValue() || + Settings::values.use_multi_core.GetValue()) { return; } auto now = Clock::now(); - const double sleep_scale = Settings::values.frame_limit / 100.0; + const double sleep_scale = Settings::values.frame_limit.GetValue() / 100.0; // Max lag caused by slow frames. Shouldn't be more than the length of a frame at the current // speed percent or it will clamp too much and prevent this from properly limiting to that diff --git a/src/core/perf_stats.h b/src/core/perf_stats.h index d9a64f072..69256b960 100644 --- a/src/core/perf_stats.h +++ b/src/core/perf_stats.h @@ -30,7 +30,6 @@ struct PerfStatsResults { class PerfStats { public: explicit PerfStats(u64 title_id); - ~PerfStats(); using Clock = std::chrono::high_resolution_clock; @@ -42,18 +41,18 @@ public: PerfStatsResults GetAndResetStats(std::chrono::microseconds current_system_time_us); /** - * Returns the Arthimetic Mean of all frametime values stored in the performance history. + * Returns the arithmetic mean of all frametime values stored in the performance history. */ - double GetMeanFrametime(); + double GetMeanFrametime() const; /** * Gets the ratio between walltime and the emulated time of the previous system frame. This is * useful for scaling inputs or outputs moving between the two time domains. */ - double GetLastFrameTimeScale(); + double GetLastFrameTimeScale() const; private: - std::mutex object_mutex{}; + mutable std::mutex object_mutex; /// Title ID for the game that is running. 0 if there is no game running yet u64 title_id{0}; @@ -61,7 +60,7 @@ private: std::size_t current_index{0}; /// Stores an hour of historical frametime data useful for processing and tracking performance /// regressions with code changes. - std::array<double, 216000> perf_history = {}; + std::array<double, 216000> perf_history{}; /// Point when the cumulative counters were reset Clock::time_point reset_point = Clock::now(); diff --git a/src/core/reporter.cpp b/src/core/reporter.cpp index 85ac81ef7..0becdf642 100644 --- a/src/core/reporter.cpp +++ b/src/core/reporter.cpp @@ -4,11 +4,12 @@ #include <ctime> #include <fstream> +#include <iomanip> #include <fmt/chrono.h> #include <fmt/format.h> #include <fmt/ostream.h> -#include <json.hpp> +#include <nlohmann/json.hpp> #include "common/file_util.h" #include "common/hex_util.h" @@ -16,17 +17,20 @@ #include "core/arm/arm_interface.h" #include "core/core.h" #include "core/hle/kernel/hle_ipc.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" #include "core/hle/result.h" #include "core/hle/service/lm/manager.h" +#include "core/memory.h" #include "core/reporter.h" #include "core/settings.h" namespace { std::string GetPath(std::string_view type, u64 title_id, std::string_view timestamp) { - return fmt::format("{}{}/{:016X}_{}.json", FileUtil::GetUserPath(FileUtil::UserPath::LogDir), - type, title_id, timestamp); + return fmt::format("{}{}/{:016X}_{}.json", + Common::FS::GetUserPath(Common::FS::UserPath::LogDir), type, title_id, + timestamp); } std::string GetTimestamp() { @@ -37,13 +41,13 @@ std::string GetTimestamp() { using namespace nlohmann; void SaveToFile(json json, const std::string& filename) { - if (!FileUtil::CreateFullPath(filename)) { + if (!Common::FS::CreateFullPath(filename)) { LOG_ERROR(Core, "Failed to create path for '{}' to save report!", filename); return; } std::ofstream file( - FileUtil::SanitizePath(filename, FileUtil::DirectorySeparator::PlatformDefault)); + Common::FS::SanitizePath(filename, Common::FS::DirectorySeparator::PlatformDefault)); file << std::setw(4) << json << std::endl; } @@ -108,14 +112,13 @@ json GetProcessorStateData(const std::string& architecture, u64 entry_point, u64 json GetProcessorStateDataAuto(Core::System& system) { const auto* process{system.CurrentProcess()}; - const auto& vm_manager{process->VMManager()}; auto& arm{system.CurrentArmInterface()}; Core::ARM_Interface::ThreadContext64 context{}; arm.SaveContext(context); return GetProcessorStateData(process->Is64BitProcess() ? "AArch64" : "AArch32", - vm_manager.GetCodeRegionBaseAddress(), context.sp, context.pc, + process->PageTable().GetCodeRegionStart(), context.sp, context.pc, context.pstate, context.cpu_registers); } @@ -147,7 +150,8 @@ json GetFullDataAuto(const std::string& timestamp, u64 title_id, Core::System& s } template <bool read_value, typename DescriptorType> -json GetHLEBufferDescriptorData(const std::vector<DescriptorType>& buffer, Memory::Memory& memory) { +json GetHLEBufferDescriptorData(const std::vector<DescriptorType>& buffer, + Core::Memory::Memory& memory) { auto buffer_out = json::array(); for (const auto& desc : buffer) { auto entry = json{ @@ -167,7 +171,7 @@ json GetHLEBufferDescriptorData(const std::vector<DescriptorType>& buffer, Memor return buffer_out; } -json GetHLERequestContextData(Kernel::HLERequestContext& ctx, Memory::Memory& memory) { +json GetHLERequestContextData(Kernel::HLERequestContext& ctx, Core::Memory::Memory& memory) { json out; auto cmd_buf = json::array(); diff --git a/src/core/reporter.h b/src/core/reporter.h index 380941b1b..86d760cf0 100644 --- a/src/core/reporter.h +++ b/src/core/reporter.h @@ -56,6 +56,7 @@ public: enum class PlayReportType { Old, + Old2, New, System, }; diff --git a/src/core/settings.cpp b/src/core/settings.cpp index c1282cb80..aadbc3932 100644 --- a/src/core/settings.cpp +++ b/src/core/settings.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <string_view> + #include "common/file_util.h" #include "core/core.h" #include "core/gdbstub/gdbstub.h" @@ -11,58 +13,24 @@ namespace Settings { -namespace NativeButton { -const std::array<const char*, NumButtons> mapping = {{ - "button_a", - "button_b", - "button_x", - "button_y", - "button_lstick", - "button_rstick", - "button_l", - "button_r", - "button_zl", - "button_zr", - "button_plus", - "button_minus", - "button_dleft", - "button_dup", - "button_dright", - "button_ddown", - "button_lstick_left", - "button_lstick_up", - "button_lstick_right", - "button_lstick_down", - "button_rstick_left", - "button_rstick_up", - "button_rstick_right", - "button_rstick_down", - "button_sl", - "button_sr", - "button_home", - "button_screenshot", -}}; -} +Values values = {}; +static bool configuring_global = true; -namespace NativeAnalog { -const std::array<const char*, NumAnalogs> mapping = {{ - "lstick", - "rstick", -}}; -} +std::string GetTimeZoneString() { + static constexpr std::array timezones{ + "auto", "default", "CET", "CST6CDT", "Cuba", "EET", "Egypt", "Eire", + "EST", "EST5EDT", "GB", "GB-Eire", "GMT", "GMT+0", "GMT-0", "GMT0", + "Greenwich", "Hongkong", "HST", "Iceland", "Iran", "Israel", "Jamaica", "Japan", + "Kwajalein", "Libya", "MET", "MST", "MST7MDT", "Navajo", "NZ", "NZ-CHAT", + "Poland", "Portugal", "PRC", "PST8PDT", "ROC", "ROK", "Singapore", "Turkey", + "UCT", "Universal", "UTC", "W-SU", "WET", "Zulu", + }; -namespace NativeMouseButton { -const std::array<const char*, NumMouseButtons> mapping = {{ - "left", - "right", - "middle", - "forward", - "back", -}}; + const auto time_zone_index = static_cast<std::size_t>(values.time_zone_index.GetValue()); + ASSERT(time_zone_index < timezones.size()); + return timezones[time_zone_index]; } -Values values = {}; - void Apply() { GDBStub::SetServerPort(values.gdbstub_port); GDBStub::ToggleServer(values.use_gdbstub); @@ -75,38 +43,119 @@ void Apply() { Service::HID::ReloadInputDevices(); } -template <typename T> -void LogSetting(const std::string& name, const T& value) { - LOG_INFO(Config, "{}: {}", name, value); -} - void LogSettings() { + const auto log_setting = [](std::string_view name, const auto& value) { + LOG_INFO(Config, "{}: {}", name, value); + }; + LOG_INFO(Config, "yuzu Configuration:"); - LogSetting("System_UseDockedMode", Settings::values.use_docked_mode); - LogSetting("System_RngSeed", Settings::values.rng_seed.value_or(0)); - LogSetting("System_CurrentUser", Settings::values.current_user); - LogSetting("System_LanguageIndex", Settings::values.language_index); - LogSetting("System_RegionIndex", Settings::values.region_index); - LogSetting("Core_UseMultiCore", Settings::values.use_multi_core); - LogSetting("Renderer_UseResolutionFactor", Settings::values.resolution_factor); - LogSetting("Renderer_UseFrameLimit", Settings::values.use_frame_limit); - LogSetting("Renderer_FrameLimit", Settings::values.frame_limit); - LogSetting("Renderer_UseDiskShaderCache", Settings::values.use_disk_shader_cache); - LogSetting("Renderer_UseAccurateGpuEmulation", Settings::values.use_accurate_gpu_emulation); - LogSetting("Renderer_UseAsynchronousGpuEmulation", - Settings::values.use_asynchronous_gpu_emulation); - LogSetting("Renderer_UseVsync", Settings::values.use_vsync); - LogSetting("Audio_OutputEngine", Settings::values.sink_id); - LogSetting("Audio_EnableAudioStretching", Settings::values.enable_audio_stretching); - LogSetting("Audio_OutputDevice", Settings::values.audio_device_id); - LogSetting("DataStorage_UseVirtualSd", Settings::values.use_virtual_sd); - LogSetting("DataStorage_NandDir", FileUtil::GetUserPath(FileUtil::UserPath::NANDDir)); - LogSetting("DataStorage_SdmcDir", FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir)); - LogSetting("Debugging_UseGdbstub", Settings::values.use_gdbstub); - LogSetting("Debugging_GdbstubPort", Settings::values.gdbstub_port); - LogSetting("Debugging_ProgramArgs", Settings::values.program_args); - LogSetting("Services_BCATBackend", Settings::values.bcat_backend); - LogSetting("Services_BCATBoxcatLocal", Settings::values.bcat_boxcat_local); + log_setting("Controls_UseDockedMode", values.use_docked_mode.GetValue()); + log_setting("System_RngSeed", values.rng_seed.GetValue().value_or(0)); + log_setting("System_CurrentUser", values.current_user); + log_setting("System_LanguageIndex", values.language_index.GetValue()); + log_setting("System_RegionIndex", values.region_index.GetValue()); + log_setting("System_TimeZoneIndex", values.time_zone_index.GetValue()); + log_setting("Core_UseMultiCore", values.use_multi_core.GetValue()); + log_setting("CPU_Accuracy", values.cpu_accuracy); + log_setting("Renderer_UseResolutionFactor", values.resolution_factor.GetValue()); + log_setting("Renderer_UseFrameLimit", values.use_frame_limit.GetValue()); + log_setting("Renderer_FrameLimit", values.frame_limit.GetValue()); + log_setting("Renderer_UseDiskShaderCache", values.use_disk_shader_cache.GetValue()); + log_setting("Renderer_GPUAccuracyLevel", values.gpu_accuracy.GetValue()); + log_setting("Renderer_UseAsynchronousGpuEmulation", + values.use_asynchronous_gpu_emulation.GetValue()); + log_setting("Renderer_UseNvdecEmulation", values.use_nvdec_emulation.GetValue()); + log_setting("Renderer_UseVsync", values.use_vsync.GetValue()); + log_setting("Renderer_UseAssemblyShaders", values.use_assembly_shaders.GetValue()); + log_setting("Renderer_UseAsynchronousShaders", values.use_asynchronous_shaders.GetValue()); + log_setting("Renderer_AnisotropicFilteringLevel", values.max_anisotropy.GetValue()); + log_setting("Audio_OutputEngine", values.sink_id); + log_setting("Audio_EnableAudioStretching", values.enable_audio_stretching.GetValue()); + log_setting("Audio_OutputDevice", values.audio_device_id); + log_setting("DataStorage_UseVirtualSd", values.use_virtual_sd); + log_setting("DataStorage_NandDir", Common::FS::GetUserPath(Common::FS::UserPath::NANDDir)); + log_setting("DataStorage_SdmcDir", Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir)); + log_setting("Debugging_UseGdbstub", values.use_gdbstub); + log_setting("Debugging_GdbstubPort", values.gdbstub_port); + log_setting("Debugging_ProgramArgs", values.program_args); + log_setting("Services_BCATBackend", values.bcat_backend); + log_setting("Services_BCATBoxcatLocal", values.bcat_boxcat_local); +} + +bool IsConfiguringGlobal() { + return configuring_global; +} + +void SetConfiguringGlobal(bool is_global) { + configuring_global = is_global; +} + +bool IsGPULevelExtreme() { + return values.gpu_accuracy.GetValue() == GPUAccuracy::Extreme; +} + +bool IsGPULevelHigh() { + return values.gpu_accuracy.GetValue() == GPUAccuracy::Extreme || + values.gpu_accuracy.GetValue() == GPUAccuracy::High; +} + +float Volume() { + if (values.audio_muted) { + return 0.0f; + } + return values.volume.GetValue(); +} + +void RestoreGlobalState() { + // If a game is running, DO NOT restore the global settings state + if (Core::System::GetInstance().IsPoweredOn()) { + return; + } + + // Audio + values.enable_audio_stretching.SetGlobal(true); + values.volume.SetGlobal(true); + + // Core + values.use_multi_core.SetGlobal(true); + + // Renderer + values.renderer_backend.SetGlobal(true); + values.vulkan_device.SetGlobal(true); + values.aspect_ratio.SetGlobal(true); + values.max_anisotropy.SetGlobal(true); + values.use_frame_limit.SetGlobal(true); + values.frame_limit.SetGlobal(true); + values.use_disk_shader_cache.SetGlobal(true); + values.gpu_accuracy.SetGlobal(true); + values.use_asynchronous_gpu_emulation.SetGlobal(true); + values.use_nvdec_emulation.SetGlobal(true); + values.use_vsync.SetGlobal(true); + values.use_assembly_shaders.SetGlobal(true); + values.use_asynchronous_shaders.SetGlobal(true); + values.use_fast_gpu_time.SetGlobal(true); + values.bg_red.SetGlobal(true); + values.bg_green.SetGlobal(true); + values.bg_blue.SetGlobal(true); + + // System + values.language_index.SetGlobal(true); + values.region_index.SetGlobal(true); + values.time_zone_index.SetGlobal(true); + values.rng_seed.SetGlobal(true); + values.custom_rtc.SetGlobal(true); + values.sound_index.SetGlobal(true); + + // Controls + values.players.SetGlobal(true); + values.use_docked_mode.SetGlobal(true); + values.vibration_enabled.SetGlobal(true); + values.motion_enabled.SetGlobal(true); +} + +void Sanitize() { + values.use_asynchronous_gpu_emulation.SetValue( + values.use_asynchronous_gpu_emulation.GetValue() || values.use_multi_core.GetValue()); } } // namespace Settings diff --git a/src/core/settings.h b/src/core/settings.h index 79ec01731..1143aba5d 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -12,386 +12,171 @@ #include <string> #include <vector> #include "common/common_types.h" +#include "input_common/settings.h" namespace Settings { -namespace NativeButton { -enum Values { - A, - B, - X, - Y, - LStick, - RStick, - L, - R, - ZL, - ZR, - Plus, - Minus, - - DLeft, - DUp, - DRight, - DDown, - - LStick_Left, - LStick_Up, - LStick_Right, - LStick_Down, - - RStick_Left, - RStick_Up, - RStick_Right, - RStick_Down, - - SL, - SR, - - Home, - Screenshot, - - NumButtons, -}; - -constexpr int BUTTON_HID_BEGIN = A; -constexpr int BUTTON_NS_BEGIN = Home; - -constexpr int BUTTON_HID_END = BUTTON_NS_BEGIN; -constexpr int BUTTON_NS_END = NumButtons; - -constexpr int NUM_BUTTONS_HID = BUTTON_HID_END - BUTTON_HID_BEGIN; -constexpr int NUM_BUTTONS_NS = BUTTON_NS_END - BUTTON_NS_BEGIN; - -extern const std::array<const char*, NumButtons> mapping; - -} // namespace NativeButton - -namespace NativeAnalog { -enum Values { - LStick, - RStick, - - NumAnalogs, +enum class RendererBackend { + OpenGL = 0, + Vulkan = 1, }; -constexpr int STICK_HID_BEGIN = LStick; -constexpr int STICK_HID_END = NumAnalogs; -constexpr int NUM_STICKS_HID = NumAnalogs; - -extern const std::array<const char*, NumAnalogs> mapping; -} // namespace NativeAnalog - -namespace NativeMouseButton { -enum Values { - Left, - Right, - Middle, - Forward, - Back, - - NumMouseButtons, +enum class GPUAccuracy : u32 { + Normal = 0, + High = 1, + Extreme = 2, }; -constexpr int MOUSE_HID_BEGIN = Left; -constexpr int MOUSE_HID_END = NumMouseButtons; -constexpr int NUM_MOUSE_HID = NumMouseButtons; - -extern const std::array<const char*, NumMouseButtons> mapping; -} // namespace NativeMouseButton - -namespace NativeKeyboard { -enum Keys { - None, - Error, - - A = 4, - B, - C, - D, - E, - F, - G, - H, - I, - J, - K, - L, - M, - N, - O, - P, - Q, - R, - S, - T, - U, - V, - W, - X, - Y, - Z, - N1, - N2, - N3, - N4, - N5, - N6, - N7, - N8, - N9, - N0, - Enter, - Escape, - Backspace, - Tab, - Space, - Minus, - Equal, - LeftBrace, - RightBrace, - Backslash, - Tilde, - Semicolon, - Apostrophe, - Grave, - Comma, - Dot, - Slash, - CapsLockKey, - - F1, - F2, - F3, - F4, - F5, - F6, - F7, - F8, - F9, - F10, - F11, - F12, - - SystemRequest, - ScrollLockKey, - Pause, - Insert, - Home, - PageUp, - Delete, - End, - PageDown, - Right, - Left, - Down, - Up, - - NumLockKey, - KPSlash, - KPAsterisk, - KPMinus, - KPPlus, - KPEnter, - KP1, - KP2, - KP3, - KP4, - KP5, - KP6, - KP7, - KP8, - KP9, - KP0, - KPDot, - - Key102, - Compose, - Power, - KPEqual, - - F13, - F14, - F15, - F16, - F17, - F18, - F19, - F20, - F21, - F22, - F23, - F24, - - Open, - Help, - Properties, - Front, - Stop, - Repeat, - Undo, - Cut, - Copy, - Paste, - Find, - Mute, - VolumeUp, - VolumeDown, - CapsLockActive, - NumLockActive, - ScrollLockActive, - KPComma, - - KPLeftParenthesis, - KPRightParenthesis, - - LeftControlKey = 0xE0, - LeftShiftKey, - LeftAltKey, - LeftMetaKey, - RightControlKey, - RightShiftKey, - RightAltKey, - RightMetaKey, - - MediaPlayPause, - MediaStopCD, - MediaPrevious, - MediaNext, - MediaEject, - MediaVolumeUp, - MediaVolumeDown, - MediaMute, - MediaWebsite, - MediaBack, - MediaForward, - MediaStop, - MediaFind, - MediaScrollUp, - MediaScrollDown, - MediaEdit, - MediaSleep, - MediaCoffee, - MediaRefresh, - MediaCalculator, - - NumKeyboardKeys, +enum class CPUAccuracy { + Accurate = 0, + Unsafe = 1, + DebugMode = 2, }; -static_assert(NumKeyboardKeys == 0xFC, "Incorrect number of keyboard keys."); - -enum Modifiers { - LeftControl, - LeftShift, - LeftAlt, - LeftMeta, - RightControl, - RightShift, - RightAlt, - RightMeta, - CapsLock, - ScrollLock, - NumLock, - - NumKeyboardMods, +template <typename Type> +class Setting final { +public: + Setting() = default; + explicit Setting(Type val) : global{val} {} + ~Setting() = default; + void SetGlobal(bool to_global) { + use_global = to_global; + } + bool UsingGlobal() const { + return use_global; + } + Type GetValue(bool need_global = false) const { + if (use_global || need_global) { + return global; + } + return local; + } + void SetValue(const Type& value) { + if (use_global) { + global = value; + } else { + local = value; + } + } + +private: + bool use_global = true; + Type global{}; + Type local{}; }; -constexpr int KEYBOARD_KEYS_HID_BEGIN = None; -constexpr int KEYBOARD_KEYS_HID_END = NumKeyboardKeys; -constexpr int NUM_KEYBOARD_KEYS_HID = NumKeyboardKeys; - -constexpr int KEYBOARD_MODS_HID_BEGIN = LeftControl; -constexpr int KEYBOARD_MODS_HID_END = NumKeyboardMods; -constexpr int NUM_KEYBOARD_MODS_HID = NumKeyboardMods; - -} // namespace NativeKeyboard - -using ButtonsRaw = std::array<std::string, NativeButton::NumButtons>; -using AnalogsRaw = std::array<std::string, NativeAnalog::NumAnalogs>; -using MouseButtonsRaw = std::array<std::string, NativeMouseButton::NumMouseButtons>; -using KeyboardKeysRaw = std::array<std::string, NativeKeyboard::NumKeyboardKeys>; -using KeyboardModsRaw = std::array<std::string, NativeKeyboard::NumKeyboardMods>; - -constexpr u32 JOYCON_BODY_NEON_RED = 0xFF3C28; -constexpr u32 JOYCON_BUTTONS_NEON_RED = 0x1E0A0A; -constexpr u32 JOYCON_BODY_NEON_BLUE = 0x0AB9E6; -constexpr u32 JOYCON_BUTTONS_NEON_BLUE = 0x001E1E; - -enum class ControllerType { - ProController, - DualJoycon, - RightJoycon, - LeftJoycon, +/** + * The InputSetting class allows for getting a reference to either the global or local members. + * This is required as we cannot easily modify the values of user-defined types within containers + * using the SetValue() member function found in the Setting class. The primary purpose of this + * class is to store an array of 10 PlayerInput structs for both the global and local (per-game) + * setting and allows for easily accessing and modifying both settings. + */ +template <typename Type> +class InputSetting final { +public: + InputSetting() = default; + explicit InputSetting(Type val) : global{val} {} + ~InputSetting() = default; + void SetGlobal(bool to_global) { + use_global = to_global; + } + bool UsingGlobal() const { + return use_global; + } + Type& GetValue(bool need_global = false) { + if (use_global || need_global) { + return global; + } + return local; + } + +private: + bool use_global = true; + Type global{}; + Type local{}; }; -struct PlayerInput { - bool connected; - ControllerType type; - ButtonsRaw buttons; - AnalogsRaw analogs; - - u32 body_color_right; - u32 button_color_right; - u32 body_color_left; - u32 button_color_left; +struct TouchFromButtonMap { + std::string name; + std::vector<std::string> buttons; }; -struct TouchscreenInput { - bool enabled; - std::string device; +struct Values { + // Audio + std::string audio_device_id; + std::string sink_id; + bool audio_muted; + Setting<bool> enable_audio_stretching; + Setting<float> volume; - u32 finger; - u32 diameter_x; - u32 diameter_y; - u32 rotation_angle; -}; + // Core + Setting<bool> use_multi_core; -enum class NANDTotalSize : u64 { - S29_1GB = 0x747C00000ULL, -}; + // Cpu + CPUAccuracy cpu_accuracy; -enum class NANDUserSize : u64 { - S26GB = 0x680000000ULL, -}; + bool cpuopt_page_tables; + bool cpuopt_block_linking; + bool cpuopt_return_stack_buffer; + bool cpuopt_fast_dispatcher; + bool cpuopt_context_elimination; + bool cpuopt_const_prop; + bool cpuopt_misc_ir; + bool cpuopt_reduce_misalign_checks; -enum class NANDSystemSize : u64 { - S2_5GB = 0xA0000000, -}; + bool cpuopt_unsafe_unfuse_fma; + bool cpuopt_unsafe_reduce_fp_error; -enum class SDMCSize : u64 { - S1GB = 0x40000000, - S2GB = 0x80000000, - S4GB = 0x100000000ULL, - S8GB = 0x200000000ULL, - S16GB = 0x400000000ULL, - S32GB = 0x800000000ULL, - S64GB = 0x1000000000ULL, - S128GB = 0x2000000000ULL, - S256GB = 0x4000000000ULL, - S1TB = 0x10000000000ULL, -}; - -enum class RendererBackend { - OpenGL = 0, - Vulkan = 1, -}; + // Renderer + Setting<RendererBackend> renderer_backend; + bool renderer_debug; + Setting<int> vulkan_device; + + Setting<u16> resolution_factor{1}; + Setting<int> aspect_ratio; + Setting<int> max_anisotropy; + Setting<bool> use_frame_limit; + Setting<u16> frame_limit; + Setting<bool> use_disk_shader_cache; + Setting<GPUAccuracy> gpu_accuracy; + Setting<bool> use_asynchronous_gpu_emulation; + Setting<bool> use_nvdec_emulation; + Setting<bool> use_vsync; + Setting<bool> use_assembly_shaders; + Setting<bool> use_asynchronous_shaders; + Setting<bool> use_fast_gpu_time; + + Setting<float> bg_red; + Setting<float> bg_green; + Setting<float> bg_blue; -struct Values { // System - bool use_docked_mode; - std::optional<u32> rng_seed; + Setting<std::optional<u32>> rng_seed; // Measured in seconds since epoch - std::optional<std::chrono::seconds> custom_rtc; + Setting<std::optional<std::chrono::seconds>> custom_rtc; // Set on game boot, reset on stop. Seconds difference between current time and `custom_rtc` std::chrono::seconds custom_rtc_differential; s32 current_user; - s32 language_index; - s32 region_index; - s32 sound_index; + Setting<s32> language_index; + Setting<s32> region_index; + Setting<s32> time_zone_index; + Setting<s32> sound_index; // Controls - std::array<PlayerInput, 10> players; + InputSetting<std::array<PlayerInput, 10>> players; + + Setting<bool> use_docked_mode; + + Setting<bool> vibration_enabled; + Setting<bool> enable_accurate_vibrations; + + Setting<bool> motion_enabled; + std::string motion_device; + std::string udp_input_address; + u16 udp_input_port; + u8 udp_pad_index; bool mouse_enabled; std::string mouse_device; @@ -405,55 +190,20 @@ struct Values { ButtonsRaw debug_pad_buttons; AnalogsRaw debug_pad_analogs; - std::string motion_device; TouchscreenInput touchscreen; - std::atomic_bool is_device_reload_pending{true}; - std::string udp_input_address; - u16 udp_input_port; - u8 udp_pad_index; - // Core - bool use_multi_core; + bool use_touch_from_button; + std::string touch_device; + int touch_from_button_map_index; + std::vector<TouchFromButtonMap> touch_from_button_maps; + + std::atomic_bool is_device_reload_pending{true}; // Data Storage bool use_virtual_sd; bool gamecard_inserted; bool gamecard_current_game; std::string gamecard_path; - NANDTotalSize nand_total_size; - NANDSystemSize nand_system_size; - NANDUserSize nand_user_size; - SDMCSize sdmc_size; - - // Renderer - RendererBackend renderer_backend; - bool renderer_debug; - int vulkan_device; - - float resolution_factor; - int aspect_ratio; - int max_anisotropy; - bool use_frame_limit; - u16 frame_limit; - bool use_disk_shader_cache; - bool use_accurate_gpu_emulation; - bool use_asynchronous_gpu_emulation; - bool use_vsync; - bool force_30fps_mode; - - float bg_red; - float bg_green; - float bg_blue; - - std::string log_filter; - - bool use_dev_keys; - - // Audio - std::string sink_id; - bool enable_audio_stretching; - std::string audio_device_id; - float volume; // Debugging bool record_frame_times; @@ -464,8 +214,14 @@ struct Values { bool dump_nso; bool reporting_services; bool quest_flag; + bool disable_macro_jit; + bool extended_logging; + + // Misceallaneous + std::string log_filter; + bool use_dev_keys; - // BCAT + // Services std::string bcat_backend; bool bcat_boxcat_local; @@ -477,8 +233,27 @@ struct Values { // Add-Ons std::map<u64, std::vector<std::string>> disabled_addons; -} extern values; +}; + +extern Values values; + +bool IsConfiguringGlobal(); +void SetConfiguringGlobal(bool is_global); + +bool IsGPULevelExtreme(); +bool IsGPULevelHigh(); + +float Volume(); + +std::string GetTimeZoneString(); void Apply(); void LogSettings(); + +// Restore the global state of all applicable settings in the Values struct +void RestoreGlobalState(); + +// Fixes settings that are known to cause issues with the emulator +void Sanitize(); + } // namespace Settings diff --git a/src/core/telemetry_session.cpp b/src/core/telemetry_session.cpp index 0f3685d1c..d11b15f38 100644 --- a/src/core/telemetry_session.cpp +++ b/src/core/telemetry_session.cpp @@ -25,6 +25,8 @@ namespace Core { +namespace Telemetry = Common::Telemetry; + static u64 GenerateTelemetryId() { u64 telemetry_id{}; @@ -56,14 +58,26 @@ static const char* TranslateRenderer(Settings::RendererBackend backend) { return "Unknown"; } +static const char* TranslateGPUAccuracyLevel(Settings::GPUAccuracy backend) { + switch (backend) { + case Settings::GPUAccuracy::Normal: + return "Normal"; + case Settings::GPUAccuracy::High: + return "High"; + case Settings::GPUAccuracy::Extreme: + return "Extreme"; + } + return "Unknown"; +} + u64 GetTelemetryId() { u64 telemetry_id{}; - const std::string filename{FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + + const std::string filename{Common::FS::GetUserPath(Common::FS::UserPath::ConfigDir) + "telemetry_id"}; - bool generate_new_id = !FileUtil::Exists(filename); + bool generate_new_id = !Common::FS::Exists(filename); if (!generate_new_id) { - FileUtil::IOFile file(filename, "rb"); + Common::FS::IOFile file(filename, "rb"); if (!file.IsOpen()) { LOG_ERROR(Core, "failed to open telemetry_id: {}", filename); return {}; @@ -76,7 +90,7 @@ u64 GetTelemetryId() { } if (generate_new_id) { - FileUtil::IOFile file(filename, "wb"); + Common::FS::IOFile file(filename, "wb"); if (!file.IsOpen()) { LOG_ERROR(Core, "failed to open telemetry_id: {}", filename); return {}; @@ -90,10 +104,10 @@ u64 GetTelemetryId() { u64 RegenerateTelemetryId() { const u64 new_telemetry_id{GenerateTelemetryId()}; - const std::string filename{FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + + const std::string filename{Common::FS::GetUserPath(Common::FS::UserPath::ConfigDir) + "telemetry_id"}; - FileUtil::IOFile file(filename, "wb"); + Common::FS::IOFile file(filename, "wb"); if (!file.IsOpen()) { LOG_ERROR(Core, "failed to open telemetry_id: {}", filename); return {}; @@ -133,7 +147,9 @@ TelemetrySession::~TelemetrySession() { } } -void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) { +void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider) { // Log one-time top-level information AddField(Telemetry::FieldType::None, "TelemetryId", GetTelemetryId()); @@ -153,9 +169,12 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) { app_loader.ReadTitle(name); if (name.empty()) { - auto [nacp, icon_file] = FileSys::PatchManager(program_id).GetControlMetadata(); - if (nacp != nullptr) { - name = nacp->GetApplicationName(); + const auto metadata = [&content_provider, &fsc, program_id] { + const FileSys::PatchManager pm{program_id, fsc, content_provider}; + return pm.GetControlMetadata(); + }(); + if (metadata.first != nullptr) { + name = metadata.first->GetApplicationName(); } } @@ -177,19 +196,29 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) { // Log user configuration information constexpr auto field_type = Telemetry::FieldType::UserConfig; AddField(field_type, "Audio_SinkId", Settings::values.sink_id); - AddField(field_type, "Audio_EnableAudioStretching", Settings::values.enable_audio_stretching); - AddField(field_type, "Core_UseMultiCore", Settings::values.use_multi_core); - AddField(field_type, "Renderer_Backend", TranslateRenderer(Settings::values.renderer_backend)); - AddField(field_type, "Renderer_ResolutionFactor", Settings::values.resolution_factor); - AddField(field_type, "Renderer_UseFrameLimit", Settings::values.use_frame_limit); - AddField(field_type, "Renderer_FrameLimit", Settings::values.frame_limit); - AddField(field_type, "Renderer_UseDiskShaderCache", Settings::values.use_disk_shader_cache); - AddField(field_type, "Renderer_UseAccurateGpuEmulation", - Settings::values.use_accurate_gpu_emulation); + AddField(field_type, "Audio_EnableAudioStretching", + Settings::values.enable_audio_stretching.GetValue()); + AddField(field_type, "Core_UseMultiCore", Settings::values.use_multi_core.GetValue()); + AddField(field_type, "Renderer_Backend", + TranslateRenderer(Settings::values.renderer_backend.GetValue())); + AddField(field_type, "Renderer_ResolutionFactor", + Settings::values.resolution_factor.GetValue()); + AddField(field_type, "Renderer_UseFrameLimit", Settings::values.use_frame_limit.GetValue()); + AddField(field_type, "Renderer_FrameLimit", Settings::values.frame_limit.GetValue()); + AddField(field_type, "Renderer_UseDiskShaderCache", + Settings::values.use_disk_shader_cache.GetValue()); + AddField(field_type, "Renderer_GPUAccuracyLevel", + TranslateGPUAccuracyLevel(Settings::values.gpu_accuracy.GetValue())); AddField(field_type, "Renderer_UseAsynchronousGpuEmulation", - Settings::values.use_asynchronous_gpu_emulation); - AddField(field_type, "Renderer_UseVsync", Settings::values.use_vsync); - AddField(field_type, "System_UseDockedMode", Settings::values.use_docked_mode); + Settings::values.use_asynchronous_gpu_emulation.GetValue()); + AddField(field_type, "Renderer_UseNvdecEmulation", + Settings::values.use_nvdec_emulation.GetValue()); + AddField(field_type, "Renderer_UseVsync", Settings::values.use_vsync.GetValue()); + AddField(field_type, "Renderer_UseAssemblyShaders", + Settings::values.use_assembly_shaders.GetValue()); + AddField(field_type, "Renderer_UseAsynchronousShaders", + Settings::values.use_asynchronous_shaders.GetValue()); + AddField(field_type, "System_UseDockedMode", Settings::values.use_docked_mode.GetValue()); } bool TelemetrySession::SubmitTestcase() { diff --git a/src/core/telemetry_session.h b/src/core/telemetry_session.h index 17ac22377..6f3d45bea 100644 --- a/src/core/telemetry_session.h +++ b/src/core/telemetry_session.h @@ -7,10 +7,18 @@ #include <string> #include "common/telemetry.h" +namespace FileSys { +class ContentProvider; +} + namespace Loader { class AppLoader; } +namespace Service::FileSystem { +class FileSystemController; +} + namespace Core { /** @@ -40,10 +48,14 @@ public: * - Title file format * - Miscellaneous settings values. * - * @param app_loader The application loader to use to retrieve - * title-specific information. + * @param app_loader The application loader to use to retrieve + * title-specific information. + * @param fsc Filesystem controller to use to retrieve info. + * @param content_provider Content provider to use to retrieve info. */ - void AddInitialInfo(Loader::AppLoader& app_loader); + void AddInitialInfo(Loader::AppLoader& app_loader, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider); /** * Wrapper around the Telemetry::FieldCollection::AddField method. @@ -52,7 +64,7 @@ public: * @param value Value for the field to add. */ template <typename T> - void AddField(Telemetry::FieldType type, const char* name, T value) { + void AddField(Common::Telemetry::FieldType type, const char* name, T value) { field_collection.AddField(type, name, std::move(value)); } @@ -63,7 +75,8 @@ public: bool SubmitTestcase(); private: - Telemetry::FieldCollection field_collection; ///< Tracks all added fields for the session + /// Tracks all added fields for the session + Common::Telemetry::FieldCollection field_collection; }; /** diff --git a/src/core/tools/freezer.cpp b/src/core/tools/freezer.cpp index 1e060f009..5c674a099 100644 --- a/src/core/tools/freezer.cpp +++ b/src/core/tools/freezer.cpp @@ -14,9 +14,9 @@ namespace Tools { namespace { -constexpr s64 MEMORY_FREEZER_TICKS = static_cast<s64>(Core::Hardware::BASE_CLOCK_RATE / 60); +constexpr auto memory_freezer_ns = std::chrono::nanoseconds{1000000000 / 60}; -u64 MemoryReadWidth(Memory::Memory& memory, u32 width, VAddr addr) { +u64 MemoryReadWidth(Core::Memory::Memory& memory, u32 width, VAddr addr) { switch (width) { case 1: return memory.Read8(addr); @@ -32,7 +32,7 @@ u64 MemoryReadWidth(Memory::Memory& memory, u32 width, VAddr addr) { } } -void MemoryWriteWidth(Memory::Memory& memory, u32 width, VAddr addr, u64 value) { +void MemoryWriteWidth(Core::Memory::Memory& memory, u32 width, VAddr addr, u64 value) { switch (width) { case 1: memory.Write8(addr, static_cast<u8>(value)); @@ -53,12 +53,14 @@ void MemoryWriteWidth(Memory::Memory& memory, u32 width, VAddr addr, u64 value) } // Anonymous namespace -Freezer::Freezer(Core::Timing::CoreTiming& core_timing_, Memory::Memory& memory_) +Freezer::Freezer(Core::Timing::CoreTiming& core_timing_, Core::Memory::Memory& memory_) : core_timing{core_timing_}, memory{memory_} { event = Core::Timing::CreateEvent( "MemoryFreezer::FrameCallback", - [this](u64 userdata, s64 cycles_late) { FrameCallback(userdata, cycles_late); }); - core_timing.ScheduleEvent(MEMORY_FREEZER_TICKS, event); + [this](std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + FrameCallback(user_data, ns_late); + }); + core_timing.ScheduleEvent(memory_freezer_ns, event); } Freezer::~Freezer() { @@ -68,7 +70,7 @@ Freezer::~Freezer() { void Freezer::SetActive(bool active) { if (!this->active.exchange(active)) { FillEntryReads(); - core_timing.ScheduleEvent(MEMORY_FREEZER_TICKS, event); + core_timing.ScheduleEvent(memory_freezer_ns, event); LOG_DEBUG(Common_Memory, "Memory freezer activated!"); } else { LOG_DEBUG(Common_Memory, "Memory freezer deactivated!"); @@ -105,28 +107,21 @@ void Freezer::Unfreeze(VAddr address) { LOG_DEBUG(Common_Memory, "Unfreezing memory for address={:016X}", address); - entries.erase( - std::remove_if(entries.begin(), entries.end(), - [&address](const Entry& entry) { return entry.address == address; }), - entries.end()); + std::erase_if(entries, [address](const Entry& entry) { return entry.address == address; }); } bool Freezer::IsFrozen(VAddr address) const { std::lock_guard lock{entries_mutex}; - return std::find_if(entries.begin(), entries.end(), [&address](const Entry& entry) { - return entry.address == address; - }) != entries.end(); + return FindEntry(address) != entries.cend(); } void Freezer::SetFrozenValue(VAddr address, u64 value) { std::lock_guard lock{entries_mutex}; - const auto iter = std::find_if(entries.begin(), entries.end(), [&address](const Entry& entry) { - return entry.address == address; - }); + const auto iter = FindEntry(address); - if (iter == entries.end()) { + if (iter == entries.cend()) { LOG_ERROR(Common_Memory, "Tried to set freeze value for address={:016X} that is not frozen!", address); return; @@ -141,11 +136,9 @@ void Freezer::SetFrozenValue(VAddr address, u64 value) { std::optional<Freezer::Entry> Freezer::GetEntry(VAddr address) const { std::lock_guard lock{entries_mutex}; - const auto iter = std::find_if(entries.begin(), entries.end(), [&address](const Entry& entry) { - return entry.address == address; - }); + const auto iter = FindEntry(address); - if (iter == entries.end()) { + if (iter == entries.cend()) { return std::nullopt; } @@ -158,7 +151,17 @@ std::vector<Freezer::Entry> Freezer::GetEntries() const { return entries; } -void Freezer::FrameCallback(u64 userdata, s64 cycles_late) { +Freezer::Entries::iterator Freezer::FindEntry(VAddr address) { + return std::find_if(entries.begin(), entries.end(), + [address](const Entry& entry) { return entry.address == address; }); +} + +Freezer::Entries::const_iterator Freezer::FindEntry(VAddr address) const { + return std::find_if(entries.begin(), entries.end(), + [address](const Entry& entry) { return entry.address == address; }); +} + +void Freezer::FrameCallback(std::uintptr_t, std::chrono::nanoseconds ns_late) { if (!IsActive()) { LOG_DEBUG(Common_Memory, "Memory freezer has been deactivated, ending callback events."); return; @@ -173,7 +176,7 @@ void Freezer::FrameCallback(u64 userdata, s64 cycles_late) { MemoryWriteWidth(memory, entry.width, entry.address, entry.value); } - core_timing.ScheduleEvent(MEMORY_FREEZER_TICKS - cycles_late, event); + core_timing.ScheduleEvent(memory_freezer_ns - ns_late, event); } void Freezer::FillEntryReads() { diff --git a/src/core/tools/freezer.h b/src/core/tools/freezer.h index 916339c6c..0fdb701a7 100644 --- a/src/core/tools/freezer.h +++ b/src/core/tools/freezer.h @@ -5,6 +5,7 @@ #pragma once #include <atomic> +#include <chrono> #include <memory> #include <mutex> #include <optional> @@ -16,7 +17,7 @@ class CoreTiming; struct EventType; } // namespace Core::Timing -namespace Memory { +namespace Core::Memory { class Memory; } @@ -38,7 +39,7 @@ public: u64 value; }; - explicit Freezer(Core::Timing::CoreTiming& core_timing_, Memory::Memory& memory_); + explicit Freezer(Core::Timing::CoreTiming& core_timing_, Core::Memory::Memory& memory_); ~Freezer(); // Enables or disables the entire memory freezer. @@ -72,17 +73,22 @@ public: std::vector<Entry> GetEntries() const; private: - void FrameCallback(u64 userdata, s64 cycles_late); + using Entries = std::vector<Entry>; + + Entries::iterator FindEntry(VAddr address); + Entries::const_iterator FindEntry(VAddr address) const; + + void FrameCallback(std::uintptr_t user_data, std::chrono::nanoseconds ns_late); void FillEntryReads(); std::atomic_bool active{false}; mutable std::mutex entries_mutex; - std::vector<Entry> entries; + Entries entries; std::shared_ptr<Core::Timing::EventType> event; Core::Timing::CoreTiming& core_timing; - Memory::Memory& memory; + Core::Memory::Memory& memory; }; } // namespace Tools diff --git a/src/input_common/CMakeLists.txt b/src/input_common/CMakeLists.txt index 2520ba321..1d1b2e08a 100644 --- a/src/input_common/CMakeLists.txt +++ b/src/input_common/CMakeLists.txt @@ -7,6 +7,18 @@ add_library(input_common STATIC main.h motion_emu.cpp motion_emu.h + motion_from_button.cpp + motion_from_button.h + motion_input.cpp + motion_input.h + settings.cpp + settings.h + touch_from_button.cpp + touch_from_button.h + gcadapter/gc_adapter.cpp + gcadapter/gc_adapter.h + gcadapter/gc_poller.cpp + gcadapter/gc_poller.h sdl/sdl.cpp sdl/sdl.h udp/client.cpp @@ -17,6 +29,39 @@ add_library(input_common STATIC udp/udp.h ) +if (MSVC) + target_compile_options(input_common PRIVATE + /W4 + /WX + + # 'expression' : signed/unsigned mismatch + /we4018 + # 'argument' : conversion from 'type1' to 'type2', possible loss of data (floating-point) + /we4244 + # 'conversion' : conversion from 'type1' to 'type2', signed/unsigned mismatch + /we4245 + # 'operator': conversion from 'type1:field_bits' to 'type2:field_bits', possible loss of data + /we4254 + # 'var' : conversion from 'size_t' to 'type', possible loss of data + /we4267 + # 'context' : truncation from 'type1' to 'type2' + /we4305 + ) +else() + target_compile_options(input_common PRIVATE + -Werror + -Werror=conversion + -Werror=ignored-qualifiers + -Werror=implicit-fallthrough + -Werror=reorder + -Werror=shadow + -Werror=sign-compare + -Werror=unused-but-set-parameter + -Werror=unused-but-set-variable + -Werror=unused-variable + ) +endif() + if(SDL2_FOUND) target_sources(input_common PRIVATE sdl/sdl_impl.cpp @@ -26,5 +71,8 @@ if(SDL2_FOUND) target_compile_definitions(input_common PRIVATE HAVE_SDL2) endif() +target_include_directories(input_common SYSTEM PRIVATE ${LIBUSB_INCLUDE_DIR}) +target_link_libraries(input_common PRIVATE ${LIBUSB_LIBRARIES}) + create_target_directory_groups(input_common) -target_link_libraries(input_common PUBLIC core PRIVATE common ${Boost_LIBRARIES}) +target_link_libraries(input_common PUBLIC core PRIVATE common Boost::boost) diff --git a/src/input_common/analog_from_button.cpp b/src/input_common/analog_from_button.cpp index 6cabdaa3c..74744d7f3 100755 --- a/src/input_common/analog_from_button.cpp +++ b/src/input_common/analog_from_button.cpp @@ -20,18 +20,22 @@ public: constexpr float SQRT_HALF = 0.707106781f; int x = 0, y = 0; - if (right->GetStatus()) + if (right->GetStatus()) { ++x; - if (left->GetStatus()) + } + if (left->GetStatus()) { --x; - if (up->GetStatus()) + } + if (up->GetStatus()) { ++y; - if (down->GetStatus()) + } + if (down->GetStatus()) { --y; + } - float coef = modifier->GetStatus() ? modifier_scale : 1.0f; - return std::make_tuple(x * coef * (y == 0 ? 1.0f : SQRT_HALF), - y * coef * (x == 0 ? 1.0f : SQRT_HALF)); + const float coef = modifier->GetStatus() ? modifier_scale : 1.0f; + return std::make_tuple(static_cast<float>(x) * coef * (y == 0 ? 1.0f : SQRT_HALF), + static_cast<float>(y) * coef * (x == 0 ? 1.0f : SQRT_HALF)); } bool GetAnalogDirectionStatus(Input::AnalogDirection direction) const override { diff --git a/src/input_common/gcadapter/gc_adapter.cpp b/src/input_common/gcadapter/gc_adapter.cpp new file mode 100644 index 000000000..d80195c82 --- /dev/null +++ b/src/input_common/gcadapter/gc_adapter.cpp @@ -0,0 +1,509 @@ +// Copyright 2014 Dolphin Emulator Project +// Licensed under GPLv2+ +// Refer to the license.txt file included. + +#include <chrono> +#include <thread> + +#ifdef _MSC_VER +#pragma warning(push) +#pragma warning(disable : 4200) // nonstandard extension used : zero-sized array in struct/union +#endif +#include <libusb.h> +#ifdef _MSC_VER +#pragma warning(pop) +#endif + +#include "common/logging/log.h" +#include "common/param_package.h" +#include "input_common/gcadapter/gc_adapter.h" +#include "input_common/settings.h" + +namespace GCAdapter { + +Adapter::Adapter() { + if (usb_adapter_handle != nullptr) { + return; + } + LOG_INFO(Input, "GC Adapter Initialization started"); + + const int init_res = libusb_init(&libusb_ctx); + if (init_res == LIBUSB_SUCCESS) { + adapter_scan_thread = std::thread(&Adapter::AdapterScanThread, this); + } else { + LOG_ERROR(Input, "libusb could not be initialized. failed with error = {}", init_res); + } +} + +Adapter::~Adapter() { + Reset(); +} + +void Adapter::AdapterInputThread() { + LOG_DEBUG(Input, "GC Adapter input thread started"); + s32 payload_size{}; + AdapterPayload adapter_payload{}; + + if (adapter_scan_thread.joinable()) { + adapter_scan_thread.join(); + } + + while (adapter_input_thread_running) { + libusb_interrupt_transfer(usb_adapter_handle, input_endpoint, adapter_payload.data(), + static_cast<s32>(adapter_payload.size()), &payload_size, 16); + if (IsPayloadCorrect(adapter_payload, payload_size)) { + UpdateControllers(adapter_payload); + UpdateVibrations(); + } + std::this_thread::yield(); + } + + if (restart_scan_thread) { + adapter_scan_thread = std::thread(&Adapter::AdapterScanThread, this); + restart_scan_thread = false; + } +} + +bool Adapter::IsPayloadCorrect(const AdapterPayload& adapter_payload, s32 payload_size) { + if (payload_size != static_cast<s32>(adapter_payload.size()) || + adapter_payload[0] != LIBUSB_DT_HID) { + LOG_DEBUG(Input, "Error reading payload (size: {}, type: {:02x})", payload_size, + adapter_payload[0]); + if (input_error_counter++ > 20) { + LOG_ERROR(Input, "GC adapter timeout, Is the adapter connected?"); + adapter_input_thread_running = false; + restart_scan_thread = true; + } + return false; + } + + input_error_counter = 0; + return true; +} + +void Adapter::UpdateControllers(const AdapterPayload& adapter_payload) { + for (std::size_t port = 0; port < pads.size(); ++port) { + const std::size_t offset = 1 + (9 * port); + const auto type = static_cast<ControllerTypes>(adapter_payload[offset] >> 4); + UpdatePadType(port, type); + if (DeviceConnected(port)) { + const u8 b1 = adapter_payload[offset + 1]; + const u8 b2 = adapter_payload[offset + 2]; + UpdateStateButtons(port, b1, b2); + UpdateStateAxes(port, adapter_payload); + if (configuring) { + UpdateYuzuSettings(port); + } + } + } +} + +void Adapter::UpdatePadType(std::size_t port, ControllerTypes pad_type) { + if (pads[port].type == pad_type) { + return; + } + // Device changed reset device and set new type + ResetDevice(port); + pads[port].type = pad_type; +} + +void Adapter::UpdateStateButtons(std::size_t port, u8 b1, u8 b2) { + if (port >= pads.size()) { + return; + } + + static constexpr std::array<PadButton, 8> b1_buttons{ + PadButton::ButtonA, PadButton::ButtonB, PadButton::ButtonX, PadButton::ButtonY, + PadButton::ButtonLeft, PadButton::ButtonRight, PadButton::ButtonDown, PadButton::ButtonUp, + }; + + static constexpr std::array<PadButton, 4> b2_buttons{ + PadButton::ButtonStart, + PadButton::TriggerZ, + PadButton::TriggerR, + PadButton::TriggerL, + }; + pads[port].buttons = 0; + for (std::size_t i = 0; i < b1_buttons.size(); ++i) { + if ((b1 & (1U << i)) != 0) { + pads[port].buttons = + static_cast<u16>(pads[port].buttons | static_cast<u16>(b1_buttons[i])); + pads[port].last_button = b1_buttons[i]; + } + } + + for (std::size_t j = 0; j < b2_buttons.size(); ++j) { + if ((b2 & (1U << j)) != 0) { + pads[port].buttons = + static_cast<u16>(pads[port].buttons | static_cast<u16>(b2_buttons[j])); + pads[port].last_button = b2_buttons[j]; + } + } +} + +void Adapter::UpdateStateAxes(std::size_t port, const AdapterPayload& adapter_payload) { + if (port >= pads.size()) { + return; + } + + const std::size_t offset = 1 + (9 * port); + static constexpr std::array<PadAxes, 6> axes{ + PadAxes::StickX, PadAxes::StickY, PadAxes::SubstickX, + PadAxes::SubstickY, PadAxes::TriggerLeft, PadAxes::TriggerRight, + }; + + for (const PadAxes axis : axes) { + const auto index = static_cast<std::size_t>(axis); + const u8 axis_value = adapter_payload[offset + 3 + index]; + if (pads[port].axis_origin[index] == 255) { + pads[port].axis_origin[index] = axis_value; + } + pads[port].axis_values[index] = + static_cast<s16>(axis_value - pads[port].axis_origin[index]); + } +} + +void Adapter::UpdateYuzuSettings(std::size_t port) { + if (port >= pads.size()) { + return; + } + + constexpr u8 axis_threshold = 50; + GCPadStatus pad_status = {.port = port}; + + if (pads[port].buttons != 0) { + pad_status.button = pads[port].last_button; + pad_queue.Push(pad_status); + } + + // Accounting for a threshold here to ensure an intentional press + for (std::size_t i = 0; i < pads[port].axis_values.size(); ++i) { + const s16 value = pads[port].axis_values[i]; + + if (value > axis_threshold || value < -axis_threshold) { + pad_status.axis = static_cast<PadAxes>(i); + pad_status.axis_value = value; + pad_status.axis_threshold = axis_threshold; + pad_queue.Push(pad_status); + } + } +} + +void Adapter::UpdateVibrations() { + // Use 8 states to keep the switching between on/off fast enough for + // a human to not notice the difference between switching from on/off + // More states = more rumble strengths = slower update time + constexpr u8 vibration_states = 8; + + vibration_counter = (vibration_counter + 1) % vibration_states; + + for (GCController& pad : pads) { + const bool vibrate = pad.rumble_amplitude > vibration_counter; + vibration_changed |= vibrate != pad.enable_vibration; + pad.enable_vibration = vibrate; + } + SendVibrations(); +} + +void Adapter::SendVibrations() { + if (!rumble_enabled || !vibration_changed) { + return; + } + s32 size{}; + constexpr u8 rumble_command = 0x11; + const u8 p1 = pads[0].enable_vibration; + const u8 p2 = pads[1].enable_vibration; + const u8 p3 = pads[2].enable_vibration; + const u8 p4 = pads[3].enable_vibration; + std::array<u8, 5> payload = {rumble_command, p1, p2, p3, p4}; + const int err = libusb_interrupt_transfer(usb_adapter_handle, output_endpoint, payload.data(), + static_cast<s32>(payload.size()), &size, 16); + if (err) { + LOG_DEBUG(Input, "Adapter libusb write failed: {}", libusb_error_name(err)); + if (output_error_counter++ > 5) { + LOG_ERROR(Input, "GC adapter output timeout, Rumble disabled"); + rumble_enabled = false; + } + return; + } + output_error_counter = 0; + vibration_changed = false; +} + +bool Adapter::RumblePlay(std::size_t port, u8 amplitude) { + pads[port].rumble_amplitude = amplitude; + + return rumble_enabled; +} + +void Adapter::AdapterScanThread() { + adapter_scan_thread_running = true; + adapter_input_thread_running = false; + if (adapter_input_thread.joinable()) { + adapter_input_thread.join(); + } + ClearLibusbHandle(); + ResetDevices(); + while (adapter_scan_thread_running && !adapter_input_thread_running) { + Setup(); + std::this_thread::sleep_for(std::chrono::seconds(1)); + } +} + +void Adapter::Setup() { + usb_adapter_handle = libusb_open_device_with_vid_pid(libusb_ctx, 0x057e, 0x0337); + + if (usb_adapter_handle == NULL) { + return; + } + if (!CheckDeviceAccess()) { + ClearLibusbHandle(); + return; + } + + libusb_device* device = libusb_get_device(usb_adapter_handle); + + LOG_INFO(Input, "GC adapter is now connected"); + // GC Adapter found and accessible, registering it + if (GetGCEndpoint(device)) { + adapter_scan_thread_running = false; + adapter_input_thread_running = true; + rumble_enabled = true; + input_error_counter = 0; + output_error_counter = 0; + adapter_input_thread = std::thread(&Adapter::AdapterInputThread, this); + } +} + +bool Adapter::CheckDeviceAccess() { + // This fixes payload problems from offbrand GCAdapters + const s32 control_transfer_error = + libusb_control_transfer(usb_adapter_handle, 0x21, 11, 0x0001, 0, nullptr, 0, 1000); + if (control_transfer_error < 0) { + LOG_ERROR(Input, "libusb_control_transfer failed with error= {}", control_transfer_error); + } + + s32 kernel_driver_error = libusb_kernel_driver_active(usb_adapter_handle, 0); + if (kernel_driver_error == 1) { + kernel_driver_error = libusb_detach_kernel_driver(usb_adapter_handle, 0); + if (kernel_driver_error != 0 && kernel_driver_error != LIBUSB_ERROR_NOT_SUPPORTED) { + LOG_ERROR(Input, "libusb_detach_kernel_driver failed with error = {}", + kernel_driver_error); + } + } + + if (kernel_driver_error && kernel_driver_error != LIBUSB_ERROR_NOT_SUPPORTED) { + libusb_close(usb_adapter_handle); + usb_adapter_handle = nullptr; + return false; + } + + const int interface_claim_error = libusb_claim_interface(usb_adapter_handle, 0); + if (interface_claim_error) { + LOG_ERROR(Input, "libusb_claim_interface failed with error = {}", interface_claim_error); + libusb_close(usb_adapter_handle); + usb_adapter_handle = nullptr; + return false; + } + + return true; +} + +bool Adapter::GetGCEndpoint(libusb_device* device) { + libusb_config_descriptor* config = nullptr; + const int config_descriptor_return = libusb_get_config_descriptor(device, 0, &config); + if (config_descriptor_return != LIBUSB_SUCCESS) { + LOG_ERROR(Input, "libusb_get_config_descriptor failed with error = {}", + config_descriptor_return); + return false; + } + + for (u8 ic = 0; ic < config->bNumInterfaces; ic++) { + const libusb_interface* interfaceContainer = &config->interface[ic]; + for (int i = 0; i < interfaceContainer->num_altsetting; i++) { + const libusb_interface_descriptor* interface = &interfaceContainer->altsetting[i]; + for (u8 e = 0; e < interface->bNumEndpoints; e++) { + const libusb_endpoint_descriptor* endpoint = &interface->endpoint[e]; + if ((endpoint->bEndpointAddress & LIBUSB_ENDPOINT_IN) != 0) { + input_endpoint = endpoint->bEndpointAddress; + } else { + output_endpoint = endpoint->bEndpointAddress; + } + } + } + } + // This transfer seems to be responsible for clearing the state of the adapter + // Used to clear the "busy" state of when the device is unexpectedly unplugged + unsigned char clear_payload = 0x13; + libusb_interrupt_transfer(usb_adapter_handle, output_endpoint, &clear_payload, + sizeof(clear_payload), nullptr, 16); + return true; +} + +void Adapter::JoinThreads() { + restart_scan_thread = false; + adapter_input_thread_running = false; + adapter_scan_thread_running = false; + + if (adapter_scan_thread.joinable()) { + adapter_scan_thread.join(); + } + + if (adapter_input_thread.joinable()) { + adapter_input_thread.join(); + } +} + +void Adapter::ClearLibusbHandle() { + if (usb_adapter_handle) { + libusb_release_interface(usb_adapter_handle, 1); + libusb_close(usb_adapter_handle); + usb_adapter_handle = nullptr; + } +} + +void Adapter::ResetDevices() { + for (std::size_t i = 0; i < pads.size(); ++i) { + ResetDevice(i); + } +} + +void Adapter::ResetDevice(std::size_t port) { + pads[port].type = ControllerTypes::None; + pads[port].enable_vibration = false; + pads[port].rumble_amplitude = 0; + pads[port].buttons = 0; + pads[port].last_button = PadButton::Undefined; + pads[port].axis_values.fill(0); + pads[port].axis_origin.fill(255); +} + +void Adapter::Reset() { + JoinThreads(); + ClearLibusbHandle(); + ResetDevices(); + + if (libusb_ctx) { + libusb_exit(libusb_ctx); + } +} + +std::vector<Common::ParamPackage> Adapter::GetInputDevices() const { + std::vector<Common::ParamPackage> devices; + for (std::size_t port = 0; port < pads.size(); ++port) { + if (!DeviceConnected(port)) { + continue; + } + std::string name = fmt::format("Gamecube Controller {}", port + 1); + devices.emplace_back(Common::ParamPackage{ + {"class", "gcpad"}, + {"display", std::move(name)}, + {"port", std::to_string(port)}, + }); + } + return devices; +} + +InputCommon::ButtonMapping Adapter::GetButtonMappingForDevice( + const Common::ParamPackage& params) const { + // This list is missing ZL/ZR since those are not considered buttons. + // We will add those afterwards + // This list also excludes any button that can't be really mapped + static constexpr std::array<std::pair<Settings::NativeButton::Values, PadButton>, 12> + switch_to_gcadapter_button = { + std::pair{Settings::NativeButton::A, PadButton::ButtonA}, + {Settings::NativeButton::B, PadButton::ButtonB}, + {Settings::NativeButton::X, PadButton::ButtonX}, + {Settings::NativeButton::Y, PadButton::ButtonY}, + {Settings::NativeButton::Plus, PadButton::ButtonStart}, + {Settings::NativeButton::DLeft, PadButton::ButtonLeft}, + {Settings::NativeButton::DUp, PadButton::ButtonUp}, + {Settings::NativeButton::DRight, PadButton::ButtonRight}, + {Settings::NativeButton::DDown, PadButton::ButtonDown}, + {Settings::NativeButton::SL, PadButton::TriggerL}, + {Settings::NativeButton::SR, PadButton::TriggerR}, + {Settings::NativeButton::R, PadButton::TriggerZ}, + }; + if (!params.Has("port")) { + return {}; + } + + InputCommon::ButtonMapping mapping{}; + for (const auto& [switch_button, gcadapter_button] : switch_to_gcadapter_button) { + Common::ParamPackage button_params({{"engine", "gcpad"}}); + button_params.Set("port", params.Get("port", 0)); + button_params.Set("button", static_cast<int>(gcadapter_button)); + mapping.insert_or_assign(switch_button, std::move(button_params)); + } + + // Add the missing bindings for ZL/ZR + static constexpr std::array<std::pair<Settings::NativeButton::Values, PadAxes>, 2> + switch_to_gcadapter_axis = { + std::pair{Settings::NativeButton::ZL, PadAxes::TriggerLeft}, + {Settings::NativeButton::ZR, PadAxes::TriggerRight}, + }; + for (const auto& [switch_button, gcadapter_axis] : switch_to_gcadapter_axis) { + Common::ParamPackage button_params({{"engine", "gcpad"}}); + button_params.Set("port", params.Get("port", 0)); + button_params.Set("button", static_cast<s32>(PadButton::Stick)); + button_params.Set("axis", static_cast<s32>(gcadapter_axis)); + button_params.Set("threshold", 0.5f); + button_params.Set("direction", "+"); + mapping.insert_or_assign(switch_button, std::move(button_params)); + } + return mapping; +} + +InputCommon::AnalogMapping Adapter::GetAnalogMappingForDevice( + const Common::ParamPackage& params) const { + if (!params.Has("port")) { + return {}; + } + + InputCommon::AnalogMapping mapping = {}; + Common::ParamPackage left_analog_params; + left_analog_params.Set("engine", "gcpad"); + left_analog_params.Set("port", params.Get("port", 0)); + left_analog_params.Set("axis_x", static_cast<int>(PadAxes::StickX)); + left_analog_params.Set("axis_y", static_cast<int>(PadAxes::StickY)); + mapping.insert_or_assign(Settings::NativeAnalog::LStick, std::move(left_analog_params)); + Common::ParamPackage right_analog_params; + right_analog_params.Set("engine", "gcpad"); + right_analog_params.Set("port", params.Get("port", 0)); + right_analog_params.Set("axis_x", static_cast<int>(PadAxes::SubstickX)); + right_analog_params.Set("axis_y", static_cast<int>(PadAxes::SubstickY)); + mapping.insert_or_assign(Settings::NativeAnalog::RStick, std::move(right_analog_params)); + return mapping; +} + +bool Adapter::DeviceConnected(std::size_t port) const { + return pads[port].type != ControllerTypes::None; +} + +void Adapter::BeginConfiguration() { + pad_queue.Clear(); + configuring = true; +} + +void Adapter::EndConfiguration() { + pad_queue.Clear(); + configuring = false; +} + +Common::SPSCQueue<GCPadStatus>& Adapter::GetPadQueue() { + return pad_queue; +} + +const Common::SPSCQueue<GCPadStatus>& Adapter::GetPadQueue() const { + return pad_queue; +} + +GCController& Adapter::GetPadState(std::size_t port) { + return pads.at(port); +} + +const GCController& Adapter::GetPadState(std::size_t port) const { + return pads.at(port); +} + +} // namespace GCAdapter diff --git a/src/input_common/gcadapter/gc_adapter.h b/src/input_common/gcadapter/gc_adapter.h new file mode 100644 index 000000000..f1256c9da --- /dev/null +++ b/src/input_common/gcadapter/gc_adapter.h @@ -0,0 +1,167 @@ +// Copyright 2014 Dolphin Emulator Project +// Licensed under GPLv2+ +// Refer to the license.txt file included. + +#pragma once +#include <algorithm> +#include <functional> +#include <mutex> +#include <thread> +#include <unordered_map> +#include "common/common_types.h" +#include "common/threadsafe_queue.h" +#include "input_common/main.h" + +struct libusb_context; +struct libusb_device; +struct libusb_device_handle; + +namespace GCAdapter { + +enum class PadButton { + Undefined = 0x0000, + ButtonLeft = 0x0001, + ButtonRight = 0x0002, + ButtonDown = 0x0004, + ButtonUp = 0x0008, + TriggerZ = 0x0010, + TriggerR = 0x0020, + TriggerL = 0x0040, + ButtonA = 0x0100, + ButtonB = 0x0200, + ButtonX = 0x0400, + ButtonY = 0x0800, + ButtonStart = 0x1000, + // Below is for compatibility with "AxisButton" type + Stick = 0x2000, +}; + +enum class PadAxes : u8 { + StickX, + StickY, + SubstickX, + SubstickY, + TriggerLeft, + TriggerRight, + Undefined, +}; + +enum class ControllerTypes { + None, + Wired, + Wireless, +}; + +struct GCPadStatus { + std::size_t port{}; + + PadButton button{PadButton::Undefined}; // Or-ed PAD_BUTTON_* and PAD_TRIGGER_* bits + + PadAxes axis{PadAxes::Undefined}; + s16 axis_value{}; + u8 axis_threshold{50}; +}; + +struct GCController { + ControllerTypes type{}; + bool enable_vibration{}; + u8 rumble_amplitude{}; + u16 buttons{}; + PadButton last_button{}; + std::array<s16, 6> axis_values{}; + std::array<u8, 6> axis_origin{}; +}; + +class Adapter { +public: + Adapter(); + ~Adapter(); + + /// Request a vibration for a controller + bool RumblePlay(std::size_t port, u8 amplitude); + + /// Used for polling + void BeginConfiguration(); + void EndConfiguration(); + + Common::SPSCQueue<GCPadStatus>& GetPadQueue(); + const Common::SPSCQueue<GCPadStatus>& GetPadQueue() const; + + GCController& GetPadState(std::size_t port); + const GCController& GetPadState(std::size_t port) const; + + /// Returns true if there is a device connected to port + bool DeviceConnected(std::size_t port) const; + + /// Used for automapping features + std::vector<Common::ParamPackage> GetInputDevices() const; + InputCommon::ButtonMapping GetButtonMappingForDevice(const Common::ParamPackage& params) const; + InputCommon::AnalogMapping GetAnalogMappingForDevice(const Common::ParamPackage& params) const; + +private: + using AdapterPayload = std::array<u8, 37>; + + void UpdatePadType(std::size_t port, ControllerTypes pad_type); + void UpdateControllers(const AdapterPayload& adapter_payload); + void UpdateYuzuSettings(std::size_t port); + void UpdateStateButtons(std::size_t port, u8 b1, u8 b2); + void UpdateStateAxes(std::size_t port, const AdapterPayload& adapter_payload); + void UpdateVibrations(); + + void AdapterInputThread(); + + void AdapterScanThread(); + + bool IsPayloadCorrect(const AdapterPayload& adapter_payload, s32 payload_size); + + // Updates vibration state of all controllers + void SendVibrations(); + + /// For use in initialization, querying devices to find the adapter + void Setup(); + + /// Resets status of all GC controller devices to a disconected state + void ResetDevices(); + + /// Resets status of device connected to a disconected state + void ResetDevice(std::size_t port); + + /// Returns true if we successfully gain access to GC Adapter + bool CheckDeviceAccess(); + + /// Captures GC Adapter endpoint address + /// Returns true if the endpoind was set correctly + bool GetGCEndpoint(libusb_device* device); + + /// For shutting down, clear all data, join all threads, release usb + void Reset(); + + // Join all threads + void JoinThreads(); + + // Release usb handles + void ClearLibusbHandle(); + + libusb_device_handle* usb_adapter_handle = nullptr; + std::array<GCController, 4> pads; + Common::SPSCQueue<GCPadStatus> pad_queue; + + std::thread adapter_input_thread; + std::thread adapter_scan_thread; + bool adapter_input_thread_running; + bool adapter_scan_thread_running; + bool restart_scan_thread; + + libusb_context* libusb_ctx; + + u8 input_endpoint{0}; + u8 output_endpoint{0}; + u8 input_error_counter{0}; + u8 output_error_counter{0}; + int vibration_counter{0}; + + bool configuring{false}; + bool rumble_enabled{true}; + bool vibration_changed{true}; +}; +} // namespace GCAdapter diff --git a/src/input_common/gcadapter/gc_poller.cpp b/src/input_common/gcadapter/gc_poller.cpp new file mode 100644 index 000000000..4d1052414 --- /dev/null +++ b/src/input_common/gcadapter/gc_poller.cpp @@ -0,0 +1,332 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <atomic> +#include <list> +#include <mutex> +#include <utility> +#include "common/assert.h" +#include "common/threadsafe_queue.h" +#include "input_common/gcadapter/gc_adapter.h" +#include "input_common/gcadapter/gc_poller.h" + +namespace InputCommon { + +class GCButton final : public Input::ButtonDevice { +public: + explicit GCButton(u32 port_, s32 button_, const GCAdapter::Adapter* adapter) + : port(port_), button(button_), gcadapter(adapter) {} + + ~GCButton() override; + + bool GetStatus() const override { + if (gcadapter->DeviceConnected(port)) { + return (gcadapter->GetPadState(port).buttons & button) != 0; + } + return false; + } + +private: + const u32 port; + const s32 button; + const GCAdapter::Adapter* gcadapter; +}; + +class GCAxisButton final : public Input::ButtonDevice { +public: + explicit GCAxisButton(u32 port_, u32 axis_, float threshold_, bool trigger_if_greater_, + const GCAdapter::Adapter* adapter) + : port(port_), axis(axis_), threshold(threshold_), trigger_if_greater(trigger_if_greater_), + gcadapter(adapter) {} + + bool GetStatus() const override { + if (gcadapter->DeviceConnected(port)) { + const float current_axis_value = gcadapter->GetPadState(port).axis_values.at(axis); + const float axis_value = current_axis_value / 128.0f; + if (trigger_if_greater) { + // TODO: Might be worthwile to set a slider for the trigger threshold. It is + // currently always set to 0.5 in configure_input_player.cpp ZL/ZR HandleClick + return axis_value > threshold; + } + return axis_value < -threshold; + } + return false; + } + +private: + const u32 port; + const u32 axis; + float threshold; + bool trigger_if_greater; + const GCAdapter::Adapter* gcadapter; +}; + +GCButtonFactory::GCButtonFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) + : adapter(std::move(adapter_)) {} + +GCButton::~GCButton() = default; + +std::unique_ptr<Input::ButtonDevice> GCButtonFactory::Create(const Common::ParamPackage& params) { + const auto button_id = params.Get("button", 0); + const auto port = static_cast<u32>(params.Get("port", 0)); + + constexpr s32 PAD_STICK_ID = static_cast<s32>(GCAdapter::PadButton::Stick); + + // button is not an axis/stick button + if (button_id != PAD_STICK_ID) { + return std::make_unique<GCButton>(port, button_id, adapter.get()); + } + + // For Axis buttons, used by the binary sticks. + if (button_id == PAD_STICK_ID) { + const int axis = params.Get("axis", 0); + const float threshold = params.Get("threshold", 0.25f); + const std::string direction_name = params.Get("direction", ""); + bool trigger_if_greater; + if (direction_name == "+") { + trigger_if_greater = true; + } else if (direction_name == "-") { + trigger_if_greater = false; + } else { + trigger_if_greater = true; + LOG_ERROR(Input, "Unknown direction {}", direction_name); + } + return std::make_unique<GCAxisButton>(port, axis, threshold, trigger_if_greater, + adapter.get()); + } + + return nullptr; +} + +Common::ParamPackage GCButtonFactory::GetNextInput() const { + Common::ParamPackage params; + GCAdapter::GCPadStatus pad; + auto& queue = adapter->GetPadQueue(); + while (queue.Pop(pad)) { + // This while loop will break on the earliest detected button + params.Set("engine", "gcpad"); + params.Set("port", static_cast<s32>(pad.port)); + if (pad.button != GCAdapter::PadButton::Undefined) { + params.Set("button", static_cast<u16>(pad.button)); + } + + // For Axis button implementation + if (pad.axis != GCAdapter::PadAxes::Undefined) { + params.Set("axis", static_cast<u8>(pad.axis)); + params.Set("button", static_cast<u16>(GCAdapter::PadButton::Stick)); + params.Set("threshold", "0.25"); + if (pad.axis_value > 0) { + params.Set("direction", "+"); + } else { + params.Set("direction", "-"); + } + break; + } + } + return params; +} + +void GCButtonFactory::BeginConfiguration() { + polling = true; + adapter->BeginConfiguration(); +} + +void GCButtonFactory::EndConfiguration() { + polling = false; + adapter->EndConfiguration(); +} + +class GCAnalog final : public Input::AnalogDevice { +public: + explicit GCAnalog(u32 port_, u32 axis_x_, u32 axis_y_, float deadzone_, + const GCAdapter::Adapter* adapter, float range_) + : port(port_), axis_x(axis_x_), axis_y(axis_y_), deadzone(deadzone_), gcadapter(adapter), + range(range_) {} + + float GetAxis(u32 axis) const { + if (gcadapter->DeviceConnected(port)) { + std::lock_guard lock{mutex}; + const auto axis_value = + static_cast<float>(gcadapter->GetPadState(port).axis_values.at(axis)); + return (axis_value) / (100.0f * range); + } + return 0.0f; + } + + std::pair<float, float> GetAnalog(u32 analog_axis_x, u32 analog_axis_y) const { + float x = GetAxis(analog_axis_x); + float y = GetAxis(analog_axis_y); + + // Make sure the coordinates are in the unit circle, + // otherwise normalize it. + float r = x * x + y * y; + if (r > 1.0f) { + r = std::sqrt(r); + x /= r; + y /= r; + } + + return {x, y}; + } + + std::tuple<float, float> GetStatus() const override { + const auto [x, y] = GetAnalog(axis_x, axis_y); + const float r = std::sqrt((x * x) + (y * y)); + if (r > deadzone) { + return {x / r * (r - deadzone) / (1 - deadzone), + y / r * (r - deadzone) / (1 - deadzone)}; + } + return {0.0f, 0.0f}; + } + + bool GetAnalogDirectionStatus(Input::AnalogDirection direction) const override { + const auto [x, y] = GetStatus(); + const float directional_deadzone = 0.5f; + switch (direction) { + case Input::AnalogDirection::RIGHT: + return x > directional_deadzone; + case Input::AnalogDirection::LEFT: + return x < -directional_deadzone; + case Input::AnalogDirection::UP: + return y > directional_deadzone; + case Input::AnalogDirection::DOWN: + return y < -directional_deadzone; + } + return false; + } + +private: + const u32 port; + const u32 axis_x; + const u32 axis_y; + const float deadzone; + const GCAdapter::Adapter* gcadapter; + const float range; + mutable std::mutex mutex; +}; + +/// An analog device factory that creates analog devices from GC Adapter +GCAnalogFactory::GCAnalogFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) + : adapter(std::move(adapter_)) {} + +/** + * Creates analog device from joystick axes + * @param params contains parameters for creating the device: + * - "port": the nth gcpad on the adapter + * - "axis_x": the index of the axis to be bind as x-axis + * - "axis_y": the index of the axis to be bind as y-axis + */ +std::unique_ptr<Input::AnalogDevice> GCAnalogFactory::Create(const Common::ParamPackage& params) { + const auto port = static_cast<u32>(params.Get("port", 0)); + const auto axis_x = static_cast<u32>(params.Get("axis_x", 0)); + const auto axis_y = static_cast<u32>(params.Get("axis_y", 1)); + const auto deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, 1.0f); + const auto range = std::clamp(params.Get("range", 1.0f), 0.50f, 1.50f); + + return std::make_unique<GCAnalog>(port, axis_x, axis_y, deadzone, adapter.get(), range); +} + +void GCAnalogFactory::BeginConfiguration() { + polling = true; + adapter->BeginConfiguration(); +} + +void GCAnalogFactory::EndConfiguration() { + polling = false; + adapter->EndConfiguration(); +} + +Common::ParamPackage GCAnalogFactory::GetNextInput() { + GCAdapter::GCPadStatus pad; + Common::ParamPackage params; + auto& queue = adapter->GetPadQueue(); + while (queue.Pop(pad)) { + if (pad.button != GCAdapter::PadButton::Undefined) { + params.Set("engine", "gcpad"); + params.Set("port", static_cast<s32>(pad.port)); + params.Set("button", static_cast<u16>(pad.button)); + return params; + } + if (pad.axis == GCAdapter::PadAxes::Undefined || + std::abs(static_cast<float>(pad.axis_value) / 128.0f) < 0.1f) { + continue; + } + // An analog device needs two axes, so we need to store the axis for later and wait for + // a second input event. The axes also must be from the same joystick. + const u8 axis = static_cast<u8>(pad.axis); + if (axis == 0 || axis == 1) { + analog_x_axis = 0; + analog_y_axis = 1; + controller_number = static_cast<s32>(pad.port); + break; + } + if (axis == 2 || axis == 3) { + analog_x_axis = 2; + analog_y_axis = 3; + controller_number = static_cast<s32>(pad.port); + break; + } + + if (analog_x_axis == -1) { + analog_x_axis = axis; + controller_number = static_cast<s32>(pad.port); + } else if (analog_y_axis == -1 && analog_x_axis != axis && + controller_number == static_cast<s32>(pad.port)) { + analog_y_axis = axis; + break; + } + } + if (analog_x_axis != -1 && analog_y_axis != -1) { + params.Set("engine", "gcpad"); + params.Set("port", controller_number); + params.Set("axis_x", analog_x_axis); + params.Set("axis_y", analog_y_axis); + analog_x_axis = -1; + analog_y_axis = -1; + controller_number = -1; + return params; + } + return params; +} + +class GCVibration final : public Input::VibrationDevice { +public: + explicit GCVibration(u32 port_, GCAdapter::Adapter* adapter) + : port(port_), gcadapter(adapter) {} + + u8 GetStatus() const override { + return gcadapter->RumblePlay(port, 0); + } + + bool SetRumblePlay(f32 amp_low, [[maybe_unused]] f32 freq_low, f32 amp_high, + [[maybe_unused]] f32 freq_high) const override { + const auto mean_amplitude = (amp_low + amp_high) * 0.5f; + const auto processed_amplitude = + static_cast<u8>((mean_amplitude + std::pow(mean_amplitude, 0.3f)) * 0.5f * 0x8); + + return gcadapter->RumblePlay(port, processed_amplitude); + } + +private: + const u32 port; + GCAdapter::Adapter* gcadapter; +}; + +/// An vibration device factory that creates vibration devices from GC Adapter +GCVibrationFactory::GCVibrationFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) + : adapter(std::move(adapter_)) {} + +/** + * Creates a vibration device from a joystick + * @param params contains parameters for creating the device: + * - "port": the nth gcpad on the adapter + */ +std::unique_ptr<Input::VibrationDevice> GCVibrationFactory::Create( + const Common::ParamPackage& params) { + const auto port = static_cast<u32>(params.Get("port", 0)); + + return std::make_unique<GCVibration>(port, adapter.get()); +} + +} // namespace InputCommon diff --git a/src/input_common/gcadapter/gc_poller.h b/src/input_common/gcadapter/gc_poller.h new file mode 100644 index 000000000..d1271e3ea --- /dev/null +++ b/src/input_common/gcadapter/gc_poller.h @@ -0,0 +1,78 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include "core/frontend/input.h" +#include "input_common/gcadapter/gc_adapter.h" + +namespace InputCommon { + +/** + * A button device factory representing a gcpad. It receives gcpad events and forward them + * to all button devices it created. + */ +class GCButtonFactory final : public Input::Factory<Input::ButtonDevice> { +public: + explicit GCButtonFactory(std::shared_ptr<GCAdapter::Adapter> adapter_); + + /** + * Creates a button device from a button press + * @param params contains parameters for creating the device: + * - "code": the code of the key to bind with the button + */ + std::unique_ptr<Input::ButtonDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput() const; + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<GCAdapter::Adapter> adapter; + bool polling = false; +}; + +/// An analog device factory that creates analog devices from GC Adapter +class GCAnalogFactory final : public Input::Factory<Input::AnalogDevice> { +public: + explicit GCAnalogFactory(std::shared_ptr<GCAdapter::Adapter> adapter_); + + std::unique_ptr<Input::AnalogDevice> Create(const Common::ParamPackage& params) override; + Common::ParamPackage GetNextInput(); + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<GCAdapter::Adapter> adapter; + int analog_x_axis = -1; + int analog_y_axis = -1; + int controller_number = -1; + bool polling = false; +}; + +/// A vibration device factory creates vibration devices from GC Adapter +class GCVibrationFactory final : public Input::Factory<Input::VibrationDevice> { +public: + explicit GCVibrationFactory(std::shared_ptr<GCAdapter::Adapter> adapter_); + + std::unique_ptr<Input::VibrationDevice> Create(const Common::ParamPackage& params) override; + +private: + std::shared_ptr<GCAdapter::Adapter> adapter; +}; + +} // namespace InputCommon diff --git a/src/input_common/keyboard.cpp b/src/input_common/keyboard.cpp index 078374be5..24a6f7a33 100644 --- a/src/input_common/keyboard.cpp +++ b/src/input_common/keyboard.cpp @@ -49,8 +49,9 @@ public: void ChangeKeyStatus(int key_code, bool pressed) { std::lock_guard guard{mutex}; for (const KeyButtonPair& pair : list) { - if (pair.key_code == key_code) + if (pair.key_code == key_code) { pair.key_button->status.store(pressed); + } } } @@ -73,10 +74,10 @@ KeyButton::~KeyButton() { } std::unique_ptr<Input::ButtonDevice> Keyboard::Create(const Common::ParamPackage& params) { - int key_code = params.Get("code", 0); + const int key_code = params.Get("code", 0); std::unique_ptr<KeyButton> button = std::make_unique<KeyButton>(key_button_list); key_button_list->AddKeyButton(key_code, button.get()); - return std::move(button); + return button; } void Keyboard::PressKey(int key_code) { diff --git a/src/input_common/main.cpp b/src/input_common/main.cpp index c98c848cf..e59ad4ff5 100644 --- a/src/input_common/main.cpp +++ b/src/input_common/main.cpp @@ -6,9 +6,14 @@ #include <thread> #include "common/param_package.h" #include "input_common/analog_from_button.h" +#include "input_common/gcadapter/gc_adapter.h" +#include "input_common/gcadapter/gc_poller.h" #include "input_common/keyboard.h" #include "input_common/main.h" #include "input_common/motion_emu.h" +#include "input_common/motion_from_button.h" +#include "input_common/touch_from_button.h" +#include "input_common/udp/client.h" #include "input_common/udp/udp.h" #ifdef HAVE_SDL2 #include "input_common/sdl/sdl.h" @@ -16,40 +21,228 @@ namespace InputCommon { -static std::shared_ptr<Keyboard> keyboard; -static std::shared_ptr<MotionEmu> motion_emu; -static std::unique_ptr<SDL::State> sdl; -static std::unique_ptr<CemuhookUDP::State> udp; +struct InputSubsystem::Impl { + void Initialize() { + gcadapter = std::make_shared<GCAdapter::Adapter>(); + gcbuttons = std::make_shared<GCButtonFactory>(gcadapter); + Input::RegisterFactory<Input::ButtonDevice>("gcpad", gcbuttons); + gcanalog = std::make_shared<GCAnalogFactory>(gcadapter); + Input::RegisterFactory<Input::AnalogDevice>("gcpad", gcanalog); + gcvibration = std::make_shared<GCVibrationFactory>(gcadapter); + Input::RegisterFactory<Input::VibrationDevice>("gcpad", gcvibration); -void Init() { - keyboard = std::make_shared<Keyboard>(); - Input::RegisterFactory<Input::ButtonDevice>("keyboard", keyboard); - Input::RegisterFactory<Input::AnalogDevice>("analog_from_button", - std::make_shared<AnalogFromButton>()); - motion_emu = std::make_shared<MotionEmu>(); - Input::RegisterFactory<Input::MotionDevice>("motion_emu", motion_emu); + keyboard = std::make_shared<Keyboard>(); + Input::RegisterFactory<Input::ButtonDevice>("keyboard", keyboard); + Input::RegisterFactory<Input::AnalogDevice>("analog_from_button", + std::make_shared<AnalogFromButton>()); + Input::RegisterFactory<Input::MotionDevice>("keyboard", + std::make_shared<MotionFromButton>()); + motion_emu = std::make_shared<MotionEmu>(); + Input::RegisterFactory<Input::MotionDevice>("motion_emu", motion_emu); + Input::RegisterFactory<Input::TouchDevice>("touch_from_button", + std::make_shared<TouchFromButtonFactory>()); - sdl = SDL::Init(); +#ifdef HAVE_SDL2 + sdl = SDL::Init(); +#endif + + udp = std::make_shared<InputCommon::CemuhookUDP::Client>(); + udpmotion = std::make_shared<UDPMotionFactory>(udp); + Input::RegisterFactory<Input::MotionDevice>("cemuhookudp", udpmotion); + udptouch = std::make_shared<UDPTouchFactory>(udp); + Input::RegisterFactory<Input::TouchDevice>("cemuhookudp", udptouch); + } + + void Shutdown() { + Input::UnregisterFactory<Input::ButtonDevice>("keyboard"); + Input::UnregisterFactory<Input::MotionDevice>("keyboard"); + keyboard.reset(); + Input::UnregisterFactory<Input::AnalogDevice>("analog_from_button"); + Input::UnregisterFactory<Input::MotionDevice>("motion_emu"); + motion_emu.reset(); + Input::UnregisterFactory<Input::TouchDevice>("touch_from_button"); +#ifdef HAVE_SDL2 + sdl.reset(); +#endif + Input::UnregisterFactory<Input::ButtonDevice>("gcpad"); + Input::UnregisterFactory<Input::AnalogDevice>("gcpad"); + Input::UnregisterFactory<Input::VibrationDevice>("gcpad"); + + gcbuttons.reset(); + gcanalog.reset(); + gcvibration.reset(); + + Input::UnregisterFactory<Input::MotionDevice>("cemuhookudp"); + Input::UnregisterFactory<Input::TouchDevice>("cemuhookudp"); + + udpmotion.reset(); + udptouch.reset(); + } + + [[nodiscard]] std::vector<Common::ParamPackage> GetInputDevices() const { + std::vector<Common::ParamPackage> devices = { + Common::ParamPackage{{"display", "Any"}, {"class", "any"}}, + Common::ParamPackage{{"display", "Keyboard/Mouse"}, {"class", "keyboard"}}, + }; +#ifdef HAVE_SDL2 + auto sdl_devices = sdl->GetInputDevices(); + devices.insert(devices.end(), sdl_devices.begin(), sdl_devices.end()); +#endif + auto udp_devices = udp->GetInputDevices(); + devices.insert(devices.end(), udp_devices.begin(), udp_devices.end()); + auto gcpad_devices = gcadapter->GetInputDevices(); + devices.insert(devices.end(), gcpad_devices.begin(), gcpad_devices.end()); + return devices; + } + + [[nodiscard]] AnalogMapping GetAnalogMappingForDevice( + const Common::ParamPackage& params) const { + if (!params.Has("class") || params.Get("class", "") == "any") { + return {}; + } + if (params.Get("class", "") == "gcpad") { + return gcadapter->GetAnalogMappingForDevice(params); + } +#ifdef HAVE_SDL2 + if (params.Get("class", "") == "sdl") { + return sdl->GetAnalogMappingForDevice(params); + } +#endif + return {}; + } + + [[nodiscard]] ButtonMapping GetButtonMappingForDevice( + const Common::ParamPackage& params) const { + if (!params.Has("class") || params.Get("class", "") == "any") { + return {}; + } + if (params.Get("class", "") == "gcpad") { + return gcadapter->GetButtonMappingForDevice(params); + } +#ifdef HAVE_SDL2 + if (params.Get("class", "") == "sdl") { + return sdl->GetButtonMappingForDevice(params); + } +#endif + return {}; + } + + [[nodiscard]] MotionMapping GetMotionMappingForDevice( + const Common::ParamPackage& params) const { + if (!params.Has("class") || params.Get("class", "") == "any") { + return {}; + } + if (params.Get("class", "") == "cemuhookudp") { + // TODO return the correct motion device + return {}; + } + return {}; + } + + std::shared_ptr<Keyboard> keyboard; + std::shared_ptr<MotionEmu> motion_emu; +#ifdef HAVE_SDL2 + std::unique_ptr<SDL::State> sdl; +#endif + std::shared_ptr<GCButtonFactory> gcbuttons; + std::shared_ptr<GCAnalogFactory> gcanalog; + std::shared_ptr<GCVibrationFactory> gcvibration; + std::shared_ptr<UDPMotionFactory> udpmotion; + std::shared_ptr<UDPTouchFactory> udptouch; + std::shared_ptr<CemuhookUDP::Client> udp; + std::shared_ptr<GCAdapter::Adapter> gcadapter; +}; + +InputSubsystem::InputSubsystem() : impl{std::make_unique<Impl>()} {} + +InputSubsystem::~InputSubsystem() = default; + +void InputSubsystem::Initialize() { + impl->Initialize(); +} + +void InputSubsystem::Shutdown() { + impl->Shutdown(); +} + +Keyboard* InputSubsystem::GetKeyboard() { + return impl->keyboard.get(); +} + +const Keyboard* InputSubsystem::GetKeyboard() const { + return impl->keyboard.get(); +} + +MotionEmu* InputSubsystem::GetMotionEmu() { + return impl->motion_emu.get(); +} + +const MotionEmu* InputSubsystem::GetMotionEmu() const { + return impl->motion_emu.get(); +} + +std::vector<Common::ParamPackage> InputSubsystem::GetInputDevices() const { + return impl->GetInputDevices(); +} + +AnalogMapping InputSubsystem::GetAnalogMappingForDevice(const Common::ParamPackage& device) const { + return impl->GetAnalogMappingForDevice(device); +} + +ButtonMapping InputSubsystem::GetButtonMappingForDevice(const Common::ParamPackage& device) const { + return impl->GetButtonMappingForDevice(device); +} + +MotionMapping InputSubsystem::GetMotionMappingForDevice(const Common::ParamPackage& device) const { + return impl->GetMotionMappingForDevice(device); +} + +GCAnalogFactory* InputSubsystem::GetGCAnalogs() { + return impl->gcanalog.get(); +} + +const GCAnalogFactory* InputSubsystem::GetGCAnalogs() const { + return impl->gcanalog.get(); +} + +GCButtonFactory* InputSubsystem::GetGCButtons() { + return impl->gcbuttons.get(); +} - udp = CemuhookUDP::Init(); +const GCButtonFactory* InputSubsystem::GetGCButtons() const { + return impl->gcbuttons.get(); } -void Shutdown() { - Input::UnregisterFactory<Input::ButtonDevice>("keyboard"); - keyboard.reset(); - Input::UnregisterFactory<Input::AnalogDevice>("analog_from_button"); - Input::UnregisterFactory<Input::MotionDevice>("motion_emu"); - motion_emu.reset(); - sdl.reset(); - udp.reset(); +UDPMotionFactory* InputSubsystem::GetUDPMotions() { + return impl->udpmotion.get(); } -Keyboard* GetKeyboard() { - return keyboard.get(); +const UDPMotionFactory* InputSubsystem::GetUDPMotions() const { + return impl->udpmotion.get(); } -MotionEmu* GetMotionEmu() { - return motion_emu.get(); +UDPTouchFactory* InputSubsystem::GetUDPTouch() { + return impl->udptouch.get(); +} + +const UDPTouchFactory* InputSubsystem::GetUDPTouch() const { + return impl->udptouch.get(); +} + +void InputSubsystem::ReloadInputDevices() { + if (!impl->udp) { + return; + } + impl->udp->ReloadUDPClient(); +} + +std::vector<std::unique_ptr<Polling::DevicePoller>> InputSubsystem::GetPollers( + Polling::DeviceType type) const { +#ifdef HAVE_SDL2 + return impl->sdl->GetPollers(type); +#else + return {}; +#endif } std::string GenerateKeyboardParam(int key_code) { @@ -73,18 +266,4 @@ std::string GenerateAnalogParamFromKeys(int key_up, int key_down, int key_left, }; return circle_pad_param.Serialize(); } - -namespace Polling { - -std::vector<std::unique_ptr<DevicePoller>> GetPollers(DeviceType type) { - std::vector<std::unique_ptr<DevicePoller>> pollers; - -#ifdef HAVE_SDL2 - pollers = sdl->GetPollers(type); -#endif - - return pollers; -} - -} // namespace Polling } // namespace InputCommon diff --git a/src/input_common/main.h b/src/input_common/main.h index 77a0ce90b..dded3f1ef 100644 --- a/src/input_common/main.h +++ b/src/input_common/main.h @@ -6,40 +6,29 @@ #include <memory> #include <string> +#include <unordered_map> #include <vector> namespace Common { class ParamPackage; } -namespace InputCommon { - -/// Initializes and registers all built-in input device factories. -void Init(); - -/// Deregisters all built-in input device factories and shuts them down. -void Shutdown(); - -class Keyboard; - -/// Gets the keyboard button device factory. -Keyboard* GetKeyboard(); - -class MotionEmu; - -/// Gets the motion emulation factory. -MotionEmu* GetMotionEmu(); +namespace Settings::NativeAnalog { +enum Values : int; +} -/// Generates a serialized param package for creating a keyboard button device -std::string GenerateKeyboardParam(int key_code); +namespace Settings::NativeButton { +enum Values : int; +} -/// Generates a serialized param package for creating an analog device taking input from keyboard -std::string GenerateAnalogParamFromKeys(int key_up, int key_down, int key_left, int key_right, - int key_modifier, float modifier_scale); +namespace Settings::NativeMotion { +enum Values : int; +} +namespace InputCommon { namespace Polling { -enum class DeviceType { Button, Analog }; +enum class DeviceType { Button, AnalogPreferred, Motion }; /** * A class that can be used to get inputs from an input device like controllers without having to @@ -49,7 +38,9 @@ class DevicePoller { public: virtual ~DevicePoller() = default; /// Setup and start polling for inputs, should be called before GetNextInput - virtual void Start() = 0; + /// If a device_id is provided, events should be filtered to only include events from this + /// device id + virtual void Start(const std::string& device_id = "") = 0; /// Stop polling virtual void Stop() = 0; /** @@ -59,8 +50,110 @@ public: */ virtual Common::ParamPackage GetNextInput() = 0; }; - -// Get all DevicePoller from all backends for a specific device type -std::vector<std::unique_ptr<DevicePoller>> GetPollers(DeviceType type); } // namespace Polling + +class GCAnalogFactory; +class GCButtonFactory; +class UDPMotionFactory; +class UDPTouchFactory; +class Keyboard; +class MotionEmu; + +/** + * Given a ParamPackage for a Device returned from `GetInputDevices`, attempt to get the default + * mapping for the device. This is currently only implemented for the SDL backend devices. + */ +using AnalogMapping = std::unordered_map<Settings::NativeAnalog::Values, Common::ParamPackage>; +using ButtonMapping = std::unordered_map<Settings::NativeButton::Values, Common::ParamPackage>; +using MotionMapping = std::unordered_map<Settings::NativeMotion::Values, Common::ParamPackage>; + +class InputSubsystem { +public: + explicit InputSubsystem(); + ~InputSubsystem(); + + InputSubsystem(const InputSubsystem&) = delete; + InputSubsystem& operator=(const InputSubsystem&) = delete; + + InputSubsystem(InputSubsystem&&) = delete; + InputSubsystem& operator=(InputSubsystem&&) = delete; + + /// Initializes and registers all built-in input device factories. + void Initialize(); + + /// Unregisters all built-in input device factories and shuts them down. + void Shutdown(); + + /// Retrieves the underlying keyboard device. + [[nodiscard]] Keyboard* GetKeyboard(); + + /// Retrieves the underlying keyboard device. + [[nodiscard]] const Keyboard* GetKeyboard() const; + + /// Retrieves the underlying motion emulation factory. + [[nodiscard]] MotionEmu* GetMotionEmu(); + + /// Retrieves the underlying motion emulation factory. + [[nodiscard]] const MotionEmu* GetMotionEmu() const; + + /** + * Returns all available input devices that this Factory can create a new device with. + * Each returned ParamPackage should have a `display` field used for display, a class field for + * backends to determine if this backend is meant to service the request and any other + * information needed to identify this in the backend later. + */ + [[nodiscard]] std::vector<Common::ParamPackage> GetInputDevices() const; + + /// Retrieves the analog mappings for the given device. + [[nodiscard]] AnalogMapping GetAnalogMappingForDevice(const Common::ParamPackage& device) const; + + /// Retrieves the button mappings for the given device. + [[nodiscard]] ButtonMapping GetButtonMappingForDevice(const Common::ParamPackage& device) const; + + /// Retrieves the motion mappings for the given device. + [[nodiscard]] MotionMapping GetMotionMappingForDevice(const Common::ParamPackage& device) const; + + /// Retrieves the underlying GameCube analog handler. + [[nodiscard]] GCAnalogFactory* GetGCAnalogs(); + + /// Retrieves the underlying GameCube analog handler. + [[nodiscard]] const GCAnalogFactory* GetGCAnalogs() const; + + /// Retrieves the underlying GameCube button handler. + [[nodiscard]] GCButtonFactory* GetGCButtons(); + + /// Retrieves the underlying GameCube button handler. + [[nodiscard]] const GCButtonFactory* GetGCButtons() const; + + /// Retrieves the underlying udp motion handler. + [[nodiscard]] UDPMotionFactory* GetUDPMotions(); + + /// Retrieves the underlying udp motion handler. + [[nodiscard]] const UDPMotionFactory* GetUDPMotions() const; + + /// Retrieves the underlying udp touch handler. + [[nodiscard]] UDPTouchFactory* GetUDPTouch(); + + /// Retrieves the underlying udp touch handler. + [[nodiscard]] const UDPTouchFactory* GetUDPTouch() const; + + /// Reloads the input devices + void ReloadInputDevices(); + + /// Get all DevicePoller from all backends for a specific device type + [[nodiscard]] std::vector<std::unique_ptr<Polling::DevicePoller>> GetPollers( + Polling::DeviceType type) const; + +private: + struct Impl; + std::unique_ptr<Impl> impl; +}; + +/// Generates a serialized param package for creating a keyboard button device +std::string GenerateKeyboardParam(int key_code); + +/// Generates a serialized param package for creating an analog device taking input from keyboard +std::string GenerateAnalogParamFromKeys(int key_up, int key_down, int key_left, int key_right, + int key_modifier, float modifier_scale); + } // namespace InputCommon diff --git a/src/input_common/motion_emu.cpp b/src/input_common/motion_emu.cpp index 868251628..d4da5596b 100644 --- a/src/input_common/motion_emu.cpp +++ b/src/input_common/motion_emu.cpp @@ -18,11 +18,11 @@ namespace InputCommon { // Implementation class of the motion emulation device class MotionEmuDevice { public: - MotionEmuDevice(int update_millisecond, float sensitivity) - : update_millisecond(update_millisecond), + explicit MotionEmuDevice(int update_millisecond_, float sensitivity_) + : update_millisecond(update_millisecond_), update_duration(std::chrono::duration_cast<std::chrono::steady_clock::duration>( std::chrono::milliseconds(update_millisecond))), - sensitivity(sensitivity), motion_emu_thread(&MotionEmuDevice::MotionEmuThread, this) {} + sensitivity(sensitivity_), motion_emu_thread(&MotionEmuDevice::MotionEmuThread, this) {} ~MotionEmuDevice() { if (motion_emu_thread.joinable()) { @@ -37,16 +37,18 @@ public: } void Tilt(int x, int y) { - auto mouse_move = Common::MakeVec(x, y) - mouse_origin; - if (is_tilting) { - std::lock_guard guard{tilt_mutex}; - if (mouse_move.x == 0 && mouse_move.y == 0) { - tilt_angle = 0; - } else { - tilt_direction = mouse_move.Cast<float>(); - tilt_angle = - std::clamp(tilt_direction.Normalize() * sensitivity, 0.0f, Common::PI * 0.5f); - } + if (!is_tilting) { + return; + } + + std::lock_guard guard{tilt_mutex}; + const auto mouse_move = Common::MakeVec(x, y) - mouse_origin; + if (mouse_move.x == 0 && mouse_move.y == 0) { + tilt_angle = 0; + } else { + tilt_direction = mouse_move.Cast<float>(); + tilt_angle = + std::clamp(tilt_direction.Normalize() * sensitivity, 0.0f, Common::PI * 0.5f); } } @@ -56,7 +58,7 @@ public: is_tilting = false; } - std::tuple<Common::Vec3<float>, Common::Vec3<float>> GetStatus() { + Input::MotionStatus GetStatus() { std::lock_guard guard{status_mutex}; return status; } @@ -76,7 +78,7 @@ private: Common::Event shutdown_event; - std::tuple<Common::Vec3<float>, Common::Vec3<float>> status; + Input::MotionStatus status; std::mutex status_mutex; // Note: always keep the thread declaration at the end so that other objects are initialized @@ -86,11 +88,10 @@ private: void MotionEmuThread() { auto update_time = std::chrono::steady_clock::now(); Common::Quaternion<float> q = Common::MakeQuaternion(Common::Vec3<float>(), 0); - Common::Quaternion<float> old_q; while (!shutdown_event.WaitUntil(update_time)) { update_time += update_duration; - old_q = q; + const Common::Quaternion<float> old_q = q; { std::lock_guard guard{tilt_mutex}; @@ -100,23 +101,32 @@ private: Common::MakeVec(-tilt_direction.y, 0.0f, tilt_direction.x), tilt_angle); } - auto inv_q = q.Inverse(); + const auto inv_q = q.Inverse(); // Set the gravity vector in world space auto gravity = Common::MakeVec(0.0f, -1.0f, 0.0f); // Find the angular rate vector in world space auto angular_rate = ((q - old_q) * inv_q).xyz * 2; - angular_rate *= 1000 / update_millisecond / Common::PI * 180; + angular_rate *= static_cast<float>(1000 / update_millisecond) / Common::PI * 180.0f; // Transform the two vectors from world space to 3DS space gravity = QuaternionRotate(inv_q, gravity); angular_rate = QuaternionRotate(inv_q, angular_rate); + // TODO: Calculate the correct rotation vector and orientation matrix + const auto matrix4x4 = q.ToMatrix(); + const auto rotation = Common::MakeVec(0.0f, 0.0f, 0.0f); + const std::array orientation{ + Common::Vec3f(matrix4x4[0], matrix4x4[1], -matrix4x4[2]), + Common::Vec3f(matrix4x4[4], matrix4x4[5], -matrix4x4[6]), + Common::Vec3f(-matrix4x4[8], -matrix4x4[9], matrix4x4[10]), + }; + // Update the sensor state { std::lock_guard guard{status_mutex}; - status = std::make_tuple(gravity, angular_rate); + status = std::make_tuple(gravity, angular_rate, rotation, orientation); } } } @@ -127,11 +137,11 @@ private: // can forward all the inputs to the implementation only when it is valid. class MotionEmuDeviceWrapper : public Input::MotionDevice { public: - MotionEmuDeviceWrapper(int update_millisecond, float sensitivity) { + explicit MotionEmuDeviceWrapper(int update_millisecond, float sensitivity) { device = std::make_shared<MotionEmuDevice>(update_millisecond, sensitivity); } - std::tuple<Common::Vec3<float>, Common::Vec3<float>> GetStatus() const override { + Input::MotionStatus GetStatus() const override { return device->GetStatus(); } @@ -139,13 +149,13 @@ public: }; std::unique_ptr<Input::MotionDevice> MotionEmu::Create(const Common::ParamPackage& params) { - int update_period = params.Get("update_period", 100); - float sensitivity = params.Get("sensitivity", 0.01f); + const int update_period = params.Get("update_period", 100); + const float sensitivity = params.Get("sensitivity", 0.01f); auto device_wrapper = std::make_unique<MotionEmuDeviceWrapper>(update_period, sensitivity); // Previously created device is disconnected here. Having two motion devices for 3DS is not // expected. current_device = device_wrapper->device; - return std::move(device_wrapper); + return device_wrapper; } void MotionEmu::BeginTilt(int x, int y) { diff --git a/src/input_common/motion_from_button.cpp b/src/input_common/motion_from_button.cpp new file mode 100644 index 000000000..29045a673 --- /dev/null +++ b/src/input_common/motion_from_button.cpp @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "input_common/motion_from_button.h" +#include "input_common/motion_input.h" + +namespace InputCommon { + +class MotionKey final : public Input::MotionDevice { +public: + using Button = std::unique_ptr<Input::ButtonDevice>; + + explicit MotionKey(Button key_) : key(std::move(key_)) {} + + Input::MotionStatus GetStatus() const override { + + if (key->GetStatus()) { + return motion.GetRandomMotion(2, 6); + } + return motion.GetRandomMotion(0, 0); + } + +private: + Button key; + InputCommon::MotionInput motion{0.0f, 0.0f, 0.0f}; +}; + +std::unique_ptr<Input::MotionDevice> MotionFromButton::Create(const Common::ParamPackage& params) { + auto key = Input::CreateDevice<Input::ButtonDevice>(params.Serialize()); + return std::make_unique<MotionKey>(std::move(key)); +} + +} // namespace InputCommon diff --git a/src/input_common/motion_from_button.h b/src/input_common/motion_from_button.h new file mode 100644 index 000000000..a959046fb --- /dev/null +++ b/src/input_common/motion_from_button.h @@ -0,0 +1,25 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "core/frontend/input.h" + +namespace InputCommon { + +/** + * An motion device factory that takes a keyboard button and uses it as a random + * motion device. + */ +class MotionFromButton final : public Input::Factory<Input::MotionDevice> { +public: + /** + * Creates an motion device from button devices + * @param params contains parameters for creating the device: + * - "key": a serialized ParamPackage for creating a button device + */ + std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override; +}; + +} // namespace InputCommon diff --git a/src/input_common/motion_input.cpp b/src/input_common/motion_input.cpp new file mode 100644 index 000000000..f77ba535d --- /dev/null +++ b/src/input_common/motion_input.cpp @@ -0,0 +1,301 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included + +#include <random> +#include "common/math_util.h" +#include "input_common/motion_input.h" + +namespace InputCommon { + +MotionInput::MotionInput(f32 new_kp, f32 new_ki, f32 new_kd) : kp(new_kp), ki(new_ki), kd(new_kd) {} + +void MotionInput::SetAcceleration(const Common::Vec3f& acceleration) { + accel = acceleration; +} + +void MotionInput::SetGyroscope(const Common::Vec3f& gyroscope) { + gyro = gyroscope - gyro_drift; + + // Auto adjust drift to minimize drift + if (!IsMoving(0.1f)) { + gyro_drift = (gyro_drift * 0.9999f) + (gyroscope * 0.0001f); + } + + if (gyro.Length2() < gyro_threshold) { + gyro = {}; + } else { + only_accelerometer = false; + } +} + +void MotionInput::SetQuaternion(const Common::Quaternion<f32>& quaternion) { + quat = quaternion; +} + +void MotionInput::SetGyroDrift(const Common::Vec3f& drift) { + gyro_drift = drift; +} + +void MotionInput::SetGyroThreshold(f32 threshold) { + gyro_threshold = threshold; +} + +void MotionInput::EnableReset(bool reset) { + reset_enabled = reset; +} + +void MotionInput::ResetRotations() { + rotations = {}; +} + +bool MotionInput::IsMoving(f32 sensitivity) const { + return gyro.Length() >= sensitivity || accel.Length() <= 0.9f || accel.Length() >= 1.1f; +} + +bool MotionInput::IsCalibrated(f32 sensitivity) const { + return real_error.Length() < sensitivity; +} + +void MotionInput::UpdateRotation(u64 elapsed_time) { + const auto sample_period = static_cast<f32>(elapsed_time) / 1000000.0f; + if (sample_period > 0.1f) { + return; + } + rotations += gyro * sample_period; +} + +void MotionInput::UpdateOrientation(u64 elapsed_time) { + if (!IsCalibrated(0.1f)) { + ResetOrientation(); + } + // Short name local variable for readability + f32 q1 = quat.w; + f32 q2 = quat.xyz[0]; + f32 q3 = quat.xyz[1]; + f32 q4 = quat.xyz[2]; + const auto sample_period = static_cast<f32>(elapsed_time) / 1000000.0f; + + // Ignore invalid elapsed time + if (sample_period > 0.1f) { + return; + } + + const auto normal_accel = accel.Normalized(); + auto rad_gyro = gyro * Common::PI * 2; + const f32 swap = rad_gyro.x; + rad_gyro.x = rad_gyro.y; + rad_gyro.y = -swap; + rad_gyro.z = -rad_gyro.z; + + // Clear gyro values if there is no gyro present + if (only_accelerometer) { + rad_gyro.x = 0; + rad_gyro.y = 0; + rad_gyro.z = 0; + } + + // Ignore drift correction if acceleration is not reliable + if (accel.Length() >= 0.75f && accel.Length() <= 1.25f) { + const f32 ax = -normal_accel.x; + const f32 ay = normal_accel.y; + const f32 az = -normal_accel.z; + + // Estimated direction of gravity + const f32 vx = 2.0f * (q2 * q4 - q1 * q3); + const f32 vy = 2.0f * (q1 * q2 + q3 * q4); + const f32 vz = q1 * q1 - q2 * q2 - q3 * q3 + q4 * q4; + + // Error is cross product between estimated direction and measured direction of gravity + const Common::Vec3f new_real_error = { + az * vx - ax * vz, + ay * vz - az * vy, + ax * vy - ay * vx, + }; + + derivative_error = new_real_error - real_error; + real_error = new_real_error; + + // Prevent integral windup + if (ki != 0.0f && !IsCalibrated(0.05f)) { + integral_error += real_error; + } else { + integral_error = {}; + } + + // Apply feedback terms + if (!only_accelerometer) { + rad_gyro += kp * real_error; + rad_gyro += ki * integral_error; + rad_gyro += kd * derivative_error; + } else { + // Give more weight to acelerometer values to compensate for the lack of gyro + rad_gyro += 35.0f * kp * real_error; + rad_gyro += 10.0f * ki * integral_error; + rad_gyro += 10.0f * kd * derivative_error; + + // Emulate gyro values for games that need them + gyro.x = -rad_gyro.y; + gyro.y = rad_gyro.x; + gyro.z = -rad_gyro.z; + UpdateRotation(elapsed_time); + } + } + + const f32 gx = rad_gyro.y; + const f32 gy = rad_gyro.x; + const f32 gz = rad_gyro.z; + + // Integrate rate of change of quaternion + const f32 pa = q2; + const f32 pb = q3; + const f32 pc = q4; + q1 = q1 + (-q2 * gx - q3 * gy - q4 * gz) * (0.5f * sample_period); + q2 = pa + (q1 * gx + pb * gz - pc * gy) * (0.5f * sample_period); + q3 = pb + (q1 * gy - pa * gz + pc * gx) * (0.5f * sample_period); + q4 = pc + (q1 * gz + pa * gy - pb * gx) * (0.5f * sample_period); + + quat.w = q1; + quat.xyz[0] = q2; + quat.xyz[1] = q3; + quat.xyz[2] = q4; + quat = quat.Normalized(); +} + +std::array<Common::Vec3f, 3> MotionInput::GetOrientation() const { + const Common::Quaternion<float> quad{ + .xyz = {-quat.xyz[1], -quat.xyz[0], -quat.w}, + .w = -quat.xyz[2], + }; + const std::array<float, 16> matrix4x4 = quad.ToMatrix(); + + return {Common::Vec3f(matrix4x4[0], matrix4x4[1], -matrix4x4[2]), + Common::Vec3f(matrix4x4[4], matrix4x4[5], -matrix4x4[6]), + Common::Vec3f(-matrix4x4[8], -matrix4x4[9], matrix4x4[10])}; +} + +Common::Vec3f MotionInput::GetAcceleration() const { + return accel; +} + +Common::Vec3f MotionInput::GetGyroscope() const { + return gyro; +} + +Common::Quaternion<f32> MotionInput::GetQuaternion() const { + return quat; +} + +Common::Vec3f MotionInput::GetRotations() const { + return rotations; +} + +Input::MotionStatus MotionInput::GetMotion() const { + const Common::Vec3f gyroscope = GetGyroscope(); + const Common::Vec3f accelerometer = GetAcceleration(); + const Common::Vec3f rotation = GetRotations(); + const std::array<Common::Vec3f, 3> orientation = GetOrientation(); + return {accelerometer, gyroscope, rotation, orientation}; +} + +Input::MotionStatus MotionInput::GetRandomMotion(int accel_magnitude, int gyro_magnitude) const { + std::random_device device; + std::mt19937 gen(device()); + std::uniform_int_distribution<s16> distribution(-1000, 1000); + const Common::Vec3f gyroscope{ + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + }; + const Common::Vec3f accelerometer{ + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + }; + constexpr Common::Vec3f rotation; + constexpr std::array orientation{ + Common::Vec3f{1.0f, 0.0f, 0.0f}, + Common::Vec3f{0.0f, 1.0f, 0.0f}, + Common::Vec3f{0.0f, 0.0f, 1.0f}, + }; + return {accelerometer * accel_magnitude, gyroscope * gyro_magnitude, rotation, orientation}; +} + +void MotionInput::ResetOrientation() { + if (!reset_enabled || only_accelerometer) { + return; + } + if (!IsMoving(0.5f) && accel.z <= -0.9f) { + ++reset_counter; + if (reset_counter > 900) { + quat.w = 0; + quat.xyz[0] = 0; + quat.xyz[1] = 0; + quat.xyz[2] = -1; + SetOrientationFromAccelerometer(); + integral_error = {}; + reset_counter = 0; + } + } else { + reset_counter = 0; + } +} + +void MotionInput::SetOrientationFromAccelerometer() { + int iterations = 0; + const f32 sample_period = 0.015f; + + const auto normal_accel = accel.Normalized(); + + while (!IsCalibrated(0.01f) && ++iterations < 100) { + // Short name local variable for readability + f32 q1 = quat.w; + f32 q2 = quat.xyz[0]; + f32 q3 = quat.xyz[1]; + f32 q4 = quat.xyz[2]; + + Common::Vec3f rad_gyro; + const f32 ax = -normal_accel.x; + const f32 ay = normal_accel.y; + const f32 az = -normal_accel.z; + + // Estimated direction of gravity + const f32 vx = 2.0f * (q2 * q4 - q1 * q3); + const f32 vy = 2.0f * (q1 * q2 + q3 * q4); + const f32 vz = q1 * q1 - q2 * q2 - q3 * q3 + q4 * q4; + + // Error is cross product between estimated direction and measured direction of gravity + const Common::Vec3f new_real_error = { + az * vx - ax * vz, + ay * vz - az * vy, + ax * vy - ay * vx, + }; + + derivative_error = new_real_error - real_error; + real_error = new_real_error; + + rad_gyro += 10.0f * kp * real_error; + rad_gyro += 5.0f * ki * integral_error; + rad_gyro += 10.0f * kd * derivative_error; + + const f32 gx = rad_gyro.y; + const f32 gy = rad_gyro.x; + const f32 gz = rad_gyro.z; + + // Integrate rate of change of quaternion + const f32 pa = q2; + const f32 pb = q3; + const f32 pc = q4; + q1 = q1 + (-q2 * gx - q3 * gy - q4 * gz) * (0.5f * sample_period); + q2 = pa + (q1 * gx + pb * gz - pc * gy) * (0.5f * sample_period); + q3 = pb + (q1 * gy - pa * gz + pc * gx) * (0.5f * sample_period); + q4 = pc + (q1 * gz + pa * gy - pb * gx) * (0.5f * sample_period); + + quat.w = q1; + quat.xyz[0] = q2; + quat.xyz[1] = q3; + quat.xyz[2] = q4; + quat = quat.Normalized(); + } +} +} // namespace InputCommon diff --git a/src/input_common/motion_input.h b/src/input_common/motion_input.h new file mode 100644 index 000000000..efe74cf19 --- /dev/null +++ b/src/input_common/motion_input.h @@ -0,0 +1,74 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included + +#pragma once + +#include "common/common_types.h" +#include "common/quaternion.h" +#include "common/vector_math.h" +#include "core/frontend/input.h" + +namespace InputCommon { + +class MotionInput { +public: + explicit MotionInput(f32 new_kp, f32 new_ki, f32 new_kd); + + MotionInput(const MotionInput&) = default; + MotionInput& operator=(const MotionInput&) = default; + + MotionInput(MotionInput&&) = default; + MotionInput& operator=(MotionInput&&) = default; + + void SetAcceleration(const Common::Vec3f& acceleration); + void SetGyroscope(const Common::Vec3f& gyroscope); + void SetQuaternion(const Common::Quaternion<f32>& quaternion); + void SetGyroDrift(const Common::Vec3f& drift); + void SetGyroThreshold(f32 threshold); + + void EnableReset(bool reset); + void ResetRotations(); + + void UpdateRotation(u64 elapsed_time); + void UpdateOrientation(u64 elapsed_time); + + [[nodiscard]] std::array<Common::Vec3f, 3> GetOrientation() const; + [[nodiscard]] Common::Vec3f GetAcceleration() const; + [[nodiscard]] Common::Vec3f GetGyroscope() const; + [[nodiscard]] Common::Vec3f GetRotations() const; + [[nodiscard]] Common::Quaternion<f32> GetQuaternion() const; + [[nodiscard]] Input::MotionStatus GetMotion() const; + [[nodiscard]] Input::MotionStatus GetRandomMotion(int accel_magnitude, + int gyro_magnitude) const; + + [[nodiscard]] bool IsMoving(f32 sensitivity) const; + [[nodiscard]] bool IsCalibrated(f32 sensitivity) const; + +private: + void ResetOrientation(); + void SetOrientationFromAccelerometer(); + + // PID constants + f32 kp; + f32 ki; + f32 kd; + + // PID errors + Common::Vec3f real_error; + Common::Vec3f integral_error; + Common::Vec3f derivative_error; + + Common::Quaternion<f32> quat{{0.0f, 0.0f, -1.0f}, 0.0f}; + Common::Vec3f rotations; + Common::Vec3f accel; + Common::Vec3f gyro; + Common::Vec3f gyro_drift; + + f32 gyro_threshold = 0.0f; + u32 reset_counter = 0; + bool reset_enabled = true; + bool only_accelerometer = true; +}; + +} // namespace InputCommon diff --git a/src/input_common/sdl/sdl.h b/src/input_common/sdl/sdl.h index 5306daa70..42bbf14d4 100644 --- a/src/input_common/sdl/sdl.h +++ b/src/input_common/sdl/sdl.h @@ -6,6 +6,7 @@ #include <memory> #include <vector> +#include "common/param_package.h" #include "input_common/main.h" namespace InputCommon::Polling { @@ -22,14 +23,24 @@ public: /// Unregisters SDL device factories and shut them down. virtual ~State() = default; - virtual Pollers GetPollers(Polling::DeviceType type) = 0; + virtual Pollers GetPollers(Polling::DeviceType) { + return {}; + } + + virtual std::vector<Common::ParamPackage> GetInputDevices() { + return {}; + } + + virtual ButtonMapping GetButtonMappingForDevice(const Common::ParamPackage&) { + return {}; + } + virtual AnalogMapping GetAnalogMappingForDevice(const Common::ParamPackage&) { + return {}; + } }; class NullState : public State { public: - Pollers GetPollers(Polling::DeviceType type) override { - return {}; - } }; std::unique_ptr<State> Init(); diff --git a/src/input_common/sdl/sdl_impl.cpp b/src/input_common/sdl/sdl_impl.cpp index a2e0c0bd2..7827e324c 100644 --- a/src/input_common/sdl/sdl_impl.cpp +++ b/src/input_common/sdl/sdl_impl.cpp @@ -3,10 +3,14 @@ // Refer to the license.txt file included. #include <algorithm> +#include <array> #include <atomic> +#include <chrono> #include <cmath> #include <functional> #include <mutex> +#include <optional> +#include <sstream> #include <string> #include <thread> #include <tuple> @@ -15,15 +19,17 @@ #include <vector> #include <SDL.h> #include "common/logging/log.h" -#include "common/math_util.h" #include "common/param_package.h" #include "common/threadsafe_queue.h" #include "core/frontend/input.h" +#include "input_common/motion_input.h" #include "input_common/sdl/sdl_impl.h" +#include "input_common/settings.h" namespace InputCommon::SDL { -static std::string GetGUID(SDL_Joystick* joystick) { +namespace { +std::string GetGUID(SDL_Joystick* joystick) { const SDL_JoystickGUID guid = SDL_JoystickGetGUID(joystick); char guid_str[33]; SDL_JoystickGetGUIDString(guid, guid_str, sizeof(guid_str)); @@ -31,7 +37,8 @@ static std::string GetGUID(SDL_Joystick* joystick) { } /// Creates a ParamPackage from an SDL_Event that can directly be used to create a ButtonDevice -static Common::ParamPackage SDLEventToButtonParamPackage(SDLState& state, const SDL_Event& event); +Common::ParamPackage SDLEventToButtonParamPackage(SDLState& state, const SDL_Event& event); +} // Anonymous namespace static int SDLEventWatcher(void* user_data, SDL_Event* event) { auto* const sdl_state = static_cast<SDLState*>(user_data); @@ -48,8 +55,10 @@ static int SDLEventWatcher(void* user_data, SDL_Event* event) { class SDLJoystick { public: - SDLJoystick(std::string guid_, int port_, SDL_Joystick* joystick) - : guid{std::move(guid_)}, port{port_}, sdl_joystick{joystick, &SDL_JoystickClose} {} + SDLJoystick(std::string guid_, int port_, SDL_Joystick* joystick, + SDL_GameController* game_controller) + : guid{std::move(guid_)}, port{port_}, sdl_joystick{joystick, &SDL_JoystickClose}, + sdl_controller{game_controller, &SDL_GameControllerClose} {} void SetButton(int button, bool value) { std::lock_guard lock{mutex}; @@ -66,14 +75,24 @@ public: state.axes.insert_or_assign(axis, value); } - float GetAxis(int axis) const { + float GetAxis(int axis, float range) const { std::lock_guard lock{mutex}; - return state.axes.at(axis) / 32767.0f; + return static_cast<float>(state.axes.at(axis)) / (32767.0f * range); } - std::tuple<float, float> GetAnalog(int axis_x, int axis_y) const { - float x = GetAxis(axis_x); - float y = GetAxis(axis_y); + bool RumblePlay(u16 amp_low, u16 amp_high) { + if (sdl_controller) { + return SDL_GameControllerRumble(sdl_controller.get(), amp_low, amp_high, 0) == 0; + } else if (sdl_joystick) { + return SDL_JoystickRumble(sdl_joystick.get(), amp_low, amp_high, 0) == 0; + } + + return false; + } + + std::tuple<float, float> GetAnalog(int axis_x, int axis_y, float range) const { + float x = GetAxis(axis_x, range); + float y = GetAxis(axis_y, range); y = -y; // 3DS uses an y-axis inverse from SDL // Make sure the coordinates are in the unit circle, @@ -88,6 +107,10 @@ public: return std::make_tuple(x, y); } + const MotionInput& GetMotion() const { + return motion; + } + void SetHat(int hat, Uint8 direction) { std::lock_guard lock{mutex}; state.hats.insert_or_assign(hat, direction); @@ -115,8 +138,13 @@ public: return sdl_joystick.get(); } - void SetSDLJoystick(SDL_Joystick* joystick) { + SDL_GameController* GetSDLGameController() const { + return sdl_controller.get(); + } + + void SetSDLJoystick(SDL_Joystick* joystick, SDL_GameController* controller) { sdl_joystick.reset(joystick); + sdl_controller.reset(controller); } private: @@ -128,21 +156,29 @@ private: std::string guid; int port; std::unique_ptr<SDL_Joystick, decltype(&SDL_JoystickClose)> sdl_joystick; + std::unique_ptr<SDL_GameController, decltype(&SDL_GameControllerClose)> sdl_controller; mutable std::mutex mutex; + + // Motion is initialized without PID values as motion input is not aviable for SDL2 + MotionInput motion{0.0f, 0.0f, 0.0f}; }; std::shared_ptr<SDLJoystick> SDLState::GetSDLJoystickByGUID(const std::string& guid, int port) { std::lock_guard lock{joystick_map_mutex}; const auto it = joystick_map.find(guid); + if (it != joystick_map.end()) { while (it->second.size() <= static_cast<std::size_t>(port)) { - auto joystick = - std::make_shared<SDLJoystick>(guid, static_cast<int>(it->second.size()), nullptr); + auto joystick = std::make_shared<SDLJoystick>(guid, static_cast<int>(it->second.size()), + nullptr, nullptr); it->second.emplace_back(std::move(joystick)); } - return it->second[port]; + + return it->second[static_cast<std::size_t>(port)]; } - auto joystick = std::make_shared<SDLJoystick>(guid, 0, nullptr); + + auto joystick = std::make_shared<SDLJoystick>(guid, 0, nullptr, nullptr); + return joystick_map[guid].emplace_back(std::move(joystick)); } @@ -152,86 +188,72 @@ std::shared_ptr<SDLJoystick> SDLState::GetSDLJoystickBySDLID(SDL_JoystickID sdl_ std::lock_guard lock{joystick_map_mutex}; const auto map_it = joystick_map.find(guid); - if (map_it != joystick_map.end()) { - const auto vec_it = - std::find_if(map_it->second.begin(), map_it->second.end(), - [&sdl_joystick](const std::shared_ptr<SDLJoystick>& joystick) { - return sdl_joystick == joystick->GetSDLJoystick(); - }); - if (vec_it != map_it->second.end()) { - // This is the common case: There is already an existing SDL_Joystick maped to a - // SDLJoystick. return the SDLJoystick - return *vec_it; - } - // Search for a SDLJoystick without a mapped SDL_Joystick... - const auto nullptr_it = std::find_if(map_it->second.begin(), map_it->second.end(), - [](const std::shared_ptr<SDLJoystick>& joystick) { - return !joystick->GetSDLJoystick(); - }); - if (nullptr_it != map_it->second.end()) { - // ... and map it - (*nullptr_it)->SetSDLJoystick(sdl_joystick); - return *nullptr_it; - } + if (map_it == joystick_map.end()) { + return nullptr; + } + + const auto vec_it = std::find_if(map_it->second.begin(), map_it->second.end(), + [&sdl_joystick](const auto& joystick) { + return joystick->GetSDLJoystick() == sdl_joystick; + }); - // There is no SDLJoystick without a mapped SDL_Joystick - // Create a new SDLJoystick - const int port = static_cast<int>(map_it->second.size()); - auto joystick = std::make_shared<SDLJoystick>(guid, port, sdl_joystick); - return map_it->second.emplace_back(std::move(joystick)); + if (vec_it == map_it->second.end()) { + return nullptr; } - auto joystick = std::make_shared<SDLJoystick>(guid, 0, sdl_joystick); - return joystick_map[guid].emplace_back(std::move(joystick)); + return *vec_it; } void SDLState::InitJoystick(int joystick_index) { SDL_Joystick* sdl_joystick = SDL_JoystickOpen(joystick_index); + SDL_GameController* sdl_gamecontroller = nullptr; + + if (SDL_IsGameController(joystick_index)) { + sdl_gamecontroller = SDL_GameControllerOpen(joystick_index); + } + if (!sdl_joystick) { - LOG_ERROR(Input, "failed to open joystick {}", joystick_index); + LOG_ERROR(Input, "Failed to open joystick {}", joystick_index); return; } + const std::string guid = GetGUID(sdl_joystick); std::lock_guard lock{joystick_map_mutex}; if (joystick_map.find(guid) == joystick_map.end()) { - auto joystick = std::make_shared<SDLJoystick>(guid, 0, sdl_joystick); + auto joystick = std::make_shared<SDLJoystick>(guid, 0, sdl_joystick, sdl_gamecontroller); joystick_map[guid].emplace_back(std::move(joystick)); return; } + auto& joystick_guid_list = joystick_map[guid]; - const auto it = std::find_if( - joystick_guid_list.begin(), joystick_guid_list.end(), - [](const std::shared_ptr<SDLJoystick>& joystick) { return !joystick->GetSDLJoystick(); }); - if (it != joystick_guid_list.end()) { - (*it)->SetSDLJoystick(sdl_joystick); + const auto joystick_it = + std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(), + [](const auto& joystick) { return !joystick->GetSDLJoystick(); }); + + if (joystick_it != joystick_guid_list.end()) { + (*joystick_it)->SetSDLJoystick(sdl_joystick, sdl_gamecontroller); return; } + const int port = static_cast<int>(joystick_guid_list.size()); - auto joystick = std::make_shared<SDLJoystick>(guid, port, sdl_joystick); + auto joystick = std::make_shared<SDLJoystick>(guid, port, sdl_joystick, sdl_gamecontroller); joystick_guid_list.emplace_back(std::move(joystick)); } void SDLState::CloseJoystick(SDL_Joystick* sdl_joystick) { const std::string guid = GetGUID(sdl_joystick); - std::shared_ptr<SDLJoystick> joystick; - { - std::lock_guard lock{joystick_map_mutex}; - // This call to guid is safe since the joystick is guaranteed to be in the map - const auto& joystick_guid_list = joystick_map[guid]; - const auto joystick_it = - std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(), - [&sdl_joystick](const std::shared_ptr<SDLJoystick>& joystick) { - return joystick->GetSDLJoystick() == sdl_joystick; - }); - joystick = *joystick_it; - } - - // Destruct SDL_Joystick outside the lock guard because SDL can internally call the - // event callback which locks the mutex again. - joystick->SetSDLJoystick(nullptr); + std::lock_guard lock{joystick_map_mutex}; + // This call to guid is safe since the joystick is guaranteed to be in the map + const auto& joystick_guid_list = joystick_map[guid]; + const auto joystick_it = std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(), + [&sdl_joystick](const auto& joystick) { + return joystick->GetSDLJoystick() == sdl_joystick; + }); + + (*joystick_it)->SetSDLJoystick(nullptr, nullptr); } void SDLState::HandleGameControllerEvent(const SDL_Event& event) { @@ -313,7 +335,7 @@ public: trigger_if_greater(trigger_if_greater_) {} bool GetStatus() const override { - const float axis_value = joystick->GetAxis(axis); + const float axis_value = joystick->GetAxis(axis, 1.0f); if (trigger_if_greater) { return axis_value > threshold; } @@ -329,22 +351,24 @@ private: class SDLAnalog final : public Input::AnalogDevice { public: - SDLAnalog(std::shared_ptr<SDLJoystick> joystick_, int axis_x_, int axis_y_, float deadzone_) - : joystick(std::move(joystick_)), axis_x(axis_x_), axis_y(axis_y_), deadzone(deadzone_) {} + explicit SDLAnalog(std::shared_ptr<SDLJoystick> joystick_, int axis_x_, int axis_y_, + float deadzone_, float range_) + : joystick(std::move(joystick_)), axis_x(axis_x_), axis_y(axis_y_), deadzone(deadzone_), + range(range_) {} std::tuple<float, float> GetStatus() const override { - const auto [x, y] = joystick->GetAnalog(axis_x, axis_y); + const auto [x, y] = joystick->GetAnalog(axis_x, axis_y, range); const float r = std::sqrt((x * x) + (y * y)); if (r > deadzone) { return std::make_tuple(x / r * (r - deadzone) / (1 - deadzone), y / r * (r - deadzone) / (1 - deadzone)); } - return std::make_tuple<float, float>(0.0f, 0.0f); + return {}; } bool GetAnalogDirectionStatus(Input::AnalogDirection direction) const override { const auto [x, y] = GetStatus(); - const float directional_deadzone = 0.4f; + const float directional_deadzone = 0.5f; switch (direction) { case Input::AnalogDirection::RIGHT: return x > directional_deadzone; @@ -363,6 +387,95 @@ private: const int axis_x; const int axis_y; const float deadzone; + const float range; +}; + +class SDLVibration final : public Input::VibrationDevice { +public: + explicit SDLVibration(std::shared_ptr<SDLJoystick> joystick_) + : joystick(std::move(joystick_)) {} + + u8 GetStatus() const override { + joystick->RumblePlay(1, 1); + return joystick->RumblePlay(0, 0); + } + + bool SetRumblePlay(f32 amp_low, [[maybe_unused]] f32 freq_low, f32 amp_high, + [[maybe_unused]] f32 freq_high) const override { + const auto process_amplitude = [](f32 amplitude) { + return static_cast<u16>((amplitude + std::pow(amplitude, 0.3f)) * 0.5f * 0xFFFF); + }; + + const auto processed_amp_low = process_amplitude(amp_low); + const auto processed_amp_high = process_amplitude(amp_high); + + return joystick->RumblePlay(processed_amp_low, processed_amp_high); + } + +private: + std::shared_ptr<SDLJoystick> joystick; +}; + +class SDLDirectionMotion final : public Input::MotionDevice { +public: + explicit SDLDirectionMotion(std::shared_ptr<SDLJoystick> joystick_, int hat_, Uint8 direction_) + : joystick(std::move(joystick_)), hat(hat_), direction(direction_) {} + + Input::MotionStatus GetStatus() const override { + if (joystick->GetHatDirection(hat, direction)) { + return joystick->GetMotion().GetRandomMotion(2, 6); + } + return joystick->GetMotion().GetRandomMotion(0, 0); + } + +private: + std::shared_ptr<SDLJoystick> joystick; + int hat; + Uint8 direction; +}; + +class SDLAxisMotion final : public Input::MotionDevice { +public: + explicit SDLAxisMotion(std::shared_ptr<SDLJoystick> joystick_, int axis_, float threshold_, + bool trigger_if_greater_) + : joystick(std::move(joystick_)), axis(axis_), threshold(threshold_), + trigger_if_greater(trigger_if_greater_) {} + + Input::MotionStatus GetStatus() const override { + const float axis_value = joystick->GetAxis(axis, 1.0f); + bool trigger = axis_value < threshold; + if (trigger_if_greater) { + trigger = axis_value > threshold; + } + + if (trigger) { + return joystick->GetMotion().GetRandomMotion(2, 6); + } + return joystick->GetMotion().GetRandomMotion(0, 0); + } + +private: + std::shared_ptr<SDLJoystick> joystick; + int axis; + float threshold; + bool trigger_if_greater; +}; + +class SDLButtonMotion final : public Input::MotionDevice { +public: + explicit SDLButtonMotion(std::shared_ptr<SDLJoystick> joystick_, int button_) + : joystick(std::move(joystick_)), button(button_) {} + + Input::MotionStatus GetStatus() const override { + if (joystick->GetButton(button)) { + return joystick->GetMotion().GetRandomMotion(2, 6); + } + return joystick->GetMotion().GetRandomMotion(0, 0); + } + +private: + std::shared_ptr<SDLJoystick> joystick; + int button; }; /// A button device factory that creates button devices from SDL joystick @@ -445,7 +558,7 @@ class SDLAnalogFactory final : public Input::Factory<Input::AnalogDevice> { public: explicit SDLAnalogFactory(SDLState& state_) : state(state_) {} /** - * Creates analog device from joystick axes + * Creates an analog device from joystick axes * @param params contains parameters for creating the device: * - "guid": the guid of the joystick to bind * - "port": the nth joystick of the same type @@ -457,14 +570,98 @@ public: const int port = params.Get("port", 0); const int axis_x = params.Get("axis_x", 0); const int axis_y = params.Get("axis_y", 1); - const float deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, .99f); - + const float deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, 1.0f); + const float range = std::clamp(params.Get("range", 1.0f), 0.50f, 1.50f); auto joystick = state.GetSDLJoystickByGUID(guid, port); // This is necessary so accessing GetAxis with axis_x and axis_y won't crash joystick->SetAxis(axis_x, 0); joystick->SetAxis(axis_y, 0); - return std::make_unique<SDLAnalog>(joystick, axis_x, axis_y, deadzone); + return std::make_unique<SDLAnalog>(joystick, axis_x, axis_y, deadzone, range); + } + +private: + SDLState& state; +}; + +/// An vibration device factory that creates vibration devices from SDL joystick +class SDLVibrationFactory final : public Input::Factory<Input::VibrationDevice> { +public: + explicit SDLVibrationFactory(SDLState& state_) : state(state_) {} + /** + * Creates a vibration device from a joystick + * @param params contains parameters for creating the device: + * - "guid": the guid of the joystick to bind + * - "port": the nth joystick of the same type + */ + std::unique_ptr<Input::VibrationDevice> Create(const Common::ParamPackage& params) override { + const std::string guid = params.Get("guid", "0"); + const int port = params.Get("port", 0); + return std::make_unique<SDLVibration>(state.GetSDLJoystickByGUID(guid, port)); + } + +private: + SDLState& state; +}; + +/// A motion device factory that creates motion devices from SDL joystick +class SDLMotionFactory final : public Input::Factory<Input::MotionDevice> { +public: + explicit SDLMotionFactory(SDLState& state_) : state(state_) {} + /** + * Creates motion device from joystick axes + * @param params contains parameters for creating the device: + * - "guid": the guid of the joystick to bind + * - "port": the nth joystick of the same type + */ + std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override { + const std::string guid = params.Get("guid", "0"); + const int port = params.Get("port", 0); + + auto joystick = state.GetSDLJoystickByGUID(guid, port); + + if (params.Has("hat")) { + const int hat = params.Get("hat", 0); + const std::string direction_name = params.Get("direction", ""); + Uint8 direction; + if (direction_name == "up") { + direction = SDL_HAT_UP; + } else if (direction_name == "down") { + direction = SDL_HAT_DOWN; + } else if (direction_name == "left") { + direction = SDL_HAT_LEFT; + } else if (direction_name == "right") { + direction = SDL_HAT_RIGHT; + } else { + direction = 0; + } + // This is necessary so accessing GetHat with hat won't crash + joystick->SetHat(hat, SDL_HAT_CENTERED); + return std::make_unique<SDLDirectionMotion>(joystick, hat, direction); + } + + if (params.Has("axis")) { + const int axis = params.Get("axis", 0); + const float threshold = params.Get("threshold", 0.5f); + const std::string direction_name = params.Get("direction", ""); + bool trigger_if_greater; + if (direction_name == "+") { + trigger_if_greater = true; + } else if (direction_name == "-") { + trigger_if_greater = false; + } else { + trigger_if_greater = true; + LOG_ERROR(Input, "Unknown direction {}", direction_name); + } + // This is necessary so accessing GetAxis with axis won't crash + joystick->SetAxis(axis, 0); + return std::make_unique<SDLAxisMotion>(joystick, axis, threshold, trigger_if_greater); + } + + const int button = params.Get("button", 0); + // This is necessary so accessing GetButton with button won't crash + joystick->SetButton(button, false); + return std::make_unique<SDLButtonMotion>(joystick, button); } private: @@ -473,15 +670,22 @@ private: SDLState::SDLState() { using namespace Input; - RegisterFactory<ButtonDevice>("sdl", std::make_shared<SDLButtonFactory>(*this)); - RegisterFactory<AnalogDevice>("sdl", std::make_shared<SDLAnalogFactory>(*this)); - - // If the frontend is going to manage the event loop, then we dont start one here - start_thread = !SDL_WasInit(SDL_INIT_JOYSTICK); + button_factory = std::make_shared<SDLButtonFactory>(*this); + analog_factory = std::make_shared<SDLAnalogFactory>(*this); + vibration_factory = std::make_shared<SDLVibrationFactory>(*this); + motion_factory = std::make_shared<SDLMotionFactory>(*this); + RegisterFactory<ButtonDevice>("sdl", button_factory); + RegisterFactory<AnalogDevice>("sdl", analog_factory); + RegisterFactory<VibrationDevice>("sdl", vibration_factory); + RegisterFactory<MotionDevice>("sdl", motion_factory); + + // If the frontend is going to manage the event loop, then we don't start one here + start_thread = SDL_WasInit(SDL_INIT_JOYSTICK) == 0; if (start_thread && SDL_Init(SDL_INIT_JOYSTICK) < 0) { LOG_CRITICAL(Input, "SDL_Init(SDL_INIT_JOYSTICK) failed with: {}", SDL_GetError()); return; } + has_gamecontroller = SDL_InitSubSystem(SDL_INIT_GAMECONTROLLER) != 0; if (SDL_SetHint(SDL_HINT_JOYSTICK_ALLOW_BACKGROUND_EVENTS, "1") == SDL_FALSE) { LOG_ERROR(Input, "Failed to set hint for background events with: {}", SDL_GetError()); } @@ -494,7 +698,7 @@ SDLState::SDLState() { using namespace std::chrono_literals; while (initialized) { SDL_PumpEvents(); - std::this_thread::sleep_for(10ms); + std::this_thread::sleep_for(1ms); } }); } @@ -509,6 +713,8 @@ SDLState::~SDLState() { using namespace Input; UnregisterFactory<ButtonDevice>("sdl"); UnregisterFactory<AnalogDevice>("sdl"); + UnregisterFactory<VibrationDevice>("sdl"); + UnregisterFactory<MotionDevice>("sdl"); CloseJoysticks(); SDL_DelEventWatch(&SDLEventWatcher, this); @@ -520,65 +726,268 @@ SDLState::~SDLState() { } } -static Common::ParamPackage SDLEventToButtonParamPackage(SDLState& state, const SDL_Event& event) { +std::vector<Common::ParamPackage> SDLState::GetInputDevices() { + std::scoped_lock lock(joystick_map_mutex); + std::vector<Common::ParamPackage> devices; + for (const auto& [key, value] : joystick_map) { + for (const auto& joystick : value) { + if (auto* const controller = joystick->GetSDLGameController()) { + std::string name = + fmt::format("{} {}", SDL_GameControllerName(controller), joystick->GetPort()); + devices.emplace_back(Common::ParamPackage{ + {"class", "sdl"}, + {"display", std::move(name)}, + {"guid", joystick->GetGUID()}, + {"port", std::to_string(joystick->GetPort())}, + }); + } else if (auto* const joy = joystick->GetSDLJoystick()) { + std::string name = fmt::format("{} {}", SDL_JoystickName(joy), joystick->GetPort()); + devices.emplace_back(Common::ParamPackage{ + {"class", "sdl"}, + {"display", std::move(name)}, + {"guid", joystick->GetGUID()}, + {"port", std::to_string(joystick->GetPort())}, + }); + } + } + } + return devices; +} + +namespace { +Common::ParamPackage BuildAnalogParamPackageForButton(int port, std::string guid, s32 axis, + float value = 0.1f) { + Common::ParamPackage params({{"engine", "sdl"}}); + params.Set("port", port); + params.Set("guid", std::move(guid)); + params.Set("axis", axis); + if (value > 0) { + params.Set("direction", "+"); + params.Set("threshold", "0.5"); + } else { + params.Set("direction", "-"); + params.Set("threshold", "-0.5"); + } + return params; +} + +Common::ParamPackage BuildButtonParamPackageForButton(int port, std::string guid, s32 button) { + Common::ParamPackage params({{"engine", "sdl"}}); + params.Set("port", port); + params.Set("guid", std::move(guid)); + params.Set("button", button); + return params; +} + +Common::ParamPackage BuildHatParamPackageForButton(int port, std::string guid, s32 hat, s32 value) { Common::ParamPackage params({{"engine", "sdl"}}); + params.Set("port", port); + params.Set("guid", std::move(guid)); + params.Set("hat", hat); + switch (value) { + case SDL_HAT_UP: + params.Set("direction", "up"); + break; + case SDL_HAT_DOWN: + params.Set("direction", "down"); + break; + case SDL_HAT_LEFT: + params.Set("direction", "left"); + break; + case SDL_HAT_RIGHT: + params.Set("direction", "right"); + break; + default: + return {}; + } + return params; +} + +Common::ParamPackage SDLEventToButtonParamPackage(SDLState& state, const SDL_Event& event) { + switch (event.type) { + case SDL_JOYAXISMOTION: { + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which)) { + return BuildAnalogParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jaxis.axis), + event.jaxis.value); + } + break; + } + case SDL_JOYBUTTONUP: { + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jbutton.which)) { + return BuildButtonParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jbutton.button)); + } + break; + } + case SDL_JOYHATMOTION: { + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jhat.which)) { + return BuildHatParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jhat.hat), + static_cast<s32>(event.jhat.value)); + } + break; + } + } + return {}; +} + +Common::ParamPackage SDLEventToMotionParamPackage(SDLState& state, const SDL_Event& event) { switch (event.type) { case SDL_JOYAXISMOTION: { - const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which); - params.Set("port", joystick->GetPort()); - params.Set("guid", joystick->GetGUID()); - params.Set("axis", event.jaxis.axis); - if (event.jaxis.value > 0) { - params.Set("direction", "+"); - params.Set("threshold", "0.5"); - } else { - params.Set("direction", "-"); - params.Set("threshold", "-0.5"); + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which)) { + return BuildAnalogParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jaxis.axis), + event.jaxis.value); } break; } case SDL_JOYBUTTONUP: { - const auto joystick = state.GetSDLJoystickBySDLID(event.jbutton.which); - params.Set("port", joystick->GetPort()); - params.Set("guid", joystick->GetGUID()); - params.Set("button", event.jbutton.button); + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jbutton.which)) { + return BuildButtonParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jbutton.button)); + } break; } case SDL_JOYHATMOTION: { - const auto joystick = state.GetSDLJoystickBySDLID(event.jhat.which); - params.Set("port", joystick->GetPort()); - params.Set("guid", joystick->GetGUID()); - params.Set("hat", event.jhat.hat); - switch (event.jhat.value) { - case SDL_HAT_UP: - params.Set("direction", "up"); - break; - case SDL_HAT_DOWN: - params.Set("direction", "down"); - break; - case SDL_HAT_LEFT: - params.Set("direction", "left"); - break; - case SDL_HAT_RIGHT: - params.Set("direction", "right"); - break; - default: - return {}; + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jhat.which)) { + return BuildHatParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jhat.hat), + static_cast<s32>(event.jhat.value)); } break; } } + return {}; +} + +Common::ParamPackage BuildParamPackageForBinding(int port, const std::string& guid, + const SDL_GameControllerButtonBind& binding) { + switch (binding.bindType) { + case SDL_CONTROLLER_BINDTYPE_NONE: + break; + case SDL_CONTROLLER_BINDTYPE_AXIS: + return BuildAnalogParamPackageForButton(port, guid, binding.value.axis); + case SDL_CONTROLLER_BINDTYPE_BUTTON: + return BuildButtonParamPackageForButton(port, guid, binding.value.button); + case SDL_CONTROLLER_BINDTYPE_HAT: + return BuildHatParamPackageForButton(port, guid, binding.value.hat.hat, + binding.value.hat.hat_mask); + } + return {}; +} + +Common::ParamPackage BuildParamPackageForAnalog(int port, const std::string& guid, int axis_x, + int axis_y) { + Common::ParamPackage params; + params.Set("engine", "sdl"); + params.Set("port", port); + params.Set("guid", guid); + params.Set("axis_x", axis_x); + params.Set("axis_y", axis_y); return params; } +} // Anonymous namespace -namespace Polling { +ButtonMapping SDLState::GetButtonMappingForDevice(const Common::ParamPackage& params) { + if (!params.Has("guid") || !params.Has("port")) { + return {}; + } + const auto joystick = GetSDLJoystickByGUID(params.Get("guid", ""), params.Get("port", 0)); + auto* controller = joystick->GetSDLGameController(); + if (controller == nullptr) { + return {}; + } + // This list is missing ZL/ZR since those are not considered buttons in SDL GameController. + // We will add those afterwards + // This list also excludes Screenshot since theres not really a mapping for that + using ButtonBindings = + std::array<std::pair<Settings::NativeButton::Values, SDL_GameControllerButton>, 17>; + static constexpr ButtonBindings switch_to_sdl_button{{ + {Settings::NativeButton::A, SDL_CONTROLLER_BUTTON_B}, + {Settings::NativeButton::B, SDL_CONTROLLER_BUTTON_A}, + {Settings::NativeButton::X, SDL_CONTROLLER_BUTTON_Y}, + {Settings::NativeButton::Y, SDL_CONTROLLER_BUTTON_X}, + {Settings::NativeButton::LStick, SDL_CONTROLLER_BUTTON_LEFTSTICK}, + {Settings::NativeButton::RStick, SDL_CONTROLLER_BUTTON_RIGHTSTICK}, + {Settings::NativeButton::L, SDL_CONTROLLER_BUTTON_LEFTSHOULDER}, + {Settings::NativeButton::R, SDL_CONTROLLER_BUTTON_RIGHTSHOULDER}, + {Settings::NativeButton::Plus, SDL_CONTROLLER_BUTTON_START}, + {Settings::NativeButton::Minus, SDL_CONTROLLER_BUTTON_BACK}, + {Settings::NativeButton::DLeft, SDL_CONTROLLER_BUTTON_DPAD_LEFT}, + {Settings::NativeButton::DUp, SDL_CONTROLLER_BUTTON_DPAD_UP}, + {Settings::NativeButton::DRight, SDL_CONTROLLER_BUTTON_DPAD_RIGHT}, + {Settings::NativeButton::DDown, SDL_CONTROLLER_BUTTON_DPAD_DOWN}, + {Settings::NativeButton::SL, SDL_CONTROLLER_BUTTON_LEFTSHOULDER}, + {Settings::NativeButton::SR, SDL_CONTROLLER_BUTTON_RIGHTSHOULDER}, + {Settings::NativeButton::Home, SDL_CONTROLLER_BUTTON_GUIDE}, + }}; + + // Add the missing bindings for ZL/ZR + using ZBindings = + std::array<std::pair<Settings::NativeButton::Values, SDL_GameControllerAxis>, 2>; + static constexpr ZBindings switch_to_sdl_axis{{ + {Settings::NativeButton::ZL, SDL_CONTROLLER_AXIS_TRIGGERLEFT}, + {Settings::NativeButton::ZR, SDL_CONTROLLER_AXIS_TRIGGERRIGHT}, + }}; + + ButtonMapping mapping; + mapping.reserve(switch_to_sdl_button.size() + switch_to_sdl_axis.size()); + + for (const auto& [switch_button, sdl_button] : switch_to_sdl_button) { + const auto& binding = SDL_GameControllerGetBindForButton(controller, sdl_button); + mapping.insert_or_assign( + switch_button, + BuildParamPackageForBinding(joystick->GetPort(), joystick->GetGUID(), binding)); + } + for (const auto& [switch_button, sdl_axis] : switch_to_sdl_axis) { + const auto& binding = SDL_GameControllerGetBindForAxis(controller, sdl_axis); + mapping.insert_or_assign( + switch_button, + BuildParamPackageForBinding(joystick->GetPort(), joystick->GetGUID(), binding)); + } + + return mapping; +} + +AnalogMapping SDLState::GetAnalogMappingForDevice(const Common::ParamPackage& params) { + if (!params.Has("guid") || !params.Has("port")) { + return {}; + } + const auto joystick = GetSDLJoystickByGUID(params.Get("guid", ""), params.Get("port", 0)); + auto* controller = joystick->GetSDLGameController(); + if (controller == nullptr) { + return {}; + } + + AnalogMapping mapping = {}; + const auto& binding_left_x = + SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_LEFTX); + const auto& binding_left_y = + SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_LEFTY); + mapping.insert_or_assign(Settings::NativeAnalog::LStick, + BuildParamPackageForAnalog(joystick->GetPort(), joystick->GetGUID(), + binding_left_x.value.axis, + binding_left_y.value.axis)); + const auto& binding_right_x = + SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_RIGHTX); + const auto& binding_right_y = + SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_RIGHTY); + mapping.insert_or_assign(Settings::NativeAnalog::RStick, + BuildParamPackageForAnalog(joystick->GetPort(), joystick->GetGUID(), + binding_right_x.value.axis, + binding_right_y.value.axis)); + return mapping; +} + +namespace Polling { class SDLPoller : public InputCommon::Polling::DevicePoller { public: explicit SDLPoller(SDLState& state_) : state(state_) {} - void Start() override { + void Start([[maybe_unused]] const std::string& device_id) override { state.event_queue.Clear(); state.polling = true; } @@ -598,70 +1007,116 @@ public: Common::ParamPackage GetNextInput() override { SDL_Event event; while (state.event_queue.Pop(event)) { - switch (event.type) { - case SDL_JOYAXISMOTION: - if (std::abs(event.jaxis.value / 32767.0) < 0.5) { - break; - } - case SDL_JOYBUTTONUP: - case SDL_JOYHATMOTION: - return SDLEventToButtonParamPackage(state, event); + const auto package = FromEvent(event); + if (package) { + return *package; } } return {}; } + [[nodiscard]] std::optional<Common::ParamPackage> FromEvent(const SDL_Event& event) const { + switch (event.type) { + case SDL_JOYAXISMOTION: + if (std::abs(event.jaxis.value / 32767.0) < 0.5) { + break; + } + [[fallthrough]]; + case SDL_JOYBUTTONUP: + case SDL_JOYHATMOTION: + return {SDLEventToButtonParamPackage(state, event)}; + } + return std::nullopt; + } }; -class SDLAnalogPoller final : public SDLPoller { +class SDLMotionPoller final : public SDLPoller { public: - explicit SDLAnalogPoller(SDLState& state_) : SDLPoller(state_) {} + explicit SDLMotionPoller(SDLState& state_) : SDLPoller(state_) {} + + Common::ParamPackage GetNextInput() override { + SDL_Event event; + while (state.event_queue.Pop(event)) { + const auto package = FromEvent(event); + if (package) { + return *package; + } + } + return {}; + } + [[nodiscard]] std::optional<Common::ParamPackage> FromEvent(const SDL_Event& event) const { + switch (event.type) { + case SDL_JOYAXISMOTION: + if (std::abs(event.jaxis.value / 32767.0) < 0.5) { + break; + } + [[fallthrough]]; + case SDL_JOYBUTTONUP: + case SDL_JOYHATMOTION: + return {SDLEventToMotionParamPackage(state, event)}; + } + return std::nullopt; + } +}; - void Start() override { - SDLPoller::Start(); +/** + * Attempts to match the press to a controller joy axis (left/right stick) and if a match + * isn't found, checks if the event matches anything from SDLButtonPoller and uses that + * instead + */ +class SDLAnalogPreferredPoller final : public SDLPoller { +public: + explicit SDLAnalogPreferredPoller(SDLState& state_) + : SDLPoller(state_), button_poller(state_) {} + void Start(const std::string& device_id) override { + SDLPoller::Start(device_id); // Reset stored axes analog_x_axis = -1; analog_y_axis = -1; - analog_axes_joystick = -1; } Common::ParamPackage GetNextInput() override { SDL_Event event; while (state.event_queue.Pop(event)) { - if (event.type != SDL_JOYAXISMOTION || std::abs(event.jaxis.value / 32767.0) < 0.5) { + // Filter out axis events that are below a threshold + if (event.type == SDL_JOYAXISMOTION && std::abs(event.jaxis.value / 32767.0) < 0.5) { continue; } - // An analog device needs two axes, so we need to store the axis for later and wait for - // a second SDL event. The axes also must be from the same joystick. - const int axis = event.jaxis.axis; - if (analog_x_axis == -1) { - analog_x_axis = axis; - analog_axes_joystick = event.jaxis.which; - } else if (analog_y_axis == -1 && analog_x_axis != axis && - analog_axes_joystick == event.jaxis.which) { - analog_y_axis = axis; + if (event.type == SDL_JOYAXISMOTION) { + const auto axis = event.jaxis.axis; + // In order to return a complete analog param, we need inputs for both axes. + // First we take the x-axis (horizontal) input, then the y-axis (vertical) input. + if (analog_x_axis == -1) { + analog_x_axis = axis; + } else if (analog_y_axis == -1 && analog_x_axis != axis) { + analog_y_axis = axis; + } + } else { + // If the press wasn't accepted as a joy axis, check for a button press + auto button_press = button_poller.FromEvent(event); + if (button_press) { + return *button_press; + } } } - Common::ParamPackage params; + if (analog_x_axis != -1 && analog_y_axis != -1) { - const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which); - params.Set("engine", "sdl"); - params.Set("port", joystick->GetPort()); - params.Set("guid", joystick->GetGUID()); - params.Set("axis_x", analog_x_axis); - params.Set("axis_y", analog_y_axis); - analog_x_axis = -1; - analog_y_axis = -1; - analog_axes_joystick = -1; - return params; + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which)) { + auto params = BuildParamPackageForAnalog(joystick->GetPort(), joystick->GetGUID(), + analog_x_axis, analog_y_axis); + analog_x_axis = -1; + analog_y_axis = -1; + return params; + } } - return params; + + return {}; } private: int analog_x_axis = -1; int analog_y_axis = -1; - SDL_JoystickID analog_axes_joystick = -1; + SDLButtonPoller button_poller; }; } // namespace Polling @@ -669,12 +1124,15 @@ SDLState::Pollers SDLState::GetPollers(InputCommon::Polling::DeviceType type) { Pollers pollers; switch (type) { - case InputCommon::Polling::DeviceType::Analog: - pollers.emplace_back(std::make_unique<Polling::SDLAnalogPoller>(*this)); + case InputCommon::Polling::DeviceType::AnalogPreferred: + pollers.emplace_back(std::make_unique<Polling::SDLAnalogPreferredPoller>(*this)); break; case InputCommon::Polling::DeviceType::Button: pollers.emplace_back(std::make_unique<Polling::SDLButtonPoller>(*this)); break; + case InputCommon::Polling::DeviceType::Motion: + pollers.emplace_back(std::make_unique<Polling::SDLMotionPoller>(*this)); + break; } return pollers; diff --git a/src/input_common/sdl/sdl_impl.h b/src/input_common/sdl/sdl_impl.h index 606a32c5b..08044b00d 100644 --- a/src/input_common/sdl/sdl_impl.h +++ b/src/input_common/sdl/sdl_impl.h @@ -21,6 +21,8 @@ namespace InputCommon::SDL { class SDLAnalogFactory; class SDLButtonFactory; +class SDLMotionFactory; +class SDLVibrationFactory; class SDLJoystick; class SDLState : public State { @@ -50,6 +52,11 @@ public: std::atomic<bool> polling = false; Common::SPSCQueue<SDL_Event> event_queue; + std::vector<Common::ParamPackage> GetInputDevices() override; + + ButtonMapping GetButtonMappingForDevice(const Common::ParamPackage& params) override; + AnalogMapping GetAnalogMappingForDevice(const Common::ParamPackage& params) override; + private: void InitJoystick(int joystick_index); void CloseJoystick(SDL_Joystick* sdl_joystick); @@ -57,12 +64,17 @@ private: /// Needs to be called before SDL_QuitSubSystem. void CloseJoysticks(); + // Set to true if SDL supports game controller subsystem + bool has_gamecontroller = false; + /// Map of GUID of a list of corresponding virtual Joysticks std::unordered_map<std::string, std::vector<std::shared_ptr<SDLJoystick>>> joystick_map; std::mutex joystick_map_mutex; std::shared_ptr<SDLButtonFactory> button_factory; std::shared_ptr<SDLAnalogFactory> analog_factory; + std::shared_ptr<SDLVibrationFactory> vibration_factory; + std::shared_ptr<SDLMotionFactory> motion_factory; bool start_thread = false; std::atomic<bool> initialized = false; diff --git a/src/input_common/settings.cpp b/src/input_common/settings.cpp new file mode 100644 index 000000000..557e7a9a0 --- /dev/null +++ b/src/input_common/settings.cpp @@ -0,0 +1,47 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "input_common/settings.h" + +namespace Settings { +namespace NativeButton { +const std::array<const char*, NumButtons> mapping = {{ + "button_a", "button_b", "button_x", "button_y", "button_lstick", + "button_rstick", "button_l", "button_r", "button_zl", "button_zr", + "button_plus", "button_minus", "button_dleft", "button_dup", "button_dright", + "button_ddown", "button_sl", "button_sr", "button_home", "button_screenshot", +}}; +} + +namespace NativeAnalog { +const std::array<const char*, NumAnalogs> mapping = {{ + "lstick", + "rstick", +}}; +} + +namespace NativeVibration { +const std::array<const char*, NumVibrations> mapping = {{ + "left_vibration_device", + "right_vibration_device", +}}; +} + +namespace NativeMotion { +const std::array<const char*, NumMotions> mapping = {{ + "motionleft", + "motionright", +}}; +} + +namespace NativeMouseButton { +const std::array<const char*, NumMouseButtons> mapping = {{ + "left", + "right", + "middle", + "forward", + "back", +}}; +} +} // namespace Settings diff --git a/src/input_common/settings.h b/src/input_common/settings.h new file mode 100644 index 000000000..75486554b --- /dev/null +++ b/src/input_common/settings.h @@ -0,0 +1,371 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <string> +#include "common/common_types.h" + +namespace Settings { +namespace NativeButton { +enum Values : int { + A, + B, + X, + Y, + LStick, + RStick, + L, + R, + ZL, + ZR, + Plus, + Minus, + + DLeft, + DUp, + DRight, + DDown, + + SL, + SR, + + Home, + Screenshot, + + NumButtons, +}; + +constexpr int BUTTON_HID_BEGIN = A; +constexpr int BUTTON_NS_BEGIN = Home; + +constexpr int BUTTON_HID_END = BUTTON_NS_BEGIN; +constexpr int BUTTON_NS_END = NumButtons; + +constexpr int NUM_BUTTONS_HID = BUTTON_HID_END - BUTTON_HID_BEGIN; +constexpr int NUM_BUTTONS_NS = BUTTON_NS_END - BUTTON_NS_BEGIN; + +extern const std::array<const char*, NumButtons> mapping; + +} // namespace NativeButton + +namespace NativeAnalog { +enum Values : int { + LStick, + RStick, + + NumAnalogs, +}; + +constexpr int STICK_HID_BEGIN = LStick; +constexpr int STICK_HID_END = NumAnalogs; +constexpr int NUM_STICKS_HID = NumAnalogs; + +extern const std::array<const char*, NumAnalogs> mapping; +} // namespace NativeAnalog + +namespace NativeVibration { +enum Values : int { + LeftVibrationDevice, + RightVibrationDevice, + + NumVibrations, +}; + +constexpr int VIBRATION_HID_BEGIN = LeftVibrationDevice; +constexpr int VIBRATION_HID_END = NumVibrations; +constexpr int NUM_VIBRATIONS_HID = NumVibrations; + +extern const std::array<const char*, NumVibrations> mapping; +}; // namespace NativeVibration + +namespace NativeMotion { +enum Values : int { + MotionLeft, + MotionRight, + + NumMotions, +}; + +constexpr int MOTION_HID_BEGIN = MotionLeft; +constexpr int MOTION_HID_END = NumMotions; +constexpr int NUM_MOTIONS_HID = NumMotions; + +extern const std::array<const char*, NumMotions> mapping; +} // namespace NativeMotion + +namespace NativeMouseButton { +enum Values { + Left, + Right, + Middle, + Forward, + Back, + + NumMouseButtons, +}; + +constexpr int MOUSE_HID_BEGIN = Left; +constexpr int MOUSE_HID_END = NumMouseButtons; +constexpr int NUM_MOUSE_HID = NumMouseButtons; + +extern const std::array<const char*, NumMouseButtons> mapping; +} // namespace NativeMouseButton + +namespace NativeKeyboard { +enum Keys { + None, + Error, + + A = 4, + B, + C, + D, + E, + F, + G, + H, + I, + J, + K, + L, + M, + N, + O, + P, + Q, + R, + S, + T, + U, + V, + W, + X, + Y, + Z, + N1, + N2, + N3, + N4, + N5, + N6, + N7, + N8, + N9, + N0, + Enter, + Escape, + Backspace, + Tab, + Space, + Minus, + Equal, + LeftBrace, + RightBrace, + Backslash, + Tilde, + Semicolon, + Apostrophe, + Grave, + Comma, + Dot, + Slash, + CapsLockKey, + + F1, + F2, + F3, + F4, + F5, + F6, + F7, + F8, + F9, + F10, + F11, + F12, + + SystemRequest, + ScrollLockKey, + Pause, + Insert, + Home, + PageUp, + Delete, + End, + PageDown, + Right, + Left, + Down, + Up, + + NumLockKey, + KPSlash, + KPAsterisk, + KPMinus, + KPPlus, + KPEnter, + KP1, + KP2, + KP3, + KP4, + KP5, + KP6, + KP7, + KP8, + KP9, + KP0, + KPDot, + + Key102, + Compose, + Power, + KPEqual, + + F13, + F14, + F15, + F16, + F17, + F18, + F19, + F20, + F21, + F22, + F23, + F24, + + Open, + Help, + Properties, + Front, + Stop, + Repeat, + Undo, + Cut, + Copy, + Paste, + Find, + Mute, + VolumeUp, + VolumeDown, + CapsLockActive, + NumLockActive, + ScrollLockActive, + KPComma, + + KPLeftParenthesis, + KPRightParenthesis, + + LeftControlKey = 0xE0, + LeftShiftKey, + LeftAltKey, + LeftMetaKey, + RightControlKey, + RightShiftKey, + RightAltKey, + RightMetaKey, + + MediaPlayPause, + MediaStopCD, + MediaPrevious, + MediaNext, + MediaEject, + MediaVolumeUp, + MediaVolumeDown, + MediaMute, + MediaWebsite, + MediaBack, + MediaForward, + MediaStop, + MediaFind, + MediaScrollUp, + MediaScrollDown, + MediaEdit, + MediaSleep, + MediaCoffee, + MediaRefresh, + MediaCalculator, + + NumKeyboardKeys, +}; + +static_assert(NumKeyboardKeys == 0xFC, "Incorrect number of keyboard keys."); + +enum Modifiers { + LeftControl, + LeftShift, + LeftAlt, + LeftMeta, + RightControl, + RightShift, + RightAlt, + RightMeta, + CapsLock, + ScrollLock, + NumLock, + + NumKeyboardMods, +}; + +constexpr int KEYBOARD_KEYS_HID_BEGIN = None; +constexpr int KEYBOARD_KEYS_HID_END = NumKeyboardKeys; +constexpr int NUM_KEYBOARD_KEYS_HID = NumKeyboardKeys; + +constexpr int KEYBOARD_MODS_HID_BEGIN = LeftControl; +constexpr int KEYBOARD_MODS_HID_END = NumKeyboardMods; +constexpr int NUM_KEYBOARD_MODS_HID = NumKeyboardMods; + +} // namespace NativeKeyboard + +using AnalogsRaw = std::array<std::string, NativeAnalog::NumAnalogs>; +using ButtonsRaw = std::array<std::string, NativeButton::NumButtons>; +using MotionsRaw = std::array<std::string, NativeMotion::NumMotions>; +using VibrationsRaw = std::array<std::string, NativeVibration::NumVibrations>; + +using MouseButtonsRaw = std::array<std::string, NativeMouseButton::NumMouseButtons>; +using KeyboardKeysRaw = std::array<std::string, NativeKeyboard::NumKeyboardKeys>; +using KeyboardModsRaw = std::array<std::string, NativeKeyboard::NumKeyboardMods>; + +constexpr u32 JOYCON_BODY_NEON_RED = 0xFF3C28; +constexpr u32 JOYCON_BUTTONS_NEON_RED = 0x1E0A0A; +constexpr u32 JOYCON_BODY_NEON_BLUE = 0x0AB9E6; +constexpr u32 JOYCON_BUTTONS_NEON_BLUE = 0x001E1E; + +enum class ControllerType { + ProController, + DualJoyconDetached, + LeftJoycon, + RightJoycon, + Handheld, +}; + +struct PlayerInput { + bool connected; + ControllerType controller_type; + ButtonsRaw buttons; + AnalogsRaw analogs; + VibrationsRaw vibrations; + MotionsRaw motions; + + bool vibration_enabled; + int vibration_strength; + + u32 body_color_left; + u32 body_color_right; + u32 button_color_left; + u32 button_color_right; +}; + +struct TouchscreenInput { + bool enabled; + std::string device; + + u32 finger; + u32 diameter_x; + u32 diameter_y; + u32 rotation_angle; +}; +} // namespace Settings diff --git a/src/input_common/touch_from_button.cpp b/src/input_common/touch_from_button.cpp new file mode 100644 index 000000000..a07124a86 --- /dev/null +++ b/src/input_common/touch_from_button.cpp @@ -0,0 +1,51 @@ +// Copyright 2020 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/frontend/framebuffer_layout.h" +#include "core/settings.h" +#include "input_common/touch_from_button.h" + +namespace InputCommon { + +class TouchFromButtonDevice final : public Input::TouchDevice { +public: + TouchFromButtonDevice() { + const auto button_index = + static_cast<std::size_t>(Settings::values.touch_from_button_map_index); + const auto& buttons = Settings::values.touch_from_button_maps[button_index].buttons; + + for (const auto& config_entry : buttons) { + const Common::ParamPackage package{config_entry}; + map.emplace_back( + Input::CreateDevice<Input::ButtonDevice>(config_entry), + std::clamp(package.Get("x", 0), 0, static_cast<int>(Layout::ScreenUndocked::Width)), + std::clamp(package.Get("y", 0), 0, + static_cast<int>(Layout::ScreenUndocked::Height))); + } + } + + std::tuple<float, float, bool> GetStatus() const override { + for (const auto& m : map) { + const bool state = std::get<0>(m)->GetStatus(); + if (state) { + const float x = static_cast<float>(std::get<1>(m)) / + static_cast<int>(Layout::ScreenUndocked::Width); + const float y = static_cast<float>(std::get<2>(m)) / + static_cast<int>(Layout::ScreenUndocked::Height); + return {x, y, true}; + } + } + return {}; + } + +private: + // A vector of the mapped button, its x and its y-coordinate + std::vector<std::tuple<std::unique_ptr<Input::ButtonDevice>, int, int>> map; +}; + +std::unique_ptr<Input::TouchDevice> TouchFromButtonFactory::Create(const Common::ParamPackage&) { + return std::make_unique<TouchFromButtonDevice>(); +} + +} // namespace InputCommon diff --git a/src/input_common/touch_from_button.h b/src/input_common/touch_from_button.h new file mode 100644 index 000000000..8b4d1aa96 --- /dev/null +++ b/src/input_common/touch_from_button.h @@ -0,0 +1,23 @@ +// Copyright 2020 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include "core/frontend/input.h" + +namespace InputCommon { + +/** + * A touch device factory that takes a list of button devices and combines them into a touch device. + */ +class TouchFromButtonFactory final : public Input::Factory<Input::TouchDevice> { +public: + /** + * Creates a touch device from a list of button devices + */ + std::unique_ptr<Input::TouchDevice> Create(const Common::ParamPackage& params) override; +}; + +} // namespace InputCommon diff --git a/src/input_common/udp/client.cpp b/src/input_common/udp/client.cpp index da5227058..c0bb90048 100644 --- a/src/input_common/udp/client.cpp +++ b/src/input_common/udp/client.cpp @@ -2,15 +2,13 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <algorithm> -#include <array> #include <chrono> #include <cstring> #include <functional> #include <thread> #include <boost/asio.hpp> -#include <boost/bind.hpp> #include "common/logging/log.h" +#include "core/settings.h" #include "input_common/udp/client.h" #include "input_common/udp/protocol.h" @@ -28,11 +26,11 @@ class Socket { public: using clock = std::chrono::system_clock; - explicit Socket(const std::string& host, u16 port, u8 pad_index, u32 client_id, - SocketCallback callback) - : callback(std::move(callback)), timer(io_service), - socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(client_id), - pad_index(pad_index) { + explicit Socket(const std::string& host, u16 port, std::size_t pad_index_, u32 client_id_, + SocketCallback callback_) + : callback(std::move(callback_)), timer(io_service), + socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(client_id_), + pad_index(pad_index_) { boost::system::error_code ec{}; auto ipv4 = boost::asio::ip::make_address_v4(host, ec); if (ec.value() != boost::system::errc::success) { @@ -65,7 +63,7 @@ public: } private: - void HandleReceive(const boost::system::error_code& error, std::size_t bytes_transferred) { + void HandleReceive(const boost::system::error_code&, std::size_t bytes_transferred) { if (auto type = Response::Validate(receive_buffer.data(), bytes_transferred)) { switch (*type) { case Type::Version: { @@ -92,16 +90,20 @@ private: StartReceive(); } - void HandleSend(const boost::system::error_code& error) { + void HandleSend(const boost::system::error_code&) { boost::system::error_code _ignored{}; // Send a request for getting port info for the pad - Request::PortInfo port_info{1, {pad_index, 0, 0, 0}}; + const Request::PortInfo port_info{1, {static_cast<u8>(pad_index), 0, 0, 0}}; const auto port_message = Request::Create(port_info, client_id); std::memcpy(&send_buffer1, &port_message, PORT_INFO_SIZE); socket.send_to(boost::asio::buffer(send_buffer1), send_endpoint, {}, _ignored); // Send a request for getting pad data for the pad - Request::PadData pad_data{Request::PadData::Flags::Id, pad_index, EMPTY_MAC_ADDRESS}; + const Request::PadData pad_data{ + Request::PadData::Flags::Id, + static_cast<u8>(pad_index), + EMPTY_MAC_ADDRESS, + }; const auto pad_message = Request::Create(pad_data, client_id); std::memcpy(send_buffer2.data(), &pad_message, PAD_DATA_SIZE); socket.send_to(boost::asio::buffer(send_buffer2), send_endpoint, {}, _ignored); @@ -114,7 +116,7 @@ private: udp::socket socket; u32 client_id{}; - u8 pad_index{}; + std::size_t pad_index{}; static constexpr std::size_t PORT_INFO_SIZE = sizeof(Message<Request::PortInfo>); static constexpr std::size_t PAD_DATA_SIZE = sizeof(Message<Request::PadData>); @@ -132,49 +134,100 @@ static void SocketLoop(Socket* socket) { socket->Loop(); } -Client::Client(std::shared_ptr<DeviceStatus> status, const std::string& host, u16 port, - u8 pad_index, u32 client_id) - : status(std::move(status)) { - StartCommunication(host, port, pad_index, client_id); +Client::Client() { + LOG_INFO(Input, "Udp Initialization started"); + for (std::size_t client = 0; client < clients.size(); client++) { + const auto pad = client % 4; + StartCommunication(client, Settings::values.udp_input_address, + Settings::values.udp_input_port, pad, 24872); + // Set motion parameters + // SetGyroThreshold value should be dependent on GyroscopeZeroDriftMode + // Real HW values are unknown, 0.0001 is an approximate to Standard + clients[client].motion.SetGyroThreshold(0.0001f); + } } Client::~Client() { - socket->Stop(); - thread.join(); + Reset(); +} + +std::vector<Common::ParamPackage> Client::GetInputDevices() const { + std::vector<Common::ParamPackage> devices; + for (std::size_t client = 0; client < clients.size(); client++) { + if (!DeviceConnected(client)) { + continue; + } + std::string name = fmt::format("UDP Controller {}", client); + devices.emplace_back(Common::ParamPackage{ + {"class", "cemuhookudp"}, + {"display", std::move(name)}, + {"port", std::to_string(client)}, + }); + } + return devices; } -void Client::ReloadSocket(const std::string& host, u16 port, u8 pad_index, u32 client_id) { - socket->Stop(); - thread.join(); - StartCommunication(host, port, pad_index, client_id); +bool Client::DeviceConnected(std::size_t pad) const { + // Use last timestamp to detect if the socket has stopped sending data + const auto now = std::chrono::system_clock::now(); + const auto time_difference = static_cast<u64>( + std::chrono::duration_cast<std::chrono::milliseconds>(now - clients[pad].last_motion_update) + .count()); + return time_difference < 1000 && clients[pad].active == 1; } -void Client::OnVersion(Response::Version data) { +void Client::ReloadUDPClient() { + for (std::size_t client = 0; client < clients.size(); client++) { + ReloadSocket(Settings::values.udp_input_address, Settings::values.udp_input_port, client); + } +} +void Client::ReloadSocket(const std::string& host, u16 port, std::size_t pad_index, u32 client_id) { + // client number must be determined from host / port and pad index + const std::size_t client = pad_index; + clients[client].socket->Stop(); + clients[client].thread.join(); + StartCommunication(client, host, port, pad_index, client_id); +} + +void Client::OnVersion([[maybe_unused]] Response::Version data) { LOG_TRACE(Input, "Version packet received: {}", data.version); } -void Client::OnPortInfo(Response::PortInfo data) { +void Client::OnPortInfo([[maybe_unused]] Response::PortInfo data) { LOG_TRACE(Input, "PortInfo packet received: {}", data.model); } void Client::OnPadData(Response::PadData data) { + // Client number must be determined from host / port and pad index + const std::size_t client = data.info.id; LOG_TRACE(Input, "PadData packet received"); - if (data.packet_counter <= packet_sequence) { + if (data.packet_counter == clients[client].packet_sequence) { LOG_WARNING( Input, "PadData packet dropped because its stale info. Current count: {} Packet count: {}", - packet_sequence, data.packet_counter); + clients[client].packet_sequence, data.packet_counter); return; } - packet_sequence = data.packet_counter; - // TODO: Check how the Switch handles motions and how the CemuhookUDP motion - // directions correspond to the ones of the Switch - Common::Vec3f accel = Common::MakeVec<float>(data.accel.x, data.accel.y, data.accel.z); - Common::Vec3f gyro = Common::MakeVec<float>(data.gyro.pitch, data.gyro.yaw, data.gyro.roll); - { - std::lock_guard guard(status->update_mutex); + clients[client].active = data.info.is_pad_active; + clients[client].packet_sequence = data.packet_counter; + const auto now = std::chrono::system_clock::now(); + const auto time_difference = + static_cast<u64>(std::chrono::duration_cast<std::chrono::microseconds>( + now - clients[client].last_motion_update) + .count()); + clients[client].last_motion_update = now; + const Common::Vec3f raw_gyroscope = {data.gyro.pitch, data.gyro.roll, -data.gyro.yaw}; + clients[client].motion.SetAcceleration({data.accel.x, -data.accel.z, data.accel.y}); + // Gyroscope values are not it the correct scale from better joy. + // Dividing by 312 allows us to make one full turn = 1 turn + // This must be a configurable valued called sensitivity + clients[client].motion.SetGyroscope(raw_gyroscope / 312.0f); + clients[client].motion.UpdateRotation(time_difference); + clients[client].motion.UpdateOrientation(time_difference); - status->motion_status = {accel, gyro}; + { + std::lock_guard guard(clients[client].status.update_mutex); + clients[client].status.motion_status = clients[client].motion.GetMotion(); // TODO: add a setting for "click" touch. Click touch refers to a device that differentiates // between a simple "tap" and a hard press that causes the touch screen to click. @@ -183,41 +236,115 @@ void Client::OnPadData(Response::PadData data) { float x = 0; float y = 0; - if (is_active && status->touch_calibration) { - const u16 min_x = status->touch_calibration->min_x; - const u16 max_x = status->touch_calibration->max_x; - const u16 min_y = status->touch_calibration->min_y; - const u16 max_y = status->touch_calibration->max_y; + if (is_active && clients[client].status.touch_calibration) { + const u16 min_x = clients[client].status.touch_calibration->min_x; + const u16 max_x = clients[client].status.touch_calibration->max_x; + const u16 min_y = clients[client].status.touch_calibration->min_y; + const u16 max_y = clients[client].status.touch_calibration->max_y; - x = (std::clamp(static_cast<u16>(data.touch_1.x), min_x, max_x) - min_x) / + x = static_cast<float>(std::clamp(static_cast<u16>(data.touch_1.x), min_x, max_x) - + min_x) / static_cast<float>(max_x - min_x); - y = (std::clamp(static_cast<u16>(data.touch_1.y), min_y, max_y) - min_y) / + y = static_cast<float>(std::clamp(static_cast<u16>(data.touch_1.y), min_y, max_y) - + min_y) / static_cast<float>(max_y - min_y); } - status->touch_status = {x, y, is_active}; + clients[client].status.touch_status = {x, y, is_active}; + + if (configuring) { + const Common::Vec3f gyroscope = clients[client].motion.GetGyroscope(); + const Common::Vec3f accelerometer = clients[client].motion.GetAcceleration(); + UpdateYuzuSettings(client, accelerometer, gyroscope, is_active); + } } } -void Client::StartCommunication(const std::string& host, u16 port, u8 pad_index, u32 client_id) { +void Client::StartCommunication(std::size_t client, const std::string& host, u16 port, + std::size_t pad_index, u32 client_id) { SocketCallback callback{[this](Response::Version version) { OnVersion(version); }, [this](Response::PortInfo info) { OnPortInfo(info); }, [this](Response::PadData data) { OnPadData(data); }}; LOG_INFO(Input, "Starting communication with UDP input server on {}:{}", host, port); - socket = std::make_unique<Socket>(host, port, pad_index, client_id, callback); - thread = std::thread{SocketLoop, this->socket.get()}; + clients[client].socket = std::make_unique<Socket>(host, port, pad_index, client_id, callback); + clients[client].thread = std::thread{SocketLoop, clients[client].socket.get()}; +} + +void Client::Reset() { + for (auto& client : clients) { + client.socket->Stop(); + client.thread.join(); + } +} + +void Client::UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc, + const Common::Vec3<float>& gyro, bool touch) { + if (gyro.Length() > 0.2f) { + LOG_DEBUG(Input, "UDP Controller {}: gyro=({}, {}, {}), accel=({}, {}, {}), touch={}", + client, gyro[0], gyro[1], gyro[2], acc[0], acc[1], acc[2], touch); + } + UDPPadStatus pad; + if (touch) { + pad.touch = PadTouch::Click; + pad_queue[client].Push(pad); + } + for (size_t i = 0; i < 3; ++i) { + if (gyro[i] > 5.0f || gyro[i] < -5.0f) { + pad.motion = static_cast<PadMotion>(i); + pad.motion_value = gyro[i]; + pad_queue[client].Push(pad); + } + if (acc[i] > 1.75f || acc[i] < -1.75f) { + pad.motion = static_cast<PadMotion>(i + 3); + pad.motion_value = acc[i]; + pad_queue[client].Push(pad); + } + } +} + +void Client::BeginConfiguration() { + for (auto& pq : pad_queue) { + pq.Clear(); + } + configuring = true; +} + +void Client::EndConfiguration() { + for (auto& pq : pad_queue) { + pq.Clear(); + } + configuring = false; } -void TestCommunication(const std::string& host, u16 port, u8 pad_index, u32 client_id, - std::function<void()> success_callback, - std::function<void()> failure_callback) { +DeviceStatus& Client::GetPadState(std::size_t pad) { + return clients[pad].status; +} + +const DeviceStatus& Client::GetPadState(std::size_t pad) const { + return clients[pad].status; +} + +std::array<Common::SPSCQueue<UDPPadStatus>, 4>& Client::GetPadQueue() { + return pad_queue; +} + +const std::array<Common::SPSCQueue<UDPPadStatus>, 4>& Client::GetPadQueue() const { + return pad_queue; +} + +void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, u32 client_id, + const std::function<void()>& success_callback, + const std::function<void()>& failure_callback) { std::thread([=] { Common::Event success_event; - SocketCallback callback{[](Response::Version version) {}, [](Response::PortInfo info) {}, - [&](Response::PadData data) { success_event.Set(); }}; + SocketCallback callback{ + .version = [](Response::Version) {}, + .port_info = [](Response::PortInfo) {}, + .pad_data = [&](Response::PadData) { success_event.Set(); }, + }; Socket socket{host, port, pad_index, client_id, std::move(callback)}; std::thread worker_thread{SocketLoop, &socket}; - bool result = success_event.WaitFor(std::chrono::seconds(8)); + const bool result = success_event.WaitFor(std::chrono::seconds(5)); socket.Stop(); worker_thread.join(); if (result) { @@ -225,16 +352,15 @@ void TestCommunication(const std::string& host, u16 port, u8 pad_index, u32 clie } else { failure_callback(); } - }) - .detach(); + }).detach(); } CalibrationConfigurationJob::CalibrationConfigurationJob( - const std::string& host, u16 port, u8 pad_index, u32 client_id, + const std::string& host, u16 port, std::size_t pad_index, u32 client_id, std::function<void(Status)> status_callback, std::function<void(u16, u16, u16, u16)> data_callback) { - std::thread([=] { + std::thread([=, this] { constexpr u16 CALIBRATION_THRESHOLD = 100; u16 min_x{UINT16_MAX}; @@ -243,14 +369,14 @@ CalibrationConfigurationJob::CalibrationConfigurationJob( u16 max_y{}; Status current_status{Status::Initialized}; - SocketCallback callback{[](Response::Version version) {}, [](Response::PortInfo info) {}, + SocketCallback callback{[](Response::Version) {}, [](Response::PortInfo) {}, [&](Response::PadData data) { if (current_status == Status::Initialized) { // Receiving data means the communication is ready now current_status = Status::Ready; status_callback(current_status); } - if (!data.touch_1.is_active) { + if (data.touch_1.is_active == 0) { return; } LOG_DEBUG(Input, "Current touch: {} {}", data.touch_1.x, @@ -280,8 +406,7 @@ CalibrationConfigurationJob::CalibrationConfigurationJob( complete_event.Wait(); socket.Stop(); worker_thread.join(); - }) - .detach(); + }).detach(); } CalibrationConfigurationJob::~CalibrationConfigurationJob() { diff --git a/src/input_common/udp/client.h b/src/input_common/udp/client.h index b8c654755..747e0c0a2 100644 --- a/src/input_common/udp/client.h +++ b/src/input_common/udp/client.h @@ -12,8 +12,12 @@ #include <thread> #include <tuple> #include "common/common_types.h" +#include "common/param_package.h" #include "common/thread.h" +#include "common/threadsafe_queue.h" #include "common/vector_math.h" +#include "core/frontend/input.h" +#include "input_common/motion_input.h" namespace InputCommon::CemuhookUDP { @@ -28,9 +32,30 @@ struct PortInfo; struct Version; } // namespace Response +enum class PadMotion { + GyroX, + GyroY, + GyroZ, + AccX, + AccY, + AccZ, + Undefined, +}; + +enum class PadTouch { + Click, + Undefined, +}; + +struct UDPPadStatus { + PadTouch touch{PadTouch::Undefined}; + PadMotion motion{PadMotion::Undefined}; + f32 motion_value{0.0f}; +}; + struct DeviceStatus { std::mutex update_mutex; - std::tuple<Common::Vec3<float>, Common::Vec3<float>> motion_status; + Input::MotionStatus motion_status; std::tuple<float, float, bool> touch_status; // calibration data for scaling the device's touch area to 3ds @@ -45,22 +70,58 @@ struct DeviceStatus { class Client { public: - explicit Client(std::shared_ptr<DeviceStatus> status, const std::string& host = DEFAULT_ADDR, - u16 port = DEFAULT_PORT, u8 pad_index = 0, u32 client_id = 24872); + // Initialize the UDP client capture and read sequence + Client(); + + // Close and release the client ~Client(); - void ReloadSocket(const std::string& host = "127.0.0.1", u16 port = 26760, u8 pad_index = 0, - u32 client_id = 24872); + + // Used for polling + void BeginConfiguration(); + void EndConfiguration(); + + std::vector<Common::ParamPackage> GetInputDevices() const; + + bool DeviceConnected(std::size_t pad) const; + void ReloadUDPClient(); + void ReloadSocket(const std::string& host = "127.0.0.1", u16 port = 26760, + std::size_t pad_index = 0, u32 client_id = 24872); + + std::array<Common::SPSCQueue<UDPPadStatus>, 4>& GetPadQueue(); + const std::array<Common::SPSCQueue<UDPPadStatus>, 4>& GetPadQueue() const; + + DeviceStatus& GetPadState(std::size_t pad); + const DeviceStatus& GetPadState(std::size_t pad) const; private: + struct ClientData { + std::unique_ptr<Socket> socket; + DeviceStatus status; + std::thread thread; + u64 packet_sequence = 0; + u8 active = 0; + + // Realtime values + // motion is initalized with PID values for drift correction on joycons + InputCommon::MotionInput motion{0.3f, 0.005f, 0.0f}; + std::chrono::time_point<std::chrono::system_clock> last_motion_update; + }; + + // For shutting down, clear all data, join all threads, release usb + void Reset(); + void OnVersion(Response::Version); void OnPortInfo(Response::PortInfo); void OnPadData(Response::PadData); - void StartCommunication(const std::string& host, u16 port, u8 pad_index, u32 client_id); + void StartCommunication(std::size_t client, const std::string& host, u16 port, + std::size_t pad_index, u32 client_id); + void UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc, + const Common::Vec3<float>& gyro, bool touch); + + bool configuring = false; - std::unique_ptr<Socket> socket; - std::shared_ptr<DeviceStatus> status; - std::thread thread; - u64 packet_sequence = 0; + std::array<ClientData, 4> clients; + std::array<Common::SPSCQueue<UDPPadStatus>, 4> pad_queue; }; /// An async job allowing configuration of the touchpad calibration. @@ -78,7 +139,7 @@ public: * @param status_callback Callback for job status updates * @param data_callback Called when calibration data is ready */ - explicit CalibrationConfigurationJob(const std::string& host, u16 port, u8 pad_index, + explicit CalibrationConfigurationJob(const std::string& host, u16 port, std::size_t pad_index, u32 client_id, std::function<void(Status)> status_callback, std::function<void(u16, u16, u16, u16)> data_callback); ~CalibrationConfigurationJob(); @@ -88,8 +149,8 @@ private: Common::Event complete_event; }; -void TestCommunication(const std::string& host, u16 port, u8 pad_index, u32 client_id, - std::function<void()> success_callback, - std::function<void()> failure_callback); +void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, u32 client_id, + const std::function<void()>& success_callback, + const std::function<void()>& failure_callback); } // namespace InputCommon::CemuhookUDP diff --git a/src/input_common/udp/protocol.h b/src/input_common/udp/protocol.h index 3ba4d1fc8..fc1aea4b9 100644 --- a/src/input_common/udp/protocol.h +++ b/src/input_common/udp/protocol.h @@ -7,7 +7,16 @@ #include <array> #include <optional> #include <type_traits> + +#ifdef _MSC_VER +#pragma warning(push) +#pragma warning(disable : 4701) +#endif #include <boost/crc.hpp> +#ifdef _MSC_VER +#pragma warning(pop) +#endif + #include "common/bit_field.h" #include "common/swap.h" @@ -93,7 +102,7 @@ static_assert(std::is_trivially_copyable_v<PadData>, /** * Creates a message with the proper header data that can be sent to the server. - * @param T data Request body to send + * @param data Request body to send * @param client_id ID of the udp client (usually not checked on the server) */ template <typename T> diff --git a/src/input_common/udp/udp.cpp b/src/input_common/udp/udp.cpp index 8c6ef1394..71a76a7aa 100644 --- a/src/input_common/udp/udp.cpp +++ b/src/input_common/udp/udp.cpp @@ -1,99 +1,142 @@ -// Copyright 2018 Citra Emulator Project +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. #include <mutex> -#include <optional> -#include <tuple> - -#include "common/param_package.h" -#include "core/frontend/input.h" -#include "core/settings.h" +#include <utility> +#include "common/assert.h" +#include "common/threadsafe_queue.h" #include "input_common/udp/client.h" #include "input_common/udp/udp.h" -namespace InputCommon::CemuhookUDP { +namespace InputCommon { -class UDPTouchDevice final : public Input::TouchDevice { +class UDPMotion final : public Input::MotionDevice { public: - explicit UDPTouchDevice(std::shared_ptr<DeviceStatus> status_) : status(std::move(status_)) {} - std::tuple<float, float, bool> GetStatus() const override { - std::lock_guard guard(status->update_mutex); - return status->touch_status; + explicit UDPMotion(std::string ip_, int port_, u32 pad_, CemuhookUDP::Client* client_) + : ip(std::move(ip_)), port(port_), pad(pad_), client(client_) {} + + Input::MotionStatus GetStatus() const override { + return client->GetPadState(pad).motion_status; } private: - std::shared_ptr<DeviceStatus> status; + const std::string ip; + const int port; + const u32 pad; + CemuhookUDP::Client* client; + mutable std::mutex mutex; }; -class UDPMotionDevice final : public Input::MotionDevice { -public: - explicit UDPMotionDevice(std::shared_ptr<DeviceStatus> status_) : status(std::move(status_)) {} - std::tuple<Common::Vec3<float>, Common::Vec3<float>> GetStatus() const override { - std::lock_guard guard(status->update_mutex); - return status->motion_status; - } +/// A motion device factory that creates motion devices from JC Adapter +UDPMotionFactory::UDPMotionFactory(std::shared_ptr<CemuhookUDP::Client> client_) + : client(std::move(client_)) {} + +/** + * Creates motion device + * @param params contains parameters for creating the device: + * - "port": the nth jcpad on the adapter + */ +std::unique_ptr<Input::MotionDevice> UDPMotionFactory::Create(const Common::ParamPackage& params) { + auto ip = params.Get("ip", "127.0.0.1"); + const auto port = params.Get("port", 26760); + const auto pad = static_cast<u32>(params.Get("pad_index", 0)); + + return std::make_unique<UDPMotion>(std::move(ip), port, pad, client.get()); +} -private: - std::shared_ptr<DeviceStatus> status; -}; +void UDPMotionFactory::BeginConfiguration() { + polling = true; + client->BeginConfiguration(); +} -class UDPTouchFactory final : public Input::Factory<Input::TouchDevice> { -public: - explicit UDPTouchFactory(std::shared_ptr<DeviceStatus> status_) : status(std::move(status_)) {} - - std::unique_ptr<Input::TouchDevice> Create(const Common::ParamPackage& params) override { - { - std::lock_guard guard(status->update_mutex); - status->touch_calibration = DeviceStatus::CalibrationData{}; - // These default values work well for DS4 but probably not other touch inputs - status->touch_calibration->min_x = params.Get("min_x", 100); - status->touch_calibration->min_y = params.Get("min_y", 50); - status->touch_calibration->max_x = params.Get("max_x", 1800); - status->touch_calibration->max_y = params.Get("max_y", 850); +void UDPMotionFactory::EndConfiguration() { + polling = false; + client->EndConfiguration(); +} + +Common::ParamPackage UDPMotionFactory::GetNextInput() { + Common::ParamPackage params; + CemuhookUDP::UDPPadStatus pad; + auto& queue = client->GetPadQueue(); + for (std::size_t pad_number = 0; pad_number < queue.size(); ++pad_number) { + while (queue[pad_number].Pop(pad)) { + if (pad.motion == CemuhookUDP::PadMotion::Undefined || std::abs(pad.motion_value) < 1) { + continue; + } + params.Set("engine", "cemuhookudp"); + params.Set("ip", "127.0.0.1"); + params.Set("port", 26760); + params.Set("pad_index", static_cast<int>(pad_number)); + params.Set("motion", static_cast<u16>(pad.motion)); + return params; } - return std::make_unique<UDPTouchDevice>(status); } + return params; +} -private: - std::shared_ptr<DeviceStatus> status; -}; - -class UDPMotionFactory final : public Input::Factory<Input::MotionDevice> { +class UDPTouch final : public Input::TouchDevice { public: - explicit UDPMotionFactory(std::shared_ptr<DeviceStatus> status_) : status(std::move(status_)) {} + explicit UDPTouch(std::string ip_, int port_, u32 pad_, CemuhookUDP::Client* client_) + : ip(std::move(ip_)), port(port_), pad(pad_), client(client_) {} - std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override { - return std::make_unique<UDPMotionDevice>(status); + std::tuple<float, float, bool> GetStatus() const override { + return client->GetPadState(pad).touch_status; } private: - std::shared_ptr<DeviceStatus> status; + const std::string ip; + const int port; + const u32 pad; + CemuhookUDP::Client* client; + mutable std::mutex mutex; }; -State::State() { - auto status = std::make_shared<DeviceStatus>(); - client = - std::make_unique<Client>(status, Settings::values.udp_input_address, - Settings::values.udp_input_port, Settings::values.udp_pad_index); - - Input::RegisterFactory<Input::TouchDevice>("cemuhookudp", - std::make_shared<UDPTouchFactory>(status)); - Input::RegisterFactory<Input::MotionDevice>("cemuhookudp", - std::make_shared<UDPMotionFactory>(status)); +/// A motion device factory that creates motion devices from JC Adapter +UDPTouchFactory::UDPTouchFactory(std::shared_ptr<CemuhookUDP::Client> client_) + : client(std::move(client_)) {} + +/** + * Creates motion device + * @param params contains parameters for creating the device: + * - "port": the nth jcpad on the adapter + */ +std::unique_ptr<Input::TouchDevice> UDPTouchFactory::Create(const Common::ParamPackage& params) { + auto ip = params.Get("ip", "127.0.0.1"); + const auto port = params.Get("port", 26760); + const auto pad = static_cast<u32>(params.Get("pad_index", 0)); + + return std::make_unique<UDPTouch>(std::move(ip), port, pad, client.get()); } -State::~State() { - Input::UnregisterFactory<Input::TouchDevice>("cemuhookudp"); - Input::UnregisterFactory<Input::MotionDevice>("cemuhookudp"); +void UDPTouchFactory::BeginConfiguration() { + polling = true; + client->BeginConfiguration(); } -void State::ReloadUDPClient() { - client->ReloadSocket(Settings::values.udp_input_address, Settings::values.udp_input_port, - Settings::values.udp_pad_index); +void UDPTouchFactory::EndConfiguration() { + polling = false; + client->EndConfiguration(); } -std::unique_ptr<State> Init() { - return std::make_unique<State>(); +Common::ParamPackage UDPTouchFactory::GetNextInput() { + Common::ParamPackage params; + CemuhookUDP::UDPPadStatus pad; + auto& queue = client->GetPadQueue(); + for (std::size_t pad_number = 0; pad_number < queue.size(); ++pad_number) { + while (queue[pad_number].Pop(pad)) { + if (pad.touch == CemuhookUDP::PadTouch::Undefined) { + continue; + } + params.Set("engine", "cemuhookudp"); + params.Set("ip", "127.0.0.1"); + params.Set("port", 26760); + params.Set("pad_index", static_cast<int>(pad_number)); + params.Set("touch", static_cast<u16>(pad.touch)); + return params; + } + } + return params; } -} // namespace InputCommon::CemuhookUDP + +} // namespace InputCommon diff --git a/src/input_common/udp/udp.h b/src/input_common/udp/udp.h index 4f83f0441..ea3fd4175 100644 --- a/src/input_common/udp/udp.h +++ b/src/input_common/udp/udp.h @@ -1,25 +1,57 @@ -// Copyright 2018 Citra Emulator Project +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. #pragma once #include <memory> +#include "core/frontend/input.h" +#include "input_common/udp/client.h" -namespace InputCommon::CemuhookUDP { +namespace InputCommon { -class Client; - -class State { +/// A motion device factory that creates motion devices from udp clients +class UDPMotionFactory final : public Input::Factory<Input::MotionDevice> { public: - State(); - ~State(); - void ReloadUDPClient(); + explicit UDPMotionFactory(std::shared_ptr<CemuhookUDP::Client> client_); + + std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput(); + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } private: - std::unique_ptr<Client> client; + std::shared_ptr<CemuhookUDP::Client> client; + bool polling = false; }; -std::unique_ptr<State> Init(); +/// A touch device factory that creates touch devices from udp clients +class UDPTouchFactory final : public Input::Factory<Input::TouchDevice> { +public: + explicit UDPTouchFactory(std::shared_ptr<CemuhookUDP::Client> client_); + + std::unique_ptr<Input::TouchDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput(); + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<CemuhookUDP::Client> client; + bool polling = false; +}; -} // namespace InputCommon::CemuhookUDP +} // namespace InputCommon diff --git a/src/tests/CMakeLists.txt b/src/tests/CMakeLists.txt index c7038b217..47ef30aa9 100644 --- a/src/tests/CMakeLists.txt +++ b/src/tests/CMakeLists.txt @@ -1,6 +1,7 @@ add_executable(tests common/bit_field.cpp common/bit_utils.cpp + common/fibers.cpp common/multi_level_queue.cpp common/param_package.cpp common/ring_buffer.cpp diff --git a/src/tests/common/fibers.cpp b/src/tests/common/fibers.cpp new file mode 100644 index 000000000..4757dd2b4 --- /dev/null +++ b/src/tests/common/fibers.cpp @@ -0,0 +1,367 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <atomic> +#include <cstdlib> +#include <functional> +#include <memory> +#include <mutex> +#include <stdexcept> +#include <thread> +#include <unordered_map> +#include <vector> + +#include <catch2/catch.hpp> + +#include "common/common_types.h" +#include "common/fiber.h" + +namespace Common { + +class ThreadIds { +public: + void Register(u32 id) { + const auto thread_id = std::this_thread::get_id(); + std::scoped_lock lock{mutex}; + if (ids.contains(thread_id)) { + throw std::logic_error{"Registering the same thread twice"}; + } + ids.emplace(thread_id, id); + } + + [[nodiscard]] u32 Get() const { + std::scoped_lock lock{mutex}; + return ids.at(std::this_thread::get_id()); + } + +private: + mutable std::mutex mutex; + std::unordered_map<std::thread::id, u32> ids; +}; + +class TestControl1 { +public: + TestControl1() = default; + + void DoWork(); + + void ExecuteThread(u32 id); + + ThreadIds thread_ids; + std::vector<std::shared_ptr<Common::Fiber>> thread_fibers; + std::vector<std::shared_ptr<Common::Fiber>> work_fibers; + std::vector<u32> items; + std::vector<u32> results; +}; + +static void WorkControl1(void* control) { + auto* test_control = static_cast<TestControl1*>(control); + test_control->DoWork(); +} + +void TestControl1::DoWork() { + const u32 id = thread_ids.Get(); + u32 value = items[id]; + for (u32 i = 0; i < id; i++) { + value++; + } + results[id] = value; + Fiber::YieldTo(work_fibers[id], thread_fibers[id]); +} + +void TestControl1::ExecuteThread(u32 id) { + thread_ids.Register(id); + auto thread_fiber = Fiber::ThreadToFiber(); + thread_fibers[id] = thread_fiber; + work_fibers[id] = std::make_shared<Fiber>(std::function<void(void*)>{WorkControl1}, this); + items[id] = rand() % 256; + Fiber::YieldTo(thread_fibers[id], work_fibers[id]); + thread_fibers[id]->Exit(); +} + +static void ThreadStart1(u32 id, TestControl1& test_control) { + test_control.ExecuteThread(id); +} + +/** This test checks for fiber setup configuration and validates that fibers are + * doing all the work required. + */ +TEST_CASE("Fibers::Setup", "[common]") { + constexpr std::size_t num_threads = 7; + TestControl1 test_control{}; + test_control.thread_fibers.resize(num_threads); + test_control.work_fibers.resize(num_threads); + test_control.items.resize(num_threads, 0); + test_control.results.resize(num_threads, 0); + std::vector<std::thread> threads; + for (u32 i = 0; i < num_threads; i++) { + threads.emplace_back(ThreadStart1, i, std::ref(test_control)); + } + for (u32 i = 0; i < num_threads; i++) { + threads[i].join(); + } + for (u32 i = 0; i < num_threads; i++) { + REQUIRE(test_control.items[i] + i == test_control.results[i]); + } +} + +class TestControl2 { +public: + TestControl2() = default; + + void DoWork1() { + trap2 = false; + while (trap.load()) + ; + for (u32 i = 0; i < 12000; i++) { + value1 += i; + } + Fiber::YieldTo(fiber1, fiber3); + const u32 id = thread_ids.Get(); + assert1 = id == 1; + value2 += 5000; + Fiber::YieldTo(fiber1, thread_fibers[id]); + } + + void DoWork2() { + while (trap2.load()) + ; + value2 = 2000; + trap = false; + Fiber::YieldTo(fiber2, fiber1); + assert3 = false; + } + + void DoWork3() { + const u32 id = thread_ids.Get(); + assert2 = id == 0; + value1 += 1000; + Fiber::YieldTo(fiber3, thread_fibers[id]); + } + + void ExecuteThread(u32 id); + + void CallFiber1() { + const u32 id = thread_ids.Get(); + Fiber::YieldTo(thread_fibers[id], fiber1); + } + + void CallFiber2() { + const u32 id = thread_ids.Get(); + Fiber::YieldTo(thread_fibers[id], fiber2); + } + + void Exit(); + + bool assert1{}; + bool assert2{}; + bool assert3{true}; + u32 value1{}; + u32 value2{}; + std::atomic<bool> trap{true}; + std::atomic<bool> trap2{true}; + ThreadIds thread_ids; + std::vector<std::shared_ptr<Common::Fiber>> thread_fibers; + std::shared_ptr<Common::Fiber> fiber1; + std::shared_ptr<Common::Fiber> fiber2; + std::shared_ptr<Common::Fiber> fiber3; +}; + +static void WorkControl2_1(void* control) { + auto* test_control = static_cast<TestControl2*>(control); + test_control->DoWork1(); +} + +static void WorkControl2_2(void* control) { + auto* test_control = static_cast<TestControl2*>(control); + test_control->DoWork2(); +} + +static void WorkControl2_3(void* control) { + auto* test_control = static_cast<TestControl2*>(control); + test_control->DoWork3(); +} + +void TestControl2::ExecuteThread(u32 id) { + thread_ids.Register(id); + auto thread_fiber = Fiber::ThreadToFiber(); + thread_fibers[id] = thread_fiber; +} + +void TestControl2::Exit() { + const u32 id = thread_ids.Get(); + thread_fibers[id]->Exit(); +} + +static void ThreadStart2_1(u32 id, TestControl2& test_control) { + test_control.ExecuteThread(id); + test_control.CallFiber1(); + test_control.Exit(); +} + +static void ThreadStart2_2(u32 id, TestControl2& test_control) { + test_control.ExecuteThread(id); + test_control.CallFiber2(); + test_control.Exit(); +} + +/** This test checks for fiber thread exchange configuration and validates that fibers are + * that a fiber has been succesfully transfered from one thread to another and that the TLS + * region of the thread is kept while changing fibers. + */ +TEST_CASE("Fibers::InterExchange", "[common]") { + TestControl2 test_control{}; + test_control.thread_fibers.resize(2); + test_control.fiber1 = + std::make_shared<Fiber>(std::function<void(void*)>{WorkControl2_1}, &test_control); + test_control.fiber2 = + std::make_shared<Fiber>(std::function<void(void*)>{WorkControl2_2}, &test_control); + test_control.fiber3 = + std::make_shared<Fiber>(std::function<void(void*)>{WorkControl2_3}, &test_control); + std::thread thread1(ThreadStart2_1, 0, std::ref(test_control)); + std::thread thread2(ThreadStart2_2, 1, std::ref(test_control)); + thread1.join(); + thread2.join(); + REQUIRE(test_control.assert1); + REQUIRE(test_control.assert2); + REQUIRE(test_control.assert3); + REQUIRE(test_control.value2 == 7000); + u32 cal_value = 0; + for (u32 i = 0; i < 12000; i++) { + cal_value += i; + } + cal_value += 1000; + REQUIRE(test_control.value1 == cal_value); +} + +class TestControl3 { +public: + TestControl3() = default; + + void DoWork1() { + value1 += 1; + Fiber::YieldTo(fiber1, fiber2); + const u32 id = thread_ids.Get(); + value3 += 1; + Fiber::YieldTo(fiber1, thread_fibers[id]); + } + + void DoWork2() { + value2 += 1; + const u32 id = thread_ids.Get(); + Fiber::YieldTo(fiber2, thread_fibers[id]); + } + + void ExecuteThread(u32 id); + + void CallFiber1() { + const u32 id = thread_ids.Get(); + Fiber::YieldTo(thread_fibers[id], fiber1); + } + + void Exit(); + + u32 value1{}; + u32 value2{}; + u32 value3{}; + ThreadIds thread_ids; + std::vector<std::shared_ptr<Common::Fiber>> thread_fibers; + std::shared_ptr<Common::Fiber> fiber1; + std::shared_ptr<Common::Fiber> fiber2; +}; + +static void WorkControl3_1(void* control) { + auto* test_control = static_cast<TestControl3*>(control); + test_control->DoWork1(); +} + +static void WorkControl3_2(void* control) { + auto* test_control = static_cast<TestControl3*>(control); + test_control->DoWork2(); +} + +void TestControl3::ExecuteThread(u32 id) { + thread_ids.Register(id); + auto thread_fiber = Fiber::ThreadToFiber(); + thread_fibers[id] = thread_fiber; +} + +void TestControl3::Exit() { + const u32 id = thread_ids.Get(); + thread_fibers[id]->Exit(); +} + +static void ThreadStart3(u32 id, TestControl3& test_control) { + test_control.ExecuteThread(id); + test_control.CallFiber1(); + test_control.Exit(); +} + +/** This test checks for one two threads racing for starting the same fiber. + * It checks execution occured in an ordered manner and by no time there were + * two contexts at the same time. + */ +TEST_CASE("Fibers::StartRace", "[common]") { + TestControl3 test_control{}; + test_control.thread_fibers.resize(2); + test_control.fiber1 = + std::make_shared<Fiber>(std::function<void(void*)>{WorkControl3_1}, &test_control); + test_control.fiber2 = + std::make_shared<Fiber>(std::function<void(void*)>{WorkControl3_2}, &test_control); + std::thread thread1(ThreadStart3, 0, std::ref(test_control)); + std::thread thread2(ThreadStart3, 1, std::ref(test_control)); + thread1.join(); + thread2.join(); + REQUIRE(test_control.value1 == 1); + REQUIRE(test_control.value2 == 1); + REQUIRE(test_control.value3 == 1); +} + +class TestControl4; + +static void WorkControl4(void* control); + +class TestControl4 { +public: + TestControl4() { + fiber1 = std::make_shared<Fiber>(std::function<void(void*)>{WorkControl4}, this); + goal_reached = false; + rewinded = false; + } + + void Execute() { + thread_fiber = Fiber::ThreadToFiber(); + Fiber::YieldTo(thread_fiber, fiber1); + thread_fiber->Exit(); + } + + void DoWork() { + fiber1->SetRewindPoint(std::function<void(void*)>{WorkControl4}, this); + if (rewinded) { + goal_reached = true; + Fiber::YieldTo(fiber1, thread_fiber); + } + rewinded = true; + fiber1->Rewind(); + } + + std::shared_ptr<Common::Fiber> fiber1; + std::shared_ptr<Common::Fiber> thread_fiber; + bool goal_reached; + bool rewinded; +}; + +static void WorkControl4(void* control) { + auto* test_control = static_cast<TestControl4*>(control); + test_control->DoWork(); +} + +TEST_CASE("Fibers::Rewind", "[common]") { + TestControl4 test_control{}; + test_control.Execute(); + REQUIRE(test_control.goal_reached); + REQUIRE(test_control.rewinded); +} + +} // namespace Common diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp index 17043346b..e54674d11 100644 --- a/src/tests/core/arm/arm_test_common.cpp +++ b/src/tests/core/arm/arm_test_common.cpp @@ -6,6 +6,7 @@ #include "common/page_table.h" #include "core/core.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" #include "core/memory.h" #include "tests/core/arm/arm_test_common.h" @@ -18,12 +19,7 @@ TestEnvironment::TestEnvironment(bool mutable_memory_) auto& system = Core::System::GetInstance(); auto process = Kernel::Process::Create(system, "", Kernel::Process::ProcessType::Userland); - page_table = &process->VMManager().page_table; - - std::fill(page_table->pointers.begin(), page_table->pointers.end(), nullptr); - page_table->special_regions.clear(); - std::fill(page_table->attributes.begin(), page_table->attributes.end(), - Common::PageType::Unmapped); + page_table = &process->PageTable().PageTableImpl(); system.Memory().MapIoRegion(*page_table, 0x00000000, 0x80000000, test_memory); system.Memory().MapIoRegion(*page_table, 0x80000000, 0x80000000, test_memory); diff --git a/src/tests/core/core_timing.cpp b/src/tests/core/core_timing.cpp index 1e3940801..b35459152 100644 --- a/src/tests/core/core_timing.cpp +++ b/src/tests/core/core_timing.cpp @@ -6,6 +6,7 @@ #include <array> #include <bitset> +#include <chrono> #include <cstdlib> #include <memory> #include <string> @@ -14,32 +15,29 @@ #include "core/core.h" #include "core/core_timing.h" +namespace { // Numbers are chosen randomly to make sure the correct one is given. -static constexpr std::array<u64, 5> CB_IDS{{42, 144, 93, 1026, UINT64_C(0xFFFF7FFFF7FFFF)}}; -static constexpr int MAX_SLICE_LENGTH = 10000; // Copied from CoreTiming internals +constexpr std::array<u64, 5> CB_IDS{{42, 144, 93, 1026, UINT64_C(0xFFFF7FFFF7FFFF)}}; +constexpr std::array<u64, 5> calls_order{{2, 0, 1, 4, 3}}; +std::array<s64, 5> delays{}; -static std::bitset<CB_IDS.size()> callbacks_ran_flags; -static u64 expected_callback = 0; -static s64 lateness = 0; +std::bitset<CB_IDS.size()> callbacks_ran_flags; +u64 expected_callback = 0; template <unsigned int IDX> -void CallbackTemplate(u64 userdata, s64 cycles_late) { +void HostCallbackTemplate(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { static_assert(IDX < CB_IDS.size(), "IDX out of range"); callbacks_ran_flags.set(IDX); - REQUIRE(CB_IDS[IDX] == userdata); - REQUIRE(CB_IDS[IDX] == expected_callback); - REQUIRE(lateness == cycles_late); -} - -static u64 callbacks_done = 0; - -void EmptyCallback(u64 userdata, s64 cycles_late) { - ++callbacks_done; + REQUIRE(CB_IDS[IDX] == user_data); + REQUIRE(CB_IDS[IDX] == CB_IDS[calls_order[expected_callback]]); + delays[IDX] = ns_late.count(); + ++expected_callback; } struct ScopeInit final { ScopeInit() { - core_timing.Initialize(); + core_timing.SetMulticore(true); + core_timing.Initialize([]() {}); } ~ScopeInit() { core_timing.Shutdown(); @@ -48,109 +46,102 @@ struct ScopeInit final { Core::Timing::CoreTiming core_timing; }; -static void AdvanceAndCheck(Core::Timing::CoreTiming& core_timing, u32 idx, u32 context = 0, - int expected_lateness = 0, int cpu_downcount = 0) { - callbacks_ran_flags = 0; - expected_callback = CB_IDS[idx]; - lateness = expected_lateness; - - // Pretend we executed X cycles of instructions. - core_timing.SwitchContext(context); - core_timing.AddTicks(core_timing.GetDowncount() - cpu_downcount); - core_timing.Advance(); - core_timing.SwitchContext((context + 1) % 4); - - REQUIRE(decltype(callbacks_ran_flags)().set(idx) == callbacks_ran_flags); +u64 TestTimerSpeed(Core::Timing::CoreTiming& core_timing) { + const u64 start = core_timing.GetGlobalTimeNs().count(); + volatile u64 placebo = 0; + for (std::size_t i = 0; i < 1000; i++) { + placebo = placebo + core_timing.GetGlobalTimeNs().count(); + } + const u64 end = core_timing.GetGlobalTimeNs().count(); + return end - start; } +} // Anonymous namespace + TEST_CASE("CoreTiming[BasicOrder]", "[core]") { ScopeInit guard; auto& core_timing = guard.core_timing; + std::vector<std::shared_ptr<Core::Timing::EventType>> events{ + Core::Timing::CreateEvent("callbackA", HostCallbackTemplate<0>), + Core::Timing::CreateEvent("callbackB", HostCallbackTemplate<1>), + Core::Timing::CreateEvent("callbackC", HostCallbackTemplate<2>), + Core::Timing::CreateEvent("callbackD", HostCallbackTemplate<3>), + Core::Timing::CreateEvent("callbackE", HostCallbackTemplate<4>), + }; - std::shared_ptr<Core::Timing::EventType> cb_a = - Core::Timing::CreateEvent("callbackA", CallbackTemplate<0>); - std::shared_ptr<Core::Timing::EventType> cb_b = - Core::Timing::CreateEvent("callbackB", CallbackTemplate<1>); - std::shared_ptr<Core::Timing::EventType> cb_c = - Core::Timing::CreateEvent("callbackC", CallbackTemplate<2>); - std::shared_ptr<Core::Timing::EventType> cb_d = - Core::Timing::CreateEvent("callbackD", CallbackTemplate<3>); - std::shared_ptr<Core::Timing::EventType> cb_e = - Core::Timing::CreateEvent("callbackE", CallbackTemplate<4>); - - // Enter slice 0 - core_timing.ResetRun(); - - // D -> B -> C -> A -> E - core_timing.SwitchContext(0); - core_timing.ScheduleEvent(1000, cb_a, CB_IDS[0]); - REQUIRE(1000 == core_timing.GetDowncount()); - core_timing.ScheduleEvent(500, cb_b, CB_IDS[1]); - REQUIRE(500 == core_timing.GetDowncount()); - core_timing.ScheduleEvent(800, cb_c, CB_IDS[2]); - REQUIRE(500 == core_timing.GetDowncount()); - core_timing.ScheduleEvent(100, cb_d, CB_IDS[3]); - REQUIRE(100 == core_timing.GetDowncount()); - core_timing.ScheduleEvent(1200, cb_e, CB_IDS[4]); - REQUIRE(100 == core_timing.GetDowncount()); - - AdvanceAndCheck(core_timing, 3, 0); - AdvanceAndCheck(core_timing, 1, 1); - AdvanceAndCheck(core_timing, 2, 2); - AdvanceAndCheck(core_timing, 0, 3); - AdvanceAndCheck(core_timing, 4, 0); -} - -TEST_CASE("CoreTiming[FairSharing]", "[core]") { + expected_callback = 0; - ScopeInit guard; - auto& core_timing = guard.core_timing; + core_timing.SyncPause(true); - std::shared_ptr<Core::Timing::EventType> empty_callback = - Core::Timing::CreateEvent("empty_callback", EmptyCallback); + const u64 one_micro = 1000U; + for (std::size_t i = 0; i < events.size(); i++) { + const u64 order = calls_order[i]; + const auto future_ns = std::chrono::nanoseconds{static_cast<s64>(i * one_micro + 100)}; - callbacks_done = 0; - u64 MAX_CALLBACKS = 10; - for (std::size_t i = 0; i < 10; i++) { - core_timing.ScheduleEvent(i * 3333U, empty_callback, 0); + core_timing.ScheduleEvent(future_ns, events[order], CB_IDS[order]); } + /// test pause + REQUIRE(callbacks_ran_flags.none()); + + core_timing.Pause(false); // No need to sync + + while (core_timing.HasPendingEvents()) + ; + + REQUIRE(callbacks_ran_flags.all()); - const s64 advances = MAX_SLICE_LENGTH / 10; - core_timing.ResetRun(); - u64 current_time = core_timing.GetTicks(); - bool keep_running{}; - do { - keep_running = false; - for (u32 active_core = 0; active_core < 4; ++active_core) { - core_timing.SwitchContext(active_core); - if (core_timing.CanCurrentContextRun()) { - core_timing.AddTicks(std::min<s64>(advances, core_timing.GetDowncount())); - core_timing.Advance(); - } - keep_running |= core_timing.CanCurrentContextRun(); - } - } while (keep_running); - u64 current_time_2 = core_timing.GetTicks(); - - REQUIRE(MAX_CALLBACKS == callbacks_done); - REQUIRE(current_time_2 == current_time + MAX_SLICE_LENGTH * 4); + for (std::size_t i = 0; i < delays.size(); i++) { + const double delay = static_cast<double>(delays[i]); + const double micro = delay / 1000.0f; + const double mili = micro / 1000.0f; + printf("HostTimer Pausing Delay[%zu]: %.3f %.6f\n", i, micro, mili); + } } -TEST_CASE("Core::Timing[PredictableLateness]", "[core]") { +TEST_CASE("CoreTiming[BasicOrderNoPausing]", "[core]") { ScopeInit guard; auto& core_timing = guard.core_timing; + std::vector<std::shared_ptr<Core::Timing::EventType>> events{ + Core::Timing::CreateEvent("callbackA", HostCallbackTemplate<0>), + Core::Timing::CreateEvent("callbackB", HostCallbackTemplate<1>), + Core::Timing::CreateEvent("callbackC", HostCallbackTemplate<2>), + Core::Timing::CreateEvent("callbackD", HostCallbackTemplate<3>), + Core::Timing::CreateEvent("callbackE", HostCallbackTemplate<4>), + }; + + core_timing.SyncPause(true); + core_timing.SyncPause(false); + + expected_callback = 0; + + const u64 start = core_timing.GetGlobalTimeNs().count(); + const u64 one_micro = 1000U; + + for (std::size_t i = 0; i < events.size(); i++) { + const u64 order = calls_order[i]; + const auto future_ns = std::chrono::nanoseconds{static_cast<s64>(i * one_micro + 100)}; + core_timing.ScheduleEvent(future_ns, events[order], CB_IDS[order]); + } - std::shared_ptr<Core::Timing::EventType> cb_a = - Core::Timing::CreateEvent("callbackA", CallbackTemplate<0>); - std::shared_ptr<Core::Timing::EventType> cb_b = - Core::Timing::CreateEvent("callbackB", CallbackTemplate<1>); + const u64 end = core_timing.GetGlobalTimeNs().count(); + const double scheduling_time = static_cast<double>(end - start); + const double timer_time = static_cast<double>(TestTimerSpeed(core_timing)); - // Enter slice 0 - core_timing.ResetRun(); + while (core_timing.HasPendingEvents()) + ; - core_timing.ScheduleEvent(100, cb_a, CB_IDS[0]); - core_timing.ScheduleEvent(200, cb_b, CB_IDS[1]); + REQUIRE(callbacks_ran_flags.all()); + + for (std::size_t i = 0; i < delays.size(); i++) { + const double delay = static_cast<double>(delays[i]); + const double micro = delay / 1000.0f; + const double mili = micro / 1000.0f; + printf("HostTimer No Pausing Delay[%zu]: %.3f %.6f\n", i, micro, mili); + } - AdvanceAndCheck(core_timing, 0, 0, 10, -10); // (100 - 10) - AdvanceAndCheck(core_timing, 1, 1, 50, -50); + const double micro = scheduling_time / 1000.0f; + const double mili = micro / 1000.0f; + printf("HostTimer No Pausing Scheduling Time: %.3f %.6f\n", micro, mili); + printf("HostTimer No Pausing Timer Time: %.3f %.6f\n", timer_time / 1000.f, + timer_time / 1000000.f); } diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index 258d58eba..abcee2a1c 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -1,13 +1,37 @@ +add_subdirectory(host_shaders) + add_library(video_core STATIC buffer_cache/buffer_block.h buffer_cache/buffer_cache.h + buffer_cache/map_interval.cpp buffer_cache/map_interval.h + cdma_pusher.cpp + cdma_pusher.h + command_classes/codecs/codec.cpp + command_classes/codecs/codec.h + command_classes/codecs/h264.cpp + command_classes/codecs/h264.h + command_classes/codecs/vp9.cpp + command_classes/codecs/vp9.h + command_classes/codecs/vp9_types.h + command_classes/host1x.cpp + command_classes/host1x.h + command_classes/nvdec.cpp + command_classes/nvdec.h + command_classes/nvdec_common.h + command_classes/sync_manager.cpp + command_classes/sync_manager.h + command_classes/vic.cpp + command_classes/vic.h + compatible_formats.cpp + compatible_formats.h dirty_flags.cpp dirty_flags.h dma_pusher.cpp dma_pusher.h engines/const_buffer_engine_interface.h engines/const_buffer_info.h + engines/engine_interface.h engines/engine_upload.cpp engines/engine_upload.h engines/fermi_2d.cpp @@ -23,6 +47,15 @@ add_library(video_core STATIC engines/shader_bytecode.h engines/shader_header.h engines/shader_type.h + macro/macro.cpp + macro/macro.h + macro/macro_hle.cpp + macro/macro_hle.h + macro/macro_interpreter.cpp + macro/macro_interpreter.h + macro/macro_jit_x64.cpp + macro/macro_jit_x64.h + fence_manager.h gpu.cpp gpu.h gpu_asynch.cpp @@ -33,8 +66,6 @@ add_library(video_core STATIC gpu_thread.h guest_driver.cpp guest_driver.h - macro_interpreter.cpp - macro_interpreter.h memory_manager.cpp memory_manager.h morton.cpp @@ -42,15 +73,17 @@ add_library(video_core STATIC query_cache.h rasterizer_accelerated.cpp rasterizer_accelerated.h - rasterizer_cache.cpp - rasterizer_cache.h rasterizer_interface.h renderer_base.cpp renderer_base.h + renderer_opengl/gl_arb_decompiler.cpp + renderer_opengl/gl_arb_decompiler.h renderer_opengl/gl_buffer_cache.cpp renderer_opengl/gl_buffer_cache.h renderer_opengl/gl_device.cpp renderer_opengl/gl_device.h + renderer_opengl/gl_fence_manager.cpp + renderer_opengl/gl_fence_manager.h renderer_opengl/gl_framebuffer_cache.cpp renderer_opengl/gl_framebuffer_cache.h renderer_opengl/gl_rasterizer.cpp @@ -84,6 +117,9 @@ add_library(video_core STATIC renderer_opengl/utils.h sampler_cache.cpp sampler_cache.h + shader_cache.h + shader_notify.cpp + shader_notify.h shader/decode/arithmetic.cpp shader/decode/arithmetic_immediate.cpp shader/decode/bfe.cpp @@ -114,6 +150,8 @@ add_library(video_core STATIC shader/decode/other.cpp shader/ast.cpp shader/ast.h + shader/async_shaders.cpp + shader/async_shaders.h shader/compiler_settings.cpp shader/compiler_settings.h shader/control_flow.cpp @@ -121,6 +159,8 @@ add_library(video_core STATIC shader/decode.cpp shader/expr.cpp shader/expr.h + shader/memory_util.cpp + shader/memory_util.h shader/node_helper.cpp shader/node_helper.h shader/node.h @@ -160,12 +200,16 @@ if (ENABLE_VULKAN) renderer_vulkan/fixed_pipeline_state.h renderer_vulkan/maxwell_to_vk.cpp renderer_vulkan/maxwell_to_vk.h + renderer_vulkan/nsight_aftermath_tracker.cpp + renderer_vulkan/nsight_aftermath_tracker.h renderer_vulkan/renderer_vulkan.h renderer_vulkan/renderer_vulkan.cpp renderer_vulkan/vk_blit_screen.cpp renderer_vulkan/vk_blit_screen.h renderer_vulkan/vk_buffer_cache.cpp renderer_vulkan/vk_buffer_cache.h + renderer_vulkan/vk_command_pool.cpp + renderer_vulkan/vk_command_pool.h renderer_vulkan/vk_compute_pass.cpp renderer_vulkan/vk_compute_pass.h renderer_vulkan/vk_compute_pipeline.cpp @@ -174,10 +218,14 @@ if (ENABLE_VULKAN) renderer_vulkan/vk_descriptor_pool.h renderer_vulkan/vk_device.cpp renderer_vulkan/vk_device.h + renderer_vulkan/vk_fence_manager.cpp + renderer_vulkan/vk_fence_manager.h renderer_vulkan/vk_graphics_pipeline.cpp renderer_vulkan/vk_graphics_pipeline.h renderer_vulkan/vk_image.cpp renderer_vulkan/vk_image.h + renderer_vulkan/vk_master_semaphore.cpp + renderer_vulkan/vk_master_semaphore.h renderer_vulkan/vk_memory_manager.cpp renderer_vulkan/vk_memory_manager.h renderer_vulkan/vk_pipeline_cache.cpp @@ -188,8 +236,8 @@ if (ENABLE_VULKAN) renderer_vulkan/vk_rasterizer.h renderer_vulkan/vk_renderpass_cache.cpp renderer_vulkan/vk_renderpass_cache.h - renderer_vulkan/vk_resource_manager.cpp - renderer_vulkan/vk_resource_manager.h + renderer_vulkan/vk_resource_pool.cpp + renderer_vulkan/vk_resource_pool.h renderer_vulkan/vk_sampler_cache.cpp renderer_vulkan/vk_sampler_cache.h renderer_vulkan/vk_scheduler.cpp @@ -213,21 +261,55 @@ if (ENABLE_VULKAN) renderer_vulkan/wrapper.cpp renderer_vulkan/wrapper.h ) - - target_include_directories(video_core PRIVATE sirit ../../externals/Vulkan-Headers/include) - target_compile_definitions(video_core PRIVATE HAS_VULKAN) endif() create_target_directory_groups(video_core) target_link_libraries(video_core PUBLIC common core) -target_link_libraries(video_core PRIVATE glad) +target_link_libraries(video_core PRIVATE glad xbyak) + +if (MSVC) + target_include_directories(video_core PRIVATE ${FFMPEG_INCLUDE_DIR}) + target_link_libraries(video_core PUBLIC ${FFMPEG_LIBRARY_DIR}/swscale.lib ${FFMPEG_LIBRARY_DIR}/avcodec.lib ${FFMPEG_LIBRARY_DIR}/avutil.lib) +else() + target_include_directories(video_core PRIVATE ${FFMPEG_INCLUDE_DIR}) + target_link_libraries(video_core PRIVATE ${FFMPEG_LIBRARIES}) +endif() + +add_dependencies(video_core host_shaders) +target_include_directories(video_core PRIVATE ${HOST_SHADERS_INCLUDE}) + if (ENABLE_VULKAN) + target_include_directories(video_core PRIVATE sirit ../../externals/Vulkan-Headers/include) + target_compile_definitions(video_core PRIVATE HAS_VULKAN) target_link_libraries(video_core PRIVATE sirit) endif() +if (ENABLE_NSIGHT_AFTERMATH) + if (NOT DEFINED ENV{NSIGHT_AFTERMATH_SDK}) + message(ERROR "Environment variable NSIGHT_AFTERMATH_SDK has to be provided") + endif() + if (NOT WIN32) + message(ERROR "Nsight Aftermath doesn't support non-Windows platforms") + endif() + target_compile_definitions(video_core PRIVATE HAS_NSIGHT_AFTERMATH) + target_include_directories(video_core PRIVATE "$ENV{NSIGHT_AFTERMATH_SDK}/include") +endif() + if (MSVC) target_compile_options(video_core PRIVATE /we4267) else() - target_compile_options(video_core PRIVATE -Werror=conversion -Wno-error=sign-conversion) + target_compile_options(video_core PRIVATE + -Werror=conversion + -Wno-error=sign-conversion + -Werror=pessimizing-move + -Werror=redundant-move + -Werror=switch + -Werror=type-limits + -Werror=unused-variable + + $<$<CXX_COMPILER_ID:GNU>:-Werror=class-memaccess> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + ) endif() diff --git a/src/video_core/buffer_cache/buffer_block.h b/src/video_core/buffer_cache/buffer_block.h index e35ee0b67..e64170e66 100644 --- a/src/video_core/buffer_cache/buffer_block.h +++ b/src/video_core/buffer_cache/buffer_block.h @@ -15,48 +15,47 @@ namespace VideoCommon { class BufferBlock { public: - bool Overlaps(const VAddr start, const VAddr end) const { + bool Overlaps(VAddr start, VAddr end) const { return (cpu_addr < end) && (cpu_addr_end > start); } - bool IsInside(const VAddr other_start, const VAddr other_end) const { + bool IsInside(VAddr other_start, VAddr other_end) const { return cpu_addr <= other_start && other_end <= cpu_addr_end; } - std::size_t GetOffset(const VAddr in_addr) { + std::size_t Offset(VAddr in_addr) const { return static_cast<std::size_t>(in_addr - cpu_addr); } - VAddr GetCpuAddr() const { + VAddr CpuAddr() const { return cpu_addr; } - VAddr GetCpuAddrEnd() const { + VAddr CpuAddrEnd() const { return cpu_addr_end; } - void SetCpuAddr(const VAddr new_addr) { + void SetCpuAddr(VAddr new_addr) { cpu_addr = new_addr; cpu_addr_end = new_addr + size; } - std::size_t GetSize() const { + std::size_t Size() const { return size; } - void SetEpoch(u64 new_epoch) { - epoch = new_epoch; + u64 Epoch() const { + return epoch; } - u64 GetEpoch() { - return epoch; + void SetEpoch(u64 new_epoch) { + epoch = new_epoch; } protected: - explicit BufferBlock(VAddr cpu_addr, const std::size_t size) : size{size} { - SetCpuAddr(cpu_addr); + explicit BufferBlock(VAddr cpu_addr_, std::size_t size_) : size{size_} { + SetCpuAddr(cpu_addr_); } - ~BufferBlock() = default; private: VAddr cpu_addr{}; diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index b57c0d4d4..e7edd733f 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -4,7 +4,7 @@ #pragma once -#include <array> +#include <list> #include <memory> #include <mutex> #include <unordered_map> @@ -12,14 +12,17 @@ #include <utility> #include <vector> -#include <boost/icl/interval_map.hpp> +#include <boost/container/small_vector.hpp> #include <boost/icl/interval_set.hpp> -#include <boost/range/iterator_range.hpp> +#include <boost/intrusive/set.hpp> #include "common/alignment.h" +#include "common/assert.h" #include "common/common_types.h" +#include "common/logging/log.h" #include "core/core.h" #include "core/memory.h" +#include "core/settings.h" #include "video_core/buffer_cache/buffer_block.h" #include "video_core/buffer_cache/map_interval.h" #include "video_core/memory_manager.h" @@ -27,105 +30,122 @@ namespace VideoCommon { -using MapInterval = std::shared_ptr<MapIntervalBase>; - -template <typename TBuffer, typename TBufferType, typename StreamBuffer> +template <typename Buffer, typename BufferType, typename StreamBuffer> class BufferCache { + using IntervalSet = boost::icl::interval_set<VAddr>; + using IntervalType = typename IntervalSet::interval_type; + using VectorMapInterval = boost::container::small_vector<MapInterval*, 1>; + + static constexpr u64 WRITE_PAGE_BIT = 11; + static constexpr u64 BLOCK_PAGE_BITS = 21; + static constexpr u64 BLOCK_PAGE_SIZE = 1ULL << BLOCK_PAGE_BITS; + public: - using BufferInfo = std::pair<const TBufferType*, u64>; + struct BufferInfo { + BufferType handle; + u64 offset; + u64 address; + }; BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4, bool is_written = false, bool use_fast_cbuf = false) { std::lock_guard lock{mutex}; - const std::optional<VAddr> cpu_addr_opt = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); - - if (!cpu_addr_opt) { - return {GetEmptyBuffer(size), 0}; + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); + if (!cpu_addr) { + return GetEmptyBuffer(size); } - VAddr cpu_addr = *cpu_addr_opt; - // Cache management is a big overhead, so only cache entries with a given size. // TODO: Figure out which size is the best for given games. constexpr std::size_t max_stream_size = 0x800; if (use_fast_cbuf || size < max_stream_size) { - if (!is_written && !IsRegionWritten(cpu_addr, cpu_addr + size - 1)) { - auto& memory_manager = system.GPU().MemoryManager(); + if (!is_written && !IsRegionWritten(*cpu_addr, *cpu_addr + size - 1)) { + const bool is_granular = gpu_memory.IsGranularRange(gpu_addr, size); if (use_fast_cbuf) { - if (memory_manager.IsGranularRange(gpu_addr, size)) { - const auto host_ptr = memory_manager.GetPointer(gpu_addr); - return ConstBufferUpload(host_ptr, size); + u8* dest; + if (is_granular) { + dest = gpu_memory.GetPointer(gpu_addr); } else { staging_buffer.resize(size); - memory_manager.ReadBlockUnsafe(gpu_addr, staging_buffer.data(), size); - return ConstBufferUpload(staging_buffer.data(), size); + dest = staging_buffer.data(); + gpu_memory.ReadBlockUnsafe(gpu_addr, dest, size); } + return ConstBufferUpload(dest, size); + } + if (is_granular) { + u8* const host_ptr = gpu_memory.GetPointer(gpu_addr); + return StreamBufferUpload(size, alignment, [host_ptr, size](u8* dest) { + std::memcpy(dest, host_ptr, size); + }); } else { - if (memory_manager.IsGranularRange(gpu_addr, size)) { - const auto host_ptr = memory_manager.GetPointer(gpu_addr); - return StreamBufferUpload(host_ptr, size, alignment); - } else { - staging_buffer.resize(size); - memory_manager.ReadBlockUnsafe(gpu_addr, staging_buffer.data(), size); - return StreamBufferUpload(staging_buffer.data(), size, alignment); - } + return StreamBufferUpload(size, alignment, [this, gpu_addr, size](u8* dest) { + gpu_memory.ReadBlockUnsafe(gpu_addr, dest, size); + }); } } } - auto block = GetBlock(cpu_addr, size); - auto map = MapAddress(block, gpu_addr, cpu_addr, size); + Buffer* const block = GetBlock(*cpu_addr, size); + MapInterval* const map = MapAddress(block, gpu_addr, *cpu_addr, size); + if (!map) { + return GetEmptyBuffer(size); + } if (is_written) { map->MarkAsModified(true, GetModifiedTicks()); - if (!map->IsWritten()) { - map->MarkAsWritten(true); - MarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1); + if (Settings::IsGPULevelHigh() && + Settings::values.use_asynchronous_gpu_emulation.GetValue()) { + MarkForAsyncFlush(map); } - } else { - if (map->IsWritten()) { - WriteBarrier(); + if (!map->is_written) { + map->is_written = true; + MarkRegionAsWritten(map->start, map->end - 1); } } - const u64 offset = static_cast<u64>(block->GetOffset(cpu_addr)); - - return {ToHandle(block), offset}; + return BufferInfo{block->Handle(), block->Offset(*cpu_addr), block->Address()}; } /// Uploads from a host memory. Returns the OpenGL buffer where it's located and its offset. BufferInfo UploadHostMemory(const void* raw_pointer, std::size_t size, std::size_t alignment = 4) { std::lock_guard lock{mutex}; - return StreamBufferUpload(raw_pointer, size, alignment); + return StreamBufferUpload(size, alignment, [raw_pointer, size](u8* dest) { + std::memcpy(dest, raw_pointer, size); + }); } - void Map(std::size_t max_size) { + /// Prepares the buffer cache for data uploading + /// @param max_size Maximum number of bytes that will be uploaded + /// @return True when a stream buffer invalidation was required, false otherwise + bool Map(std::size_t max_size) { std::lock_guard lock{mutex}; + bool invalidated; std::tie(buffer_ptr, buffer_offset_base, invalidated) = stream_buffer->Map(max_size, 4); buffer_offset = buffer_offset_base; + + return invalidated; } - /// Finishes the upload stream, returns true on bindings invalidation. - bool Unmap() { + /// Finishes the upload stream + void Unmap() { std::lock_guard lock{mutex}; - stream_buffer->Unmap(buffer_offset - buffer_offset_base); - return std::exchange(invalidated, false); } + /// Function called at the end of each frame, inteded for deferred operations void TickFrame() { ++epoch; + while (!pending_destruction.empty()) { // Delay at least 4 frames before destruction. // This is due to triple buffering happening on some drivers. static constexpr u64 epochs_to_destroy = 5; - if (pending_destruction.front()->GetEpoch() + epochs_to_destroy > epoch) { + if (pending_destruction.front()->Epoch() + epochs_to_destroy > epoch) { break; } - pending_destruction.pop_front(); + pending_destruction.pop(); } } @@ -133,117 +153,193 @@ public: void FlushRegion(VAddr addr, std::size_t size) { std::lock_guard lock{mutex}; - std::vector<MapInterval> objects = GetMapsInRange(addr, size); - std::sort(objects.begin(), objects.end(), [](const MapInterval& a, const MapInterval& b) { - return a->GetModificationTick() < b->GetModificationTick(); - }); - for (auto& object : objects) { - if (object->IsModified() && object->IsRegistered()) { + VectorMapInterval objects = GetMapsInRange(addr, size); + std::sort(objects.begin(), objects.end(), + [](MapInterval* lhs, MapInterval* rhs) { return lhs->ticks < rhs->ticks; }); + for (MapInterval* object : objects) { + if (object->is_modified && object->is_registered) { + mutex.unlock(); FlushMap(object); + mutex.lock(); } } } + bool MustFlushRegion(VAddr addr, std::size_t size) { + std::lock_guard lock{mutex}; + + const VectorMapInterval objects = GetMapsInRange(addr, size); + return std::any_of(objects.cbegin(), objects.cend(), [](const MapInterval* map) { + return map->is_modified && map->is_registered; + }); + } + /// Mark the specified region as being invalidated void InvalidateRegion(VAddr addr, u64 size) { std::lock_guard lock{mutex}; - std::vector<MapInterval> objects = GetMapsInRange(addr, size); - for (auto& object : objects) { - if (object->IsRegistered()) { + for (auto& object : GetMapsInRange(addr, size)) { + if (object->is_registered) { Unregister(object); } } } - virtual const TBufferType* GetEmptyBuffer(std::size_t size) = 0; + void OnCPUWrite(VAddr addr, std::size_t size) { + std::lock_guard lock{mutex}; -protected: - explicit BufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system, - std::unique_ptr<StreamBuffer> stream_buffer) - : rasterizer{rasterizer}, system{system}, stream_buffer{std::move(stream_buffer)}, - stream_buffer_handle{this->stream_buffer->GetHandle()} {} + for (MapInterval* object : GetMapsInRange(addr, size)) { + if (object->is_memory_marked && object->is_registered) { + UnmarkMemory(object); + object->is_sync_pending = true; + marked_for_unregister.emplace_back(object); + } + } + } - ~BufferCache() = default; + void SyncGuestHost() { + std::lock_guard lock{mutex}; + + for (auto& object : marked_for_unregister) { + if (object->is_registered) { + object->is_sync_pending = false; + Unregister(object); + } + } + marked_for_unregister.clear(); + } + + void CommitAsyncFlushes() { + if (uncommitted_flushes) { + auto commit_list = std::make_shared<std::list<MapInterval*>>(); + for (MapInterval* map : *uncommitted_flushes) { + if (map->is_registered && map->is_modified) { + // TODO(Blinkhawk): Implement backend asynchronous flushing + // AsyncFlushMap(map) + commit_list->push_back(map); + } + } + if (!commit_list->empty()) { + committed_flushes.push_back(commit_list); + } else { + committed_flushes.emplace_back(); + } + } else { + committed_flushes.emplace_back(); + } + uncommitted_flushes.reset(); + } - virtual const TBufferType* ToHandle(const TBuffer& storage) = 0; + bool ShouldWaitAsyncFlushes() const { + return !committed_flushes.empty() && committed_flushes.front() != nullptr; + } - virtual void WriteBarrier() = 0; + bool HasUncommittedFlushes() const { + return uncommitted_flushes != nullptr; + } + + void PopAsyncFlushes() { + if (committed_flushes.empty()) { + return; + } + auto& flush_list = committed_flushes.front(); + if (!flush_list) { + committed_flushes.pop_front(); + return; + } + for (MapInterval* map : *flush_list) { + if (map->is_registered) { + // TODO(Blinkhawk): Replace this for reading the asynchronous flush + FlushMap(map); + } + } + committed_flushes.pop_front(); + } - virtual TBuffer CreateBlock(VAddr cpu_addr, std::size_t size) = 0; + virtual BufferInfo GetEmptyBuffer(std::size_t size) = 0; - virtual void UploadBlockData(const TBuffer& buffer, std::size_t offset, std::size_t size, - const u8* data) = 0; +protected: + explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::MemoryManager& gpu_memory_, Core::Memory::Memory& cpu_memory_, + std::unique_ptr<StreamBuffer> stream_buffer_) + : rasterizer{rasterizer_}, gpu_memory{gpu_memory_}, cpu_memory{cpu_memory_}, + stream_buffer{std::move(stream_buffer_)}, stream_buffer_handle{stream_buffer->Handle()} {} - virtual void DownloadBlockData(const TBuffer& buffer, std::size_t offset, std::size_t size, - u8* data) = 0; + ~BufferCache() = default; - virtual void CopyBlock(const TBuffer& src, const TBuffer& dst, std::size_t src_offset, - std::size_t dst_offset, std::size_t size) = 0; + virtual std::shared_ptr<Buffer> CreateBlock(VAddr cpu_addr, std::size_t size) = 0; virtual BufferInfo ConstBufferUpload(const void* raw_pointer, std::size_t size) { return {}; } /// Register an object into the cache - void Register(const MapInterval& new_map, bool inherit_written = false) { - const VAddr cpu_addr = new_map->GetStart(); + MapInterval* Register(MapInterval new_map, bool inherit_written = false) { + const VAddr cpu_addr = new_map.start; if (!cpu_addr) { LOG_CRITICAL(HW_GPU, "Failed to register buffer with unmapped gpu_address 0x{:016x}", - new_map->GetGpuAddress()); - return; + new_map.gpu_addr); + return nullptr; } - const std::size_t size = new_map->GetEnd() - new_map->GetStart(); - new_map->MarkAsRegistered(true); - const IntervalType interval{new_map->GetStart(), new_map->GetEnd()}; - mapped_addresses.insert({interval, new_map}); + const std::size_t size = new_map.end - new_map.start; + new_map.is_registered = true; rasterizer.UpdatePagesCachedCount(cpu_addr, size, 1); + new_map.is_memory_marked = true; if (inherit_written) { - MarkRegionAsWritten(new_map->GetStart(), new_map->GetEnd() - 1); - new_map->MarkAsWritten(true); + MarkRegionAsWritten(new_map.start, new_map.end - 1); + new_map.is_written = true; } + MapInterval* const storage = mapped_addresses_allocator.Allocate(); + *storage = new_map; + mapped_addresses.insert(*storage); + return storage; } - /// Unregisters an object from the cache - void Unregister(MapInterval& map) { - const std::size_t size = map->GetEnd() - map->GetStart(); - rasterizer.UpdatePagesCachedCount(map->GetStart(), size, -1); - map->MarkAsRegistered(false); - if (map->IsWritten()) { - UnmarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1); + void UnmarkMemory(MapInterval* map) { + if (!map->is_memory_marked) { + return; } - const IntervalType delete_interval{map->GetStart(), map->GetEnd()}; - mapped_addresses.erase(delete_interval); + const std::size_t size = map->end - map->start; + rasterizer.UpdatePagesCachedCount(map->start, size, -1); + map->is_memory_marked = false; } -private: - MapInterval CreateMap(const VAddr start, const VAddr end, const GPUVAddr gpu_addr) { - return std::make_shared<MapIntervalBase>(start, end, gpu_addr); + /// Unregisters an object from the cache + void Unregister(MapInterval* map) { + UnmarkMemory(map); + map->is_registered = false; + if (map->is_sync_pending) { + map->is_sync_pending = false; + marked_for_unregister.remove(map); + } + if (map->is_written) { + UnmarkRegionAsWritten(map->start, map->end - 1); + } + const auto it = mapped_addresses.find(*map); + ASSERT(it != mapped_addresses.end()); + mapped_addresses.erase(it); + mapped_addresses_allocator.Release(map); } - MapInterval MapAddress(const TBuffer& block, const GPUVAddr gpu_addr, const VAddr cpu_addr, - const std::size_t size) { - - std::vector<MapInterval> overlaps = GetMapsInRange(cpu_addr, size); +private: + MapInterval* MapAddress(Buffer* block, GPUVAddr gpu_addr, VAddr cpu_addr, std::size_t size) { + const VectorMapInterval overlaps = GetMapsInRange(cpu_addr, size); if (overlaps.empty()) { - auto& memory_manager = system.GPU().MemoryManager(); const VAddr cpu_addr_end = cpu_addr + size; - MapInterval new_map = CreateMap(cpu_addr, cpu_addr_end, gpu_addr); - if (memory_manager.IsGranularRange(gpu_addr, size)) { - u8* host_ptr = memory_manager.GetPointer(gpu_addr); - UploadBlockData(block, block->GetOffset(cpu_addr), size, host_ptr); + if (gpu_memory.IsGranularRange(gpu_addr, size)) { + u8* const host_ptr = gpu_memory.GetPointer(gpu_addr); + block->Upload(block->Offset(cpu_addr), size, host_ptr); } else { staging_buffer.resize(size); - memory_manager.ReadBlockUnsafe(gpu_addr, staging_buffer.data(), size); - UploadBlockData(block, block->GetOffset(cpu_addr), size, staging_buffer.data()); + gpu_memory.ReadBlockUnsafe(gpu_addr, staging_buffer.data(), size); + block->Upload(block->Offset(cpu_addr), size, staging_buffer.data()); } - Register(new_map); - return new_map; + return Register(MapInterval(cpu_addr, cpu_addr_end, gpu_addr)); } const VAddr cpu_addr_end = cpu_addr + size; if (overlaps.size() == 1) { - MapInterval& current_map = overlaps[0]; + MapInterval* const current_map = overlaps[0]; if (current_map->IsInside(cpu_addr, cpu_addr_end)) { return current_map; } @@ -253,57 +349,70 @@ private: bool write_inheritance = false; bool modified_inheritance = false; // Calculate new buffer parameters - for (auto& overlap : overlaps) { - new_start = std::min(overlap->GetStart(), new_start); - new_end = std::max(overlap->GetEnd(), new_end); - write_inheritance |= overlap->IsWritten(); - modified_inheritance |= overlap->IsModified(); + for (MapInterval* overlap : overlaps) { + new_start = std::min(overlap->start, new_start); + new_end = std::max(overlap->end, new_end); + write_inheritance |= overlap->is_written; + modified_inheritance |= overlap->is_modified; } GPUVAddr new_gpu_addr = gpu_addr + new_start - cpu_addr; for (auto& overlap : overlaps) { Unregister(overlap); } UpdateBlock(block, new_start, new_end, overlaps); - MapInterval new_map = CreateMap(new_start, new_end, new_gpu_addr); + + const MapInterval new_map{new_start, new_end, new_gpu_addr}; + MapInterval* const map = Register(new_map, write_inheritance); + if (!map) { + return nullptr; + } if (modified_inheritance) { - new_map->MarkAsModified(true, GetModifiedTicks()); + map->MarkAsModified(true, GetModifiedTicks()); + if (Settings::IsGPULevelHigh() && + Settings::values.use_asynchronous_gpu_emulation.GetValue()) { + MarkForAsyncFlush(map); + } } - Register(new_map, write_inheritance); - return new_map; + return map; } - void UpdateBlock(const TBuffer& block, VAddr start, VAddr end, - std::vector<MapInterval>& overlaps) { + void UpdateBlock(Buffer* block, VAddr start, VAddr end, const VectorMapInterval& overlaps) { const IntervalType base_interval{start, end}; IntervalSet interval_set{}; interval_set.add(base_interval); for (auto& overlap : overlaps) { - const IntervalType subtract{overlap->GetStart(), overlap->GetEnd()}; + const IntervalType subtract{overlap->start, overlap->end}; interval_set.subtract(subtract); } for (auto& interval : interval_set) { - std::size_t size = interval.upper() - interval.lower(); - if (size > 0) { - staging_buffer.resize(size); - system.Memory().ReadBlockUnsafe(interval.lower(), staging_buffer.data(), size); - UploadBlockData(block, block->GetOffset(interval.lower()), size, - staging_buffer.data()); + const std::size_t size = interval.upper() - interval.lower(); + if (size == 0) { + continue; } + staging_buffer.resize(size); + cpu_memory.ReadBlockUnsafe(interval.lower(), staging_buffer.data(), size); + block->Upload(block->Offset(interval.lower()), size, staging_buffer.data()); } } - std::vector<MapInterval> GetMapsInRange(VAddr addr, std::size_t size) { + VectorMapInterval GetMapsInRange(VAddr addr, std::size_t size) { + VectorMapInterval result; if (size == 0) { - return {}; + return result; } - std::vector<MapInterval> objects{}; - const IntervalType interval{addr, addr + size}; - for (auto& pair : boost::make_iterator_range(mapped_addresses.equal_range(interval))) { - objects.push_back(pair.second); + const VAddr addr_end = addr + size; + auto it = mapped_addresses.lower_bound(addr); + if (it != mapped_addresses.begin()) { + --it; } - - return objects; + while (it != mapped_addresses.end() && it->start < addr_end) { + if (it->Overlaps(addr, addr_end)) { + result.push_back(&*it); + } + ++it; + } + return result; } /// Returns a ticks counter used for tracking when cached objects were last modified @@ -311,24 +420,28 @@ private: return ++modified_ticks; } - void FlushMap(MapInterval map) { - std::size_t size = map->GetEnd() - map->GetStart(); - TBuffer block = blocks[map->GetStart() >> block_page_bits]; + void FlushMap(MapInterval* map) { + const auto it = blocks.find(map->start >> BLOCK_PAGE_BITS); + ASSERT_OR_EXECUTE(it != blocks.end(), return;); + + std::shared_ptr<Buffer> block = it->second; + + const std::size_t size = map->end - map->start; staging_buffer.resize(size); - DownloadBlockData(block, block->GetOffset(map->GetStart()), size, staging_buffer.data()); - system.Memory().WriteBlockUnsafe(map->GetStart(), staging_buffer.data(), size); + block->Download(block->Offset(map->start), size, staging_buffer.data()); + cpu_memory.WriteBlockUnsafe(map->start, staging_buffer.data(), size); map->MarkAsModified(false, 0); } - BufferInfo StreamBufferUpload(const void* raw_pointer, std::size_t size, - std::size_t alignment) { + template <typename Callable> + BufferInfo StreamBufferUpload(std::size_t size, std::size_t alignment, Callable&& callable) { AlignBuffer(alignment); const std::size_t uploaded_offset = buffer_offset; - std::memcpy(buffer_ptr, raw_pointer, size); + callable(buffer_ptr); buffer_ptr += size; buffer_offset += size; - return {&stream_buffer_handle, uploaded_offset}; + return BufferInfo{stream_buffer->Handle(), uploaded_offset, stream_buffer->Address()}; } void AlignBuffer(std::size_t alignment) { @@ -338,151 +451,148 @@ private: buffer_offset = offset_aligned; } - TBuffer EnlargeBlock(TBuffer buffer) { - const std::size_t old_size = buffer->GetSize(); - const std::size_t new_size = old_size + block_page_size; - const VAddr cpu_addr = buffer->GetCpuAddr(); - TBuffer new_buffer = CreateBlock(cpu_addr, new_size); - CopyBlock(buffer, new_buffer, 0, 0, old_size); - buffer->SetEpoch(epoch); - pending_destruction.push_back(buffer); + std::shared_ptr<Buffer> EnlargeBlock(std::shared_ptr<Buffer> buffer) { + const std::size_t old_size = buffer->Size(); + const std::size_t new_size = old_size + BLOCK_PAGE_SIZE; + const VAddr cpu_addr = buffer->CpuAddr(); + std::shared_ptr<Buffer> new_buffer = CreateBlock(cpu_addr, new_size); + new_buffer->CopyFrom(*buffer, 0, 0, old_size); + QueueDestruction(std::move(buffer)); + const VAddr cpu_addr_end = cpu_addr + new_size - 1; - u64 page_start = cpu_addr >> block_page_bits; - const u64 page_end = cpu_addr_end >> block_page_bits; - while (page_start <= page_end) { - blocks[page_start] = new_buffer; - ++page_start; + const u64 page_end = cpu_addr_end >> BLOCK_PAGE_BITS; + for (u64 page_start = cpu_addr >> BLOCK_PAGE_BITS; page_start <= page_end; ++page_start) { + blocks.insert_or_assign(page_start, new_buffer); } + return new_buffer; } - TBuffer MergeBlocks(TBuffer first, TBuffer second) { - const std::size_t size_1 = first->GetSize(); - const std::size_t size_2 = second->GetSize(); - const VAddr first_addr = first->GetCpuAddr(); - const VAddr second_addr = second->GetCpuAddr(); + std::shared_ptr<Buffer> MergeBlocks(std::shared_ptr<Buffer> first, + std::shared_ptr<Buffer> second) { + const std::size_t size_1 = first->Size(); + const std::size_t size_2 = second->Size(); + const VAddr first_addr = first->CpuAddr(); + const VAddr second_addr = second->CpuAddr(); const VAddr new_addr = std::min(first_addr, second_addr); const std::size_t new_size = size_1 + size_2; - TBuffer new_buffer = CreateBlock(new_addr, new_size); - CopyBlock(first, new_buffer, 0, new_buffer->GetOffset(first_addr), size_1); - CopyBlock(second, new_buffer, 0, new_buffer->GetOffset(second_addr), size_2); - first->SetEpoch(epoch); - second->SetEpoch(epoch); - pending_destruction.push_back(first); - pending_destruction.push_back(second); + + std::shared_ptr<Buffer> new_buffer = CreateBlock(new_addr, new_size); + new_buffer->CopyFrom(*first, 0, new_buffer->Offset(first_addr), size_1); + new_buffer->CopyFrom(*second, 0, new_buffer->Offset(second_addr), size_2); + QueueDestruction(std::move(first)); + QueueDestruction(std::move(second)); + const VAddr cpu_addr_end = new_addr + new_size - 1; - u64 page_start = new_addr >> block_page_bits; - const u64 page_end = cpu_addr_end >> block_page_bits; - while (page_start <= page_end) { - blocks[page_start] = new_buffer; - ++page_start; + const u64 page_end = cpu_addr_end >> BLOCK_PAGE_BITS; + for (u64 page_start = new_addr >> BLOCK_PAGE_BITS; page_start <= page_end; ++page_start) { + blocks.insert_or_assign(page_start, new_buffer); } return new_buffer; } - TBuffer GetBlock(const VAddr cpu_addr, const std::size_t size) { - TBuffer found{}; + Buffer* GetBlock(VAddr cpu_addr, std::size_t size) { + std::shared_ptr<Buffer> found; + const VAddr cpu_addr_end = cpu_addr + size - 1; - u64 page_start = cpu_addr >> block_page_bits; - const u64 page_end = cpu_addr_end >> block_page_bits; - while (page_start <= page_end) { + const u64 page_end = cpu_addr_end >> BLOCK_PAGE_BITS; + for (u64 page_start = cpu_addr >> BLOCK_PAGE_BITS; page_start <= page_end; ++page_start) { auto it = blocks.find(page_start); if (it == blocks.end()) { if (found) { found = EnlargeBlock(found); - } else { - const VAddr start_addr = (page_start << block_page_bits); - found = CreateBlock(start_addr, block_page_size); - blocks[page_start] = found; - } - } else { - if (found) { - if (found == it->second) { - ++page_start; - continue; - } - found = MergeBlocks(found, it->second); - } else { - found = it->second; + continue; } + const VAddr start_addr = page_start << BLOCK_PAGE_BITS; + found = CreateBlock(start_addr, BLOCK_PAGE_SIZE); + blocks.insert_or_assign(page_start, found); + continue; + } + if (!found) { + found = it->second; + continue; + } + if (found != it->second) { + found = MergeBlocks(std::move(found), it->second); } - ++page_start; } - return found; + return found.get(); } - void MarkRegionAsWritten(const VAddr start, const VAddr end) { - u64 page_start = start >> write_page_bit; - const u64 page_end = end >> write_page_bit; - while (page_start <= page_end) { - auto it = written_pages.find(page_start); - if (it != written_pages.end()) { - it->second = it->second + 1; - } else { - written_pages[page_start] = 1; + void MarkRegionAsWritten(VAddr start, VAddr end) { + const u64 page_end = end >> WRITE_PAGE_BIT; + for (u64 page_start = start >> WRITE_PAGE_BIT; page_start <= page_end; ++page_start) { + if (const auto [it, inserted] = written_pages.emplace(page_start, 1); !inserted) { + ++it->second; } - page_start++; } } - void UnmarkRegionAsWritten(const VAddr start, const VAddr end) { - u64 page_start = start >> write_page_bit; - const u64 page_end = end >> write_page_bit; - while (page_start <= page_end) { + void UnmarkRegionAsWritten(VAddr start, VAddr end) { + const u64 page_end = end >> WRITE_PAGE_BIT; + for (u64 page_start = start >> WRITE_PAGE_BIT; page_start <= page_end; ++page_start) { auto it = written_pages.find(page_start); if (it != written_pages.end()) { if (it->second > 1) { - it->second = it->second - 1; + --it->second; } else { written_pages.erase(it); } } - page_start++; } } - bool IsRegionWritten(const VAddr start, const VAddr end) const { - u64 page_start = start >> write_page_bit; - const u64 page_end = end >> write_page_bit; - while (page_start <= page_end) { + bool IsRegionWritten(VAddr start, VAddr end) const { + const u64 page_end = end >> WRITE_PAGE_BIT; + for (u64 page_start = start >> WRITE_PAGE_BIT; page_start <= page_end; ++page_start) { if (written_pages.count(page_start) > 0) { return true; } - page_start++; } return false; } + void QueueDestruction(std::shared_ptr<Buffer> buffer) { + buffer->SetEpoch(epoch); + pending_destruction.push(std::move(buffer)); + } + + void MarkForAsyncFlush(MapInterval* map) { + if (!uncommitted_flushes) { + uncommitted_flushes = std::make_shared<std::unordered_set<MapInterval*>>(); + } + uncommitted_flushes->insert(map); + } + VideoCore::RasterizerInterface& rasterizer; - Core::System& system; + Tegra::MemoryManager& gpu_memory; + Core::Memory::Memory& cpu_memory; std::unique_ptr<StreamBuffer> stream_buffer; - TBufferType stream_buffer_handle{}; - - bool invalidated = false; + BufferType stream_buffer_handle; u8* buffer_ptr = nullptr; u64 buffer_offset = 0; u64 buffer_offset_base = 0; - using IntervalSet = boost::icl::interval_set<VAddr>; - using IntervalCache = boost::icl::interval_map<VAddr, MapInterval>; - using IntervalType = typename IntervalCache::interval_type; - IntervalCache mapped_addresses; + MapIntervalAllocator mapped_addresses_allocator; + boost::intrusive::set<MapInterval, boost::intrusive::compare<MapIntervalCompare>> + mapped_addresses; - static constexpr u64 write_page_bit = 11; std::unordered_map<u64, u32> written_pages; + std::unordered_map<u64, std::shared_ptr<Buffer>> blocks; - static constexpr u64 block_page_bits = 21; - static constexpr u64 block_page_size = 1ULL << block_page_bits; - std::unordered_map<u64, TBuffer> blocks; - - std::list<TBuffer> pending_destruction; + std::queue<std::shared_ptr<Buffer>> pending_destruction; u64 epoch = 0; u64 modified_ticks = 0; std::vector<u8> staging_buffer; + std::list<MapInterval*> marked_for_unregister; + + std::shared_ptr<std::unordered_set<MapInterval*>> uncommitted_flushes; + std::list<std::shared_ptr<std::list<MapInterval*>>> committed_flushes; + std::recursive_mutex mutex; }; diff --git a/src/video_core/buffer_cache/map_interval.cpp b/src/video_core/buffer_cache/map_interval.cpp new file mode 100644 index 000000000..62587e18a --- /dev/null +++ b/src/video_core/buffer_cache/map_interval.cpp @@ -0,0 +1,33 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <array> +#include <cstddef> +#include <memory> + +#include "video_core/buffer_cache/map_interval.h" + +namespace VideoCommon { + +MapIntervalAllocator::MapIntervalAllocator() { + FillFreeList(first_chunk); +} + +MapIntervalAllocator::~MapIntervalAllocator() = default; + +void MapIntervalAllocator::AllocateNewChunk() { + *new_chunk = std::make_unique<Chunk>(); + FillFreeList(**new_chunk); + new_chunk = &(*new_chunk)->next; +} + +void MapIntervalAllocator::FillFreeList(Chunk& chunk) { + const std::size_t old_size = free_list.size(); + free_list.resize(old_size + chunk.data.size()); + std::transform(chunk.data.rbegin(), chunk.data.rend(), free_list.begin() + old_size, + [](MapInterval& interval) { return &interval; }); +} + +} // namespace VideoCommon diff --git a/src/video_core/buffer_cache/map_interval.h b/src/video_core/buffer_cache/map_interval.h index b0956029d..fe0bcd1d8 100644 --- a/src/video_core/buffer_cache/map_interval.h +++ b/src/video_core/buffer_cache/map_interval.h @@ -4,86 +4,89 @@ #pragma once +#include <array> +#include <cstddef> +#include <memory> +#include <vector> + +#include <boost/intrusive/set_hook.hpp> + #include "common/common_types.h" #include "video_core/gpu.h" namespace VideoCommon { -class MapIntervalBase { -public: - MapIntervalBase(const VAddr start, const VAddr end, const GPUVAddr gpu_addr) - : start{start}, end{end}, gpu_addr{gpu_addr} {} +struct MapInterval : public boost::intrusive::set_base_hook<boost::intrusive::optimize_size<true>> { + MapInterval() = default; - void SetCpuAddress(VAddr new_cpu_addr) { - cpu_addr = new_cpu_addr; - } + /*implicit*/ MapInterval(VAddr start_) noexcept : start{start_} {} - VAddr GetCpuAddress() const { - return cpu_addr; - } + explicit MapInterval(VAddr start_, VAddr end_, GPUVAddr gpu_addr_) noexcept + : start{start_}, end{end_}, gpu_addr{gpu_addr_} {} - GPUVAddr GetGpuAddress() const { - return gpu_addr; + bool IsInside(VAddr other_start, VAddr other_end) const noexcept { + return start <= other_start && other_end <= end; } - bool IsInside(const VAddr other_start, const VAddr other_end) const { - return (start <= other_start && other_end <= end); + bool Overlaps(VAddr other_start, VAddr other_end) const noexcept { + return start < other_end && other_start < end; } - bool operator==(const MapIntervalBase& rhs) const { - return std::tie(start, end) == std::tie(rhs.start, rhs.end); - } - - bool operator!=(const MapIntervalBase& rhs) const { - return !operator==(rhs); - } - - void MarkAsRegistered(const bool registered) { - is_registered = registered; + void MarkAsModified(bool is_modified_, u64 ticks_) noexcept { + is_modified = is_modified_; + ticks = ticks_; } - bool IsRegistered() const { - return is_registered; - } + boost::intrusive::set_member_hook<> member_hook_; + VAddr start = 0; + VAddr end = 0; + GPUVAddr gpu_addr = 0; + u64 ticks = 0; + bool is_written = false; + bool is_modified = false; + bool is_registered = false; + bool is_memory_marked = false; + bool is_sync_pending = false; +}; - VAddr GetStart() const { - return start; +struct MapIntervalCompare { + constexpr bool operator()(const MapInterval& lhs, const MapInterval& rhs) const noexcept { + return lhs.start < rhs.start; } +}; - VAddr GetEnd() const { - return end; +class MapIntervalAllocator { +public: + MapIntervalAllocator(); + ~MapIntervalAllocator(); + + MapInterval* Allocate() { + if (free_list.empty()) { + AllocateNewChunk(); + } + MapInterval* const interval = free_list.back(); + free_list.pop_back(); + return interval; } - void MarkAsModified(const bool is_modified_, const u64 tick) { - is_modified = is_modified_; - ticks = tick; + void Release(MapInterval* interval) { + free_list.push_back(interval); } - bool IsModified() const { - return is_modified; - } +private: + struct Chunk { + std::unique_ptr<Chunk> next; + std::array<MapInterval, 0x8000> data; + }; - u64 GetModificationTick() const { - return ticks; - } + void AllocateNewChunk(); - void MarkAsWritten(const bool is_written_) { - is_written = is_written_; - } + void FillFreeList(Chunk& chunk); - bool IsWritten() const { - return is_written; - } + std::vector<MapInterval*> free_list; + std::unique_ptr<Chunk>* new_chunk = &first_chunk.next; -private: - VAddr start; - VAddr end; - GPUVAddr gpu_addr; - VAddr cpu_addr{}; - bool is_written{}; - bool is_modified{}; - bool is_registered{}; - u64 ticks{}; + Chunk first_chunk; }; } // namespace VideoCommon diff --git a/src/video_core/cdma_pusher.cpp b/src/video_core/cdma_pusher.cpp new file mode 100644 index 000000000..b60f86260 --- /dev/null +++ b/src/video_core/cdma_pusher.cpp @@ -0,0 +1,171 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include "command_classes/host1x.h" +#include "command_classes/nvdec.h" +#include "command_classes/vic.h" +#include "common/bit_util.h" +#include "video_core/cdma_pusher.h" +#include "video_core/command_classes/nvdec_common.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra { +CDmaPusher::CDmaPusher(GPU& gpu) + : gpu(gpu), nvdec_processor(std::make_shared<Nvdec>(gpu)), + vic_processor(std::make_unique<Vic>(gpu, nvdec_processor)), + host1x_processor(std::make_unique<Host1x>(gpu)), + nvdec_sync(std::make_unique<SyncptIncrManager>(gpu)), + vic_sync(std::make_unique<SyncptIncrManager>(gpu)) {} + +CDmaPusher::~CDmaPusher() = default; + +void CDmaPusher::Push(ChCommandHeaderList&& entries) { + cdma_queue.push(std::move(entries)); +} + +void CDmaPusher::DispatchCalls() { + while (!cdma_queue.empty()) { + Step(); + } +} + +void CDmaPusher::Step() { + const auto entries{cdma_queue.front()}; + cdma_queue.pop(); + + std::vector<u32> values(entries.size()); + std::memcpy(values.data(), entries.data(), entries.size() * sizeof(u32)); + + for (const u32 value : values) { + if (mask != 0) { + const u32 lbs = Common::CountTrailingZeroes32(mask); + mask &= ~(1U << lbs); + ExecuteCommand(static_cast<u32>(offset + lbs), value); + continue; + } else if (count != 0) { + --count; + ExecuteCommand(static_cast<u32>(offset), value); + if (incrementing) { + ++offset; + } + continue; + } + const auto mode = static_cast<ChSubmissionMode>((value >> 28) & 0xf); + switch (mode) { + case ChSubmissionMode::SetClass: { + mask = value & 0x3f; + offset = (value >> 16) & 0xfff; + current_class = static_cast<ChClassId>((value >> 6) & 0x3ff); + break; + } + case ChSubmissionMode::Incrementing: + case ChSubmissionMode::NonIncrementing: + count = value & 0xffff; + offset = (value >> 16) & 0xfff; + incrementing = mode == ChSubmissionMode::Incrementing; + break; + case ChSubmissionMode::Mask: + mask = value & 0xffff; + offset = (value >> 16) & 0xfff; + break; + case ChSubmissionMode::Immediate: { + const u32 data = value & 0xfff; + offset = (value >> 16) & 0xfff; + ExecuteCommand(static_cast<u32>(offset), data); + break; + } + default: + UNIMPLEMENTED_MSG("ChSubmission mode {} is not implemented!", static_cast<u32>(mode)); + break; + } + } +} + +void CDmaPusher::ExecuteCommand(u32 offset, u32 data) { + switch (current_class) { + case ChClassId::NvDec: + ThiStateWrite(nvdec_thi_state, offset, {data}); + switch (static_cast<ThiMethod>(offset)) { + case ThiMethod::IncSyncpt: { + LOG_DEBUG(Service_NVDRV, "NVDEC Class IncSyncpt Method"); + const auto syncpoint_id = static_cast<u32>(data & 0xFF); + const auto cond = static_cast<u32>((data >> 8) & 0xFF); + if (cond == 0) { + nvdec_sync->Increment(syncpoint_id); + } else { + nvdec_sync->IncrementWhenDone(static_cast<u32>(current_class), syncpoint_id); + nvdec_sync->SignalDone(syncpoint_id); + } + break; + } + case ThiMethod::SetMethod1: + LOG_DEBUG(Service_NVDRV, "NVDEC method 0x{:X}", + static_cast<u32>(nvdec_thi_state.method_0)); + nvdec_processor->ProcessMethod( + static_cast<Tegra::Nvdec::Method>(nvdec_thi_state.method_0), {data}); + break; + default: + break; + } + break; + case ChClassId::GraphicsVic: + ThiStateWrite(vic_thi_state, static_cast<u32>(offset), {data}); + switch (static_cast<ThiMethod>(offset)) { + case ThiMethod::IncSyncpt: { + LOG_DEBUG(Service_NVDRV, "VIC Class IncSyncpt Method"); + const auto syncpoint_id = static_cast<u32>(data & 0xFF); + const auto cond = static_cast<u32>((data >> 8) & 0xFF); + if (cond == 0) { + vic_sync->Increment(syncpoint_id); + } else { + vic_sync->IncrementWhenDone(static_cast<u32>(current_class), syncpoint_id); + vic_sync->SignalDone(syncpoint_id); + } + break; + } + case ThiMethod::SetMethod1: + LOG_DEBUG(Service_NVDRV, "VIC method 0x{:X}, Args=({})", + static_cast<u32>(vic_thi_state.method_0), data); + vic_processor->ProcessMethod(static_cast<Tegra::Vic::Method>(vic_thi_state.method_0), + {data}); + break; + default: + break; + } + break; + case ChClassId::Host1x: + // This device is mainly for syncpoint synchronization + LOG_DEBUG(Service_NVDRV, "Host1X Class Method"); + host1x_processor->ProcessMethod(static_cast<Tegra::Host1x::Method>(offset), {data}); + break; + default: + UNIMPLEMENTED_MSG("Current class not implemented {:X}", static_cast<u32>(current_class)); + break; + } +} + +void CDmaPusher::ThiStateWrite(ThiRegisters& state, u32 offset, const std::vector<u32>& arguments) { + u8* const state_offset = reinterpret_cast<u8*>(&state) + sizeof(u32) * offset; + std::memcpy(state_offset, arguments.data(), sizeof(u32) * arguments.size()); +} + +} // namespace Tegra diff --git a/src/video_core/cdma_pusher.h b/src/video_core/cdma_pusher.h new file mode 100644 index 000000000..982f309c5 --- /dev/null +++ b/src/video_core/cdma_pusher.h @@ -0,0 +1,138 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <unordered_map> +#include <vector> +#include <queue> + +#include "common/bit_field.h" +#include "common/common_types.h" +#include "video_core/command_classes/sync_manager.h" + +namespace Tegra { + +class GPU; +class Nvdec; +class Vic; +class Host1x; + +enum class ChSubmissionMode : u32 { + SetClass = 0, + Incrementing = 1, + NonIncrementing = 2, + Mask = 3, + Immediate = 4, + Restart = 5, + Gather = 6, +}; + +enum class ChClassId : u32 { + NoClass = 0x0, + Host1x = 0x1, + VideoEncodeMpeg = 0x20, + VideoEncodeNvEnc = 0x21, + VideoStreamingVi = 0x30, + VideoStreamingIsp = 0x32, + VideoStreamingIspB = 0x34, + VideoStreamingViI2c = 0x36, + GraphicsVic = 0x5d, + Graphics3D = 0x60, + GraphicsGpu = 0x61, + Tsec = 0xe0, + TsecB = 0xe1, + NvJpg = 0xc0, + NvDec = 0xf0 +}; + +enum class ChMethod : u32 { + Empty = 0, + SetMethod = 0x10, + SetData = 0x11, +}; + +union ChCommandHeader { + u32 raw; + BitField<0, 16, u32> value; + BitField<16, 12, ChMethod> method_offset; + BitField<28, 4, ChSubmissionMode> submission_mode; +}; +static_assert(sizeof(ChCommandHeader) == sizeof(u32), "ChCommand header is an invalid size"); + +struct ChCommand { + ChClassId class_id{}; + int method_offset{}; + std::vector<u32> arguments; +}; + +using ChCommandHeaderList = std::vector<Tegra::ChCommandHeader>; +using ChCommandList = std::vector<Tegra::ChCommand>; + +struct ThiRegisters { + u32_le increment_syncpt{}; + INSERT_PADDING_WORDS(1); + u32_le increment_syncpt_error{}; + u32_le ctx_switch_incremement_syncpt{}; + INSERT_PADDING_WORDS(4); + u32_le ctx_switch{}; + INSERT_PADDING_WORDS(1); + u32_le ctx_syncpt_eof{}; + INSERT_PADDING_WORDS(5); + u32_le method_0{}; + u32_le method_1{}; + INSERT_PADDING_WORDS(12); + u32_le int_status{}; + u32_le int_mask{}; +}; + +enum class ThiMethod : u32 { + IncSyncpt = offsetof(ThiRegisters, increment_syncpt) / sizeof(u32), + SetMethod0 = offsetof(ThiRegisters, method_0) / sizeof(u32), + SetMethod1 = offsetof(ThiRegisters, method_1) / sizeof(u32), +}; + +class CDmaPusher { +public: + explicit CDmaPusher(GPU& gpu); + ~CDmaPusher(); + + /// Push NVDEC command buffer entries into queue + void Push(ChCommandHeaderList&& entries); + + /// Process queued command buffer entries + void DispatchCalls(); + + /// Process one queue element + void Step(); + + /// Invoke command class devices to execute the command based on the current state + void ExecuteCommand(u32 offset, u32 data); + +private: + /// Write arguments value to the ThiRegisters member at the specified offset + void ThiStateWrite(ThiRegisters& state, u32 offset, const std::vector<u32>& arguments); + + GPU& gpu; + + std::shared_ptr<Tegra::Nvdec> nvdec_processor; + std::unique_ptr<Tegra::Vic> vic_processor; + std::unique_ptr<Tegra::Host1x> host1x_processor; + std::unique_ptr<SyncptIncrManager> nvdec_sync; + std::unique_ptr<SyncptIncrManager> vic_sync; + ChClassId current_class{}; + ThiRegisters vic_thi_state{}; + ThiRegisters nvdec_thi_state{}; + + s32 count{}; + s32 offset{}; + s32 mask{}; + bool incrementing{}; + + // Queue of command lists to be processed + std::queue<ChCommandHeaderList> cdma_queue; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/codec.cpp b/src/video_core/command_classes/codecs/codec.cpp new file mode 100644 index 000000000..1adf3cd13 --- /dev/null +++ b/src/video_core/command_classes/codecs/codec.cpp @@ -0,0 +1,115 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> +#include <fstream> +#include <vector> +#include "common/assert.h" +#include "video_core/command_classes/codecs/codec.h" +#include "video_core/command_classes/codecs/h264.h" +#include "video_core/command_classes/codecs/vp9.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +extern "C" { +#include <libavutil/opt.h> +} + +namespace Tegra { + +Codec::Codec(GPU& gpu_) + : gpu(gpu_), h264_decoder(std::make_unique<Decoder::H264>(gpu)), + vp9_decoder(std::make_unique<Decoder::VP9>(gpu)) {} + +Codec::~Codec() { + if (!initialized) { + return; + } + // Free libav memory + avcodec_send_packet(av_codec_ctx, nullptr); + avcodec_receive_frame(av_codec_ctx, av_frame); + avcodec_flush_buffers(av_codec_ctx); + + av_frame_unref(av_frame); + av_free(av_frame); + avcodec_close(av_codec_ctx); +} + +void Codec::SetTargetCodec(NvdecCommon::VideoCodec codec) { + LOG_INFO(Service_NVDRV, "NVDEC video codec initialized to {}", static_cast<u32>(codec)); + current_codec = codec; +} + +void Codec::StateWrite(u32 offset, u64 arguments) { + u8* const state_offset = reinterpret_cast<u8*>(&state) + offset * sizeof(u64); + std::memcpy(state_offset, &arguments, sizeof(u64)); +} + +void Codec::Decode() { + bool is_first_frame = false; + + if (!initialized) { + if (current_codec == NvdecCommon::VideoCodec::H264) { + av_codec = avcodec_find_decoder(AV_CODEC_ID_H264); + } else if (current_codec == NvdecCommon::VideoCodec::Vp9) { + av_codec = avcodec_find_decoder(AV_CODEC_ID_VP9); + } else { + LOG_ERROR(Service_NVDRV, "Unknown video codec {}", static_cast<u32>(current_codec)); + return; + } + + av_codec_ctx = avcodec_alloc_context3(av_codec); + av_frame = av_frame_alloc(); + av_opt_set(av_codec_ctx->priv_data, "tune", "zerolatency", 0); + + // TODO(ameerj): libavcodec gpu hw acceleration + + const auto av_error = avcodec_open2(av_codec_ctx, av_codec, nullptr); + if (av_error < 0) { + LOG_ERROR(Service_NVDRV, "avcodec_open2() Failed."); + av_frame_unref(av_frame); + av_free(av_frame); + avcodec_close(av_codec_ctx); + return; + } + initialized = true; + is_first_frame = true; + } + bool vp9_hidden_frame = false; + + AVPacket packet{}; + av_init_packet(&packet); + std::vector<u8> frame_data; + + if (current_codec == NvdecCommon::VideoCodec::H264) { + frame_data = h264_decoder->ComposeFrameHeader(state, is_first_frame); + } else if (current_codec == NvdecCommon::VideoCodec::Vp9) { + frame_data = vp9_decoder->ComposeFrameHeader(state); + vp9_hidden_frame = vp9_decoder->WasFrameHidden(); + } + + packet.data = frame_data.data(); + packet.size = static_cast<int>(frame_data.size()); + + avcodec_send_packet(av_codec_ctx, &packet); + + if (!vp9_hidden_frame) { + // Only receive/store visible frames + avcodec_receive_frame(av_codec_ctx, av_frame); + } +} + +AVFrame* Codec::GetCurrentFrame() { + return av_frame; +} + +const AVFrame* Codec::GetCurrentFrame() const { + return av_frame; +} + +NvdecCommon::VideoCodec Codec::GetCurrentCodec() const { + return current_codec; +} + +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/codec.h b/src/video_core/command_classes/codecs/codec.h new file mode 100644 index 000000000..5bbe6a332 --- /dev/null +++ b/src/video_core/command_classes/codecs/codec.h @@ -0,0 +1,66 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include "common/common_types.h" +#include "video_core/command_classes/nvdec_common.h" + +extern "C" { +#if defined(__GNUC__) || defined(__clang__) +#pragma GCC diagnostic ignored "-Wconversion" +#endif +#include <libavcodec/avcodec.h> +#if defined(__GNUC__) || defined(__clang__) +#pragma GCC diagnostic pop +#endif +} + +namespace Tegra { +class GPU; +struct VicRegisters; + +namespace Decoder { +class H264; +class VP9; +} // namespace Decoder + +class Codec { +public: + explicit Codec(GPU& gpu); + ~Codec(); + + /// Sets NVDEC video stream codec + void SetTargetCodec(NvdecCommon::VideoCodec codec); + + /// Populate NvdecRegisters state with argument value at the provided offset + void StateWrite(u32 offset, u64 arguments); + + /// Call decoders to construct headers, decode AVFrame with ffmpeg + void Decode(); + + /// Returns most recently decoded frame + [[nodiscard]] AVFrame* GetCurrentFrame(); + [[nodiscard]] const AVFrame* GetCurrentFrame() const; + + /// Returns the value of current_codec + [[nodiscard]] NvdecCommon::VideoCodec GetCurrentCodec() const; + +private: + bool initialized{}; + NvdecCommon::VideoCodec current_codec{NvdecCommon::VideoCodec::None}; + + AVCodec* av_codec{nullptr}; + AVCodecContext* av_codec_ctx{nullptr}; + AVFrame* av_frame{nullptr}; + + GPU& gpu; + std::unique_ptr<Decoder::H264> h264_decoder; + std::unique_ptr<Decoder::VP9> vp9_decoder; + + NvdecCommon::NvdecRegisters state{}; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/h264.cpp b/src/video_core/command_classes/codecs/h264.cpp new file mode 100644 index 000000000..33e063e20 --- /dev/null +++ b/src/video_core/command_classes/codecs/h264.cpp @@ -0,0 +1,293 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include <array> +#include "common/bit_util.h" +#include "video_core/command_classes/codecs/h264.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra::Decoder { +namespace { +// ZigZag LUTs from libavcodec. +constexpr std::array<u8, 64> zig_zag_direct{ + 0, 1, 8, 16, 9, 2, 3, 10, 17, 24, 32, 25, 18, 11, 4, 5, 12, 19, 26, 33, 40, 48, + 41, 34, 27, 20, 13, 6, 7, 14, 21, 28, 35, 42, 49, 56, 57, 50, 43, 36, 29, 22, 15, 23, + 30, 37, 44, 51, 58, 59, 52, 45, 38, 31, 39, 46, 53, 60, 61, 54, 47, 55, 62, 63, +}; + +constexpr std::array<u8, 16> zig_zag_scan{ + 0 + 0 * 4, 1 + 0 * 4, 0 + 1 * 4, 0 + 2 * 4, 1 + 1 * 4, 2 + 0 * 4, 3 + 0 * 4, 2 + 1 * 4, + 1 + 2 * 4, 0 + 3 * 4, 1 + 3 * 4, 2 + 2 * 4, 3 + 1 * 4, 3 + 2 * 4, 2 + 3 * 4, 3 + 3 * 4, +}; +} // Anonymous namespace + +H264::H264(GPU& gpu_) : gpu(gpu_) {} + +H264::~H264() = default; + +const std::vector<u8>& H264::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, + bool is_first_frame) { + H264DecoderContext context{}; + gpu.MemoryManager().ReadBlock(state.picture_info_offset, &context, sizeof(H264DecoderContext)); + + const s32 frame_number = static_cast<s32>((context.h264_parameter_set.flags >> 46) & 0x1ffff); + if (!is_first_frame && frame_number != 0) { + frame.resize(context.frame_data_size); + + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.data(), frame.size()); + } else { + /// Encode header + H264BitWriter writer{}; + writer.WriteU(1, 24); + writer.WriteU(0, 1); + writer.WriteU(3, 2); + writer.WriteU(7, 5); + writer.WriteU(100, 8); + writer.WriteU(0, 8); + writer.WriteU(31, 8); + writer.WriteUe(0); + const auto chroma_format_idc = + static_cast<u32>((context.h264_parameter_set.flags >> 12) & 3); + writer.WriteUe(chroma_format_idc); + if (chroma_format_idc == 3) { + writer.WriteBit(false); + } + + writer.WriteUe(0); + writer.WriteUe(0); + writer.WriteBit(false); // QpprimeYZeroTransformBypassFlag + writer.WriteBit(false); // Scaling matrix present flag + + const auto order_cnt_type = static_cast<u32>((context.h264_parameter_set.flags >> 14) & 3); + writer.WriteUe(static_cast<u32>((context.h264_parameter_set.flags >> 8) & 0xf)); + writer.WriteUe(order_cnt_type); + if (order_cnt_type == 0) { + writer.WriteUe(context.h264_parameter_set.log2_max_pic_order_cnt); + } else if (order_cnt_type == 1) { + writer.WriteBit(context.h264_parameter_set.delta_pic_order_always_zero_flag != 0); + + writer.WriteSe(0); + writer.WriteSe(0); + writer.WriteUe(0); + } + + const s32 pic_height = context.h264_parameter_set.pic_height_in_map_units / + (context.h264_parameter_set.frame_mbs_only_flag ? 1 : 2); + + writer.WriteUe(16); + writer.WriteBit(false); + writer.WriteUe(context.h264_parameter_set.pic_width_in_mbs - 1); + writer.WriteUe(pic_height - 1); + writer.WriteBit(context.h264_parameter_set.frame_mbs_only_flag != 0); + + if (!context.h264_parameter_set.frame_mbs_only_flag) { + writer.WriteBit(((context.h264_parameter_set.flags >> 0) & 1) != 0); + } + + writer.WriteBit(((context.h264_parameter_set.flags >> 1) & 1) != 0); + writer.WriteBit(false); // Frame cropping flag + writer.WriteBit(false); // VUI parameter present flag + + writer.End(); + + // H264 PPS + writer.WriteU(1, 24); + writer.WriteU(0, 1); + writer.WriteU(3, 2); + writer.WriteU(8, 5); + + writer.WriteUe(0); + writer.WriteUe(0); + + writer.WriteBit(context.h264_parameter_set.entropy_coding_mode_flag != 0); + writer.WriteBit(false); + writer.WriteUe(0); + writer.WriteUe(context.h264_parameter_set.num_refidx_l0_default_active); + writer.WriteUe(context.h264_parameter_set.num_refidx_l1_default_active); + writer.WriteBit(((context.h264_parameter_set.flags >> 2) & 1) != 0); + writer.WriteU(static_cast<s32>((context.h264_parameter_set.flags >> 32) & 0x3), 2); + s32 pic_init_qp = static_cast<s32>((context.h264_parameter_set.flags >> 16) & 0x3f); + pic_init_qp = (pic_init_qp << 26) >> 26; + writer.WriteSe(pic_init_qp); + writer.WriteSe(0); + s32 chroma_qp_index_offset = + static_cast<s32>((context.h264_parameter_set.flags >> 22) & 0x1f); + chroma_qp_index_offset = (chroma_qp_index_offset << 27) >> 27; + + writer.WriteSe(chroma_qp_index_offset); + writer.WriteBit(context.h264_parameter_set.deblocking_filter_control_flag != 0); + writer.WriteBit(((context.h264_parameter_set.flags >> 3) & 1) != 0); + writer.WriteBit(context.h264_parameter_set.redundant_pic_count_flag != 0); + writer.WriteBit(context.h264_parameter_set.transform_8x8_mode_flag != 0); + + writer.WriteBit(true); + + for (s32 index = 0; index < 6; index++) { + writer.WriteBit(true); + const auto matrix_x4 = + std::vector<u8>(context.scaling_matrix_4.begin(), context.scaling_matrix_4.end()); + writer.WriteScalingList(matrix_x4, index * 16, 16); + } + + if (context.h264_parameter_set.transform_8x8_mode_flag) { + for (s32 index = 0; index < 2; index++) { + writer.WriteBit(true); + const auto matrix_x8 = std::vector<u8>(context.scaling_matrix_8.begin(), + context.scaling_matrix_8.end()); + + writer.WriteScalingList(matrix_x8, index * 64, 64); + } + } + + s32 chroma_qp_index_offset2 = + static_cast<s32>((context.h264_parameter_set.flags >> 27) & 0x1f); + chroma_qp_index_offset2 = (chroma_qp_index_offset2 << 27) >> 27; + + writer.WriteSe(chroma_qp_index_offset2); + + writer.End(); + + const auto& encoded_header = writer.GetByteArray(); + frame.resize(encoded_header.size() + context.frame_data_size); + std::memcpy(frame.data(), encoded_header.data(), encoded_header.size()); + + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, + frame.data() + encoded_header.size(), + context.frame_data_size); + } + + return frame; +} + +H264BitWriter::H264BitWriter() = default; + +H264BitWriter::~H264BitWriter() = default; + +void H264BitWriter::WriteU(s32 value, s32 value_sz) { + WriteBits(value, value_sz); +} + +void H264BitWriter::WriteSe(s32 value) { + WriteExpGolombCodedInt(value); +} + +void H264BitWriter::WriteUe(u32 value) { + WriteExpGolombCodedUInt(value); +} + +void H264BitWriter::End() { + WriteBit(true); + Flush(); +} + +void H264BitWriter::WriteBit(bool state) { + WriteBits(state ? 1 : 0, 1); +} + +void H264BitWriter::WriteScalingList(const std::vector<u8>& list, s32 start, s32 count) { + std::vector<u8> scan(count); + if (count == 16) { + std::memcpy(scan.data(), zig_zag_scan.data(), scan.size()); + } else { + std::memcpy(scan.data(), zig_zag_direct.data(), scan.size()); + } + u8 last_scale = 8; + + for (s32 index = 0; index < count; index++) { + const u8 value = list[start + scan[index]]; + const s32 delta_scale = static_cast<s32>(value - last_scale); + + WriteSe(delta_scale); + + last_scale = value; + } +} + +std::vector<u8>& H264BitWriter::GetByteArray() { + return byte_array; +} + +const std::vector<u8>& H264BitWriter::GetByteArray() const { + return byte_array; +} + +void H264BitWriter::WriteBits(s32 value, s32 bit_count) { + s32 value_pos = 0; + + s32 remaining = bit_count; + + while (remaining > 0) { + s32 copy_size = remaining; + + const s32 free_bits = GetFreeBufferBits(); + + if (copy_size > free_bits) { + copy_size = free_bits; + } + + const s32 mask = (1 << copy_size) - 1; + + const s32 src_shift = (bit_count - value_pos) - copy_size; + const s32 dst_shift = (buffer_size - buffer_pos) - copy_size; + + buffer |= ((value >> src_shift) & mask) << dst_shift; + + value_pos += copy_size; + buffer_pos += copy_size; + remaining -= copy_size; + } +} + +void H264BitWriter::WriteExpGolombCodedInt(s32 value) { + const s32 sign = value <= 0 ? 0 : 1; + if (value < 0) { + value = -value; + } + value = (value << 1) - sign; + WriteExpGolombCodedUInt(value); +} + +void H264BitWriter::WriteExpGolombCodedUInt(u32 value) { + const s32 size = 32 - Common::CountLeadingZeroes32(static_cast<s32>(value + 1)); + WriteBits(1, size); + + value -= (1U << (size - 1)) - 1; + WriteBits(static_cast<s32>(value), size - 1); +} + +s32 H264BitWriter::GetFreeBufferBits() { + if (buffer_pos == buffer_size) { + Flush(); + } + + return buffer_size - buffer_pos; +} + +void H264BitWriter::Flush() { + if (buffer_pos == 0) { + return; + } + byte_array.push_back(static_cast<u8>(buffer)); + + buffer = 0; + buffer_pos = 0; +} +} // namespace Tegra::Decoder diff --git a/src/video_core/command_classes/codecs/h264.h b/src/video_core/command_classes/codecs/h264.h new file mode 100644 index 000000000..273449495 --- /dev/null +++ b/src/video_core/command_classes/codecs/h264.h @@ -0,0 +1,118 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#pragma once + +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "video_core/command_classes/nvdec_common.h" + +namespace Tegra { +class GPU; +namespace Decoder { + +class H264BitWriter { +public: + H264BitWriter(); + ~H264BitWriter(); + + /// The following Write methods are based on clause 9.1 in the H.264 specification. + /// WriteSe and WriteUe write in the Exp-Golomb-coded syntax + void WriteU(s32 value, s32 value_sz); + void WriteSe(s32 value); + void WriteUe(u32 value); + + /// Finalize the bitstream + void End(); + + /// append a bit to the stream, equivalent value to the state parameter + void WriteBit(bool state); + + /// Based on section 7.3.2.1.1.1 and Table 7-4 in the H.264 specification + /// Writes the scaling matrices of the sream + void WriteScalingList(const std::vector<u8>& list, s32 start, s32 count); + + /// Return the bitstream as a vector. + [[nodiscard]] std::vector<u8>& GetByteArray(); + [[nodiscard]] const std::vector<u8>& GetByteArray() const; + +private: + void WriteBits(s32 value, s32 bit_count); + void WriteExpGolombCodedInt(s32 value); + void WriteExpGolombCodedUInt(u32 value); + [[nodiscard]] s32 GetFreeBufferBits(); + void Flush(); + + s32 buffer_size{8}; + + s32 buffer{}; + s32 buffer_pos{}; + std::vector<u8> byte_array; +}; + +class H264 { +public: + explicit H264(GPU& gpu); + ~H264(); + + /// Compose the H264 header of the frame for FFmpeg decoding + [[nodiscard]] const std::vector<u8>& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, + bool is_first_frame = false); + +private: + struct H264ParameterSet { + u32 log2_max_pic_order_cnt{}; + u32 delta_pic_order_always_zero_flag{}; + u32 frame_mbs_only_flag{}; + u32 pic_width_in_mbs{}; + u32 pic_height_in_map_units{}; + INSERT_PADDING_WORDS(1); + u32 entropy_coding_mode_flag{}; + u32 bottom_field_pic_order_flag{}; + u32 num_refidx_l0_default_active{}; + u32 num_refidx_l1_default_active{}; + u32 deblocking_filter_control_flag{}; + u32 redundant_pic_count_flag{}; + u32 transform_8x8_mode_flag{}; + INSERT_PADDING_WORDS(9); + u64 flags{}; + u32 frame_number{}; + u32 frame_number2{}; + }; + static_assert(sizeof(H264ParameterSet) == 0x68, "H264ParameterSet is an invalid size"); + + struct H264DecoderContext { + INSERT_PADDING_BYTES(0x48); + u32 frame_data_size{}; + INSERT_PADDING_BYTES(0xc); + H264ParameterSet h264_parameter_set{}; + INSERT_PADDING_BYTES(0x100); + std::array<u8, 0x60> scaling_matrix_4; + std::array<u8, 0x80> scaling_matrix_8; + }; + static_assert(sizeof(H264DecoderContext) == 0x2a0, "H264DecoderContext is an invalid size"); + + std::vector<u8> frame; + GPU& gpu; +}; + +} // namespace Decoder +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp new file mode 100644 index 000000000..ab44fdc9e --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -0,0 +1,1040 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> // for std::memcpy +#include <numeric> +#include "video_core/command_classes/codecs/vp9.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra::Decoder { +namespace { +// Default compressed header probabilities once frame context resets +constexpr Vp9EntropyProbs default_probs{ + .y_mode_prob{ + 65, 32, 18, 144, 162, 194, 41, 51, 98, 132, 68, 18, 165, 217, 196, 45, 40, 78, + 173, 80, 19, 176, 240, 193, 64, 35, 46, 221, 135, 38, 194, 248, 121, 96, 85, 29, + }, + .partition_prob{ + 199, 122, 141, 0, 147, 63, 159, 0, 148, 133, 118, 0, 121, 104, 114, 0, + 174, 73, 87, 0, 92, 41, 83, 0, 82, 99, 50, 0, 53, 39, 39, 0, + 177, 58, 59, 0, 68, 26, 63, 0, 52, 79, 25, 0, 17, 14, 12, 0, + 222, 34, 30, 0, 72, 16, 44, 0, 58, 32, 12, 0, 10, 7, 6, 0, + }, + .coef_probs{ + 195, 29, 183, 0, 84, 49, 136, 0, 8, 42, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 31, 107, 169, 0, 35, 99, 159, 0, 17, 82, 140, 0, 8, 66, 114, 0, + 2, 44, 76, 0, 1, 19, 32, 0, 40, 132, 201, 0, 29, 114, 187, 0, 13, 91, 157, 0, + 7, 75, 127, 0, 3, 58, 95, 0, 1, 28, 47, 0, 69, 142, 221, 0, 42, 122, 201, 0, + 15, 91, 159, 0, 6, 67, 121, 0, 1, 42, 77, 0, 1, 17, 31, 0, 102, 148, 228, 0, + 67, 117, 204, 0, 17, 82, 154, 0, 6, 59, 114, 0, 2, 39, 75, 0, 1, 15, 29, 0, + 156, 57, 233, 0, 119, 57, 212, 0, 58, 48, 163, 0, 29, 40, 124, 0, 12, 30, 81, 0, + 3, 12, 31, 0, 191, 107, 226, 0, 124, 117, 204, 0, 25, 99, 155, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 29, 148, 210, 0, 37, 126, 194, 0, 8, 93, 157, 0, + 2, 68, 118, 0, 1, 39, 69, 0, 1, 17, 33, 0, 41, 151, 213, 0, 27, 123, 193, 0, + 3, 82, 144, 0, 1, 58, 105, 0, 1, 32, 60, 0, 1, 13, 26, 0, 59, 159, 220, 0, + 23, 126, 198, 0, 4, 88, 151, 0, 1, 66, 114, 0, 1, 38, 71, 0, 1, 18, 34, 0, + 114, 136, 232, 0, 51, 114, 207, 0, 11, 83, 155, 0, 3, 56, 105, 0, 1, 33, 65, 0, + 1, 17, 34, 0, 149, 65, 234, 0, 121, 57, 215, 0, 61, 49, 166, 0, 28, 36, 114, 0, + 12, 25, 76, 0, 3, 16, 42, 0, 214, 49, 220, 0, 132, 63, 188, 0, 42, 65, 137, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 137, 221, 0, 104, 131, 216, 0, + 49, 111, 192, 0, 21, 87, 155, 0, 2, 49, 87, 0, 1, 16, 28, 0, 89, 163, 230, 0, + 90, 137, 220, 0, 29, 100, 183, 0, 10, 70, 135, 0, 2, 42, 81, 0, 1, 17, 33, 0, + 108, 167, 237, 0, 55, 133, 222, 0, 15, 97, 179, 0, 4, 72, 135, 0, 1, 45, 85, 0, + 1, 19, 38, 0, 124, 146, 240, 0, 66, 124, 224, 0, 17, 88, 175, 0, 4, 58, 122, 0, + 1, 36, 75, 0, 1, 18, 37, 0, 141, 79, 241, 0, 126, 70, 227, 0, 66, 58, 182, 0, + 30, 44, 136, 0, 12, 34, 96, 0, 2, 20, 47, 0, 229, 99, 249, 0, 143, 111, 235, 0, + 46, 109, 192, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 82, 158, 236, 0, + 94, 146, 224, 0, 25, 117, 191, 0, 9, 87, 149, 0, 3, 56, 99, 0, 1, 33, 57, 0, + 83, 167, 237, 0, 68, 145, 222, 0, 10, 103, 177, 0, 2, 72, 131, 0, 1, 41, 79, 0, + 1, 20, 39, 0, 99, 167, 239, 0, 47, 141, 224, 0, 10, 104, 178, 0, 2, 73, 133, 0, + 1, 44, 85, 0, 1, 22, 47, 0, 127, 145, 243, 0, 71, 129, 228, 0, 17, 93, 177, 0, + 3, 61, 124, 0, 1, 41, 84, 0, 1, 21, 52, 0, 157, 78, 244, 0, 140, 72, 231, 0, + 69, 58, 184, 0, 31, 44, 137, 0, 14, 38, 105, 0, 8, 23, 61, 0, 125, 34, 187, 0, + 52, 41, 133, 0, 6, 31, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 37, 109, 153, 0, 51, 102, 147, 0, 23, 87, 128, 0, 8, 67, 101, 0, 1, 41, 63, 0, + 1, 19, 29, 0, 31, 154, 185, 0, 17, 127, 175, 0, 6, 96, 145, 0, 2, 73, 114, 0, + 1, 51, 82, 0, 1, 28, 45, 0, 23, 163, 200, 0, 10, 131, 185, 0, 2, 93, 148, 0, + 1, 67, 111, 0, 1, 41, 69, 0, 1, 14, 24, 0, 29, 176, 217, 0, 12, 145, 201, 0, + 3, 101, 156, 0, 1, 69, 111, 0, 1, 39, 63, 0, 1, 14, 23, 0, 57, 192, 233, 0, + 25, 154, 215, 0, 6, 109, 167, 0, 3, 78, 118, 0, 1, 48, 69, 0, 1, 21, 29, 0, + 202, 105, 245, 0, 108, 106, 216, 0, 18, 90, 144, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 33, 172, 219, 0, 64, 149, 206, 0, 14, 117, 177, 0, 5, 90, 141, 0, + 2, 61, 95, 0, 1, 37, 57, 0, 33, 179, 220, 0, 11, 140, 198, 0, 1, 89, 148, 0, + 1, 60, 104, 0, 1, 33, 57, 0, 1, 12, 21, 0, 30, 181, 221, 0, 8, 141, 198, 0, + 1, 87, 145, 0, 1, 58, 100, 0, 1, 31, 55, 0, 1, 12, 20, 0, 32, 186, 224, 0, + 7, 142, 198, 0, 1, 86, 143, 0, 1, 58, 100, 0, 1, 31, 55, 0, 1, 12, 22, 0, + 57, 192, 227, 0, 20, 143, 204, 0, 3, 96, 154, 0, 1, 68, 112, 0, 1, 42, 69, 0, + 1, 19, 32, 0, 212, 35, 215, 0, 113, 47, 169, 0, 29, 48, 105, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 74, 129, 203, 0, 106, 120, 203, 0, 49, 107, 178, 0, + 19, 84, 144, 0, 4, 50, 84, 0, 1, 15, 25, 0, 71, 172, 217, 0, 44, 141, 209, 0, + 15, 102, 173, 0, 6, 76, 133, 0, 2, 51, 89, 0, 1, 24, 42, 0, 64, 185, 231, 0, + 31, 148, 216, 0, 8, 103, 175, 0, 3, 74, 131, 0, 1, 46, 81, 0, 1, 18, 30, 0, + 65, 196, 235, 0, 25, 157, 221, 0, 5, 105, 174, 0, 1, 67, 120, 0, 1, 38, 69, 0, + 1, 15, 30, 0, 65, 204, 238, 0, 30, 156, 224, 0, 7, 107, 177, 0, 2, 70, 124, 0, + 1, 42, 73, 0, 1, 18, 34, 0, 225, 86, 251, 0, 144, 104, 235, 0, 42, 99, 181, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 175, 239, 0, 112, 165, 229, 0, + 29, 136, 200, 0, 12, 103, 162, 0, 6, 77, 123, 0, 2, 53, 84, 0, 75, 183, 239, 0, + 30, 155, 221, 0, 3, 106, 171, 0, 1, 74, 128, 0, 1, 44, 76, 0, 1, 17, 28, 0, + 73, 185, 240, 0, 27, 159, 222, 0, 2, 107, 172, 0, 1, 75, 127, 0, 1, 42, 73, 0, + 1, 17, 29, 0, 62, 190, 238, 0, 21, 159, 222, 0, 2, 107, 172, 0, 1, 72, 122, 0, + 1, 40, 71, 0, 1, 18, 32, 0, 61, 199, 240, 0, 27, 161, 226, 0, 4, 113, 180, 0, + 1, 76, 129, 0, 1, 46, 80, 0, 1, 23, 41, 0, 7, 27, 153, 0, 5, 30, 95, 0, + 1, 16, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 50, 75, 127, 0, + 57, 75, 124, 0, 27, 67, 108, 0, 10, 54, 86, 0, 1, 33, 52, 0, 1, 12, 18, 0, + 43, 125, 151, 0, 26, 108, 148, 0, 7, 83, 122, 0, 2, 59, 89, 0, 1, 38, 60, 0, + 1, 17, 27, 0, 23, 144, 163, 0, 13, 112, 154, 0, 2, 75, 117, 0, 1, 50, 81, 0, + 1, 31, 51, 0, 1, 14, 23, 0, 18, 162, 185, 0, 6, 123, 171, 0, 1, 78, 125, 0, + 1, 51, 86, 0, 1, 31, 54, 0, 1, 14, 23, 0, 15, 199, 227, 0, 3, 150, 204, 0, + 1, 91, 146, 0, 1, 55, 95, 0, 1, 30, 53, 0, 1, 11, 20, 0, 19, 55, 240, 0, + 19, 59, 196, 0, 3, 52, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 41, 166, 207, 0, 104, 153, 199, 0, 31, 123, 181, 0, 14, 101, 152, 0, 5, 72, 106, 0, + 1, 36, 52, 0, 35, 176, 211, 0, 12, 131, 190, 0, 2, 88, 144, 0, 1, 60, 101, 0, + 1, 36, 60, 0, 1, 16, 28, 0, 28, 183, 213, 0, 8, 134, 191, 0, 1, 86, 142, 0, + 1, 56, 96, 0, 1, 30, 53, 0, 1, 12, 20, 0, 20, 190, 215, 0, 4, 135, 192, 0, + 1, 84, 139, 0, 1, 53, 91, 0, 1, 28, 49, 0, 1, 11, 20, 0, 13, 196, 216, 0, + 2, 137, 192, 0, 1, 86, 143, 0, 1, 57, 99, 0, 1, 32, 56, 0, 1, 13, 24, 0, + 211, 29, 217, 0, 96, 47, 156, 0, 22, 43, 87, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 78, 120, 193, 0, 111, 116, 186, 0, 46, 102, 164, 0, 15, 80, 128, 0, + 2, 49, 76, 0, 1, 18, 28, 0, 71, 161, 203, 0, 42, 132, 192, 0, 10, 98, 150, 0, + 3, 69, 109, 0, 1, 44, 70, 0, 1, 18, 29, 0, 57, 186, 211, 0, 30, 140, 196, 0, + 4, 93, 146, 0, 1, 62, 102, 0, 1, 38, 65, 0, 1, 16, 27, 0, 47, 199, 217, 0, + 14, 145, 196, 0, 1, 88, 142, 0, 1, 57, 98, 0, 1, 36, 62, 0, 1, 15, 26, 0, + 26, 219, 229, 0, 5, 155, 207, 0, 1, 94, 151, 0, 1, 60, 104, 0, 1, 36, 62, 0, + 1, 16, 28, 0, 233, 29, 248, 0, 146, 47, 220, 0, 43, 52, 140, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 100, 163, 232, 0, 179, 161, 222, 0, 63, 142, 204, 0, + 37, 113, 174, 0, 26, 89, 137, 0, 18, 68, 97, 0, 85, 181, 230, 0, 32, 146, 209, 0, + 7, 100, 164, 0, 3, 71, 121, 0, 1, 45, 77, 0, 1, 18, 30, 0, 65, 187, 230, 0, + 20, 148, 207, 0, 2, 97, 159, 0, 1, 68, 116, 0, 1, 40, 70, 0, 1, 14, 29, 0, + 40, 194, 227, 0, 8, 147, 204, 0, 1, 94, 155, 0, 1, 65, 112, 0, 1, 39, 66, 0, + 1, 14, 26, 0, 16, 208, 228, 0, 3, 151, 207, 0, 1, 98, 160, 0, 1, 67, 117, 0, + 1, 41, 74, 0, 1, 17, 31, 0, 17, 38, 140, 0, 7, 34, 80, 0, 1, 17, 29, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 37, 75, 128, 0, 41, 76, 128, 0, + 26, 66, 116, 0, 12, 52, 94, 0, 2, 32, 55, 0, 1, 10, 16, 0, 50, 127, 154, 0, + 37, 109, 152, 0, 16, 82, 121, 0, 5, 59, 85, 0, 1, 35, 54, 0, 1, 13, 20, 0, + 40, 142, 167, 0, 17, 110, 157, 0, 2, 71, 112, 0, 1, 44, 72, 0, 1, 27, 45, 0, + 1, 11, 17, 0, 30, 175, 188, 0, 9, 124, 169, 0, 1, 74, 116, 0, 1, 48, 78, 0, + 1, 30, 49, 0, 1, 11, 18, 0, 10, 222, 223, 0, 2, 150, 194, 0, 1, 83, 128, 0, + 1, 48, 79, 0, 1, 27, 45, 0, 1, 11, 17, 0, 36, 41, 235, 0, 29, 36, 193, 0, + 10, 27, 111, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 165, 222, 0, + 177, 162, 215, 0, 110, 135, 195, 0, 57, 113, 168, 0, 23, 83, 120, 0, 10, 49, 61, 0, + 85, 190, 223, 0, 36, 139, 200, 0, 5, 90, 146, 0, 1, 60, 103, 0, 1, 38, 65, 0, + 1, 18, 30, 0, 72, 202, 223, 0, 23, 141, 199, 0, 2, 86, 140, 0, 1, 56, 97, 0, + 1, 36, 61, 0, 1, 16, 27, 0, 55, 218, 225, 0, 13, 145, 200, 0, 1, 86, 141, 0, + 1, 57, 99, 0, 1, 35, 61, 0, 1, 13, 22, 0, 15, 235, 212, 0, 1, 132, 184, 0, + 1, 84, 139, 0, 1, 57, 97, 0, 1, 34, 56, 0, 1, 14, 23, 0, 181, 21, 201, 0, + 61, 37, 123, 0, 10, 38, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 47, 106, 172, 0, 95, 104, 173, 0, 42, 93, 159, 0, 18, 77, 131, 0, 4, 50, 81, 0, + 1, 17, 23, 0, 62, 147, 199, 0, 44, 130, 189, 0, 28, 102, 154, 0, 18, 75, 115, 0, + 2, 44, 65, 0, 1, 12, 19, 0, 55, 153, 210, 0, 24, 130, 194, 0, 3, 93, 146, 0, + 1, 61, 97, 0, 1, 31, 50, 0, 1, 10, 16, 0, 49, 186, 223, 0, 17, 148, 204, 0, + 1, 96, 142, 0, 1, 53, 83, 0, 1, 26, 44, 0, 1, 11, 17, 0, 13, 217, 212, 0, + 2, 136, 180, 0, 1, 78, 124, 0, 1, 50, 83, 0, 1, 29, 49, 0, 1, 14, 23, 0, + 197, 13, 247, 0, 82, 17, 222, 0, 25, 17, 162, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 126, 186, 247, 0, 234, 191, 243, 0, 176, 177, 234, 0, 104, 158, 220, 0, + 66, 128, 186, 0, 55, 90, 137, 0, 111, 197, 242, 0, 46, 158, 219, 0, 9, 104, 171, 0, + 2, 65, 125, 0, 1, 44, 80, 0, 1, 17, 91, 0, 104, 208, 245, 0, 39, 168, 224, 0, + 3, 109, 162, 0, 1, 79, 124, 0, 1, 50, 102, 0, 1, 43, 102, 0, 84, 220, 246, 0, + 31, 177, 231, 0, 2, 115, 180, 0, 1, 79, 134, 0, 1, 55, 77, 0, 1, 60, 79, 0, + 43, 243, 240, 0, 8, 180, 217, 0, 1, 115, 166, 0, 1, 84, 121, 0, 1, 51, 67, 0, + 1, 16, 6, 0, + }, + .switchable_interp_prob{235, 162, 36, 255, 34, 3, 149, 144}, + .inter_mode_prob{ + 2, 173, 34, 0, 7, 145, 85, 0, 7, 166, 63, 0, 7, 94, + 66, 0, 8, 64, 46, 0, 17, 81, 31, 0, 25, 29, 30, 0, + }, + .intra_inter_prob{9, 102, 187, 225}, + .comp_inter_prob{9, 102, 187, 225, 0}, + .single_ref_prob{33, 16, 77, 74, 142, 142, 172, 170, 238, 247}, + .comp_ref_prob{50, 126, 123, 221, 226}, + .tx_32x32_prob{3, 136, 37, 5, 52, 13}, + .tx_16x16_prob{20, 152, 15, 101}, + .tx_8x8_prob{100, 66}, + .skip_probs{192, 128, 64}, + .joints{32, 64, 96}, + .sign{128, 128}, + .classes{ + 224, 144, 192, 168, 192, 176, 192, 198, 198, 245, + 216, 128, 176, 160, 176, 176, 192, 198, 198, 208, + }, + .class_0{216, 208}, + .prob_bits{ + 136, 140, 148, 160, 176, 192, 224, 234, 234, 240, + 136, 140, 148, 160, 176, 192, 224, 234, 234, 240, + }, + .class_0_fr{128, 128, 64, 96, 112, 64, 128, 128, 64, 96, 112, 64}, + .fr{64, 96, 64, 64, 96, 64}, + .class_0_hp{160, 160}, + .high_precision{128, 128}, +}; + +constexpr std::array<s32, 256> norm_lut{ + 0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, +}; + +constexpr std::array<s32, 254> map_lut{ + 20, 21, 22, 23, 24, 25, 0, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, + 1, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 2, 50, 51, 52, 53, 54, + 55, 56, 57, 58, 59, 60, 61, 3, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, + 73, 4, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 5, 86, 87, 88, 89, + 90, 91, 92, 93, 94, 95, 96, 97, 6, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, + 108, 109, 7, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 8, 122, 123, 124, + 125, 126, 127, 128, 129, 130, 131, 132, 133, 9, 134, 135, 136, 137, 138, 139, 140, 141, 142, + 143, 144, 145, 10, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 11, 158, 159, + 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 12, 170, 171, 172, 173, 174, 175, 176, 177, + 178, 179, 180, 181, 13, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 14, 194, + 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 15, 206, 207, 208, 209, 210, 211, 212, + 213, 214, 215, 216, 217, 16, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 17, + 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 18, 242, 243, 244, 245, 246, 247, + 248, 249, 250, 251, 252, 253, 19, +}; + +// 6.2.14 Tile size calculation + +[[nodiscard]] s32 CalcMinLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 min_log2 = 0; + + while ((64 << min_log2) < sb64_cols) { + min_log2++; + } + + return min_log2; +} + +[[nodiscard]] s32 CalcMaxLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 max_log2 = 1; + + while ((sb64_cols >> max_log2) >= 4) { + max_log2++; + } + + return max_log2 - 1; +} + +// Recenters probability. Based on section 6.3.6 of VP9 Specification +[[nodiscard]] s32 RecenterNonNeg(s32 new_prob, s32 old_prob) { + if (new_prob > old_prob * 2) { + return new_prob; + } + + if (new_prob >= old_prob) { + return (new_prob - old_prob) * 2; + } + + return (old_prob - new_prob) * 2 - 1; +} + +// Adjusts old_prob depending on new_prob. Based on section 6.3.5 of VP9 Specification +[[nodiscard]] s32 RemapProbability(s32 new_prob, s32 old_prob) { + new_prob--; + old_prob--; + + std::size_t index{}; + + if (old_prob * 2 <= 0xff) { + index = static_cast<std::size_t>(std::max(0, RecenterNonNeg(new_prob, old_prob) - 1)); + } else { + index = static_cast<std::size_t>( + std::max(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1)); + } + + return map_lut[index]; +} +} // Anonymous namespace + +VP9::VP9(GPU& gpu) : gpu(gpu) {} + +VP9::~VP9() = default; + +void VP9::WriteProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const bool update = new_prob != old_prob; + + writer.Write(update, diff_update_probability); + + if (update) { + WriteProbabilityDelta(writer, new_prob, old_prob); + } +} +template <typename T, std::size_t N> +void VP9::WriteProbabilityUpdate(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob) { + for (std::size_t offset = 0; offset < new_prob.size(); ++offset) { + WriteProbabilityUpdate(writer, new_prob[offset], old_prob[offset]); + } +} + +template <typename T, std::size_t N> +void VP9::WriteProbabilityUpdateAligned4(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob) { + for (std::size_t offset = 0; offset < new_prob.size(); offset += 4) { + WriteProbabilityUpdate(writer, new_prob[offset + 0], old_prob[offset + 0]); + WriteProbabilityUpdate(writer, new_prob[offset + 1], old_prob[offset + 1]); + WriteProbabilityUpdate(writer, new_prob[offset + 2], old_prob[offset + 2]); + } +} + +void VP9::WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const int delta = RemapProbability(new_prob, old_prob); + + EncodeTermSubExp(writer, delta); +} + +void VP9::EncodeTermSubExp(VpxRangeEncoder& writer, s32 value) { + if (WriteLessThan(writer, value, 16)) { + writer.Write(value, 4); + } else if (WriteLessThan(writer, value, 32)) { + writer.Write(value - 16, 4); + } else if (WriteLessThan(writer, value, 64)) { + writer.Write(value - 32, 5); + } else { + value -= 64; + + constexpr s32 size = 8; + + const s32 mask = (1 << size) - 191; + + const s32 delta = value - mask; + + if (delta < 0) { + writer.Write(value, size - 1); + } else { + writer.Write(delta / 2 + mask, size - 1); + writer.Write(delta & 1, 1); + } + } +} + +bool VP9::WriteLessThan(VpxRangeEncoder& writer, s32 value, s32 test) { + const bool is_lt = value < test; + writer.Write(!is_lt); + return is_lt; +} + +void VP9::WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, + const std::array<u8, 2304>& new_prob, + const std::array<u8, 2304>& old_prob) { + // Note: There's 1 byte added on each packet for alignment, + // this byte is ignored when doing updates. + constexpr s32 block_bytes = 2 * 2 * 6 * 6 * 4; + + const auto needs_update = [&](s32 base_index) -> bool { + s32 index = base_index; + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 6; k++) { + for (s32 l = 0; l < 6; l++) { + if (new_prob[index + 0] != old_prob[index + 0] || + new_prob[index + 1] != old_prob[index + 1] || + new_prob[index + 2] != old_prob[index + 2]) { + return true; + } + + index += 4; + } + } + } + } + return false; + }; + + for (s32 block_index = 0; block_index < 4; block_index++) { + const s32 base_index = block_index * block_bytes; + const bool update = needs_update(base_index); + writer.Write(update); + + if (update) { + s32 index = base_index; + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 6; k++) { + for (s32 l = 0; l < 6; l++) { + if (k != 0 || l < 3) { + WriteProbabilityUpdate(writer, new_prob[index + 0], + old_prob[index + 0]); + WriteProbabilityUpdate(writer, new_prob[index + 1], + old_prob[index + 1]); + WriteProbabilityUpdate(writer, new_prob[index + 2], + old_prob[index + 2]); + } + index += 4; + } + } + } + } + } + + if (block_index == tx_mode) { + break; + } + } +} + +void VP9::WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const bool update = new_prob != old_prob; + writer.Write(update, diff_update_probability); + + if (update) { + writer.Write(new_prob >> 1, 7); + } +} + +Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) { + PictureInfo picture_info{}; + gpu.MemoryManager().ReadBlock(state.picture_info_offset, &picture_info, sizeof(PictureInfo)); + Vp9PictureInfo vp9_info = picture_info.Convert(); + + InsertEntropy(state.vp9_entropy_probs_offset, vp9_info.entropy); + + // surface_luma_offset[0:3] contains the address of the reference frame offsets in the following + // order: last, golden, altref, current. It may be worthwhile to track the updates done here + // to avoid buffering frame data needed for reference frame updating in the header composition. + std::memcpy(vp9_info.frame_offsets.data(), state.surface_luma_offset.data(), 4 * sizeof(u64)); + + return vp9_info; +} + +void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) { + EntropyProbs entropy{}; + gpu.MemoryManager().ReadBlock(offset, &entropy, sizeof(EntropyProbs)); + entropy.Convert(dst); +} + +Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) { + Vp9FrameContainer frame{}; + { + gpu.SyncGuestHost(); + frame.info = GetVp9PictureInfo(state); + + frame.bit_stream.resize(frame.info.bitstream_size); + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.bit_stream.data(), + frame.info.bitstream_size); + } + // Buffer two frames, saving the last show frame info + if (!next_next_frame.bit_stream.empty()) { + Vp9FrameContainer temp{ + .info = frame.info, + .bit_stream = frame.bit_stream, + }; + next_next_frame.info.show_frame = frame.info.last_frame_shown; + frame.info = next_next_frame.info; + frame.bit_stream = next_next_frame.bit_stream; + next_next_frame = std::move(temp); + + if (!next_frame.bit_stream.empty()) { + Vp9FrameContainer temp2{ + .info = frame.info, + .bit_stream = frame.bit_stream, + }; + next_frame.info.show_frame = frame.info.last_frame_shown; + frame.info = next_frame.info; + frame.bit_stream = next_frame.bit_stream; + next_frame = std::move(temp2); + } else { + next_frame.info = frame.info; + next_frame.bit_stream = frame.bit_stream; + } + } else { + next_next_frame.info = frame.info; + next_next_frame.bit_stream = frame.bit_stream; + } + return frame; +} + +std::vector<u8> VP9::ComposeCompressedHeader() { + VpxRangeEncoder writer{}; + + if (!current_frame_info.lossless) { + if (static_cast<u32>(current_frame_info.transform_mode) >= 3) { + writer.Write(3, 2); + writer.Write(current_frame_info.transform_mode == 4); + } else { + writer.Write(current_frame_info.transform_mode, 2); + } + } + + if (current_frame_info.transform_mode == 4) { + // tx_mode_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_8x8_prob, + prev_frame_probs.tx_8x8_prob); + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_16x16_prob, + prev_frame_probs.tx_16x16_prob); + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_32x32_prob, + prev_frame_probs.tx_32x32_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.tx_8x8_prob = current_frame_info.entropy.tx_8x8_prob; + prev_frame_probs.tx_16x16_prob = current_frame_info.entropy.tx_16x16_prob; + prev_frame_probs.tx_32x32_prob = current_frame_info.entropy.tx_32x32_prob; + } + } + // read_coef_probs() in the spec + WriteCoefProbabilityUpdate(writer, current_frame_info.transform_mode, + current_frame_info.entropy.coef_probs, prev_frame_probs.coef_probs); + // read_skip_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.skip_probs, + prev_frame_probs.skip_probs); + + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.coef_probs = current_frame_info.entropy.coef_probs; + prev_frame_probs.skip_probs = current_frame_info.entropy.skip_probs; + } + + if (!current_frame_info.intra_only) { + // read_inter_probs() in the spec + WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.inter_mode_prob, + prev_frame_probs.inter_mode_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.inter_mode_prob = current_frame_info.entropy.inter_mode_prob; + } + + if (current_frame_info.interp_filter == 4) { + // read_interp_filter_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.switchable_interp_prob, + prev_frame_probs.switchable_interp_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.switchable_interp_prob = + current_frame_info.entropy.switchable_interp_prob; + } + } + + // read_is_inter_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.intra_inter_prob, + prev_frame_probs.intra_inter_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.intra_inter_prob = current_frame_info.entropy.intra_inter_prob; + } + // frame_reference_mode() in the spec + if ((current_frame_info.ref_frame_sign_bias[1] & 1) != + (current_frame_info.ref_frame_sign_bias[2] & 1) || + (current_frame_info.ref_frame_sign_bias[1] & 1) != + (current_frame_info.ref_frame_sign_bias[3] & 1)) { + if (current_frame_info.reference_mode >= 1) { + writer.Write(1, 1); + writer.Write(current_frame_info.reference_mode == 2); + } else { + writer.Write(0, 1); + } + } + + // frame_reference_mode_probs() in the spec + if (current_frame_info.reference_mode == 2) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_inter_prob, + prev_frame_probs.comp_inter_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.comp_inter_prob = current_frame_info.entropy.comp_inter_prob; + } + } + + if (current_frame_info.reference_mode != 1) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.single_ref_prob, + prev_frame_probs.single_ref_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.single_ref_prob = current_frame_info.entropy.single_ref_prob; + } + } + + if (current_frame_info.reference_mode != 0) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_ref_prob, + prev_frame_probs.comp_ref_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.comp_ref_prob = current_frame_info.entropy.comp_ref_prob; + } + } + + // read_y_mode_probs + for (std::size_t index = 0; index < current_frame_info.entropy.y_mode_prob.size(); + ++index) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.y_mode_prob[index], + prev_frame_probs.y_mode_prob[index]); + } + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.y_mode_prob = current_frame_info.entropy.y_mode_prob; + } + // read_partition_probs + WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.partition_prob, + prev_frame_probs.partition_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.partition_prob = current_frame_info.entropy.partition_prob; + } + + // mv_probs + for (s32 i = 0; i < 3; i++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.joints[i], + prev_frame_probs.joints[i]); + } + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.joints = current_frame_info.entropy.joints; + } + + for (s32 i = 0; i < 2; i++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.sign[i], + prev_frame_probs.sign[i]); + + for (s32 j = 0; j < 10; j++) { + const int index = i * 10 + j; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.classes[index], + prev_frame_probs.classes[index]); + } + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0[i], + prev_frame_probs.class_0[i]); + + for (s32 j = 0; j < 10; j++) { + const int index = i * 10 + j; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.prob_bits[index], + prev_frame_probs.prob_bits[index]); + } + } + + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 3; k++) { + const int index = i * 2 * 3 + j * 3 + k; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_fr[index], + prev_frame_probs.class_0_fr[index]); + } + } + + for (s32 j = 0; j < 3; j++) { + const int index = i * 3 + j; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.fr[index], + prev_frame_probs.fr[index]); + } + } + + if (current_frame_info.allow_high_precision_mv) { + for (s32 index = 0; index < 2; index++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_hp[index], + prev_frame_probs.class_0_hp[index]); + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.high_precision[index], + prev_frame_probs.high_precision[index]); + } + } + + // save previous probs + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.sign = current_frame_info.entropy.sign; + prev_frame_probs.classes = current_frame_info.entropy.classes; + prev_frame_probs.class_0 = current_frame_info.entropy.class_0; + prev_frame_probs.prob_bits = current_frame_info.entropy.prob_bits; + prev_frame_probs.class_0_fr = current_frame_info.entropy.class_0_fr; + prev_frame_probs.fr = current_frame_info.entropy.fr; + prev_frame_probs.class_0_hp = current_frame_info.entropy.class_0_hp; + prev_frame_probs.high_precision = current_frame_info.entropy.high_precision; + } + } + + writer.End(); + return writer.GetBuffer(); +} + +VpxBitStreamWriter VP9::ComposeUncompressedHeader() { + VpxBitStreamWriter uncomp_writer{}; + + uncomp_writer.WriteU(2, 2); // Frame marker. + uncomp_writer.WriteU(0, 2); // Profile. + uncomp_writer.WriteBit(false); // Show existing frame. + uncomp_writer.WriteBit(!current_frame_info.is_key_frame); // is key frame? + uncomp_writer.WriteBit(current_frame_info.show_frame); // show frame? + uncomp_writer.WriteBit(current_frame_info.error_resilient_mode); // error reslience + + if (current_frame_info.is_key_frame) { + uncomp_writer.WriteU(frame_sync_code, 24); + uncomp_writer.WriteU(0, 3); // Color space. + uncomp_writer.WriteU(0, 1); // Color range. + uncomp_writer.WriteU(current_frame_info.frame_size.width - 1, 16); + uncomp_writer.WriteU(current_frame_info.frame_size.height - 1, 16); + uncomp_writer.WriteBit(false); // Render and frame size different. + + // Reset context + prev_frame_probs = default_probs; + swap_next_golden = false; + loop_filter_ref_deltas.fill(0); + loop_filter_mode_deltas.fill(0); + + // allow frames offsets to stabilize before checking for golden frames + grace_period = 4; + + // On key frames, all frame slots are set to the current frame, + // so the value of the selected slot doesn't really matter. + frame_ctxs.fill({current_frame_number, false, default_probs}); + + // intra only, meaning the frame can be recreated with no other references + current_frame_info.intra_only = true; + + } else { + + if (!current_frame_info.show_frame) { + uncomp_writer.WriteBit(current_frame_info.intra_only); + if (!current_frame_info.last_frame_was_key) { + swap_next_golden = !swap_next_golden; + } + } else { + current_frame_info.intra_only = false; + } + if (!current_frame_info.error_resilient_mode) { + uncomp_writer.WriteU(0, 2); // Reset frame context. + } + + // Last, Golden, Altref frames + std::array<s32, 3> ref_frame_index{0, 1, 2}; + + // Set when next frame is hidden + // altref and golden references are swapped + if (swap_next_golden) { + ref_frame_index = std::array<s32, 3>{0, 2, 1}; + } + + // update Last Frame + u64 refresh_frame_flags = 1; + + // golden frame may refresh, determined if the next golden frame offset is changed + bool golden_refresh = false; + if (grace_period <= 0) { + for (s32 index = 1; index < 3; ++index) { + if (current_frame_info.frame_offsets[index] != + next_frame.info.frame_offsets[index]) { + current_frame_info.refresh_frame[index] = true; + golden_refresh = true; + grace_period = 3; + } + } + } + + if (current_frame_info.show_frame && + (!next_frame.info.show_frame || next_frame.info.is_key_frame)) { + // Update golden frame + refresh_frame_flags = swap_next_golden ? 2 : 4; + } + + if (!current_frame_info.show_frame) { + // Update altref + refresh_frame_flags = swap_next_golden ? 2 : 4; + } else if (golden_refresh) { + refresh_frame_flags = 3; + } + + if (current_frame_info.intra_only) { + uncomp_writer.WriteU(frame_sync_code, 24); + uncomp_writer.WriteU(static_cast<s32>(refresh_frame_flags), 8); + uncomp_writer.WriteU(current_frame_info.frame_size.width - 1, 16); + uncomp_writer.WriteU(current_frame_info.frame_size.height - 1, 16); + uncomp_writer.WriteBit(false); // Render and frame size different. + } else { + uncomp_writer.WriteU(static_cast<s32>(refresh_frame_flags), 8); + + for (s32 index = 1; index < 4; index++) { + uncomp_writer.WriteU(ref_frame_index[index - 1], 3); + uncomp_writer.WriteU(current_frame_info.ref_frame_sign_bias[index], 1); + } + + uncomp_writer.WriteBit(true); // Frame size with refs. + uncomp_writer.WriteBit(false); // Render and frame size different. + uncomp_writer.WriteBit(current_frame_info.allow_high_precision_mv); + uncomp_writer.WriteBit(current_frame_info.interp_filter == 4); + + if (current_frame_info.interp_filter != 4) { + uncomp_writer.WriteU(current_frame_info.interp_filter, 2); + } + } + } + + if (!current_frame_info.error_resilient_mode) { + uncomp_writer.WriteBit(true); // Refresh frame context. where do i get this info from? + uncomp_writer.WriteBit(true); // Frame parallel decoding mode. + } + + int frame_ctx_idx = 0; + if (!current_frame_info.show_frame) { + frame_ctx_idx = 1; + } + + uncomp_writer.WriteU(frame_ctx_idx, 2); // Frame context index. + prev_frame_probs = + frame_ctxs[frame_ctx_idx].probs; // reference probabilities for compressed header + frame_ctxs[frame_ctx_idx] = {current_frame_number, false, current_frame_info.entropy}; + + uncomp_writer.WriteU(current_frame_info.first_level, 6); + uncomp_writer.WriteU(current_frame_info.sharpness_level, 3); + uncomp_writer.WriteBit(current_frame_info.mode_ref_delta_enabled); + + if (current_frame_info.mode_ref_delta_enabled) { + // check if ref deltas are different, update accordingly + std::array<bool, 4> update_loop_filter_ref_deltas; + std::array<bool, 2> update_loop_filter_mode_deltas; + + bool loop_filter_delta_update = false; + + for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { + const s8 old_deltas = loop_filter_ref_deltas[index]; + const s8 new_deltas = current_frame_info.ref_deltas[index]; + const bool differing_delta = old_deltas != new_deltas; + + update_loop_filter_ref_deltas[index] = differing_delta; + loop_filter_delta_update |= differing_delta; + } + + for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { + const s8 old_deltas = loop_filter_mode_deltas[index]; + const s8 new_deltas = current_frame_info.mode_deltas[index]; + const bool differing_delta = old_deltas != new_deltas; + + update_loop_filter_mode_deltas[index] = differing_delta; + loop_filter_delta_update |= differing_delta; + } + + uncomp_writer.WriteBit(loop_filter_delta_update); + + if (loop_filter_delta_update) { + for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { + uncomp_writer.WriteBit(update_loop_filter_ref_deltas[index]); + + if (update_loop_filter_ref_deltas[index]) { + uncomp_writer.WriteS(current_frame_info.ref_deltas[index], 6); + } + } + + for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { + uncomp_writer.WriteBit(update_loop_filter_mode_deltas[index]); + + if (update_loop_filter_mode_deltas[index]) { + uncomp_writer.WriteS(current_frame_info.mode_deltas[index], 6); + } + } + // save new deltas + loop_filter_ref_deltas = current_frame_info.ref_deltas; + loop_filter_mode_deltas = current_frame_info.mode_deltas; + } + } + + uncomp_writer.WriteU(current_frame_info.base_q_index, 8); + + uncomp_writer.WriteDeltaQ(current_frame_info.y_dc_delta_q); + uncomp_writer.WriteDeltaQ(current_frame_info.uv_dc_delta_q); + uncomp_writer.WriteDeltaQ(current_frame_info.uv_ac_delta_q); + + uncomp_writer.WriteBit(false); // Segmentation enabled (TODO). + + const s32 min_tile_cols_log2 = CalcMinLog2TileCols(current_frame_info.frame_size.width); + const s32 max_tile_cols_log2 = CalcMaxLog2TileCols(current_frame_info.frame_size.width); + + const s32 tile_cols_log2_diff = current_frame_info.log2_tile_cols - min_tile_cols_log2; + const s32 tile_cols_log2_inc_mask = (1 << tile_cols_log2_diff) - 1; + + // If it's less than the maximum, we need to add an extra 0 on the bitstream + // to indicate that it should stop reading. + if (current_frame_info.log2_tile_cols < max_tile_cols_log2) { + uncomp_writer.WriteU(tile_cols_log2_inc_mask << 1, tile_cols_log2_diff + 1); + } else { + uncomp_writer.WriteU(tile_cols_log2_inc_mask, tile_cols_log2_diff); + } + + const bool tile_rows_log2_is_nonzero = current_frame_info.log2_tile_rows != 0; + + uncomp_writer.WriteBit(tile_rows_log2_is_nonzero); + + if (tile_rows_log2_is_nonzero) { + uncomp_writer.WriteBit(current_frame_info.log2_tile_rows > 1); + } + + return uncomp_writer; +} + +const std::vector<u8>& VP9::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state) { + std::vector<u8> bitstream; + { + Vp9FrameContainer curr_frame = GetCurrentFrame(state); + current_frame_info = curr_frame.info; + bitstream = std::move(curr_frame.bit_stream); + } + + // The uncompressed header routine sets PrevProb parameters needed for the compressed header + auto uncomp_writer = ComposeUncompressedHeader(); + std::vector<u8> compressed_header = ComposeCompressedHeader(); + + uncomp_writer.WriteU(static_cast<s32>(compressed_header.size()), 16); + uncomp_writer.Flush(); + std::vector<u8> uncompressed_header = uncomp_writer.GetByteArray(); + + // Write headers and frame to buffer + frame.resize(uncompressed_header.size() + compressed_header.size() + bitstream.size()); + std::memcpy(frame.data(), uncompressed_header.data(), uncompressed_header.size()); + std::memcpy(frame.data() + uncompressed_header.size(), compressed_header.data(), + compressed_header.size()); + std::memcpy(frame.data() + uncompressed_header.size() + compressed_header.size(), + bitstream.data(), bitstream.size()); + + // keep track of frame number + current_frame_number++; + grace_period--; + + // don't display hidden frames + hidden = !current_frame_info.show_frame; + return frame; +} + +VpxRangeEncoder::VpxRangeEncoder() { + Write(false); +} + +VpxRangeEncoder::~VpxRangeEncoder() = default; + +void VpxRangeEncoder::Write(s32 value, s32 value_size) { + for (s32 bit = value_size - 1; bit >= 0; bit--) { + Write(((value >> bit) & 1) != 0); + } +} + +void VpxRangeEncoder::Write(bool bit) { + Write(bit, half_probability); +} + +void VpxRangeEncoder::Write(bool bit, s32 probability) { + u32 local_range = range; + const u32 split = 1 + (((local_range - 1) * static_cast<u32>(probability)) >> 8); + local_range = split; + + if (bit) { + low_value += split; + local_range = range - split; + } + + s32 shift = norm_lut[local_range]; + local_range <<= shift; + count += shift; + + if (count >= 0) { + const s32 offset = shift - count; + + if (((low_value << (offset - 1)) >> 31) != 0) { + const s32 current_pos = static_cast<s32>(base_stream.GetPosition()); + base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); + while (PeekByte() == 0xff) { + base_stream.WriteByte(0); + + base_stream.Seek(-2, Common::SeekOrigin::FromCurrentPos); + } + base_stream.WriteByte(static_cast<u8>((PeekByte() + 1))); + base_stream.Seek(current_pos, Common::SeekOrigin::SetOrigin); + } + base_stream.WriteByte(static_cast<u8>((low_value >> (24 - offset)))); + + low_value <<= offset; + shift = count; + low_value &= 0xffffff; + count -= 8; + } + + low_value <<= shift; + range = local_range; +} + +void VpxRangeEncoder::End() { + for (std::size_t index = 0; index < 32; ++index) { + Write(false); + } +} + +u8 VpxRangeEncoder::PeekByte() { + const u8 value = base_stream.ReadByte(); + base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); + + return value; +} + +VpxBitStreamWriter::VpxBitStreamWriter() = default; + +VpxBitStreamWriter::~VpxBitStreamWriter() = default; + +void VpxBitStreamWriter::WriteU(u32 value, u32 value_size) { + WriteBits(value, value_size); +} + +void VpxBitStreamWriter::WriteS(s32 value, u32 value_size) { + const bool sign = value < 0; + if (sign) { + value = -value; + } + + WriteBits(static_cast<u32>(value << 1) | (sign ? 1 : 0), value_size + 1); +} + +void VpxBitStreamWriter::WriteDeltaQ(u32 value) { + const bool delta_coded = value != 0; + WriteBit(delta_coded); + + if (delta_coded) { + WriteBits(value, 4); + } +} + +void VpxBitStreamWriter::WriteBits(u32 value, u32 bit_count) { + s32 value_pos = 0; + s32 remaining = bit_count; + + while (remaining > 0) { + s32 copy_size = remaining; + + const s32 free = GetFreeBufferBits(); + + if (copy_size > free) { + copy_size = free; + } + + const s32 mask = (1 << copy_size) - 1; + + const s32 src_shift = (bit_count - value_pos) - copy_size; + const s32 dst_shift = (buffer_size - buffer_pos) - copy_size; + + buffer |= ((value >> src_shift) & mask) << dst_shift; + + value_pos += copy_size; + buffer_pos += copy_size; + remaining -= copy_size; + } +} + +void VpxBitStreamWriter::WriteBit(bool state) { + WriteBits(state ? 1 : 0, 1); +} + +s32 VpxBitStreamWriter::GetFreeBufferBits() { + if (buffer_pos == buffer_size) { + Flush(); + } + + return buffer_size - buffer_pos; +} + +void VpxBitStreamWriter::Flush() { + if (buffer_pos == 0) { + return; + } + byte_array.push_back(static_cast<u8>(buffer)); + buffer = 0; + buffer_pos = 0; +} + +std::vector<u8>& VpxBitStreamWriter::GetByteArray() { + return byte_array; +} + +const std::vector<u8>& VpxBitStreamWriter::GetByteArray() const { + return byte_array; +} + +} // namespace Tegra::Decoder diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h new file mode 100644 index 000000000..e2504512c --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9.h @@ -0,0 +1,196 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <vector> + +#include "common/common_types.h" +#include "common/stream.h" +#include "video_core/command_classes/codecs/vp9_types.h" +#include "video_core/command_classes/nvdec_common.h" + +namespace Tegra { +class GPU; +enum class FrameType { KeyFrame = 0, InterFrame = 1 }; +namespace Decoder { + +/// The VpxRangeEncoder, and VpxBitStreamWriter classes are used to compose the +/// VP9 header bitstreams. + +class VpxRangeEncoder { +public: + VpxRangeEncoder(); + ~VpxRangeEncoder(); + + VpxRangeEncoder(const VpxRangeEncoder&) = delete; + VpxRangeEncoder& operator=(const VpxRangeEncoder&) = delete; + + VpxRangeEncoder(VpxRangeEncoder&&) = default; + VpxRangeEncoder& operator=(VpxRangeEncoder&&) = default; + + /// Writes the rightmost value_size bits from value into the stream + void Write(s32 value, s32 value_size); + + /// Writes a single bit with half probability + void Write(bool bit); + + /// Writes a bit to the base_stream encoded with probability + void Write(bool bit, s32 probability); + + /// Signal the end of the bitstream + void End(); + + [[nodiscard]] std::vector<u8>& GetBuffer() { + return base_stream.GetBuffer(); + } + + [[nodiscard]] const std::vector<u8>& GetBuffer() const { + return base_stream.GetBuffer(); + } + +private: + u8 PeekByte(); + Common::Stream base_stream{}; + u32 low_value{}; + u32 range{0xff}; + s32 count{-24}; + s32 half_probability{128}; +}; + +class VpxBitStreamWriter { +public: + VpxBitStreamWriter(); + ~VpxBitStreamWriter(); + + VpxBitStreamWriter(const VpxBitStreamWriter&) = delete; + VpxBitStreamWriter& operator=(const VpxBitStreamWriter&) = delete; + + VpxBitStreamWriter(VpxBitStreamWriter&&) = default; + VpxBitStreamWriter& operator=(VpxBitStreamWriter&&) = default; + + /// Write an unsigned integer value + void WriteU(u32 value, u32 value_size); + + /// Write a signed integer value + void WriteS(s32 value, u32 value_size); + + /// Based on 6.2.10 of VP9 Spec, writes a delta coded value + void WriteDeltaQ(u32 value); + + /// Write a single bit. + void WriteBit(bool state); + + /// Pushes current buffer into buffer_array, resets buffer + void Flush(); + + /// Returns byte_array + [[nodiscard]] std::vector<u8>& GetByteArray(); + + /// Returns const byte_array + [[nodiscard]] const std::vector<u8>& GetByteArray() const; + +private: + /// Write bit_count bits from value into buffer + void WriteBits(u32 value, u32 bit_count); + + /// Gets next available position in buffer, invokes Flush() if buffer is full + s32 GetFreeBufferBits(); + + s32 buffer_size{8}; + + s32 buffer{}; + s32 buffer_pos{}; + std::vector<u8> byte_array; +}; + +class VP9 { +public: + explicit VP9(GPU& gpu); + ~VP9(); + + VP9(const VP9&) = delete; + VP9& operator=(const VP9&) = delete; + + VP9(VP9&&) = default; + VP9& operator=(VP9&&) = delete; + + /// Composes the VP9 frame from the GPU state information. Based on the official VP9 spec + /// documentation + [[nodiscard]] const std::vector<u8>& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state); + + /// Returns true if the most recent frame was a hidden frame. + [[nodiscard]] bool WasFrameHidden() const { + return hidden; + } + +private: + /// Generates compressed header probability updates in the bitstream writer + template <typename T, std::size_t N> + void WriteProbabilityUpdate(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob); + + /// Generates compressed header probability updates in the bitstream writer + /// If probs are not equal, WriteProbabilityDelta is invoked + void WriteProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); + + /// Generates compressed header probability deltas in the bitstream writer + void WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); + + /// Inverse of 6.3.4 Decode term subexp + void EncodeTermSubExp(VpxRangeEncoder& writer, s32 value); + + /// Writes if the value is less than the test value + bool WriteLessThan(VpxRangeEncoder& writer, s32 value, s32 test); + + /// Writes probability updates for the Coef probabilities + void WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, + const std::array<u8, 2304>& new_prob, + const std::array<u8, 2304>& old_prob); + + /// Write probabilities for 4-byte aligned structures + template <typename T, std::size_t N> + void WriteProbabilityUpdateAligned4(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob); + + /// Write motion vector probability updates. 6.3.17 in the spec + void WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); + + /// Returns VP9 information from NVDEC provided offset and size + [[nodiscard]] Vp9PictureInfo GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state); + + /// Read and convert NVDEC provided entropy probs to Vp9EntropyProbs struct + void InsertEntropy(u64 offset, Vp9EntropyProbs& dst); + + /// Returns frame to be decoded after buffering + [[nodiscard]] Vp9FrameContainer GetCurrentFrame(const NvdecCommon::NvdecRegisters& state); + + /// Use NVDEC providied information to compose the headers for the current frame + [[nodiscard]] std::vector<u8> ComposeCompressedHeader(); + [[nodiscard]] VpxBitStreamWriter ComposeUncompressedHeader(); + + GPU& gpu; + std::vector<u8> frame; + + std::array<s8, 4> loop_filter_ref_deltas{}; + std::array<s8, 2> loop_filter_mode_deltas{}; + + bool hidden = false; + s64 current_frame_number = -2; // since we buffer 2 frames + s32 grace_period = 6; // frame offsets need to stabilize + std::array<FrameContexts, 4> frame_ctxs{}; + Vp9FrameContainer next_frame{}; + Vp9FrameContainer next_next_frame{}; + bool swap_next_golden{}; + + Vp9PictureInfo current_frame_info{}; + Vp9EntropyProbs prev_frame_probs{}; + + s32 diff_update_probability = 252; + s32 frame_sync_code = 0x498342; +}; + +} // namespace Decoder +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/vp9_types.h b/src/video_core/command_classes/codecs/vp9_types.h new file mode 100644 index 000000000..4f0b05d22 --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9_types.h @@ -0,0 +1,366 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <cstring> +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Tegra { +class GPU; + +namespace Decoder { +struct Vp9FrameDimensions { + s16 width{}; + s16 height{}; + s16 luma_pitch{}; + s16 chroma_pitch{}; +}; +static_assert(sizeof(Vp9FrameDimensions) == 0x8, "Vp9 Vp9FrameDimensions is an invalid size"); + +enum FrameFlags : u32 { + IsKeyFrame = 1 << 0, + LastFrameIsKeyFrame = 1 << 1, + FrameSizeChanged = 1 << 2, + ErrorResilientMode = 1 << 3, + LastShowFrame = 1 << 4, + IntraOnly = 1 << 5, +}; + +enum class MvJointType { + MvJointZero = 0, /* Zero vector */ + MvJointHnzvz = 1, /* Vert zero, hor nonzero */ + MvJointHzvnz = 2, /* Hor zero, vert nonzero */ + MvJointHnzvnz = 3, /* Both components nonzero */ +}; +enum class MvClassType { + MvClass0 = 0, /* (0, 2] integer pel */ + MvClass1 = 1, /* (2, 4] integer pel */ + MvClass2 = 2, /* (4, 8] integer pel */ + MvClass3 = 3, /* (8, 16] integer pel */ + MvClass4 = 4, /* (16, 32] integer pel */ + MvClass5 = 5, /* (32, 64] integer pel */ + MvClass6 = 6, /* (64, 128] integer pel */ + MvClass7 = 7, /* (128, 256] integer pel */ + MvClass8 = 8, /* (256, 512] integer pel */ + MvClass9 = 9, /* (512, 1024] integer pel */ + MvClass10 = 10, /* (1024,2048] integer pel */ +}; + +enum class BlockSize { + Block4x4 = 0, + Block4x8 = 1, + Block8x4 = 2, + Block8x8 = 3, + Block8x16 = 4, + Block16x8 = 5, + Block16x16 = 6, + Block16x32 = 7, + Block32x16 = 8, + Block32x32 = 9, + Block32x64 = 10, + Block64x32 = 11, + Block64x64 = 12, + BlockSizes = 13, + BlockInvalid = BlockSizes +}; + +enum class PredictionMode { + DcPred = 0, // Average of above and left pixels + VPred = 1, // Vertical + HPred = 2, // Horizontal + D45Pred = 3, // Directional 45 deg = round(arctan(1 / 1) * 180 / pi) + D135Pred = 4, // Directional 135 deg = 180 - 45 + D117Pred = 5, // Directional 117 deg = 180 - 63 + D153Pred = 6, // Directional 153 deg = 180 - 27 + D207Pred = 7, // Directional 207 deg = 180 + 27 + D63Pred = 8, // Directional 63 deg = round(arctan(2 / 1) * 180 / pi) + TmPred = 9, // True-motion + NearestMv = 10, + NearMv = 11, + ZeroMv = 12, + NewMv = 13, + MbModeCount = 14 +}; + +enum class TxSize { + Tx4x4 = 0, // 4x4 transform + Tx8x8 = 1, // 8x8 transform + Tx16x16 = 2, // 16x16 transform + Tx32x32 = 3, // 32x32 transform + TxSizes = 4 +}; + +enum class TxMode { + Only4X4 = 0, // Only 4x4 transform used + Allow8X8 = 1, // Allow block transform size up to 8x8 + Allow16X16 = 2, // Allow block transform size up to 16x16 + Allow32X32 = 3, // Allow block transform size up to 32x32 + TxModeSelect = 4, // Transform specified for each block + TxModes = 5 +}; + +enum class reference_mode { + SingleReference = 0, + CompoundReference = 1, + ReferenceModeSelect = 2, + ReferenceModes = 3 +}; + +struct Segmentation { + u8 enabled{}; + u8 update_map{}; + u8 temporal_update{}; + u8 abs_delta{}; + std::array<u32, 8> feature_mask{}; + std::array<std::array<s16, 4>, 8> feature_data{}; +}; +static_assert(sizeof(Segmentation) == 0x64, "Segmentation is an invalid size"); + +struct LoopFilter { + u8 mode_ref_delta_enabled{}; + std::array<s8, 4> ref_deltas{}; + std::array<s8, 2> mode_deltas{}; +}; +static_assert(sizeof(LoopFilter) == 0x7, "LoopFilter is an invalid size"); + +struct Vp9EntropyProbs { + std::array<u8, 36> y_mode_prob{}; + std::array<u8, 64> partition_prob{}; + std::array<u8, 2304> coef_probs{}; + std::array<u8, 8> switchable_interp_prob{}; + std::array<u8, 28> inter_mode_prob{}; + std::array<u8, 4> intra_inter_prob{}; + std::array<u8, 5> comp_inter_prob{}; + std::array<u8, 10> single_ref_prob{}; + std::array<u8, 5> comp_ref_prob{}; + std::array<u8, 6> tx_32x32_prob{}; + std::array<u8, 4> tx_16x16_prob{}; + std::array<u8, 2> tx_8x8_prob{}; + std::array<u8, 3> skip_probs{}; + std::array<u8, 3> joints{}; + std::array<u8, 2> sign{}; + std::array<u8, 20> classes{}; + std::array<u8, 2> class_0{}; + std::array<u8, 20> prob_bits{}; + std::array<u8, 12> class_0_fr{}; + std::array<u8, 6> fr{}; + std::array<u8, 2> class_0_hp{}; + std::array<u8, 2> high_precision{}; +}; +static_assert(sizeof(Vp9EntropyProbs) == 0x9F4, "Vp9EntropyProbs is an invalid size"); + +struct Vp9PictureInfo { + bool is_key_frame{}; + bool intra_only{}; + bool last_frame_was_key{}; + bool frame_size_changed{}; + bool error_resilient_mode{}; + bool last_frame_shown{}; + bool show_frame{}; + std::array<s8, 4> ref_frame_sign_bias{}; + s32 base_q_index{}; + s32 y_dc_delta_q{}; + s32 uv_dc_delta_q{}; + s32 uv_ac_delta_q{}; + bool lossless{}; + s32 transform_mode{}; + bool allow_high_precision_mv{}; + s32 interp_filter{}; + s32 reference_mode{}; + s8 comp_fixed_ref{}; + std::array<s8, 2> comp_var_ref{}; + s32 log2_tile_cols{}; + s32 log2_tile_rows{}; + bool segment_enabled{}; + bool segment_map_update{}; + bool segment_map_temporal_update{}; + s32 segment_abs_delta{}; + std::array<u32, 8> segment_feature_enable{}; + std::array<std::array<s16, 4>, 8> segment_feature_data{}; + bool mode_ref_delta_enabled{}; + bool use_prev_in_find_mv_refs{}; + std::array<s8, 4> ref_deltas{}; + std::array<s8, 2> mode_deltas{}; + Vp9EntropyProbs entropy{}; + Vp9FrameDimensions frame_size{}; + u8 first_level{}; + u8 sharpness_level{}; + u32 bitstream_size{}; + std::array<u64, 4> frame_offsets{}; + std::array<bool, 4> refresh_frame{}; +}; + +struct Vp9FrameContainer { + Vp9PictureInfo info{}; + std::vector<u8> bit_stream; +}; + +struct PictureInfo { + INSERT_PADDING_WORDS(12); + u32 bitstream_size{}; + INSERT_PADDING_WORDS(5); + Vp9FrameDimensions last_frame_size{}; + Vp9FrameDimensions golden_frame_size{}; + Vp9FrameDimensions alt_frame_size{}; + Vp9FrameDimensions current_frame_size{}; + u32 vp9_flags{}; + std::array<s8, 4> ref_frame_sign_bias{}; + u8 first_level{}; + u8 sharpness_level{}; + u8 base_q_index{}; + u8 y_dc_delta_q{}; + u8 uv_ac_delta_q{}; + u8 uv_dc_delta_q{}; + u8 lossless{}; + u8 tx_mode{}; + u8 allow_high_precision_mv{}; + u8 interp_filter{}; + u8 reference_mode{}; + s8 comp_fixed_ref{}; + std::array<s8, 2> comp_var_ref{}; + u8 log2_tile_cols{}; + u8 log2_tile_rows{}; + Segmentation segmentation{}; + LoopFilter loop_filter{}; + INSERT_PADDING_BYTES(5); + u32 surface_params{}; + INSERT_PADDING_WORDS(3); + + [[nodiscard]] Vp9PictureInfo Convert() const { + return { + .is_key_frame = (vp9_flags & FrameFlags::IsKeyFrame) != 0, + .intra_only = (vp9_flags & FrameFlags::IntraOnly) != 0, + .last_frame_was_key = (vp9_flags & FrameFlags::LastFrameIsKeyFrame) != 0, + .frame_size_changed = (vp9_flags & FrameFlags::FrameSizeChanged) != 0, + .error_resilient_mode = (vp9_flags & FrameFlags::ErrorResilientMode) != 0, + .last_frame_shown = (vp9_flags & FrameFlags::LastShowFrame) != 0, + .ref_frame_sign_bias = ref_frame_sign_bias, + .base_q_index = base_q_index, + .y_dc_delta_q = y_dc_delta_q, + .uv_dc_delta_q = uv_dc_delta_q, + .uv_ac_delta_q = uv_ac_delta_q, + .lossless = lossless != 0, + .transform_mode = tx_mode, + .allow_high_precision_mv = allow_high_precision_mv != 0, + .interp_filter = interp_filter, + .reference_mode = reference_mode, + .comp_fixed_ref = comp_fixed_ref, + .comp_var_ref = comp_var_ref, + .log2_tile_cols = log2_tile_cols, + .log2_tile_rows = log2_tile_rows, + .segment_enabled = segmentation.enabled != 0, + .segment_map_update = segmentation.update_map != 0, + .segment_map_temporal_update = segmentation.temporal_update != 0, + .segment_abs_delta = segmentation.abs_delta, + .segment_feature_enable = segmentation.feature_mask, + .segment_feature_data = segmentation.feature_data, + .mode_ref_delta_enabled = loop_filter.mode_ref_delta_enabled != 0, + .use_prev_in_find_mv_refs = !(vp9_flags == (FrameFlags::ErrorResilientMode)) && + !(vp9_flags == (FrameFlags::FrameSizeChanged)) && + !(vp9_flags == (FrameFlags::IntraOnly)) && + (vp9_flags == (FrameFlags::LastShowFrame)) && + !(vp9_flags == (FrameFlags::LastFrameIsKeyFrame)), + .ref_deltas = loop_filter.ref_deltas, + .mode_deltas = loop_filter.mode_deltas, + .frame_size = current_frame_size, + .first_level = first_level, + .sharpness_level = sharpness_level, + .bitstream_size = bitstream_size, + }; + } +}; +static_assert(sizeof(PictureInfo) == 0x100, "PictureInfo is an invalid size"); + +struct EntropyProbs { + INSERT_PADDING_BYTES(1024); + std::array<std::array<u8, 4>, 7> inter_mode_prob{}; + std::array<u8, 4> intra_inter_prob{}; + INSERT_PADDING_BYTES(80); + std::array<std::array<u8, 1>, 2> tx_8x8_prob{}; + std::array<std::array<u8, 2>, 2> tx_16x16_prob{}; + std::array<std::array<u8, 3>, 2> tx_32x32_prob{}; + std::array<u8, 4> y_mode_prob_e8{}; + std::array<std::array<u8, 8>, 4> y_mode_prob_e0e7{}; + INSERT_PADDING_BYTES(64); + std::array<std::array<u8, 4>, 16> partition_prob{}; + INSERT_PADDING_BYTES(10); + std::array<std::array<u8, 2>, 4> switchable_interp_prob{}; + std::array<u8, 5> comp_inter_prob{}; + std::array<u8, 4> skip_probs{}; + std::array<u8, 3> joints{}; + std::array<u8, 2> sign{}; + std::array<std::array<u8, 1>, 2> class_0{}; + std::array<std::array<u8, 3>, 2> fr{}; + std::array<u8, 2> class_0_hp{}; + std::array<u8, 2> high_precision{}; + std::array<std::array<u8, 10>, 2> classes{}; + std::array<std::array<std::array<u8, 3>, 2>, 2> class_0_fr{}; + std::array<std::array<u8, 10>, 2> pred_bits{}; + std::array<std::array<u8, 2>, 5> single_ref_prob{}; + std::array<u8, 5> comp_ref_prob{}; + INSERT_PADDING_BYTES(17); + std::array<std::array<std::array<std::array<std::array<std::array<u8, 4>, 6>, 6>, 2>, 2>, 4> + coef_probs{}; + + void Convert(Vp9EntropyProbs& fc) { + std::memcpy(fc.inter_mode_prob.data(), inter_mode_prob.data(), fc.inter_mode_prob.size()); + + std::memcpy(fc.intra_inter_prob.data(), intra_inter_prob.data(), + fc.intra_inter_prob.size()); + + std::memcpy(fc.tx_8x8_prob.data(), tx_8x8_prob.data(), fc.tx_8x8_prob.size()); + std::memcpy(fc.tx_16x16_prob.data(), tx_16x16_prob.data(), fc.tx_16x16_prob.size()); + std::memcpy(fc.tx_32x32_prob.data(), tx_32x32_prob.data(), fc.tx_32x32_prob.size()); + + for (s32 i = 0; i < 4; i++) { + for (s32 j = 0; j < 9; j++) { + fc.y_mode_prob[j + 9 * i] = j < 8 ? y_mode_prob_e0e7[i][j] : y_mode_prob_e8[i]; + } + } + + std::memcpy(fc.partition_prob.data(), partition_prob.data(), fc.partition_prob.size()); + + std::memcpy(fc.switchable_interp_prob.data(), switchable_interp_prob.data(), + fc.switchable_interp_prob.size()); + std::memcpy(fc.comp_inter_prob.data(), comp_inter_prob.data(), fc.comp_inter_prob.size()); + std::memcpy(fc.skip_probs.data(), skip_probs.data(), fc.skip_probs.size()); + + std::memcpy(fc.joints.data(), joints.data(), fc.joints.size()); + + std::memcpy(fc.sign.data(), sign.data(), fc.sign.size()); + std::memcpy(fc.class_0.data(), class_0.data(), fc.class_0.size()); + std::memcpy(fc.fr.data(), fr.data(), fc.fr.size()); + std::memcpy(fc.class_0_hp.data(), class_0_hp.data(), fc.class_0_hp.size()); + std::memcpy(fc.high_precision.data(), high_precision.data(), fc.high_precision.size()); + std::memcpy(fc.classes.data(), classes.data(), fc.classes.size()); + std::memcpy(fc.class_0_fr.data(), class_0_fr.data(), fc.class_0_fr.size()); + std::memcpy(fc.prob_bits.data(), pred_bits.data(), fc.prob_bits.size()); + std::memcpy(fc.single_ref_prob.data(), single_ref_prob.data(), fc.single_ref_prob.size()); + std::memcpy(fc.comp_ref_prob.data(), comp_ref_prob.data(), fc.comp_ref_prob.size()); + + std::memcpy(fc.coef_probs.data(), coef_probs.data(), fc.coef_probs.size()); + } +}; +static_assert(sizeof(EntropyProbs) == 0xEA0, "EntropyProbs is an invalid size"); + +enum class Ref { Last, Golden, AltRef }; + +struct RefPoolElement { + s64 frame{}; + Ref ref{}; + bool refresh{}; +}; + +struct FrameContexts { + s64 from{}; + bool adapted{}; + Vp9EntropyProbs probs{}; +}; + +}; // namespace Decoder +}; // namespace Tegra diff --git a/src/video_core/command_classes/host1x.cpp b/src/video_core/command_classes/host1x.cpp new file mode 100644 index 000000000..c4dd4881a --- /dev/null +++ b/src/video_core/command_classes/host1x.cpp @@ -0,0 +1,39 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "video_core/command_classes/host1x.h" +#include "video_core/gpu.h" + +Tegra::Host1x::Host1x(GPU& gpu_) : gpu(gpu_) {} + +Tegra::Host1x::~Host1x() = default; + +void Tegra::Host1x::StateWrite(u32 offset, u32 arguments) { + u8* const state_offset = reinterpret_cast<u8*>(&state) + offset * sizeof(u32); + std::memcpy(state_offset, &arguments, sizeof(u32)); +} + +void Tegra::Host1x::ProcessMethod(Method method, const std::vector<u32>& arguments) { + StateWrite(static_cast<u32>(method), arguments[0]); + switch (method) { + case Method::WaitSyncpt: + Execute(arguments[0]); + break; + case Method::LoadSyncptPayload32: + syncpoint_value = arguments[0]; + break; + case Method::WaitSyncpt32: + Execute(arguments[0]); + break; + default: + UNIMPLEMENTED_MSG("Host1x method 0x{:X}", static_cast<u32>(method)); + break; + } +} + +void Tegra::Host1x::Execute(u32 data) { + // This method waits on a valid syncpoint. + // TODO: Implement when proper Async is in place +} diff --git a/src/video_core/command_classes/host1x.h b/src/video_core/command_classes/host1x.h new file mode 100644 index 000000000..013eaa0c1 --- /dev/null +++ b/src/video_core/command_classes/host1x.h @@ -0,0 +1,78 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Tegra { +class GPU; +class Nvdec; + +class Host1x { +public: + struct Host1xClassRegisters { + u32 incr_syncpt{}; + u32 incr_syncpt_ctrl{}; + u32 incr_syncpt_error{}; + INSERT_PADDING_WORDS(5); + u32 wait_syncpt{}; + u32 wait_syncpt_base{}; + u32 wait_syncpt_incr{}; + u32 load_syncpt_base{}; + u32 incr_syncpt_base{}; + u32 clear{}; + u32 wait{}; + u32 wait_with_interrupt{}; + u32 delay_use{}; + u32 tick_count_high{}; + u32 tick_count_low{}; + u32 tick_ctrl{}; + INSERT_PADDING_WORDS(23); + u32 ind_ctrl{}; + u32 ind_off2{}; + u32 ind_off{}; + std::array<u32, 31> ind_data{}; + INSERT_PADDING_WORDS(1); + u32 load_syncpoint_payload32{}; + u32 stall_ctrl{}; + u32 wait_syncpt32{}; + u32 wait_syncpt_base32{}; + u32 load_syncpt_base32{}; + u32 incr_syncpt_base32{}; + u32 stall_count_high{}; + u32 stall_count_low{}; + u32 xref_ctrl{}; + u32 channel_xref_high{}; + u32 channel_xref_low{}; + }; + static_assert(sizeof(Host1xClassRegisters) == 0x164, "Host1xClassRegisters is an invalid size"); + + enum class Method : u32 { + WaitSyncpt = offsetof(Host1xClassRegisters, wait_syncpt) / 4, + LoadSyncptPayload32 = offsetof(Host1xClassRegisters, load_syncpoint_payload32) / 4, + WaitSyncpt32 = offsetof(Host1xClassRegisters, wait_syncpt32) / 4, + }; + + explicit Host1x(GPU& gpu); + ~Host1x(); + + /// Writes the method into the state, Invoke Execute() if encountered + void ProcessMethod(Method method, const std::vector<u32>& arguments); + +private: + /// For Host1x, execute is waiting on a syncpoint previously written into the state + void Execute(u32 data); + + /// Write argument into the provided offset + void StateWrite(u32 offset, u32 arguments); + + u32 syncpoint_value{}; + Host1xClassRegisters state{}; + GPU& gpu; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/nvdec.cpp b/src/video_core/command_classes/nvdec.cpp new file mode 100644 index 000000000..8ca7a7b06 --- /dev/null +++ b/src/video_core/command_classes/nvdec.cpp @@ -0,0 +1,52 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "video_core/command_classes/nvdec.h" +#include "video_core/gpu.h" + +namespace Tegra { + +Nvdec::Nvdec(GPU& gpu_) : gpu(gpu_), codec(std::make_unique<Codec>(gpu)) {} + +Nvdec::~Nvdec() = default; + +void Nvdec::ProcessMethod(Method method, const std::vector<u32>& arguments) { + if (method == Method::SetVideoCodec) { + codec->StateWrite(static_cast<u32>(method), arguments[0]); + } else { + codec->StateWrite(static_cast<u32>(method), static_cast<u64>(arguments[0]) << 8); + } + + switch (method) { + case Method::SetVideoCodec: + codec->SetTargetCodec(static_cast<NvdecCommon::VideoCodec>(arguments[0])); + break; + case Method::Execute: + Execute(); + break; + } +} + +AVFrame* Nvdec::GetFrame() { + return codec->GetCurrentFrame(); +} + +const AVFrame* Nvdec::GetFrame() const { + return codec->GetCurrentFrame(); +} + +void Nvdec::Execute() { + switch (codec->GetCurrentCodec()) { + case NvdecCommon::VideoCodec::H264: + case NvdecCommon::VideoCodec::Vp9: + codec->Decode(); + break; + default: + UNIMPLEMENTED_MSG("Unknown codec {}", static_cast<u32>(codec->GetCurrentCodec())); + break; + } +} + +} // namespace Tegra diff --git a/src/video_core/command_classes/nvdec.h b/src/video_core/command_classes/nvdec.h new file mode 100644 index 000000000..eec4443f9 --- /dev/null +++ b/src/video_core/command_classes/nvdec.h @@ -0,0 +1,39 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <vector> +#include "common/common_types.h" +#include "video_core/command_classes/codecs/codec.h" + +namespace Tegra { +class GPU; + +class Nvdec { +public: + enum class Method : u32 { + SetVideoCodec = 0x80, + Execute = 0xc0, + }; + + explicit Nvdec(GPU& gpu); + ~Nvdec(); + + /// Writes the method into the state, Invoke Execute() if encountered + void ProcessMethod(Method method, const std::vector<u32>& arguments); + + /// Return most recently decoded frame + [[nodiscard]] AVFrame* GetFrame(); + [[nodiscard]] const AVFrame* GetFrame() const; + +private: + /// Invoke codec to decode a frame + void Execute(); + + GPU& gpu; + std::unique_ptr<Codec> codec; +}; +} // namespace Tegra diff --git a/src/video_core/command_classes/nvdec_common.h b/src/video_core/command_classes/nvdec_common.h new file mode 100644 index 000000000..01b5e086d --- /dev/null +++ b/src/video_core/command_classes/nvdec_common.h @@ -0,0 +1,48 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Tegra::NvdecCommon { + +struct NvdecRegisters { + INSERT_PADDING_WORDS(256); + u64 set_codec_id{}; + INSERT_PADDING_WORDS(254); + u64 set_platform_id{}; + u64 picture_info_offset{}; + u64 frame_bitstream_offset{}; + u64 frame_number{}; + u64 h264_slice_data_offsets{}; + u64 h264_mv_dump_offset{}; + INSERT_PADDING_WORDS(6); + u64 frame_stats_offset{}; + u64 h264_last_surface_luma_offset{}; + u64 h264_last_surface_chroma_offset{}; + std::array<u64, 17> surface_luma_offset{}; + std::array<u64, 17> surface_chroma_offset{}; + INSERT_PADDING_WORDS(132); + u64 vp9_entropy_probs_offset{}; + u64 vp9_backward_updates_offset{}; + u64 vp9_last_frame_segmap_offset{}; + u64 vp9_curr_frame_segmap_offset{}; + INSERT_PADDING_WORDS(2); + u64 vp9_last_frame_mvs_offset{}; + u64 vp9_curr_frame_mvs_offset{}; + INSERT_PADDING_WORDS(2); +}; +static_assert(sizeof(NvdecRegisters) == (0xBC0), "NvdecRegisters is incorrect size"); + +enum class VideoCodec : u32 { + None = 0x0, + H264 = 0x3, + Vp8 = 0x5, + H265 = 0x7, + Vp9 = 0x9, +}; + +} // namespace Tegra::NvdecCommon diff --git a/src/video_core/command_classes/sync_manager.cpp b/src/video_core/command_classes/sync_manager.cpp new file mode 100644 index 000000000..19dc9e0ab --- /dev/null +++ b/src/video_core/command_classes/sync_manager.cpp @@ -0,0 +1,60 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include <algorithm> +#include "sync_manager.h" +#include "video_core/gpu.h" + +namespace Tegra { +SyncptIncrManager::SyncptIncrManager(GPU& gpu_) : gpu(gpu_) {} +SyncptIncrManager::~SyncptIncrManager() = default; + +void SyncptIncrManager::Increment(u32 id) { + increments.emplace_back(0, 0, id, true); + IncrementAllDone(); +} + +u32 SyncptIncrManager::IncrementWhenDone(u32 class_id, u32 id) { + const u32 handle = current_id++; + increments.emplace_back(handle, class_id, id); + return handle; +} + +void SyncptIncrManager::SignalDone(u32 handle) { + const auto done_incr = + std::find_if(increments.begin(), increments.end(), + [handle](const SyncptIncr& incr) { return incr.id == handle; }); + if (done_incr != increments.cend()) { + done_incr->complete = true; + } + IncrementAllDone(); +} + +void SyncptIncrManager::IncrementAllDone() { + std::size_t done_count = 0; + for (; done_count < increments.size(); ++done_count) { + if (!increments[done_count].complete) { + break; + } + gpu.IncrementSyncPoint(increments[done_count].syncpt_id); + } + increments.erase(increments.begin(), increments.begin() + done_count); +} +} // namespace Tegra diff --git a/src/video_core/command_classes/sync_manager.h b/src/video_core/command_classes/sync_manager.h new file mode 100644 index 000000000..2c321ec58 --- /dev/null +++ b/src/video_core/command_classes/sync_manager.h @@ -0,0 +1,64 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#pragma once + +#include <mutex> +#include <vector> +#include "common/common_types.h" + +namespace Tegra { +class GPU; +struct SyncptIncr { + u32 id; + u32 class_id; + u32 syncpt_id; + bool complete; + + SyncptIncr(u32 id_, u32 class_id_, u32 syncpt_id_, bool done = false) + : id(id_), class_id(class_id_), syncpt_id(syncpt_id_), complete(done) {} +}; + +class SyncptIncrManager { +public: + explicit SyncptIncrManager(GPU& gpu); + ~SyncptIncrManager(); + + /// Add syncpoint id and increment all + void Increment(u32 id); + + /// Returns a handle to increment later + u32 IncrementWhenDone(u32 class_id, u32 id); + + /// IncrememntAllDone, including handle + void SignalDone(u32 handle); + + /// Increment all sequential pending increments that are already done. + void IncrementAllDone(); + +private: + std::vector<SyncptIncr> increments; + std::mutex increment_lock; + u32 current_id{}; + + GPU& gpu; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/vic.cpp b/src/video_core/command_classes/vic.cpp new file mode 100644 index 000000000..5b52da277 --- /dev/null +++ b/src/video_core/command_classes/vic.cpp @@ -0,0 +1,180 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> +#include "common/assert.h" +#include "video_core/command_classes/nvdec.h" +#include "video_core/command_classes/vic.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" +#include "video_core/texture_cache/surface_params.h" + +extern "C" { +#include <libswscale/swscale.h> +} + +namespace Tegra { + +Vic::Vic(GPU& gpu_, std::shared_ptr<Nvdec> nvdec_processor_) + : gpu(gpu_), nvdec_processor(std::move(nvdec_processor_)) {} +Vic::~Vic() = default; + +void Vic::VicStateWrite(u32 offset, u32 arguments) { + u8* const state_offset = reinterpret_cast<u8*>(&vic_state) + offset * sizeof(u32); + std::memcpy(state_offset, &arguments, sizeof(u32)); +} + +void Vic::ProcessMethod(Method method, const std::vector<u32>& arguments) { + LOG_DEBUG(HW_GPU, "Vic method 0x{:X}", static_cast<u32>(method)); + VicStateWrite(static_cast<u32>(method), arguments[0]); + const u64 arg = static_cast<u64>(arguments[0]) << 8; + switch (method) { + case Method::Execute: + Execute(); + break; + case Method::SetConfigStructOffset: + config_struct_address = arg; + break; + case Method::SetOutputSurfaceLumaOffset: + output_surface_luma_address = arg; + break; + case Method::SetOutputSurfaceChromaUOffset: + output_surface_chroma_u_address = arg; + break; + case Method::SetOutputSurfaceChromaVOffset: + output_surface_chroma_v_address = arg; + break; + default: + break; + } +} + +void Vic::Execute() { + if (output_surface_luma_address == 0) { + LOG_ERROR(Service_NVDRV, "VIC Luma address not set. Recieved 0x{:X}", + vic_state.output_surface.luma_offset); + return; + } + const VicConfig config{gpu.MemoryManager().Read<u64>(config_struct_address + 0x20)}; + const VideoPixelFormat pixel_format = + static_cast<VideoPixelFormat>(config.pixel_format.Value()); + switch (pixel_format) { + case VideoPixelFormat::BGRA8: + case VideoPixelFormat::RGBA8: { + LOG_TRACE(Service_NVDRV, "Writing RGB Frame"); + const auto* frame = nvdec_processor->GetFrame(); + + if (!frame || frame->width == 0 || frame->height == 0) { + return; + } + if (scaler_ctx == nullptr || frame->width != scaler_width || + frame->height != scaler_height) { + const AVPixelFormat target_format = + (pixel_format == VideoPixelFormat::RGBA8) ? AV_PIX_FMT_RGBA : AV_PIX_FMT_BGRA; + + sws_freeContext(scaler_ctx); + scaler_ctx = nullptr; + + // FFmpeg returns all frames in YUV420, convert it into expected format + scaler_ctx = + sws_getContext(frame->width, frame->height, AV_PIX_FMT_YUV420P, frame->width, + frame->height, target_format, 0, nullptr, nullptr, nullptr); + + scaler_width = frame->width; + scaler_height = frame->height; + } + // Get Converted frame + const std::size_t linear_size = frame->width * frame->height * 4; + + using AVMallocPtr = std::unique_ptr<u8, decltype(&av_free)>; + AVMallocPtr converted_frame_buffer{static_cast<u8*>(av_malloc(linear_size)), av_free}; + + const int converted_stride{frame->width * 4}; + u8* const converted_frame_buf_addr{converted_frame_buffer.get()}; + + sws_scale(scaler_ctx, frame->data, frame->linesize, 0, frame->height, + &converted_frame_buf_addr, &converted_stride); + + const u32 blk_kind = static_cast<u32>(config.block_linear_kind); + if (blk_kind != 0) { + // swizzle pitch linear to block linear + const u32 block_height = static_cast<u32>(config.block_linear_height_log2); + const auto size = Tegra::Texture::CalculateSize(true, 4, frame->width, frame->height, 1, + block_height, 0); + std::vector<u8> swizzled_data(size); + Tegra::Texture::CopySwizzledData(frame->width, frame->height, 1, 4, 4, + swizzled_data.data(), converted_frame_buffer.get(), + false, block_height, 0, 1); + + gpu.MemoryManager().WriteBlock(output_surface_luma_address, swizzled_data.data(), size); + gpu.Maxwell3D().OnMemoryWrite(); + } else { + // send pitch linear frame + gpu.MemoryManager().WriteBlock(output_surface_luma_address, converted_frame_buf_addr, + linear_size); + gpu.Maxwell3D().OnMemoryWrite(); + } + break; + } + case VideoPixelFormat::Yuv420: { + LOG_TRACE(Service_NVDRV, "Writing YUV420 Frame"); + + const auto* frame = nvdec_processor->GetFrame(); + + if (!frame || frame->width == 0 || frame->height == 0) { + return; + } + + const std::size_t surface_width = config.surface_width_minus1 + 1; + const std::size_t surface_height = config.surface_height_minus1 + 1; + const std::size_t half_width = surface_width / 2; + const std::size_t half_height = config.surface_height_minus1 / 2; + const std::size_t aligned_width = (surface_width + 0xff) & ~0xff; + + const auto* luma_ptr = frame->data[0]; + const auto* chroma_b_ptr = frame->data[1]; + const auto* chroma_r_ptr = frame->data[2]; + const auto stride = frame->linesize[0]; + const auto half_stride = frame->linesize[1]; + + std::vector<u8> luma_buffer(aligned_width * surface_height); + std::vector<u8> chroma_buffer(aligned_width * half_height); + + // Populate luma buffer + for (std::size_t y = 0; y < surface_height - 1; ++y) { + std::size_t src = y * stride; + std::size_t dst = y * aligned_width; + + std::size_t size = surface_width; + + for (std::size_t offset = 0; offset < size; ++offset) { + luma_buffer[dst + offset] = luma_ptr[src + offset]; + } + } + gpu.MemoryManager().WriteBlock(output_surface_luma_address, luma_buffer.data(), + luma_buffer.size()); + + // Populate chroma buffer from both channels with interleaving. + for (std::size_t y = 0; y < half_height; ++y) { + std::size_t src = y * half_stride; + std::size_t dst = y * aligned_width; + + for (std::size_t x = 0; x < half_width; ++x) { + chroma_buffer[dst + x * 2] = chroma_b_ptr[src + x]; + chroma_buffer[dst + x * 2 + 1] = chroma_r_ptr[src + x]; + } + } + gpu.MemoryManager().WriteBlock(output_surface_chroma_u_address, chroma_buffer.data(), + chroma_buffer.size()); + gpu.Maxwell3D().OnMemoryWrite(); + break; + } + default: + UNIMPLEMENTED_MSG("Unknown video pixel format {}", config.pixel_format.Value()); + break; + } +} + +} // namespace Tegra diff --git a/src/video_core/command_classes/vic.h b/src/video_core/command_classes/vic.h new file mode 100644 index 000000000..8c4e284a1 --- /dev/null +++ b/src/video_core/command_classes/vic.h @@ -0,0 +1,110 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <vector> +#include "common/bit_field.h" +#include "common/common_types.h" + +struct SwsContext; + +namespace Tegra { +class GPU; +class Nvdec; + +struct PlaneOffsets { + u32 luma_offset{}; + u32 chroma_u_offset{}; + u32 chroma_v_offset{}; +}; + +struct VicRegisters { + INSERT_PADDING_WORDS(64); + u32 nop{}; + INSERT_PADDING_WORDS(15); + u32 pm_trigger{}; + INSERT_PADDING_WORDS(47); + u32 set_application_id{}; + u32 set_watchdog_timer{}; + INSERT_PADDING_WORDS(17); + u32 context_save_area{}; + u32 context_switch{}; + INSERT_PADDING_WORDS(43); + u32 execute{}; + INSERT_PADDING_WORDS(63); + std::array<std::array<PlaneOffsets, 8>, 8> surfacex_slots{}; + u32 picture_index{}; + u32 control_params{}; + u32 config_struct_offset{}; + u32 filter_struct_offset{}; + u32 palette_offset{}; + u32 hist_offset{}; + u32 context_id{}; + u32 fce_ucode_size{}; + PlaneOffsets output_surface{}; + u32 fce_ucode_offset{}; + INSERT_PADDING_WORDS(4); + std::array<u32, 8> slot_context_id{}; + INSERT_PADDING_WORDS(16); +}; +static_assert(sizeof(VicRegisters) == 0x7A0, "VicRegisters is an invalid size"); + +class Vic { +public: + enum class Method : u32 { + Execute = 0xc0, + SetControlParams = 0x1c1, + SetConfigStructOffset = 0x1c2, + SetOutputSurfaceLumaOffset = 0x1c8, + SetOutputSurfaceChromaUOffset = 0x1c9, + SetOutputSurfaceChromaVOffset = 0x1ca + }; + + explicit Vic(GPU& gpu, std::shared_ptr<Nvdec> nvdec_processor); + ~Vic(); + + /// Write to the device state. + void ProcessMethod(Method method, const std::vector<u32>& arguments); + +private: + void Execute(); + + void VicStateWrite(u32 offset, u32 arguments); + VicRegisters vic_state{}; + + enum class VideoPixelFormat : u64_le { + RGBA8 = 0x1f, + BGRA8 = 0x20, + Yuv420 = 0x44, + }; + + union VicConfig { + u64_le raw{}; + BitField<0, 7, u64_le> pixel_format; + BitField<7, 2, u64_le> chroma_loc_horiz; + BitField<9, 2, u64_le> chroma_loc_vert; + BitField<11, 4, u64_le> block_linear_kind; + BitField<15, 4, u64_le> block_linear_height_log2; + BitField<19, 3, u64_le> reserved0; + BitField<22, 10, u64_le> reserved1; + BitField<32, 14, u64_le> surface_width_minus1; + BitField<46, 14, u64_le> surface_height_minus1; + }; + + GPU& gpu; + std::shared_ptr<Tegra::Nvdec> nvdec_processor; + + GPUVAddr config_struct_address{}; + GPUVAddr output_surface_luma_address{}; + GPUVAddr output_surface_chroma_u_address{}; + GPUVAddr output_surface_chroma_v_address{}; + + SwsContext* scaler_ctx{}; + s32 scaler_width{}; + s32 scaler_height{}; +}; + +} // namespace Tegra diff --git a/src/video_core/compatible_formats.cpp b/src/video_core/compatible_formats.cpp new file mode 100644 index 000000000..b06c32c84 --- /dev/null +++ b/src/video_core/compatible_formats.cpp @@ -0,0 +1,155 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> +#include <bitset> +#include <cstddef> + +#include "video_core/compatible_formats.h" +#include "video_core/surface.h" + +namespace VideoCore::Surface { + +namespace { + +// Compatibility table taken from Table 3.X.2 in: +// https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_texture_view.txt + +constexpr std::array VIEW_CLASS_128_BITS = { + PixelFormat::R32G32B32A32_FLOAT, + PixelFormat::R32G32B32A32_UINT, + PixelFormat::R32G32B32A32_SINT, +}; + +constexpr std::array VIEW_CLASS_96_BITS = { + PixelFormat::R32G32B32_FLOAT, +}; +// Missing formats: +// PixelFormat::RGB32UI, +// PixelFormat::RGB32I, + +constexpr std::array VIEW_CLASS_64_BITS = { + PixelFormat::R32G32_FLOAT, PixelFormat::R32G32_UINT, + PixelFormat::R32G32_SINT, PixelFormat::R16G16B16A16_FLOAT, + PixelFormat::R16G16B16A16_UNORM, PixelFormat::R16G16B16A16_SNORM, + PixelFormat::R16G16B16A16_UINT, PixelFormat::R16G16B16A16_SINT, +}; + +// TODO: How should we handle 48 bits? + +constexpr std::array VIEW_CLASS_32_BITS = { + PixelFormat::R16G16_FLOAT, PixelFormat::B10G11R11_FLOAT, PixelFormat::R32_FLOAT, + PixelFormat::A2B10G10R10_UNORM, PixelFormat::R16G16_UINT, PixelFormat::R32_UINT, + PixelFormat::R16G16_SINT, PixelFormat::R32_SINT, PixelFormat::A8B8G8R8_UNORM, + PixelFormat::R16G16_UNORM, PixelFormat::A8B8G8R8_SNORM, PixelFormat::R16G16_SNORM, + PixelFormat::A8B8G8R8_SRGB, PixelFormat::E5B9G9R9_FLOAT, PixelFormat::B8G8R8A8_UNORM, + PixelFormat::B8G8R8A8_SRGB, PixelFormat::A8B8G8R8_UINT, PixelFormat::A8B8G8R8_SINT, + PixelFormat::A2B10G10R10_UINT, +}; + +// TODO: How should we handle 24 bits? + +constexpr std::array VIEW_CLASS_16_BITS = { + PixelFormat::R16_FLOAT, PixelFormat::R8G8_UINT, PixelFormat::R16_UINT, + PixelFormat::R16_SINT, PixelFormat::R8G8_UNORM, PixelFormat::R16_UNORM, + PixelFormat::R8G8_SNORM, PixelFormat::R16_SNORM, PixelFormat::R8G8_SINT, +}; + +constexpr std::array VIEW_CLASS_8_BITS = { + PixelFormat::R8_UINT, + PixelFormat::R8_UNORM, + PixelFormat::R8_SINT, + PixelFormat::R8_SNORM, +}; + +constexpr std::array VIEW_CLASS_RGTC1_RED = { + PixelFormat::BC4_UNORM, + PixelFormat::BC4_SNORM, +}; + +constexpr std::array VIEW_CLASS_RGTC2_RG = { + PixelFormat::BC5_UNORM, + PixelFormat::BC5_SNORM, +}; + +constexpr std::array VIEW_CLASS_BPTC_UNORM = { + PixelFormat::BC7_UNORM, + PixelFormat::BC7_SRGB, +}; + +constexpr std::array VIEW_CLASS_BPTC_FLOAT = { + PixelFormat::BC6H_SFLOAT, + PixelFormat::BC6H_UFLOAT, +}; + +// Compatibility table taken from Table 4.X.1 in: +// https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_copy_image.txt + +constexpr std::array COPY_CLASS_128_BITS = { + PixelFormat::R32G32B32A32_UINT, PixelFormat::R32G32B32A32_FLOAT, PixelFormat::R32G32B32A32_SINT, + PixelFormat::BC2_UNORM, PixelFormat::BC2_SRGB, PixelFormat::BC3_UNORM, + PixelFormat::BC3_SRGB, PixelFormat::BC5_UNORM, PixelFormat::BC5_SNORM, + PixelFormat::BC7_UNORM, PixelFormat::BC7_SRGB, PixelFormat::BC6H_SFLOAT, + PixelFormat::BC6H_UFLOAT, +}; +// Missing formats: +// PixelFormat::RGBA32I +// COMPRESSED_RG_RGTC2 + +constexpr std::array COPY_CLASS_64_BITS = { + PixelFormat::R16G16B16A16_FLOAT, PixelFormat::R16G16B16A16_UINT, + PixelFormat::R16G16B16A16_UNORM, PixelFormat::R16G16B16A16_SNORM, + PixelFormat::R16G16B16A16_SINT, PixelFormat::R32G32_UINT, + PixelFormat::R32G32_FLOAT, PixelFormat::R32G32_SINT, + PixelFormat::BC1_RGBA_UNORM, PixelFormat::BC1_RGBA_SRGB, +}; +// Missing formats: +// COMPRESSED_RGB_S3TC_DXT1_EXT +// COMPRESSED_SRGB_S3TC_DXT1_EXT +// COMPRESSED_RGBA_S3TC_DXT1_EXT +// COMPRESSED_SIGNED_RED_RGTC1 + +void Enable(FormatCompatibility::Table& compatiblity, size_t format_a, size_t format_b) { + compatiblity[format_a][format_b] = true; + compatiblity[format_b][format_a] = true; +} + +void Enable(FormatCompatibility::Table& compatibility, PixelFormat format_a, PixelFormat format_b) { + Enable(compatibility, static_cast<size_t>(format_a), static_cast<size_t>(format_b)); +} + +template <typename Range> +void EnableRange(FormatCompatibility::Table& compatibility, const Range& range) { + for (auto it_a = range.begin(); it_a != range.end(); ++it_a) { + for (auto it_b = it_a; it_b != range.end(); ++it_b) { + Enable(compatibility, *it_a, *it_b); + } + } +} + +} // Anonymous namespace + +FormatCompatibility::FormatCompatibility() { + for (size_t i = 0; i < MaxPixelFormat; ++i) { + // Identity is allowed + Enable(view, i, i); + } + + EnableRange(view, VIEW_CLASS_128_BITS); + EnableRange(view, VIEW_CLASS_96_BITS); + EnableRange(view, VIEW_CLASS_64_BITS); + EnableRange(view, VIEW_CLASS_32_BITS); + EnableRange(view, VIEW_CLASS_16_BITS); + EnableRange(view, VIEW_CLASS_8_BITS); + EnableRange(view, VIEW_CLASS_RGTC1_RED); + EnableRange(view, VIEW_CLASS_RGTC2_RG); + EnableRange(view, VIEW_CLASS_BPTC_UNORM); + EnableRange(view, VIEW_CLASS_BPTC_FLOAT); + + copy = view; + EnableRange(copy, COPY_CLASS_128_BITS); + EnableRange(copy, COPY_CLASS_64_BITS); +} + +} // namespace VideoCore::Surface diff --git a/src/video_core/compatible_formats.h b/src/video_core/compatible_formats.h new file mode 100644 index 000000000..51766349b --- /dev/null +++ b/src/video_core/compatible_formats.h @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <bitset> +#include <cstddef> + +#include "video_core/surface.h" + +namespace VideoCore::Surface { + +class FormatCompatibility { +public: + using Table = std::array<std::bitset<MaxPixelFormat>, MaxPixelFormat>; + + explicit FormatCompatibility(); + + bool TestView(PixelFormat format_a, PixelFormat format_b) const noexcept { + return view[static_cast<size_t>(format_a)][static_cast<size_t>(format_b)]; + } + + bool TestCopy(PixelFormat format_a, PixelFormat format_b) const noexcept { + return copy[static_cast<size_t>(format_a)][static_cast<size_t>(format_b)]; + } + +private: + Table view; + Table copy; +}; + +} // namespace VideoCore::Surface diff --git a/src/video_core/dma_pusher.cpp b/src/video_core/dma_pusher.cpp index 713c14182..d8801b1f5 100644 --- a/src/video_core/dma_pusher.cpp +++ b/src/video_core/dma_pusher.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/cityhash.h" #include "common/microprofile.h" #include "core/core.h" #include "core/memory.h" @@ -12,7 +13,7 @@ namespace Tegra { -DmaPusher::DmaPusher(GPU& gpu) : gpu(gpu) {} +DmaPusher::DmaPusher(Core::System& system, GPU& gpu) : gpu{gpu}, system{system} {} DmaPusher::~DmaPusher() = default; @@ -21,17 +22,22 @@ MICROPROFILE_DEFINE(DispatchCalls, "GPU", "Execute command buffer", MP_RGB(128, void DmaPusher::DispatchCalls() { MICROPROFILE_SCOPE(DispatchCalls); + gpu.SyncGuestHost(); // On entering GPU code, assume all memory may be touched by the ARM core. gpu.Maxwell3D().OnMemoryWrite(); dma_pushbuffer_subindex = 0; - while (Core::System::GetInstance().IsPoweredOn()) { + dma_state.is_last_call = true; + + while (system.IsPoweredOn()) { if (!Step()) { break; } } gpu.FlushCommands(); + gpu.SyncGuestHost(); + gpu.OnCommandListEnd(); } bool DmaPusher::Step() { @@ -40,44 +46,59 @@ bool DmaPusher::Step() { return false; } - const CommandList& command_list{dma_pushbuffer.front()}; - ASSERT_OR_EXECUTE(!command_list.empty(), { - // Somehow the command_list is empty, in order to avoid a crash - // We ignore it and assume its size is 0. - dma_pushbuffer.pop(); - dma_pushbuffer_subindex = 0; - return true; - }); - const CommandListHeader command_list_header{command_list[dma_pushbuffer_subindex++]}; - GPUVAddr dma_get = command_list_header.addr; - GPUVAddr dma_put = dma_get + command_list_header.size * sizeof(u32); - bool non_main = command_list_header.is_non_main; - - if (dma_pushbuffer_subindex >= command_list.size()) { - // We've gone through the current list, remove it from the queue - dma_pushbuffer.pop(); - dma_pushbuffer_subindex = 0; - } + CommandList& command_list{dma_pushbuffer.front()}; - if (command_list_header.size == 0) { - return true; - } + ASSERT_OR_EXECUTE( + command_list.command_lists.size() || command_list.prefetch_command_list.size(), { + // Somehow the command_list is empty, in order to avoid a crash + // We ignore it and assume its size is 0. + dma_pushbuffer.pop(); + dma_pushbuffer_subindex = 0; + return true; + }); - // Push buffer non-empty, read a word - command_headers.resize(command_list_header.size); - gpu.MemoryManager().ReadBlockUnsafe(dma_get, command_headers.data(), - command_list_header.size * sizeof(u32)); + if (command_list.prefetch_command_list.size()) { + // Prefetched command list from nvdrv, used for things like synchronization + command_headers = std::move(command_list.prefetch_command_list); + dma_pushbuffer.pop(); + } else { + const CommandListHeader command_list_header{ + command_list.command_lists[dma_pushbuffer_subindex++]}; + const GPUVAddr dma_get = command_list_header.addr; + + if (dma_pushbuffer_subindex >= command_list.command_lists.size()) { + // We've gone through the current list, remove it from the queue + dma_pushbuffer.pop(); + dma_pushbuffer_subindex = 0; + } - for (const CommandHeader& command_header : command_headers) { + if (command_list_header.size == 0) { + return true; + } - // now, see if we're in the middle of a command - if (dma_state.length_pending) { - // Second word of long non-inc methods command - method count - dma_state.length_pending = 0; - dma_state.method_count = command_header.method_count_; - } else if (dma_state.method_count) { + // Push buffer non-empty, read a word + command_headers.resize(command_list_header.size); + gpu.MemoryManager().ReadBlockUnsafe(dma_get, command_headers.data(), + command_list_header.size * sizeof(u32)); + } + for (std::size_t index = 0; index < command_headers.size();) { + const CommandHeader& command_header = command_headers[index]; + + if (dma_state.method_count) { // Data word of methods command - CallMethod(command_header.argument); + if (dma_state.non_incrementing) { + const u32 max_write = static_cast<u32>( + std::min<std::size_t>(index + dma_state.method_count, command_headers.size()) - + index); + CallMultiMethod(&command_header.argument, max_write); + dma_state.method_count -= max_write; + dma_state.is_last_call = true; + index += max_write; + continue; + } else { + dma_state.is_last_call = dma_state.method_count <= 1; + CallMethod(command_header.argument); + } if (!dma_state.non_incrementing) { dma_state.method++; @@ -117,11 +138,7 @@ bool DmaPusher::Step() { break; } } - } - - if (!non_main) { - // TODO (degasus): This is dead code, as dma_mget is never read. - dma_mget = dma_put; + index++; } return true; @@ -134,7 +151,22 @@ void DmaPusher::SetState(const CommandHeader& command_header) { } void DmaPusher::CallMethod(u32 argument) const { - gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count}); + if (dma_state.method < non_puller_methods) { + gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count}); + } else { + subchannels[dma_state.subchannel]->CallMethod(dma_state.method, argument, + dma_state.is_last_call); + } +} + +void DmaPusher::CallMultiMethod(const u32* base_start, u32 num_methods) const { + if (dma_state.method < non_puller_methods) { + gpu.CallMultiMethod(dma_state.method, dma_state.subchannel, base_start, num_methods, + dma_state.method_count); + } else { + subchannels[dma_state.subchannel]->CallMultiMethod(dma_state.method, base_start, + num_methods, dma_state.method_count); + } } } // namespace Tegra diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h index 6ab06518f..96ac267f7 100644 --- a/src/video_core/dma_pusher.h +++ b/src/video_core/dma_pusher.h @@ -4,14 +4,22 @@ #pragma once +#include <array> #include <vector> #include <queue> #include "common/bit_field.h" #include "common/common_types.h" +#include "video_core/engines/engine_interface.h" + +namespace Core { +class System; +} namespace Tegra { +class GPU; + enum class SubmissionMode : u32 { IncreasingOld = 0, Increasing = 1, @@ -21,6 +29,31 @@ enum class SubmissionMode : u32 { IncreaseOnce = 5 }; +// Note that, traditionally, methods are treated as 4-byte addressable locations, and hence +// their numbers are written down multiplied by 4 in Docs. Here we are not multiply by 4. +// So the values you see in docs might be multiplied by 4. +enum class BufferMethods : u32 { + BindObject = 0x0, + Nop = 0x2, + SemaphoreAddressHigh = 0x4, + SemaphoreAddressLow = 0x5, + SemaphoreSequence = 0x6, + SemaphoreTrigger = 0x7, + NotifyIntr = 0x8, + WrcacheFlush = 0x9, + Unk28 = 0xA, + UnkCacheFlush = 0xB, + RefCnt = 0x14, + SemaphoreAcquire = 0x1A, + SemaphoreRelease = 0x1B, + FenceValue = 0x1C, + FenceAction = 0x1D, + WaitForInterrupt = 0x1E, + Unk7c = 0x1F, + Yield = 0x20, + NonPullerMethods = 0x40, +}; + struct CommandListHeader { union { u64 raw; @@ -43,9 +76,23 @@ union CommandHeader { static_assert(std::is_standard_layout_v<CommandHeader>, "CommandHeader is not standard layout"); static_assert(sizeof(CommandHeader) == sizeof(u32), "CommandHeader has incorrect size!"); -class GPU; - -using CommandList = std::vector<Tegra::CommandListHeader>; +inline CommandHeader BuildCommandHeader(BufferMethods method, u32 arg_count, SubmissionMode mode) { + CommandHeader result{}; + result.method.Assign(static_cast<u32>(method)); + result.arg_count.Assign(arg_count); + result.mode.Assign(mode); + return result; +} + +struct CommandList final { + CommandList() = default; + explicit CommandList(std::size_t size) : command_lists(size) {} + explicit CommandList(std::vector<Tegra::CommandHeader>&& prefetch_command_list) + : prefetch_command_list{std::move(prefetch_command_list)} {} + + std::vector<Tegra::CommandListHeader> command_lists; + std::vector<Tegra::CommandHeader> prefetch_command_list; +}; /** * The DmaPusher class implements DMA submission to FIFOs, providing an area of memory that the @@ -54,9 +101,9 @@ using CommandList = std::vector<Tegra::CommandListHeader>; * See https://envytools.readthedocs.io/en/latest/hw/fifo/dma-pusher.html#fifo-dma-pusher for * details on this implementation. */ -class DmaPusher { +class DmaPusher final { public: - explicit DmaPusher(GPU& gpu); + explicit DmaPusher(Core::System& system, GPU& gpu); ~DmaPusher(); void Push(CommandList&& entries) { @@ -65,14 +112,19 @@ public: void DispatchCalls(); + void BindSubchannel(Tegra::Engines::EngineInterface* engine, u32 subchannel_id) { + subchannels[subchannel_id] = engine; + } + private: + static constexpr u32 non_puller_methods = 0x40; + static constexpr u32 max_subchannels = 8; bool Step(); void SetState(const CommandHeader& command_header); void CallMethod(u32 argument) const; - - GPU& gpu; + void CallMultiMethod(const u32* base_start, u32 num_methods) const; std::vector<CommandHeader> command_headers; ///< Buffer for list of commands fetched at once @@ -85,13 +137,18 @@ private: u32 method_count; ///< Current method count u32 length_pending; ///< Large NI command length pending bool non_incrementing; ///< Current command's NI flag + bool is_last_call; }; DmaState dma_state{}; bool dma_increment_once{}; - GPUVAddr dma_mget{}; ///< main pushbuffer last read address bool ib_enable{true}; ///< IB mode enabled + + std::array<Tegra::Engines::EngineInterface*, max_subchannels> subchannels{}; + + GPU& gpu; + Core::System& system; }; } // namespace Tegra diff --git a/src/video_core/engines/const_buffer_engine_interface.h b/src/video_core/engines/const_buffer_engine_interface.h index ebe139504..f46e81bb7 100644 --- a/src/video_core/engines/const_buffer_engine_interface.h +++ b/src/video_core/engines/const_buffer_engine_interface.h @@ -93,6 +93,7 @@ public: virtual SamplerDescriptor AccessBoundSampler(ShaderType stage, u64 offset) const = 0; virtual SamplerDescriptor AccessBindlessSampler(ShaderType stage, u64 const_buffer, u64 offset) const = 0; + virtual SamplerDescriptor AccessSampler(u32 handle) const = 0; virtual u32 GetBoundBuffer() const = 0; virtual VideoCore::GuestDriverProfile& AccessGuestDriverProfile() = 0; diff --git a/src/video_core/engines/engine_interface.h b/src/video_core/engines/engine_interface.h new file mode 100644 index 000000000..18a9db7e6 --- /dev/null +++ b/src/video_core/engines/engine_interface.h @@ -0,0 +1,22 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <type_traits> +#include "common/common_types.h" + +namespace Tegra::Engines { + +class EngineInterface { +public: + /// Write the value to the register identified by method. + virtual void CallMethod(u32 method, u32 method_argument, bool is_last_call) = 0; + + /// Write multiple values to the register identified by method. + virtual void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) = 0; +}; + +} // namespace Tegra::Engines diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp index 85d308e26..9409c4075 100644 --- a/src/video_core/engines/fermi_2d.cpp +++ b/src/video_core/engines/fermi_2d.cpp @@ -10,15 +10,21 @@ namespace Tegra::Engines { -Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {} +Fermi2D::Fermi2D() = default; -void Fermi2D::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +Fermi2D::~Fermi2D() = default; + +void Fermi2D::BindRasterizer(VideoCore::RasterizerInterface& rasterizer_) { + rasterizer = &rasterizer_; +} + +void Fermi2D::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Fermi2D register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; - switch (method_call.method) { + switch (method) { // Trigger the surface copy on the last register write. This is blit_src_y, but this is 64-bit, // so trigger on the second 32-bit write. case FERMI2D_REG_INDEX(blit_src_y) + 1: { @@ -28,7 +34,13 @@ void Fermi2D::CallMethod(const GPU::MethodCall& method_call) { } } -std::pair<u32, u32> DelimitLine(u32 src_1, u32 src_2, u32 dst_1, u32 dst_2, u32 src_line) { +void Fermi2D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) { + for (std::size_t i = 0; i < amount; i++) { + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); + } +} + +static std::pair<u32, u32> DelimitLine(u32 src_1, u32 src_2, u32 dst_1, u32 dst_2, u32 src_line) { const u32 line_a = src_2 - src_1; const u32 line_b = dst_2 - dst_1; const u32 excess = std::max<s32>(0, line_a - src_line + src_1); @@ -75,13 +87,13 @@ void Fermi2D::HandleSurfaceCopy() { const Common::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2}; const Common::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y, dst_blit_x2, dst_blit_y2}; - Config copy_config; - copy_config.operation = regs.operation; - copy_config.filter = regs.blit_control.filter; - copy_config.src_rect = src_rect; - copy_config.dst_rect = dst_rect; - - if (!rasterizer.AccelerateSurfaceCopy(regs.src, regs.dst, copy_config)) { + const Config copy_config{ + .operation = regs.operation, + .filter = regs.blit_control.filter, + .src_rect = src_rect, + .dst_rect = dst_rect, + }; + if (!rasterizer->AccelerateSurfaceCopy(regs.src, regs.dst, copy_config)) { UNIMPLEMENTED(); } } diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h index dba342c70..0909709ec 100644 --- a/src/video_core/engines/fermi_2d.h +++ b/src/video_core/engines/fermi_2d.h @@ -10,6 +10,7 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/math_util.h" +#include "video_core/engines/engine_interface.h" #include "video_core/gpu.h" namespace Tegra { @@ -31,13 +32,20 @@ namespace Tegra::Engines { #define FERMI2D_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::Fermi2D::Regs, field_name) / sizeof(u32)) -class Fermi2D final { +class Fermi2D final : public EngineInterface { public: - explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer); - ~Fermi2D() = default; + explicit Fermi2D(); + ~Fermi2D(); + + /// Binds a rasterizer to this engine. + void BindRasterizer(VideoCore::RasterizerInterface& rasterizer); /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; + + /// Write multiple values to the register identified by method. + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; enum class Origin : u32 { Center = 0, @@ -137,14 +145,14 @@ public: } regs{}; struct Config { - Operation operation; - Filter filter; + Operation operation{}; + Filter filter{}; Common::Rectangle<u32> src_rect; Common::Rectangle<u32> dst_rect; }; private: - VideoCore::RasterizerInterface& rasterizer; + VideoCore::RasterizerInterface* rasterizer; /// Performs the copy from the source surface to the destination surface as configured in the /// registers. diff --git a/src/video_core/engines/kepler_compute.cpp b/src/video_core/engines/kepler_compute.cpp index 368c75a66..898370739 100644 --- a/src/video_core/engines/kepler_compute.cpp +++ b/src/video_core/engines/kepler_compute.cpp @@ -16,28 +16,28 @@ namespace Tegra::Engines { -KeplerCompute::KeplerCompute(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - MemoryManager& memory_manager) - : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager}, upload_state{ - memory_manager, - regs.upload} {} +KeplerCompute::KeplerCompute(Core::System& system_, MemoryManager& memory_manager_) + : system{system_}, memory_manager{memory_manager_}, upload_state{memory_manager, regs.upload} {} KeplerCompute::~KeplerCompute() = default; -void KeplerCompute::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +void KeplerCompute::BindRasterizer(VideoCore::RasterizerInterface& rasterizer_) { + rasterizer = &rasterizer_; +} + +void KeplerCompute::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid KeplerCompute register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; - switch (method_call.method) { + switch (method) { case KEPLER_COMPUTE_REG_INDEX(exec_upload): { upload_state.ProcessExec(regs.exec_upload.linear != 0); break; } case KEPLER_COMPUTE_REG_INDEX(data_upload): { - const bool is_last_call = method_call.IsLastCall(); - upload_state.ProcessData(method_call.argument, is_last_call); + upload_state.ProcessData(method_argument, is_last_call); if (is_last_call) { system.GPU().Maxwell3D().OnMemoryWrite(); } @@ -51,6 +51,13 @@ void KeplerCompute::CallMethod(const GPU::MethodCall& method_call) { } } +void KeplerCompute::CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) { + for (std::size_t i = 0; i < amount; i++) { + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); + } +} + Texture::FullTextureInfo KeplerCompute::GetTexture(std::size_t offset) const { const std::bitset<8> cbuf_mask = launch_description.const_buffer_enable_mask.Value(); ASSERT(cbuf_mask[regs.tex_cb_index]); @@ -86,8 +93,11 @@ SamplerDescriptor KeplerCompute::AccessBindlessSampler(ShaderType stage, u64 con ASSERT(stage == ShaderType::Compute); const auto& tex_info_buffer = launch_description.const_buffer_config[const_buffer]; const GPUVAddr tex_info_address = tex_info_buffer.Address() + offset; + return AccessSampler(memory_manager.Read<u32>(tex_info_address)); +} - const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(tex_info_address)}; +SamplerDescriptor KeplerCompute::AccessSampler(u32 handle) const { + const Texture::TextureHandle tex_handle{handle}; const Texture::FullTextureInfo tex_info = GetTextureInfo(tex_handle); SamplerDescriptor result = SamplerDescriptor::FromTIC(tex_info.tic); result.is_shadow.Assign(tex_info.tsc.depth_compare_enabled.Value()); @@ -95,11 +105,11 @@ SamplerDescriptor KeplerCompute::AccessBindlessSampler(ShaderType stage, u64 con } VideoCore::GuestDriverProfile& KeplerCompute::AccessGuestDriverProfile() { - return rasterizer.AccessGuestDriverProfile(); + return rasterizer->AccessGuestDriverProfile(); } const VideoCore::GuestDriverProfile& KeplerCompute::AccessGuestDriverProfile() const { - return rasterizer.AccessGuestDriverProfile(); + return rasterizer->AccessGuestDriverProfile(); } void KeplerCompute::ProcessLaunch() { @@ -110,7 +120,7 @@ void KeplerCompute::ProcessLaunch() { const GPUVAddr code_addr = regs.code_loc.Address() + launch_description.program_start; LOG_TRACE(HW_GPU, "Compute invocation launched at address 0x{:016x}", code_addr); - rasterizer.DispatchCompute(code_addr); + rasterizer->DispatchCompute(code_addr); } Texture::TICEntry KeplerCompute::GetTICEntry(u32 tic_index) const { diff --git a/src/video_core/engines/kepler_compute.h b/src/video_core/engines/kepler_compute.h index eeb79c56f..7f2500aab 100644 --- a/src/video_core/engines/kepler_compute.h +++ b/src/video_core/engines/kepler_compute.h @@ -11,6 +11,7 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/engines/const_buffer_engine_interface.h" +#include "video_core/engines/engine_interface.h" #include "video_core/engines/engine_upload.h" #include "video_core/engines/shader_type.h" #include "video_core/gpu.h" @@ -39,12 +40,14 @@ namespace Tegra::Engines { #define KEPLER_COMPUTE_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::KeplerCompute::Regs, field_name) / sizeof(u32)) -class KeplerCompute final : public ConstBufferEngineInterface { +class KeplerCompute final : public ConstBufferEngineInterface, public EngineInterface { public: - explicit KeplerCompute(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - MemoryManager& memory_manager); + explicit KeplerCompute(Core::System& system, MemoryManager& memory_manager); ~KeplerCompute(); + /// Binds a rasterizer to this engine. + void BindRasterizer(VideoCore::RasterizerInterface& rasterizer); + static constexpr std::size_t NumConstBuffers = 8; struct Regs { @@ -200,7 +203,11 @@ public: "KeplerCompute LaunchParams has wrong size"); /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; + + /// Write multiple values to the register identified by method. + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; Texture::FullTextureInfo GetTexture(std::size_t offset) const; @@ -214,6 +221,8 @@ public: SamplerDescriptor AccessBindlessSampler(ShaderType stage, u64 const_buffer, u64 offset) const override; + SamplerDescriptor AccessSampler(u32 handle) const override; + u32 GetBoundBuffer() const override { return regs.tex_cb_index; } @@ -223,11 +232,6 @@ public: const VideoCore::GuestDriverProfile& AccessGuestDriverProfile() const override; private: - Core::System& system; - VideoCore::RasterizerInterface& rasterizer; - MemoryManager& memory_manager; - Upload::State upload_state; - void ProcessLaunch(); /// Retrieves information about a specific TIC entry from the TIC buffer. @@ -235,6 +239,11 @@ private: /// Retrieves information about a specific TSC entry from the TSC buffer. Texture::TSCEntry GetTSCEntry(u32 tsc_index) const; + + Core::System& system; + MemoryManager& memory_manager; + VideoCore::RasterizerInterface* rasterizer = nullptr; + Upload::State upload_state; }; #define ASSERT_REG_POSITION(field_name, position) \ diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp index 597872e43..dc71b2eec 100644 --- a/src/video_core/engines/kepler_memory.cpp +++ b/src/video_core/engines/kepler_memory.cpp @@ -19,20 +19,19 @@ KeplerMemory::KeplerMemory(Core::System& system, MemoryManager& memory_manager) KeplerMemory::~KeplerMemory() = default; -void KeplerMemory::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, +void KeplerMemory::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < Regs::NUM_REGS, "Invalid KeplerMemory register, increase the size of the Regs structure"); - regs.reg_array[method_call.method] = method_call.argument; + regs.reg_array[method] = method_argument; - switch (method_call.method) { + switch (method) { case KEPLERMEMORY_REG_INDEX(exec): { upload_state.ProcessExec(regs.exec.linear != 0); break; } case KEPLERMEMORY_REG_INDEX(data): { - const bool is_last_call = method_call.IsLastCall(); - upload_state.ProcessData(method_call.argument, is_last_call); + upload_state.ProcessData(method_argument, is_last_call); if (is_last_call) { system.GPU().Maxwell3D().OnMemoryWrite(); } @@ -41,4 +40,11 @@ void KeplerMemory::CallMethod(const GPU::MethodCall& method_call) { } } +void KeplerMemory::CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) { + for (std::size_t i = 0; i < amount; i++) { + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); + } +} + } // namespace Tegra::Engines diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h index 396fb6e86..5b7f71a00 100644 --- a/src/video_core/engines/kepler_memory.h +++ b/src/video_core/engines/kepler_memory.h @@ -10,6 +10,7 @@ #include "common/bit_field.h" #include "common/common_funcs.h" #include "common/common_types.h" +#include "video_core/engines/engine_interface.h" #include "video_core/engines/engine_upload.h" #include "video_core/gpu.h" @@ -32,13 +33,17 @@ namespace Tegra::Engines { #define KEPLERMEMORY_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::KeplerMemory::Regs, field_name) / sizeof(u32)) -class KeplerMemory final { +class KeplerMemory final : public EngineInterface { public: KeplerMemory(Core::System& system, MemoryManager& memory_manager); ~KeplerMemory(); /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; + + /// Write multiple values to the register identified by method. + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; struct Regs { static constexpr size_t NUM_REGS = 0x7F; diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index ba63b44b4..6287df633 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -22,15 +22,19 @@ using VideoCore::QueryType; /// First register id that is actually a Macro call. constexpr u32 MacroRegistersStart = 0xE00; -Maxwell3D::Maxwell3D(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - MemoryManager& memory_manager) - : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager}, - macro_interpreter{*this}, upload_state{memory_manager, regs.upload} { +Maxwell3D::Maxwell3D(Core::System& system_, MemoryManager& memory_manager_) + : system{system_}, memory_manager{memory_manager_}, macro_engine{GetMacroEngine(*this)}, + upload_state{memory_manager, regs.upload} { dirty.flags.flip(); - InitializeRegisterDefaults(); } +Maxwell3D::~Maxwell3D() = default; + +void Maxwell3D::BindRasterizer(VideoCore::RasterizerInterface& rasterizer_) { + rasterizer = &rasterizer_; +} + void Maxwell3D::InitializeRegisterDefaults() { // Initializes registers to their default values - what games expect them to be at boot. This is // for certain registers that may not be explicitly set by games. @@ -44,6 +48,12 @@ void Maxwell3D::InitializeRegisterDefaults() { viewport.depth_range_near = 0.0f; viewport.depth_range_far = 1.0f; } + for (auto& viewport : regs.viewport_transform) { + viewport.swizzle.x.Assign(Regs::ViewportSwizzle::PositiveX); + viewport.swizzle.y.Assign(Regs::ViewportSwizzle::PositiveY); + viewport.swizzle.z.Assign(Regs::ViewportSwizzle::PositiveZ); + viewport.swizzle.w.Assign(Regs::ViewportSwizzle::PositiveW); + } // Doom and Bomberman seems to use the uninitialized registers and just enable blend // so initialize blend registers with sane values @@ -92,11 +102,19 @@ void Maxwell3D::InitializeRegisterDefaults() { color_mask.A.Assign(1); } + for (auto& format : regs.vertex_attrib_format) { + format.constant.Assign(1); + } + // NVN games expect these values to be enabled at boot regs.rasterize_enable = 1; regs.rt_separate_frag_data = 1; regs.framebuffer_srgb = 1; + regs.line_width_aliased = 1.0f; + regs.line_width_smooth = 1.0f; regs.front_face = Maxwell3D::Regs::FrontFace::ClockWise; + regs.polygon_mode_back = Maxwell3D::Regs::PolygonMode::Fill; + regs.polygon_mode_front = Maxwell3D::Regs::PolygonMode::Fill; shadow_state = regs; @@ -106,7 +124,113 @@ void Maxwell3D::InitializeRegisterDefaults() { mme_inline[MAXWELL3D_REG_INDEX(index_array.count)] = true; } -void Maxwell3D::CallMacroMethod(u32 method, std::size_t num_parameters, const u32* parameters) { +void Maxwell3D::ProcessMacro(u32 method, const u32* base_start, u32 amount, bool is_last_call) { + if (executing_macro == 0) { + // A macro call must begin by writing the macro method's register, not its argument. + ASSERT_MSG((method % 2) == 0, + "Can't start macro execution by writing to the ARGS register"); + executing_macro = method; + } + + macro_params.insert(macro_params.end(), base_start, base_start + amount); + + // Call the macro when there are no more parameters in the command buffer + if (is_last_call) { + CallMacroMethod(executing_macro, macro_params); + macro_params.clear(); + } +} + +u32 Maxwell3D::ProcessShadowRam(u32 method, u32 argument) { + // Keep track of the register value in shadow_state when requested. + const auto control = shadow_state.shadow_ram_control; + if (control == Regs::ShadowRamControl::Track || + control == Regs::ShadowRamControl::TrackWithFilter) { + shadow_state.reg_array[method] = argument; + return argument; + } + if (control == Regs::ShadowRamControl::Replay) { + return shadow_state.reg_array[method]; + } + return argument; +} + +void Maxwell3D::ProcessDirtyRegisters(u32 method, u32 argument) { + if (regs.reg_array[method] == argument) { + return; + } + regs.reg_array[method] = argument; + + for (const auto& table : dirty.tables) { + dirty.flags[table[method]] = true; + } +} + +void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argument, + bool is_last_call) { + switch (method) { + case MAXWELL3D_REG_INDEX(wait_for_idle): + return rasterizer->WaitForIdle(); + case MAXWELL3D_REG_INDEX(shadow_ram_control): + shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(nonshadow_argument); + return; + case MAXWELL3D_REG_INDEX(macros.data): + return macro_engine->AddCode(regs.macros.upload_address, argument); + case MAXWELL3D_REG_INDEX(macros.bind): + return ProcessMacroBind(argument); + case MAXWELL3D_REG_INDEX(firmware[4]): + return ProcessFirmwareCall4(); + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[3]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[4]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[5]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[6]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[7]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[8]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[9]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[10]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[11]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[12]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[13]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[14]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): + return StartCBData(method); + case MAXWELL3D_REG_INDEX(cb_bind[0]): + return ProcessCBBind(0); + case MAXWELL3D_REG_INDEX(cb_bind[1]): + return ProcessCBBind(1); + case MAXWELL3D_REG_INDEX(cb_bind[2]): + return ProcessCBBind(2); + case MAXWELL3D_REG_INDEX(cb_bind[3]): + return ProcessCBBind(3); + case MAXWELL3D_REG_INDEX(cb_bind[4]): + return ProcessCBBind(4); + case MAXWELL3D_REG_INDEX(draw.vertex_end_gl): + return DrawArrays(); + case MAXWELL3D_REG_INDEX(clear_buffers): + return ProcessClearBuffers(); + case MAXWELL3D_REG_INDEX(query.query_get): + return ProcessQueryGet(); + case MAXWELL3D_REG_INDEX(condition.mode): + return ProcessQueryCondition(); + case MAXWELL3D_REG_INDEX(counter_reset): + return ProcessCounterReset(); + case MAXWELL3D_REG_INDEX(sync_info): + return ProcessSyncPoint(); + case MAXWELL3D_REG_INDEX(exec_upload): + return upload_state.ProcessExec(regs.exec_upload.linear != 0); + case MAXWELL3D_REG_INDEX(data_upload): + upload_state.ProcessData(argument, is_last_call); + if (is_last_call) { + OnMemoryWrite(); + } + return; + } +} + +void Maxwell3D::CallMacroMethod(u32 method, const std::vector<u32>& parameters) { // Reset the current macro. executing_macro = 0; @@ -115,18 +239,16 @@ void Maxwell3D::CallMacroMethod(u32 method, std::size_t num_parameters, const u3 ((method - MacroRegistersStart) >> 1) % static_cast<u32>(macro_positions.size()); // Execute the current macro. - macro_interpreter.Execute(macro_positions[entry], num_parameters, parameters); + macro_engine->Execute(*this, macro_positions[entry], parameters); if (mme_draw.current_mode != MMEDrawMode::Undefined) { FlushMMEInlineDraw(); } } -void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { - const u32 method = method_call.method; - +void Maxwell3D::CallMethod(u32 method, u32 method_argument, bool is_last_call) { if (method == cb_data_state.current) { - regs.reg_array[method] = method_call.argument; - ProcessCBData(method_call.argument); + regs.reg_array[method] = method_argument; + ProcessCBData(method_argument); return; } else if (cb_data_state.current != null_cb_data) { FinishCBData(); @@ -141,61 +263,27 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { // Methods after 0xE00 are special, they're actually triggers for some microcode that was // uploaded to the GPU during initialization. if (method >= MacroRegistersStart) { - // We're trying to execute a macro - if (executing_macro == 0) { - // A macro call must begin by writing the macro method's register, not its argument. - ASSERT_MSG((method % 2) == 0, - "Can't start macro execution by writing to the ARGS register"); - executing_macro = method; - } - - macro_params.push_back(method_call.argument); - - // Call the macro when there are no more parameters in the command buffer - if (method_call.IsLastCall()) { - CallMacroMethod(executing_macro, macro_params.size(), macro_params.data()); - macro_params.clear(); - } + ProcessMacro(method, &method_argument, 1, is_last_call); return; } ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Maxwell3D register, increase the size of the Regs structure"); - u32 arg = method_call.argument; - // Keep track of the register value in shadow_state when requested. - if (shadow_state.shadow_ram_control == Regs::ShadowRamControl::Track || - shadow_state.shadow_ram_control == Regs::ShadowRamControl::TrackWithFilter) { - shadow_state.reg_array[method] = arg; - } else if (shadow_state.shadow_ram_control == Regs::ShadowRamControl::Replay) { - arg = shadow_state.reg_array[method]; - } - - if (regs.reg_array[method] != arg) { - regs.reg_array[method] = arg; + const u32 argument = ProcessShadowRam(method, method_argument); + ProcessDirtyRegisters(method, argument); + ProcessMethodCall(method, argument, method_argument, is_last_call); +} - for (const auto& table : dirty.tables) { - dirty.flags[table[method]] = true; - } +void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) { + // Methods after 0xE00 are special, they're actually triggers for some microcode that was + // uploaded to the GPU during initialization. + if (method >= MacroRegistersStart) { + ProcessMacro(method, base_start, amount, amount == methods_pending); + return; } - switch (method) { - case MAXWELL3D_REG_INDEX(shadow_ram_control): { - shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(method_call.argument); - break; - } - case MAXWELL3D_REG_INDEX(macros.data): { - ProcessMacroUpload(arg); - break; - } - case MAXWELL3D_REG_INDEX(macros.bind): { - ProcessMacroBind(arg); - break; - } - case MAXWELL3D_REG_INDEX(firmware[4]): { - ProcessFirmwareCall4(); - break; - } case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]): case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]): case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]): @@ -211,67 +299,13 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { case MAXWELL3D_REG_INDEX(const_buffer.cb_data[12]): case MAXWELL3D_REG_INDEX(const_buffer.cb_data[13]): case MAXWELL3D_REG_INDEX(const_buffer.cb_data[14]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): { - StartCBData(method); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[0]): { - ProcessCBBind(0); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[1]): { - ProcessCBBind(1); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[2]): { - ProcessCBBind(2); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[3]): { - ProcessCBBind(3); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[4]): { - ProcessCBBind(4); - break; - } - case MAXWELL3D_REG_INDEX(draw.vertex_end_gl): { - DrawArrays(); - break; - } - case MAXWELL3D_REG_INDEX(clear_buffers): { - ProcessClearBuffers(); - break; - } - case MAXWELL3D_REG_INDEX(query.query_get): { - ProcessQueryGet(); - break; - } - case MAXWELL3D_REG_INDEX(condition.mode): { - ProcessQueryCondition(); - break; - } - case MAXWELL3D_REG_INDEX(counter_reset): { - ProcessCounterReset(); - break; - } - case MAXWELL3D_REG_INDEX(sync_info): { - ProcessSyncPoint(); - break; - } - case MAXWELL3D_REG_INDEX(exec_upload): { - upload_state.ProcessExec(regs.exec_upload.linear != 0); - break; - } - case MAXWELL3D_REG_INDEX(data_upload): { - const bool is_last_call = method_call.IsLastCall(); - upload_state.ProcessData(arg, is_last_call); - if (is_last_call) { - OnMemoryWrite(); - } + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): + ProcessCBMultiData(method, base_start, amount); break; - } default: + for (std::size_t i = 0; i < amount; i++) { + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); + } break; } } @@ -300,16 +334,15 @@ void Maxwell3D::StepInstance(const MMEDrawMode expected_mode, const u32 count) { StepInstance(expected_mode, count); } -void Maxwell3D::CallMethodFromMME(const GPU::MethodCall& method_call) { - const u32 method = method_call.method; +void Maxwell3D::CallMethodFromMME(u32 method, u32 method_argument) { if (mme_inline[method]) { - regs.reg_array[method] = method_call.argument; + regs.reg_array[method] = method_argument; if (method == MAXWELL3D_REG_INDEX(vertex_buffer.count) || method == MAXWELL3D_REG_INDEX(index_array.count)) { const MMEDrawMode expected_mode = method == MAXWELL3D_REG_INDEX(vertex_buffer.count) ? MMEDrawMode::Array : MMEDrawMode::Indexed; - StepInstance(expected_mode, method_call.argument); + StepInstance(expected_mode, method_argument); } else if (method == MAXWELL3D_REG_INDEX(draw.vertex_begin_gl)) { mme_draw.instance_mode = (regs.draw.instance_next != 0) || (regs.draw.instance_cont != 0); @@ -321,7 +354,7 @@ void Maxwell3D::CallMethodFromMME(const GPU::MethodCall& method_call) { if (mme_draw.current_mode != MMEDrawMode::Undefined) { FlushMMEInlineDraw(); } - CallMethod(method_call); + CallMethod(method, method_argument, true); } } @@ -337,7 +370,7 @@ void Maxwell3D::FlushMMEInlineDraw() { const bool is_indexed = mme_draw.current_mode == MMEDrawMode::Indexed; if (ShouldExecute()) { - rasterizer.Draw(is_indexed, true); + rasterizer->Draw(is_indexed, true); } // TODO(bunnei): Below, we reset vertex count so that we can use these registers to determine if @@ -358,9 +391,7 @@ void Maxwell3D::FlushMMEInlineDraw() { } void Maxwell3D::ProcessMacroUpload(u32 data) { - ASSERT_MSG(regs.macros.upload_address < macro_memory.size(), - "upload_address exceeded macro_memory size!"); - macro_memory[regs.macros.upload_address++] = data; + macro_engine->AddCode(regs.macros.upload_address++, data); } void Maxwell3D::ProcessMacroBind(u32 data) { @@ -395,12 +426,17 @@ void Maxwell3D::StampQueryResult(u64 payload, bool long_query) { void Maxwell3D::ProcessQueryGet() { // TODO(Subv): Support the other query units. - ASSERT_MSG(regs.query.query_get.unit == Regs::QueryUnit::Crop, - "Units other than CROP are unimplemented"); + if (regs.query.query_get.unit != Regs::QueryUnit::Crop) { + LOG_DEBUG(HW_GPU, "Units other than CROP are unimplemented"); + } switch (regs.query.query_get.operation) { case Regs::QueryOperation::Release: - StampQueryResult(regs.query.query_sequence, regs.query.query_get.short_query == 0); + if (regs.query.query_get.fence == 1) { + rasterizer->SignalSemaphore(regs.query.QueryAddress(), regs.query.query_sequence); + } else { + StampQueryResult(regs.query.query_sequence, regs.query.query_get.short_query == 0); + } break; case Regs::QueryOperation::Acquire: // TODO(Blinkhawk): Under this operation, the GPU waits for the CPU to write a value that @@ -465,11 +501,11 @@ void Maxwell3D::ProcessQueryCondition() { void Maxwell3D::ProcessCounterReset() { switch (regs.counter_reset) { case Regs::CounterReset::SampleCnt: - rasterizer.ResetCounter(QueryType::SamplesPassed); + rasterizer->ResetCounter(QueryType::SamplesPassed); break; default: - LOG_WARNING(Render_OpenGL, "Unimplemented counter reset={}", - static_cast<int>(regs.counter_reset)); + LOG_DEBUG(Render_OpenGL, "Unimplemented counter reset={}", + static_cast<int>(regs.counter_reset)); break; } } @@ -479,7 +515,7 @@ void Maxwell3D::ProcessSyncPoint() { const u32 increment = regs.sync_info.increment.Value(); [[maybe_unused]] const u32 cache_flush = regs.sync_info.unknown.Value(); if (increment) { - system.GPU().IncrementSyncPoint(sync_point); + rasterizer->SignalSyncPoint(sync_point); } } @@ -502,7 +538,7 @@ void Maxwell3D::DrawArrays() { const bool is_indexed{regs.index_array.count && !regs.vertex_buffer.count}; if (ShouldExecute()) { - rasterizer.Draw(is_indexed, false); + rasterizer->Draw(is_indexed, false); } // TODO(bunnei): Below, we reset vertex count so that we can use these registers to determine if @@ -522,12 +558,12 @@ std::optional<u64> Maxwell3D::GetQueryResult() { return 0; case Regs::QuerySelect::SamplesPassed: // Deferred. - rasterizer.Query(regs.query.QueryAddress(), VideoCore::QueryType::SamplesPassed, - system.GPU().GetTicks()); - return {}; + rasterizer->Query(regs.query.QueryAddress(), VideoCore::QueryType::SamplesPassed, + system.GPU().GetTicks()); + return std::nullopt; default: - UNIMPLEMENTED_MSG("Unimplemented query select type {}", - static_cast<u32>(regs.query.query_get.select.Value())); + LOG_DEBUG(HW_GPU, "Unimplemented query select type {}", + static_cast<u32>(regs.query.query_get.select.Value())); return 1; } } @@ -562,6 +598,28 @@ void Maxwell3D::StartCBData(u32 method) { ProcessCBData(regs.const_buffer.cb_data[cb_data_state.id]); } +void Maxwell3D::ProcessCBMultiData(u32 method, const u32* start_base, u32 amount) { + if (cb_data_state.current != method) { + if (cb_data_state.current != null_cb_data) { + FinishCBData(); + } + constexpr u32 first_cb_data = MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]); + cb_data_state.start_pos = regs.const_buffer.cb_pos; + cb_data_state.id = method - first_cb_data; + cb_data_state.current = method; + cb_data_state.counter = 0; + } + const std::size_t id = cb_data_state.id; + const std::size_t size = amount; + std::size_t i = 0; + for (; i < size; i++) { + cb_data_state.buffer[id][cb_data_state.counter] = start_base[i]; + cb_data_state.counter++; + } + // Increment the current buffer position. + regs.const_buffer.cb_pos = regs.const_buffer.cb_pos + 4 * amount; +} + void Maxwell3D::FinishCBData() { // Write the input value to the current const buffer at the current position. const GPUVAddr buffer_address = regs.const_buffer.BufferAddress(); @@ -628,7 +686,7 @@ void Maxwell3D::ProcessClearBuffers() { regs.clear_buffers.R == regs.clear_buffers.B && regs.clear_buffers.R == regs.clear_buffers.A); - rasterizer.Clear(); + rasterizer->Clear(); } u32 Maxwell3D::AccessConstBuffer32(ShaderType stage, u64 const_buffer, u64 offset) const { @@ -650,8 +708,11 @@ SamplerDescriptor Maxwell3D::AccessBindlessSampler(ShaderType stage, u64 const_b const auto& shader = state.shader_stages[static_cast<std::size_t>(stage)]; const auto& tex_info_buffer = shader.const_buffers[const_buffer]; const GPUVAddr tex_info_address = tex_info_buffer.address + offset; + return AccessSampler(memory_manager.Read<u32>(tex_info_address)); +} - const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(tex_info_address)}; +SamplerDescriptor Maxwell3D::AccessSampler(u32 handle) const { + const Texture::TextureHandle tex_handle{handle}; const Texture::FullTextureInfo tex_info = GetTextureInfo(tex_handle); SamplerDescriptor result = SamplerDescriptor::FromTIC(tex_info.tic); result.is_shadow.Assign(tex_info.tsc.depth_compare_enabled.Value()); @@ -659,11 +720,11 @@ SamplerDescriptor Maxwell3D::AccessBindlessSampler(ShaderType stage, u64 const_b } VideoCore::GuestDriverProfile& Maxwell3D::AccessGuestDriverProfile() { - return rasterizer.AccessGuestDriverProfile(); + return rasterizer->AccessGuestDriverProfile(); } const VideoCore::GuestDriverProfile& Maxwell3D::AccessGuestDriverProfile() const { - return rasterizer.AccessGuestDriverProfile(); + return rasterizer->AccessGuestDriverProfile(); } } // namespace Tegra::Engines diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 5cf6a4cc3..1cbe8fe67 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -19,10 +19,11 @@ #include "common/math_util.h" #include "video_core/engines/const_buffer_engine_interface.h" #include "video_core/engines/const_buffer_info.h" +#include "video_core/engines/engine_interface.h" #include "video_core/engines/engine_upload.h" #include "video_core/engines/shader_type.h" #include "video_core/gpu.h" -#include "video_core/macro_interpreter.h" +#include "video_core/macro/macro.h" #include "video_core/textures/texture.h" namespace Core { @@ -48,11 +49,13 @@ namespace Tegra::Engines { #define MAXWELL3D_REG_INDEX(field_name) \ (offsetof(Tegra::Engines::Maxwell3D::Regs, field_name) / sizeof(u32)) -class Maxwell3D final : public ConstBufferEngineInterface { +class Maxwell3D final : public ConstBufferEngineInterface, public EngineInterface { public: - explicit Maxwell3D(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - MemoryManager& memory_manager); - ~Maxwell3D() = default; + explicit Maxwell3D(Core::System& system, MemoryManager& memory_manager); + ~Maxwell3D(); + + /// Binds a rasterizer to this engine. + void BindRasterizer(VideoCore::RasterizerInterface& rasterizer); /// Register structure of the Maxwell3D engine. /// TODO(Subv): This structure will need to be made bigger as more registers are discovered. @@ -575,6 +578,17 @@ public: Replay = 3, }; + enum class ViewportSwizzle : u32 { + PositiveX = 0, + NegativeX = 1, + PositiveY = 2, + NegativeY = 3, + PositiveZ = 4, + NegativeZ = 5, + PositiveW = 6, + NegativeW = 7, + }; + struct RenderTargetConfig { u32 address_high; u32 address_low; @@ -586,6 +600,7 @@ public: BitField<4, 3, u32> block_height; BitField<8, 3, u32> block_depth; BitField<12, 1, InvMemoryLayout> type; + BitField<16, 1, u32> is_3d; } memory_layout; union { BitField<0, 16, u32> layers; @@ -618,7 +633,14 @@ public: f32 translate_x; f32 translate_y; f32 translate_z; - INSERT_UNION_PADDING_WORDS(2); + union { + u32 raw; + BitField<0, 3, ViewportSwizzle> x; + BitField<4, 3, ViewportSwizzle> y; + BitField<8, 3, ViewportSwizzle> z; + BitField<12, 3, ViewportSwizzle> w; + } swizzle; + INSERT_UNION_PADDING_WORDS(1); Common::Rectangle<f32> GetRect() const { return { @@ -627,7 +649,7 @@ public: GetX() + GetWidth(), // right GetY() // bottom }; - }; + } f32 GetX() const { return std::max(0.0f, translate_x - std::fabs(scale_x)); @@ -709,7 +731,9 @@ public: union { struct { - INSERT_UNION_PADDING_WORDS(0x45); + INSERT_UNION_PADDING_WORDS(0x44); + + u32 wait_for_idle; struct { u32 upload_address; @@ -1149,7 +1173,7 @@ public: /// Returns whether the vertex array specified by index is supposed to be /// accessed per instance or not. - bool IsInstancingEnabled(u32 index) const { + bool IsInstancingEnabled(std::size_t index) const { return is_instanced[index]; } } instanced_arrays; @@ -1179,6 +1203,7 @@ public: BitField<0, 1, u32> depth_range_0_1; BitField<3, 1, u32> depth_clamp_near; BitField<4, 1, u32> depth_clamp_far; + BitField<11, 1, u32> depth_clamp_disabled; } view_volume_clip_control; INSERT_UNION_PADDING_WORDS(0x1F); @@ -1259,7 +1284,8 @@ public: GPUVAddr LimitAddress() const { return static_cast<GPUVAddr>((static_cast<GPUVAddr>(limit_high) << 32) | - limit_low); + limit_low) + + 1; } } vertex_array_limit[NumVertexArrays]; @@ -1356,10 +1382,14 @@ public: u32 GetRegisterValue(u32 method) const; /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; + + /// Write multiple values to the register identified by method. + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; /// Write the value to the register identified by method. - void CallMethodFromMME(const GPU::MethodCall& method_call); + void CallMethodFromMME(u32 method, u32 method_argument); void FlushMMEInlineDraw(); @@ -1376,6 +1406,8 @@ public: SamplerDescriptor AccessBindlessSampler(ShaderType stage, u64 const_buffer, u64 offset) const override; + SamplerDescriptor AccessSampler(u32 handle) const override; + u32 GetBoundBuffer() const override { return regs.tex_cb_index; } @@ -1384,17 +1416,16 @@ public: const VideoCore::GuestDriverProfile& AccessGuestDriverProfile() const override; - /// Memory for macro code - it's undetermined how big this is, however 1MB is much larger than - /// we've seen used. - using MacroMemory = std::array<u32, 0x40000>; + bool ShouldExecute() const { + return execute_on; + } - /// Gets a reference to macro memory. - const MacroMemory& GetMacroMemory() const { - return macro_memory; + VideoCore::RasterizerInterface& Rasterizer() { + return *rasterizer; } - bool ShouldExecute() const { - return execute_on; + const VideoCore::RasterizerInterface& Rasterizer() const { + return *rasterizer; } /// Notify a memory write has happened. @@ -1430,27 +1461,31 @@ public: private: void InitializeRegisterDefaults(); - Core::System& system; + void ProcessMacro(u32 method, const u32* base_start, u32 amount, bool is_last_call); - VideoCore::RasterizerInterface& rasterizer; + u32 ProcessShadowRam(u32 method, u32 argument); + void ProcessDirtyRegisters(u32 method, u32 argument); + + void ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argument, bool is_last_call); + + Core::System& system; MemoryManager& memory_manager; + VideoCore::RasterizerInterface* rasterizer = nullptr; + /// Start offsets of each macro in macro_memory std::array<u32, 0x80> macro_positions = {}; std::array<bool, Regs::NUM_REGS> mme_inline{}; - /// Memory for macro code - MacroMemory macro_memory; - /// Macro method that is currently being executed / being fed parameters. u32 executing_macro = 0; /// Parameters that have been submitted to the macro call so far. std::vector<u32> macro_params; /// Interpreter for the macro codes uploaded to the GPU. - MacroInterpreter macro_interpreter; + std::unique_ptr<MacroEngine> macro_engine; static constexpr u32 null_cb_data = 0xFFFFFFFF; struct { @@ -1479,7 +1514,7 @@ private: * @param num_parameters Number of arguments * @param parameters Arguments to the method call */ - void CallMacroMethod(u32 method, std::size_t num_parameters, const u32* parameters); + void CallMacroMethod(u32 method, const std::vector<u32>& parameters); /// Handles writes to the macro uploading register. void ProcessMacroUpload(u32 data); @@ -1511,6 +1546,7 @@ private: /// Handles a write to the CB_DATA[i] register. void StartCBData(u32 method); void ProcessCBData(u32 value); + void ProcessCBMultiData(u32 method, const u32* start_base, u32 amount); void FinishCBData(); /// Handles a write to the CB_BIND register. @@ -1530,6 +1566,7 @@ private: static_assert(offsetof(Maxwell3D::Regs, field_name) == position * 4, \ "Field " #field_name " has invalid position") +ASSERT_REG_POSITION(wait_for_idle, 0x44); ASSERT_REG_POSITION(macros, 0x45); ASSERT_REG_POSITION(shadow_ram_control, 0x49); ASSERT_REG_POSITION(upload, 0x60); diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index c2610f992..8fa359d0a 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -14,43 +14,45 @@ namespace Tegra::Engines { +using namespace Texture; + MaxwellDMA::MaxwellDMA(Core::System& system, MemoryManager& memory_manager) : system{system}, memory_manager{memory_manager} {} -void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) { - ASSERT_MSG(method_call.method < Regs::NUM_REGS, - "Invalid MaxwellDMA register, increase the size of the Regs structure"); - - regs.reg_array[method_call.method] = method_call.argument; +void MaxwellDMA::CallMethod(u32 method, u32 method_argument, bool is_last_call) { + ASSERT_MSG(method < NUM_REGS, "Invalid MaxwellDMA register"); -#define MAXWELLDMA_REG_INDEX(field_name) \ - (offsetof(Tegra::Engines::MaxwellDMA::Regs, field_name) / sizeof(u32)) + regs.reg_array[method] = method_argument; - switch (method_call.method) { - case MAXWELLDMA_REG_INDEX(exec): { - HandleCopy(); - break; - } + if (method == offsetof(Regs, launch_dma) / sizeof(u32)) { + Launch(); } - -#undef MAXWELLDMA_REG_INDEX } -void MaxwellDMA::HandleCopy() { - LOG_TRACE(HW_GPU, "Requested a DMA copy"); +void MaxwellDMA::CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) { + for (size_t i = 0; i < amount; ++i) { + CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); + } +} - const GPUVAddr source = regs.src_address.Address(); - const GPUVAddr dest = regs.dst_address.Address(); +void MaxwellDMA::Launch() { + LOG_TRACE(Render_OpenGL, "DMA copy 0x{:x} -> 0x{:x}", static_cast<GPUVAddr>(regs.offset_in), + static_cast<GPUVAddr>(regs.offset_out)); // TODO(Subv): Perform more research and implement all features of this engine. - ASSERT(regs.exec.enable_swizzle == 0); - ASSERT(regs.exec.query_mode == Regs::QueryMode::None); - ASSERT(regs.exec.query_intr == Regs::QueryIntr::None); - ASSERT(regs.exec.copy_mode == Regs::CopyMode::Unk2); - ASSERT(regs.dst_params.pos_x == 0); - ASSERT(regs.dst_params.pos_y == 0); - - if (!regs.exec.is_dst_linear && !regs.exec.is_src_linear) { + const LaunchDMA& launch = regs.launch_dma; + ASSERT(launch.remap_enable == 0); + ASSERT(launch.semaphore_type == LaunchDMA::SemaphoreType::NONE); + ASSERT(launch.interrupt_type == LaunchDMA::InterruptType::NONE); + ASSERT(launch.data_transfer_type == LaunchDMA::DataTransferType::NON_PIPELINED); + ASSERT(regs.dst_params.origin.x == 0); + ASSERT(regs.dst_params.origin.y == 0); + + const bool is_src_pitch = launch.src_memory_layout == LaunchDMA::MemoryLayout::PITCH; + const bool is_dst_pitch = launch.dst_memory_layout == LaunchDMA::MemoryLayout::PITCH; + + if (!is_src_pitch && !is_dst_pitch) { // If both the source and the destination are in block layout, assert. UNREACHABLE_MSG("Tiled->Tiled DMA transfers are not yet implemented"); return; @@ -59,99 +61,154 @@ void MaxwellDMA::HandleCopy() { // All copies here update the main memory, so mark all rasterizer states as invalid. system.GPU().Maxwell3D().OnMemoryWrite(); - if (regs.exec.is_dst_linear && regs.exec.is_src_linear) { - // When the enable_2d bit is disabled, the copy is performed as if we were copying a 1D - // buffer of length `x_count`, otherwise we copy a 2D image of dimensions (x_count, - // y_count). - if (!regs.exec.enable_2d) { - memory_manager.CopyBlock(dest, source, regs.x_count); - return; - } + if (is_src_pitch && is_dst_pitch) { + CopyPitchToPitch(); + } else { + ASSERT(launch.multi_line_enable == 1); - // If both the source and the destination are in linear layout, perform a line-by-line - // copy. We're going to take a subrect of size (x_count, y_count) from the source - // rectangle. There is no need to manually flush/invalidate the regions because - // CopyBlock does that for us. - for (u32 line = 0; line < regs.y_count; ++line) { - const GPUVAddr source_line = source + line * regs.src_pitch; - const GPUVAddr dest_line = dest + line * regs.dst_pitch; - memory_manager.CopyBlock(dest_line, source_line, regs.x_count); + if (!is_src_pitch && is_dst_pitch) { + CopyBlockLinearToPitch(); + } else { + CopyPitchToBlockLinear(); } + } +} + +void MaxwellDMA::CopyPitchToPitch() { + // When `multi_line_enable` bit is disabled the copy is performed as if we were copying a 1D + // buffer of length `line_length_in`. + // Otherwise we copy a 2D image of dimensions (line_length_in, line_count). + if (!regs.launch_dma.multi_line_enable) { + memory_manager.CopyBlock(regs.offset_out, regs.offset_in, regs.line_length_in); return; } - ASSERT(regs.exec.enable_2d == 1); + // Perform a line-by-line copy. + // We're going to take a subrect of size (line_length_in, line_count) from the source rectangle. + // There is no need to manually flush/invalidate the regions because CopyBlock does that for us. + for (u32 line = 0; line < regs.line_count; ++line) { + const GPUVAddr source_line = regs.offset_in + static_cast<size_t>(line) * regs.pitch_in; + const GPUVAddr dest_line = regs.offset_out + static_cast<size_t>(line) * regs.pitch_out; + memory_manager.CopyBlock(dest_line, source_line, regs.line_length_in); + } +} - if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) { - ASSERT(regs.src_params.BlockDepth() == 0); - // If the input is tiled and the output is linear, deswizzle the input and copy it over. - const u32 bytes_per_pixel = regs.dst_pitch / regs.x_count; - const std::size_t src_size = Texture::CalculateSize( - true, bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y, - regs.src_params.size_z, regs.src_params.BlockHeight(), regs.src_params.BlockDepth()); +void MaxwellDMA::CopyBlockLinearToPitch() { + UNIMPLEMENTED_IF(regs.src_params.block_size.depth != 0); + UNIMPLEMENTED_IF(regs.src_params.layer != 0); - const std::size_t src_layer_size = Texture::CalculateSize( - true, bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y, 1, - regs.src_params.BlockHeight(), regs.src_params.BlockDepth()); + // Optimized path for micro copies. + const size_t dst_size = static_cast<size_t>(regs.pitch_out) * regs.line_count; + if (dst_size < GOB_SIZE && regs.pitch_out <= GOB_SIZE_X) { + FastCopyBlockLinearToPitch(); + return; + } - const std::size_t dst_size = regs.dst_pitch * regs.y_count; + // Deswizzle the input and copy it over. + const u32 bytes_per_pixel = regs.pitch_out / regs.line_length_in; + const Parameters& src_params = regs.src_params; + const u32 width = src_params.width; + const u32 height = src_params.height; + const u32 depth = src_params.depth; + const u32 block_height = src_params.block_size.height; + const u32 block_depth = src_params.block_size.depth; + const size_t src_size = + CalculateSize(true, bytes_per_pixel, width, height, depth, block_height, block_depth); + + if (read_buffer.size() < src_size) { + read_buffer.resize(src_size); + } + if (write_buffer.size() < dst_size) { + write_buffer.resize(dst_size); + } - if (read_buffer.size() < src_size) { - read_buffer.resize(src_size); - } + memory_manager.ReadBlock(regs.offset_in, read_buffer.data(), src_size); + memory_manager.ReadBlock(regs.offset_out, write_buffer.data(), dst_size); - if (write_buffer.size() < dst_size) { - write_buffer.resize(dst_size); - } + UnswizzleSubrect(regs.line_length_in, regs.line_count, regs.pitch_out, width, bytes_per_pixel, + block_height, src_params.origin.x, src_params.origin.y, write_buffer.data(), + read_buffer.data()); - memory_manager.ReadBlock(source, read_buffer.data(), src_size); - memory_manager.ReadBlock(dest, write_buffer.data(), dst_size); + memory_manager.WriteBlock(regs.offset_out, write_buffer.data(), dst_size); +} - Texture::UnswizzleSubrect( - regs.x_count, regs.y_count, regs.dst_pitch, regs.src_params.size_x, bytes_per_pixel, - read_buffer.data() + src_layer_size * regs.src_params.pos_z, write_buffer.data(), - regs.src_params.BlockHeight(), regs.src_params.pos_x, regs.src_params.pos_y); +void MaxwellDMA::CopyPitchToBlockLinear() { + const auto& dst_params = regs.dst_params; + const u32 bytes_per_pixel = regs.pitch_in / regs.line_length_in; + const u32 width = dst_params.width; + const u32 height = dst_params.height; + const u32 depth = dst_params.depth; + const u32 block_height = dst_params.block_size.height; + const u32 block_depth = dst_params.block_size.depth; + const size_t dst_size = + CalculateSize(true, bytes_per_pixel, width, height, depth, block_height, block_depth); + const size_t dst_layer_size = + CalculateSize(true, bytes_per_pixel, width, height, 1, block_height, block_depth); + + const size_t src_size = static_cast<size_t>(regs.pitch_in) * regs.line_count; + + if (read_buffer.size() < src_size) { + read_buffer.resize(src_size); + } + if (write_buffer.size() < dst_size) { + write_buffer.resize(dst_size); + } - memory_manager.WriteBlock(dest, write_buffer.data(), dst_size); + if (Settings::IsGPULevelExtreme()) { + memory_manager.ReadBlock(regs.offset_in, read_buffer.data(), src_size); + memory_manager.ReadBlock(regs.offset_out, write_buffer.data(), dst_size); } else { - ASSERT(regs.dst_params.BlockDepth() == 0); - - const u32 bytes_per_pixel = regs.src_pitch / regs.x_count; - - const std::size_t dst_size = Texture::CalculateSize( - true, bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y, - regs.dst_params.size_z, regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth()); - - const std::size_t dst_layer_size = Texture::CalculateSize( - true, bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y, 1, - regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth()); + memory_manager.ReadBlockUnsafe(regs.offset_in, read_buffer.data(), src_size); + memory_manager.ReadBlockUnsafe(regs.offset_out, write_buffer.data(), dst_size); + } - const std::size_t src_size = regs.src_pitch * regs.y_count; + // If the input is linear and the output is tiled, swizzle the input and copy it over. + if (regs.dst_params.block_size.depth > 0) { + ASSERT(dst_params.layer == 0); + SwizzleSliceToVoxel(regs.line_length_in, regs.line_count, regs.pitch_in, width, height, + bytes_per_pixel, block_height, block_depth, dst_params.origin.x, + dst_params.origin.y, write_buffer.data(), read_buffer.data()); + } else { + SwizzleSubrect(regs.line_length_in, regs.line_count, regs.pitch_in, width, bytes_per_pixel, + write_buffer.data() + dst_layer_size * dst_params.layer, read_buffer.data(), + block_height, dst_params.origin.x, dst_params.origin.y); + } - if (read_buffer.size() < src_size) { - read_buffer.resize(src_size); - } + memory_manager.WriteBlock(regs.offset_out, write_buffer.data(), dst_size); +} - if (write_buffer.size() < dst_size) { - write_buffer.resize(dst_size); - } +void MaxwellDMA::FastCopyBlockLinearToPitch() { + const u32 bytes_per_pixel = regs.pitch_out / regs.line_length_in; + const size_t src_size = GOB_SIZE; + const size_t dst_size = static_cast<size_t>(regs.pitch_out) * regs.line_count; + u32 pos_x = regs.src_params.origin.x; + u32 pos_y = regs.src_params.origin.y; + const u64 offset = GetGOBOffset(regs.src_params.width, regs.src_params.height, pos_x, pos_y, + regs.src_params.block_size.height, bytes_per_pixel); + const u32 x_in_gob = 64 / bytes_per_pixel; + pos_x = pos_x % x_in_gob; + pos_y = pos_y % 8; + + if (read_buffer.size() < src_size) { + read_buffer.resize(src_size); + } + if (write_buffer.size() < dst_size) { + write_buffer.resize(dst_size); + } - if (Settings::values.use_accurate_gpu_emulation) { - memory_manager.ReadBlock(source, read_buffer.data(), src_size); - memory_manager.ReadBlock(dest, write_buffer.data(), dst_size); - } else { - memory_manager.ReadBlockUnsafe(source, read_buffer.data(), src_size); - memory_manager.ReadBlockUnsafe(dest, write_buffer.data(), dst_size); - } + if (Settings::IsGPULevelExtreme()) { + memory_manager.ReadBlock(regs.offset_in + offset, read_buffer.data(), src_size); + memory_manager.ReadBlock(regs.offset_out, write_buffer.data(), dst_size); + } else { + memory_manager.ReadBlockUnsafe(regs.offset_in + offset, read_buffer.data(), src_size); + memory_manager.ReadBlockUnsafe(regs.offset_out, write_buffer.data(), dst_size); + } - // If the input is linear and the output is tiled, swizzle the input and copy it over. - Texture::SwizzleSubrect( - regs.x_count, regs.y_count, regs.src_pitch, regs.dst_params.size_x, bytes_per_pixel, - write_buffer.data() + dst_layer_size * regs.dst_params.pos_z, read_buffer.data(), - regs.dst_params.BlockHeight(), regs.dst_params.pos_x, regs.dst_params.pos_y); + UnswizzleSubrect(regs.line_length_in, regs.line_count, regs.pitch_out, regs.src_params.width, + bytes_per_pixel, regs.src_params.block_size.height, pos_x, pos_y, + write_buffer.data(), read_buffer.data()); - memory_manager.WriteBlock(dest, write_buffer.data(), dst_size); - } + memory_manager.WriteBlock(regs.offset_out, write_buffer.data(), dst_size); } } // namespace Tegra::Engines diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h index 4f40d1d1f..50f445efc 100644 --- a/src/video_core/engines/maxwell_dma.h +++ b/src/video_core/engines/maxwell_dma.h @@ -10,6 +10,7 @@ #include "common/bit_field.h" #include "common/common_funcs.h" #include "common/common_types.h" +#include "video_core/engines/engine_interface.h" #include "video_core/gpu.h" namespace Core { @@ -23,156 +24,190 @@ class MemoryManager; namespace Tegra::Engines { /** - * This Engine is known as GK104_Copy. Documentation can be found in: + * This engine is known as gk104_copy. Documentation can be found in: + * https://github.com/NVIDIA/open-gpu-doc/blob/master/classes/dma-copy/clb0b5.h * https://github.com/envytools/envytools/blob/master/rnndb/fifo/gk104_copy.xml */ -class MaxwellDMA final { +class MaxwellDMA final : public EngineInterface { public: - explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager); - ~MaxwellDMA() = default; - - /// Write the value to the register identified by method. - void CallMethod(const GPU::MethodCall& method_call); + struct PackedGPUVAddr { + u32 upper; + u32 lower; + + constexpr operator GPUVAddr() const noexcept { + return (static_cast<GPUVAddr>(upper & 0xff) << 32) | lower; + } + }; + + union BlockSize { + BitField<0, 4, u32> width; + BitField<4, 4, u32> height; + BitField<8, 4, u32> depth; + BitField<12, 4, u32> gob_height; + }; + static_assert(sizeof(BlockSize) == 4); + + union Origin { + BitField<0, 16, u32> x; + BitField<16, 16, u32> y; + }; + static_assert(sizeof(Origin) == 4); + + struct Parameters { + BlockSize block_size; + u32 width; + u32 height; + u32 depth; + u32 layer; + Origin origin; + }; + static_assert(sizeof(Parameters) == 24); + + struct Semaphore { + PackedGPUVAddr address; + u32 payload; + }; + static_assert(sizeof(Semaphore) == 12); + + struct RenderEnable { + enum class Mode : u32 { + FALSE = 0, + TRUE = 1, + CONDITIONAL = 2, + RENDER_IF_EQUAL = 3, + RENDER_IF_NOT_EQUAL = 4, + }; - struct Regs { - static constexpr std::size_t NUM_REGS = 0x1D6; + PackedGPUVAddr address; + BitField<0, 3, Mode> mode; + }; + static_assert(sizeof(RenderEnable) == 12); + + enum class PhysModeTarget : u32 { + LOCAL_FB = 0, + COHERENT_SYSMEM = 1, + NONCOHERENT_SYSMEM = 2, + }; + using PhysMode = BitField<0, 2, PhysModeTarget>; + + union LaunchDMA { + enum class DataTransferType : u32 { + NONE = 0, + PIPELINED = 1, + NON_PIPELINED = 2, + }; - struct Parameters { - union { - BitField<0, 4, u32> block_depth; - BitField<4, 4, u32> block_height; - BitField<8, 4, u32> block_width; - }; - u32 size_x; - u32 size_y; - u32 size_z; - u32 pos_z; - union { - BitField<0, 16, u32> pos_x; - BitField<16, 16, u32> pos_y; - }; + enum class SemaphoreType : u32 { + NONE = 0, + RELEASE_ONE_WORD_SEMAPHORE = 1, + RELEASE_FOUR_WORD_SEMAPHORE = 2, + }; - u32 BlockHeight() const { - return block_height.Value(); - } + enum class InterruptType : u32 { + NONE = 0, + BLOCKING = 1, + NON_BLOCKING = 2, + }; - u32 BlockDepth() const { - return block_depth.Value(); - } + enum class MemoryLayout : u32 { + BLOCKLINEAR = 0, + PITCH = 1, }; - static_assert(sizeof(Parameters) == 24, "Parameters has wrong size"); + enum class Type : u32 { + VIRTUAL = 0, + PHYSICAL = 1, + }; - enum class ComponentMode : u32 { - Src0 = 0, - Src1 = 1, - Src2 = 2, - Src3 = 3, - Const0 = 4, - Const1 = 5, - Zero = 6, + enum class SemaphoreReduction : u32 { + IMIN = 0, + IMAX = 1, + IXOR = 2, + IAND = 3, + IOR = 4, + IADD = 5, + INC = 6, + DEC = 7, + FADD = 0xA, }; - enum class CopyMode : u32 { - None = 0, - Unk1 = 1, - Unk2 = 2, + enum class SemaphoreReductionSign : u32 { + SIGNED = 0, + UNSIGNED = 1, }; - enum class QueryMode : u32 { - None = 0, - Short = 1, - Long = 2, + enum class BypassL2 : u32 { + USE_PTE_SETTING = 0, + FORCE_VOLATILE = 1, }; - enum class QueryIntr : u32 { - None = 0, - Block = 1, - NonBlock = 2, + BitField<0, 2, DataTransferType> data_transfer_type; + BitField<2, 1, u32> flush_enable; + BitField<3, 2, SemaphoreType> semaphore_type; + BitField<5, 2, InterruptType> interrupt_type; + BitField<7, 1, MemoryLayout> src_memory_layout; + BitField<8, 1, MemoryLayout> dst_memory_layout; + BitField<9, 1, u32> multi_line_enable; + BitField<10, 1, u32> remap_enable; + BitField<11, 1, u32> rmwdisable; + BitField<12, 1, Type> src_type; + BitField<13, 1, Type> dst_type; + BitField<14, 4, SemaphoreReduction> semaphore_reduction; + BitField<18, 1, SemaphoreReductionSign> semaphore_reduction_sign; + BitField<19, 1, u32> reduction_enable; + BitField<20, 1, BypassL2> bypass_l2; + }; + static_assert(sizeof(LaunchDMA) == 4); + + struct RemapConst { + enum Swizzle : u32 { + SRC_X = 0, + SRC_Y = 1, + SRC_Z = 2, + SRC_W = 3, + CONST_A = 4, + CONST_B = 5, + NO_WRITE = 6, }; + PackedGPUVAddr address; + union { - struct { - INSERT_UNION_PADDING_WORDS(0xC0); - - struct { - union { - BitField<0, 2, CopyMode> copy_mode; - BitField<2, 1, u32> flush; - - BitField<3, 2, QueryMode> query_mode; - BitField<5, 2, QueryIntr> query_intr; - - BitField<7, 1, u32> is_src_linear; - BitField<8, 1, u32> is_dst_linear; - - BitField<9, 1, u32> enable_2d; - BitField<10, 1, u32> enable_swizzle; - }; - } exec; - - INSERT_UNION_PADDING_WORDS(0x3F); - - struct { - u32 address_high; - u32 address_low; - - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - } src_address; - - struct { - u32 address_high; - u32 address_low; - - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - } dst_address; - - u32 src_pitch; - u32 dst_pitch; - u32 x_count; - u32 y_count; - - INSERT_UNION_PADDING_WORDS(0xB8); - - u32 const0; - u32 const1; - union { - BitField<0, 4, ComponentMode> component0; - BitField<4, 4, ComponentMode> component1; - BitField<8, 4, ComponentMode> component2; - BitField<12, 4, ComponentMode> component3; - BitField<16, 2, u32> component_size; - BitField<20, 3, u32> src_num_components; - BitField<24, 3, u32> dst_num_components; - - u32 SrcBytePerPixel() const { - return src_num_components.Value() * component_size.Value(); - } - u32 DstBytePerPixel() const { - return dst_num_components.Value() * component_size.Value(); - } - } swizzle_config; + BitField<0, 3, Swizzle> dst_x; + BitField<4, 3, Swizzle> dst_y; + BitField<8, 3, Swizzle> dst_z; + BitField<12, 3, Swizzle> dst_w; + BitField<16, 2, u32> component_size_minus_one; + BitField<20, 2, u32> num_src_components_minus_one; + BitField<24, 2, u32> num_dst_components_minus_one; + }; + }; + static_assert(sizeof(RemapConst) == 12); - Parameters dst_params; + explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager); + ~MaxwellDMA() = default; - INSERT_UNION_PADDING_WORDS(1); + /// Write the value to the register identified by method. + void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; - Parameters src_params; - - INSERT_UNION_PADDING_WORDS(0x13); - }; - std::array<u32, NUM_REGS> reg_array; - }; - } regs{}; + /// Write multiple values to the register identified by method. + void CallMultiMethod(u32 method, const u32* base_start, u32 amount, + u32 methods_pending) override; private: + /// Performs the copy from the source buffer to the destination buffer as configured in the + /// registers. + void Launch(); + + void CopyPitchToPitch(); + + void CopyBlockLinearToPitch(); + + void CopyPitchToBlockLinear(); + + void FastCopyBlockLinearToPitch(); + Core::System& system; MemoryManager& memory_manager; @@ -180,28 +215,58 @@ private: std::vector<u8> read_buffer; std::vector<u8> write_buffer; - /// Performs the copy from the source buffer to the destination buffer as configured in the - /// registers. - void HandleCopy(); -}; + static constexpr std::size_t NUM_REGS = 0x800; + struct Regs { + union { + struct { + u32 reserved[0x40]; + u32 nop; + u32 reserved01[0xf]; + u32 pm_trigger; + u32 reserved02[0x3f]; + Semaphore semaphore; + u32 reserved03[0x2]; + RenderEnable render_enable; + PhysMode src_phys_mode; + PhysMode dst_phys_mode; + u32 reserved04[0x26]; + LaunchDMA launch_dma; + u32 reserved05[0x3f]; + PackedGPUVAddr offset_in; + PackedGPUVAddr offset_out; + u32 pitch_in; + u32 pitch_out; + u32 line_length_in; + u32 line_count; + u32 reserved06[0xb8]; + RemapConst remap_const; + Parameters dst_params; + u32 reserved07[0x1]; + Parameters src_params; + u32 reserved08[0x275]; + u32 pm_trigger_end; + u32 reserved09[0x3ba]; + }; + std::array<u32, NUM_REGS> reg_array; + }; + } regs{}; #define ASSERT_REG_POSITION(field_name, position) \ static_assert(offsetof(MaxwellDMA::Regs, field_name) == position * 4, \ "Field " #field_name " has invalid position") -ASSERT_REG_POSITION(exec, 0xC0); -ASSERT_REG_POSITION(src_address, 0x100); -ASSERT_REG_POSITION(dst_address, 0x102); -ASSERT_REG_POSITION(src_pitch, 0x104); -ASSERT_REG_POSITION(dst_pitch, 0x105); -ASSERT_REG_POSITION(x_count, 0x106); -ASSERT_REG_POSITION(y_count, 0x107); -ASSERT_REG_POSITION(const0, 0x1C0); -ASSERT_REG_POSITION(const1, 0x1C1); -ASSERT_REG_POSITION(swizzle_config, 0x1C2); -ASSERT_REG_POSITION(dst_params, 0x1C3); -ASSERT_REG_POSITION(src_params, 0x1CA); + ASSERT_REG_POSITION(launch_dma, 0xC0); + ASSERT_REG_POSITION(offset_in, 0x100); + ASSERT_REG_POSITION(offset_out, 0x102); + ASSERT_REG_POSITION(pitch_in, 0x104); + ASSERT_REG_POSITION(pitch_out, 0x105); + ASSERT_REG_POSITION(line_length_in, 0x106); + ASSERT_REG_POSITION(line_count, 0x107); + ASSERT_REG_POSITION(remap_const, 0x1C0); + ASSERT_REG_POSITION(dst_params, 0x1C3); + ASSERT_REG_POSITION(src_params, 0x1CA); #undef ASSERT_REG_POSITION +}; } // namespace Tegra::Engines diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index 5e9cfba22..37d17efdc 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -32,31 +32,31 @@ struct Register { constexpr Register() = default; - constexpr Register(u64 value) : value(value) {} + constexpr Register(u64 value_) : value(value_) {} - constexpr operator u64() const { + [[nodiscard]] constexpr operator u64() const { return value; } template <typename T> - constexpr u64 operator-(const T& oth) const { + [[nodiscard]] constexpr u64 operator-(const T& oth) const { return value - oth; } template <typename T> - constexpr u64 operator&(const T& oth) const { + [[nodiscard]] constexpr u64 operator&(const T& oth) const { return value & oth; } - constexpr u64 operator&(const Register& oth) const { + [[nodiscard]] constexpr u64 operator&(const Register& oth) const { return value & oth.value; } - constexpr u64 operator~() const { + [[nodiscard]] constexpr u64 operator~() const { return ~value; } - u64 GetSwizzledIndex(u64 elem) const { + [[nodiscard]] u64 GetSwizzledIndex(u64 elem) const { elem = (value + elem) & 3; return (value & ~3) + elem; } @@ -75,7 +75,7 @@ enum class AttributeSize : u64 { union Attribute { Attribute() = default; - constexpr explicit Attribute(u64 value) : value(value) {} + constexpr explicit Attribute(u64 value_) : value(value_) {} enum class Index : u64 { LayerViewportPointSize = 6, @@ -107,7 +107,7 @@ union Attribute { BitField<31, 1, u64> patch; BitField<47, 3, AttributeSize> size; - bool IsPhysical() const { + [[nodiscard]] bool IsPhysical() const { return patch == 0 && element == 0 && static_cast<u64>(index.Value()) == 0; } } fmt20; @@ -124,7 +124,7 @@ union Attribute { union Sampler { Sampler() = default; - constexpr explicit Sampler(u64 value) : value(value) {} + constexpr explicit Sampler(u64 value_) : value(value_) {} enum class Index : u64 { Sampler_0 = 8, @@ -137,7 +137,7 @@ union Sampler { union Image { Image() = default; - constexpr explicit Image(u64 value) : value{value} {} + constexpr explicit Image(u64 value_) : value{value_} {} BitField<36, 13, u64> index; u64 value; @@ -168,18 +168,22 @@ enum class Pred : u64 { }; enum class PredCondition : u64 { - LessThan = 1, - Equal = 2, - LessEqual = 3, - GreaterThan = 4, - NotEqual = 5, - GreaterEqual = 6, - LessThanWithNan = 9, - LessEqualWithNan = 11, - GreaterThanWithNan = 12, - NotEqualWithNan = 13, - GreaterEqualWithNan = 14, - // TODO(Subv): Other condition types + F = 0, // Always false + LT = 1, // Ordered less than + EQ = 2, // Ordered equal + LE = 3, // Ordered less than or equal + GT = 4, // Ordered greater than + NE = 5, // Ordered not equal + GE = 6, // Ordered greater than or equal + NUM = 7, // Ordered + NAN_ = 8, // Unordered + LTU = 9, // Unordered less than + EQU = 10, // Unordered equal + LEU = 11, // Unordered less than or equal + GTU = 12, // Unordered greater than + NEU = 13, // Unordered not equal + GEU = 14, // Unordered greater than or equal + T = 15, // Always true }; enum class PredOperation : u64 { @@ -501,14 +505,14 @@ struct IpaMode { IpaInterpMode interpolation_mode; IpaSampleMode sampling_mode; - bool operator==(const IpaMode& a) const { + [[nodiscard]] bool operator==(const IpaMode& a) const { return std::tie(interpolation_mode, sampling_mode) == std::tie(a.interpolation_mode, a.sampling_mode); } - bool operator!=(const IpaMode& a) const { + [[nodiscard]] bool operator!=(const IpaMode& a) const { return !operator==(a); } - bool operator<(const IpaMode& a) const { + [[nodiscard]] bool operator<(const IpaMode& a) const { return std::tie(interpolation_mode, sampling_mode) < std::tie(a.interpolation_mode, a.sampling_mode); } @@ -654,7 +658,12 @@ union Instruction { return *this; } - constexpr Instruction(u64 value) : value{value} {} + constexpr Instruction(u64 value_) : value{value_} {} + constexpr Instruction(const Instruction& instr) : value(instr.value) {} + + [[nodiscard]] constexpr bool Bit(u64 offset) const { + return ((value >> offset) & 1) != 0; + } BitField<0, 8, Register> gpr0; BitField<8, 8, Register> gpr8; @@ -737,34 +746,34 @@ union Instruction { BitField<28, 8, u64> imm_lut28; BitField<48, 8, u64> imm_lut48; - u32 GetImmLut28() const { + [[nodiscard]] u32 GetImmLut28() const { return static_cast<u32>(imm_lut28); } - u32 GetImmLut48() const { + [[nodiscard]] u32 GetImmLut48() const { return static_cast<u32>(imm_lut48); } } lop3; - u16 GetImm20_16() const { + [[nodiscard]] u16 GetImm20_16() const { return static_cast<u16>(imm20_16); } - u32 GetImm20_19() const { + [[nodiscard]] u32 GetImm20_19() const { u32 imm{static_cast<u32>(imm20_19)}; imm <<= 12; imm |= negate_imm ? 0x80000000 : 0; return imm; } - u32 GetImm20_32() const { + [[nodiscard]] u32 GetImm20_32() const { return static_cast<u32>(imm20_32); } - s32 GetSignedImm20_20() const { - u32 immediate = static_cast<u32>(imm20_19 | (negate_imm << 19)); + [[nodiscard]] s32 GetSignedImm20_20() const { + const auto immediate = static_cast<u32>(imm20_19 | (negate_imm << 19)); // Sign extend the 20-bit value. - u32 mask = 1U << (20 - 1); + const auto mask = 1U << (20 - 1); return static_cast<s32>((immediate ^ mask) - mask); } } alu; @@ -813,15 +822,17 @@ union Instruction { } alu_integer; union { + BitField<43, 1, u64> x; + } iadd; + + union { BitField<39, 1, u64> ftz; BitField<32, 1, u64> saturate; BitField<49, 2, HalfMerge> merge; - BitField<43, 1, u64> negate_a; BitField<44, 1, u64> abs_a; BitField<47, 2, HalfType> type_a; - BitField<31, 1, u64> negate_b; BitField<30, 1, u64> abs_b; BitField<28, 2, HalfType> type_b; @@ -846,7 +857,7 @@ union Instruction { BitField<56, 1, u64> second_negate; BitField<30, 9, u64> second; - u32 PackImmediates() const { + [[nodiscard]] u32 PackImmediates() const { // Immediates are half floats shifted. constexpr u32 imm_shift = 6; return static_cast<u32>((first << imm_shift) | (second << (16 + imm_shift))); @@ -1022,7 +1033,7 @@ union Instruction { BitField<28, 2, AtomicType> type; BitField<30, 22, s64> offset; - s32 GetImmediateOffset() const { + [[nodiscard]] s32 GetImmediateOffset() const { return static_cast<s32>(offset << 2); } } atoms; @@ -1204,7 +1215,7 @@ union Instruction { BitField<39, 4, u64> rounding; // H0, H1 extract for F16 missing BitField<41, 1, u64> selector; // Guessed as some games set it, TODO: reverse this value - F2fRoundingOp GetRoundingMode() const { + [[nodiscard]] F2fRoundingOp GetRoundingMode() const { constexpr u64 rounding_mask = 0x0B; return static_cast<F2fRoundingOp>(rounding.Value() & rounding_mask); } @@ -1228,15 +1239,15 @@ union Instruction { BitField<54, 1, u64> aoffi_flag; BitField<55, 3, TextureProcessMode> process_mode; - bool IsComponentEnabled(std::size_t component) const { - return ((1ull << component) & component_mask) != 0; + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { + return ((1ULL << component) & component_mask) != 0; } - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { return process_mode; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return dc_flag != 0; @@ -1260,15 +1271,15 @@ union Instruction { BitField<36, 1, u64> aoffi_flag; BitField<37, 3, TextureProcessMode> process_mode; - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { return ((1ULL << component) & component_mask) != 0; } - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { return process_mode; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return dc_flag != 0; @@ -1288,7 +1299,7 @@ union Instruction { BitField<31, 4, u64> component_mask; BitField<49, 1, u64> nodep_flag; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NODEP: return nodep_flag != 0; @@ -1298,7 +1309,7 @@ union Instruction { return false; } - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { return ((1ULL << component) & component_mask) != 0; } } txq; @@ -1310,11 +1321,11 @@ union Instruction { BitField<35, 1, u64> ndv_flag; BitField<49, 1, u64> nodep_flag; - bool IsComponentEnabled(std::size_t component) const { - return ((1ull << component) & component_mask) != 0; + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { + return ((1ULL << component) & component_mask) != 0; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NDV: return (ndv_flag != 0); @@ -1336,7 +1347,7 @@ union Instruction { BitField<54, 2, u64> offset_mode; BitField<56, 2, u64> component; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NDV: return ndv_flag != 0; @@ -1362,7 +1373,7 @@ union Instruction { BitField<33, 2, u64> offset_mode; BitField<37, 2, u64> component; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NDV: return ndv_flag != 0; @@ -1388,7 +1399,7 @@ union Instruction { BitField<52, 2, u64> component; BitField<55, 1, u64> fp16_flag; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return dc_flag != 0; @@ -1411,16 +1422,20 @@ union Instruction { BitField<53, 4, u64> texture_info; BitField<59, 1, u64> fp32_flag; - TextureType GetTextureType() const { + [[nodiscard]] TextureType GetTextureType() const { // The TEXS instruction has a weird encoding for the texture type. - if (texture_info == 0) + if (texture_info == 0) { return TextureType::Texture1D; - if (texture_info >= 1 && texture_info <= 9) + } + if (texture_info >= 1 && texture_info <= 9) { return TextureType::Texture2D; - if (texture_info >= 10 && texture_info <= 11) + } + if (texture_info >= 10 && texture_info <= 11) { return TextureType::Texture3D; - if (texture_info >= 12 && texture_info <= 13) + } + if (texture_info >= 12 && texture_info <= 13) { return TextureType::TextureCube; + } LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}", static_cast<u32>(texture_info.Value())); @@ -1428,7 +1443,7 @@ union Instruction { return TextureType::Texture1D; } - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { switch (texture_info) { case 0: case 2: @@ -1447,7 +1462,7 @@ union Instruction { return TextureProcessMode::None; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return (texture_info >= 4 && texture_info <= 6) || texture_info == 9; @@ -1459,16 +1474,16 @@ union Instruction { return false; } - bool IsArrayTexture() const { + [[nodiscard]] bool IsArrayTexture() const { // TEXS only supports Texture2D arrays. return texture_info >= 7 && texture_info <= 9; } - bool HasTwoDestinations() const { + [[nodiscard]] bool HasTwoDestinations() const { return gpr28.Value() != Register::ZeroIndex; } - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { static constexpr std::array<std::array<u32, 8>, 4> mask_lut{{ {}, {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc}, @@ -1495,7 +1510,7 @@ union Instruction { BitField<54, 1, u64> cl; BitField<55, 1, u64> process_mode; - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { return process_mode == 0 ? TextureProcessMode::LZ : TextureProcessMode::LL; } } tld; @@ -1505,9 +1520,9 @@ union Instruction { BitField<53, 4, u64> texture_info; BitField<59, 1, u64> fp32_flag; - TextureType GetTextureType() const { + [[nodiscard]] TextureType GetTextureType() const { // The TLDS instruction has a weird encoding for the texture type. - if (texture_info >= 0 && texture_info <= 1) { + if (texture_info <= 1) { return TextureType::Texture1D; } if (texture_info == 2 || texture_info == 8 || texture_info == 12 || @@ -1524,13 +1539,14 @@ union Instruction { return TextureType::Texture1D; } - TextureProcessMode GetTextureProcessMode() const { - if (texture_info == 1 || texture_info == 5 || texture_info == 12) + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { + if (texture_info == 1 || texture_info == 5 || texture_info == 12) { return TextureProcessMode::LL; + } return TextureProcessMode::LZ; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::AOFFI: return texture_info == 12 || texture_info == 4; @@ -1544,7 +1560,7 @@ union Instruction { return false; } - bool IsArrayTexture() const { + [[nodiscard]] bool IsArrayTexture() const { // TEXS only supports Texture2D arrays. return texture_info == 8; } @@ -1556,7 +1572,7 @@ union Instruction { BitField<35, 1, u64> aoffi_flag; BitField<49, 1, u64> nodep_flag; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::AOFFI: return aoffi_flag != 0; @@ -1580,7 +1596,7 @@ union Instruction { BitField<20, 3, StoreType> store_data_layout; BitField<20, 4, u64> component_mask_selector; - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { ASSERT(mode == SurfaceDataMode::P); constexpr u8 R = 0b0001; constexpr u8 G = 0b0010; @@ -1593,7 +1609,7 @@ union Instruction { return std::bitset<4>{mask.at(component_mask_selector)}.test(component); } - StoreType GetStoreDataLayout() const { + [[nodiscard]] StoreType GetStoreDataLayout() const { ASSERT(mode == SurfaceDataMode::D_BA); return store_data_layout; } @@ -1611,14 +1627,15 @@ union Instruction { BitField<20, 24, u64> target; BitField<5, 1, u64> constant_buffer; - s32 GetBranchTarget() const { + [[nodiscard]] s32 GetBranchTarget() const { // Sign extend the branch target offset - u32 mask = 1U << (24 - 1); - u32 value = static_cast<u32>(target); + const auto mask = 1U << (24 - 1); + const auto target_value = static_cast<u32>(target); + constexpr auto instruction_size = static_cast<s32>(sizeof(Instruction)); + // The branch offset is relative to the next instruction and is stored in bytes, so // divide it by the size of an instruction and add 1 to it. - return static_cast<s32>((value ^ mask) - mask) / static_cast<s32>(sizeof(Instruction)) + - 1; + return static_cast<s32>((target_value ^ mask) - mask) / instruction_size + 1; } } bra; @@ -1626,14 +1643,15 @@ union Instruction { BitField<20, 24, u64> target; BitField<5, 1, u64> constant_buffer; - s32 GetBranchExtend() const { + [[nodiscard]] s32 GetBranchExtend() const { // Sign extend the branch target offset - u32 mask = 1U << (24 - 1); - u32 value = static_cast<u32>(target); + const auto mask = 1U << (24 - 1); + const auto target_value = static_cast<u32>(target); + constexpr auto instruction_size = static_cast<s32>(sizeof(Instruction)); + // The branch offset is relative to the next instruction and is stored in bytes, so // divide it by the size of an instruction and add 1 to it. - return static_cast<s32>((value ^ mask) - mask) / static_cast<s32>(sizeof(Instruction)) + - 1; + return static_cast<s32>((target_value ^ mask) - mask) / instruction_size + 1; } } brx; @@ -1686,7 +1704,7 @@ union Instruction { BitField<50, 1, u64> is_op_b_register; BitField<51, 3, VmnmxOperation> operation; - VmnmxType SourceFormatA() const { + [[nodiscard]] VmnmxType SourceFormatA() const { switch (src_format_a) { case 0b11: return VmnmxType::Bits32; @@ -1697,7 +1715,7 @@ union Instruction { } } - VmnmxType SourceFormatB() const { + [[nodiscard]] VmnmxType SourceFormatB() const { switch (src_format_b) { case 0b11: return VmnmxType::Bits32; @@ -1728,7 +1746,7 @@ union Instruction { BitField<20, 14, u64> shifted_offset; BitField<34, 5, u64> index; - u64 GetOffset() const { + [[nodiscard]] u64 GetOffset() const { return shifted_offset * 4; } } cbuf34; @@ -1737,7 +1755,7 @@ union Instruction { BitField<20, 16, s64> offset; BitField<36, 5, u64> index; - s64 GetOffset() const { + [[nodiscard]] s64 GetOffset() const { return offset; } } cbuf36; @@ -1867,7 +1885,9 @@ public: HSETP2_C, HSETP2_R, HSETP2_IMM, + HSET2_C, HSET2_R, + HSET2_IMM, POPC_C, POPC_R, POPC_IMM, @@ -1880,6 +1900,7 @@ public: ICMP_IMM, FCMP_RR, FCMP_RC, + FCMP_IMMR, MUFU, // Multi-Function Operator RRO_C, // Range Reduction Operator RRO_R, @@ -1983,29 +2004,29 @@ public: /// Returns whether an opcode has an execution predicate field or not (ie, whether it can be /// conditionally executed). - static bool IsPredicatedInstruction(Id opcode) { + [[nodiscard]] static bool IsPredicatedInstruction(Id opcode) { // TODO(Subv): Add the rest of unpredicated instructions. return opcode != Id::SSY && opcode != Id::PBK; } class Matcher { public: - constexpr Matcher(const char* const name, u16 mask, u16 expected, Id id, Type type) - : name{name}, mask{mask}, expected{expected}, id{id}, type{type} {} + constexpr Matcher(const char* const name_, u16 mask_, u16 expected_, Id id_, Type type_) + : name{name_}, mask{mask_}, expected{expected_}, id{id_}, type{type_} {} - constexpr const char* GetName() const { + [[nodiscard]] constexpr const char* GetName() const { return name; } - constexpr u16 GetMask() const { + [[nodiscard]] constexpr u16 GetMask() const { return mask; } - constexpr Id GetId() const { + [[nodiscard]] constexpr Id GetId() const { return id; } - constexpr Type GetType() const { + [[nodiscard]] constexpr Type GetType() const { return type; } @@ -2014,7 +2035,7 @@ public: * @param instruction The instruction to test * @returns true if the given instruction matches. */ - constexpr bool Matches(u16 instruction) const { + [[nodiscard]] constexpr bool Matches(u16 instruction) const { return (instruction & mask) == expected; } @@ -2026,7 +2047,8 @@ public: Type type; }; - static std::optional<std::reference_wrapper<const Matcher>> Decode(Instruction instr) { + using DecodeResult = std::optional<std::reference_wrapper<const Matcher>>; + [[nodiscard]] static DecodeResult Decode(Instruction instr) { static const auto table{GetDecodeTable()}; const auto matches_instruction = [instr](const auto& matcher) { @@ -2048,7 +2070,7 @@ private: * A '0' in a bitstring indicates that a zero must be present at that bit position. * A '1' in a bitstring indicates that a one must be present at that bit position. */ - static constexpr auto GetMaskAndExpect(const char* const bitstring) { + [[nodiscard]] static constexpr auto GetMaskAndExpect(const char* const bitstring) { u16 mask = 0, expect = 0; for (std::size_t i = 0; i < opcode_bitsize; i++) { const std::size_t bit_position = opcode_bitsize - i - 1; @@ -2070,14 +2092,14 @@ private: public: /// Creates a matcher that can match and parse instructions based on bitstring. - static constexpr auto GetMatcher(const char* const bitstring, Id op, Type type, - const char* const name) { + [[nodiscard]] static constexpr auto GetMatcher(const char* const bitstring, Id op, + Type type, const char* const name) { const auto [mask, expected] = GetMaskAndExpect(bitstring); return Matcher(name, mask, expected, op, type); } }; - static std::vector<Matcher> GetDecodeTable() { + [[nodiscard]] static std::vector<Matcher> GetDecodeTable() { std::vector<Matcher> table = { #define INST(bitstring, op, type, name) Detail::GetMatcher(bitstring, op, type, name) INST("111000110011----", Id::KIL, Type::Flow, "KIL"), @@ -2187,9 +2209,12 @@ private: INST("0111111-1-------", Id::HSETP2_C, Type::HalfSetPredicate, "HSETP2_C"), INST("0101110100100---", Id::HSETP2_R, Type::HalfSetPredicate, "HSETP2_R"), INST("0111111-0-------", Id::HSETP2_IMM, Type::HalfSetPredicate, "HSETP2_IMM"), + INST("0111110-1-------", Id::HSET2_C, Type::HalfSet, "HSET2_C"), INST("0101110100011---", Id::HSET2_R, Type::HalfSet, "HSET2_R"), + INST("0111110-0-------", Id::HSET2_IMM, Type::HalfSet, "HSET2_IMM"), INST("010110111010----", Id::FCMP_RR, Type::Arithmetic, "FCMP_RR"), INST("010010111010----", Id::FCMP_RC, Type::Arithmetic, "FCMP_RC"), + INST("0011011-1010----", Id::FCMP_IMMR, Type::Arithmetic, "FCMP_IMMR"), INST("0101000010000---", Id::MUFU, Type::Arithmetic, "MUFU"), INST("0100110010010---", Id::RRO_C, Type::Arithmetic, "RRO_C"), INST("0101110010010---", Id::RRO_R, Type::Arithmetic, "RRO_R"), diff --git a/src/video_core/engines/shader_header.h b/src/video_core/engines/shader_header.h index 72e2a33d5..ceec05459 100644 --- a/src/video_core/engines/shader_header.h +++ b/src/video_core/engines/shader_header.h @@ -41,30 +41,30 @@ struct Header { BitField<26, 1, u32> does_load_or_store; BitField<27, 1, u32> does_fp64; BitField<28, 4, u32> stream_out_mask; - } common0{}; + } common0; union { BitField<0, 24, u32> shader_local_memory_low_size; BitField<24, 8, u32> per_patch_attribute_count; - } common1{}; + } common1; union { BitField<0, 24, u32> shader_local_memory_high_size; BitField<24, 8, u32> threads_per_input_primitive; - } common2{}; + } common2; union { BitField<0, 24, u32> shader_local_memory_crs_size; BitField<24, 4, OutputTopology> output_topology; BitField<28, 4, u32> reserved; - } common3{}; + } common3; union { BitField<0, 12, u32> max_output_vertices; BitField<12, 8, u32> store_req_start; // NOTE: not used by geometry shaders. BitField<20, 4, u32> reserved; BitField<24, 8, u32> store_req_end; // NOTE: not used by geometry shaders. - } common4{}; + } common4; union { struct { @@ -145,7 +145,7 @@ struct Header { } } ps; - std::array<u32, 0xF> raw{}; + std::array<u32, 0xF> raw; }; u64 GetLocalMemorySize() const { @@ -153,7 +153,6 @@ struct Header { (common2.shader_local_memory_high_size << 24)); } }; - static_assert(sizeof(Header) == 0x50, "Incorrect structure size"); } // namespace Tegra::Shader diff --git a/src/video_core/fence_manager.h b/src/video_core/fence_manager.h new file mode 100644 index 000000000..de6991ef6 --- /dev/null +++ b/src/video_core/fence_manager.h @@ -0,0 +1,164 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <algorithm> +#include <queue> + +#include "common/common_types.h" +#include "core/core.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" +#include "video_core/rasterizer_interface.h" + +namespace VideoCommon { + +class FenceBase { +public: + FenceBase(u32 payload, bool is_stubbed) + : address{}, payload{payload}, is_semaphore{false}, is_stubbed{is_stubbed} {} + + FenceBase(GPUVAddr address, u32 payload, bool is_stubbed) + : address{address}, payload{payload}, is_semaphore{true}, is_stubbed{is_stubbed} {} + + GPUVAddr GetAddress() const { + return address; + } + + u32 GetPayload() const { + return payload; + } + + bool IsSemaphore() const { + return is_semaphore; + } + +private: + GPUVAddr address; + u32 payload; + bool is_semaphore; + +protected: + bool is_stubbed; +}; + +template <typename TFence, typename TTextureCache, typename TTBufferCache, typename TQueryCache> +class FenceManager { +public: + void SignalSemaphore(GPUVAddr addr, u32 value) { + TryReleasePendingFences(); + const bool should_flush = ShouldFlush(); + CommitAsyncFlushes(); + TFence new_fence = CreateFence(addr, value, !should_flush); + fences.push(new_fence); + QueueFence(new_fence); + if (should_flush) { + rasterizer.FlushCommands(); + } + rasterizer.SyncGuestHost(); + } + + void SignalSyncPoint(u32 value) { + TryReleasePendingFences(); + const bool should_flush = ShouldFlush(); + CommitAsyncFlushes(); + TFence new_fence = CreateFence(value, !should_flush); + fences.push(new_fence); + QueueFence(new_fence); + if (should_flush) { + rasterizer.FlushCommands(); + } + rasterizer.SyncGuestHost(); + } + + void WaitPendingFences() { + while (!fences.empty()) { + TFence& current_fence = fences.front(); + if (ShouldWait()) { + WaitFence(current_fence); + } + PopAsyncFlushes(); + if (current_fence->IsSemaphore()) { + gpu_memory.template Write<u32>(current_fence->GetAddress(), + current_fence->GetPayload()); + } else { + gpu.IncrementSyncPoint(current_fence->GetPayload()); + } + fences.pop(); + } + } + +protected: + explicit FenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + TTextureCache& texture_cache_, TTBufferCache& buffer_cache_, + TQueryCache& query_cache_) + : rasterizer{rasterizer_}, gpu{gpu_}, gpu_memory{gpu.MemoryManager()}, + texture_cache{texture_cache_}, buffer_cache{buffer_cache_}, query_cache{query_cache_} {} + + virtual ~FenceManager() = default; + + /// Creates a Sync Point Fence Interface, does not create a backend fence if 'is_stubbed' is + /// true + virtual TFence CreateFence(u32 value, bool is_stubbed) = 0; + /// Creates a Semaphore Fence Interface, does not create a backend fence if 'is_stubbed' is true + virtual TFence CreateFence(GPUVAddr addr, u32 value, bool is_stubbed) = 0; + /// Queues a fence into the backend if the fence isn't stubbed. + virtual void QueueFence(TFence& fence) = 0; + /// Notifies that the backend fence has been signaled/reached in host GPU. + virtual bool IsFenceSignaled(TFence& fence) const = 0; + /// Waits until a fence has been signalled by the host GPU. + virtual void WaitFence(TFence& fence) = 0; + + VideoCore::RasterizerInterface& rasterizer; + Tegra::GPU& gpu; + Tegra::MemoryManager& gpu_memory; + TTextureCache& texture_cache; + TTBufferCache& buffer_cache; + TQueryCache& query_cache; + +private: + void TryReleasePendingFences() { + while (!fences.empty()) { + TFence& current_fence = fences.front(); + if (ShouldWait() && !IsFenceSignaled(current_fence)) { + return; + } + PopAsyncFlushes(); + if (current_fence->IsSemaphore()) { + gpu_memory.template Write<u32>(current_fence->GetAddress(), + current_fence->GetPayload()); + } else { + gpu.IncrementSyncPoint(current_fence->GetPayload()); + } + fences.pop(); + } + } + + bool ShouldWait() const { + return texture_cache.ShouldWaitAsyncFlushes() || buffer_cache.ShouldWaitAsyncFlushes() || + query_cache.ShouldWaitAsyncFlushes(); + } + + bool ShouldFlush() const { + return texture_cache.HasUncommittedFlushes() || buffer_cache.HasUncommittedFlushes() || + query_cache.HasUncommittedFlushes(); + } + + void PopAsyncFlushes() { + texture_cache.PopAsyncFlushes(); + buffer_cache.PopAsyncFlushes(); + query_cache.PopAsyncFlushes(); + } + + void CommitAsyncFlushes() { + texture_cache.CommitAsyncFlushes(); + buffer_cache.CommitAsyncFlushes(); + query_cache.CommitAsyncFlushes(); + } + + std::queue<TFence> fences; +}; + +} // namespace VideoCommon diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index 8acf2eda2..ebd149c3a 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <chrono> + #include "common/assert.h" #include "common/microprofile.h" #include "core/core.h" @@ -9,6 +11,7 @@ #include "core/core_timing_util.h" #include "core/frontend/emu_window.h" #include "core/memory.h" +#include "core/settings.h" #include "video_core/engines/fermi_2d.h" #include "video_core/engines/kepler_compute.h" #include "video_core/engines/kepler_memory.h" @@ -17,26 +20,36 @@ #include "video_core/gpu.h" #include "video_core/memory_manager.h" #include "video_core/renderer_base.h" +#include "video_core/shader_notify.h" #include "video_core/video_core.h" namespace Tegra { MICROPROFILE_DEFINE(GPU_wait, "GPU", "Wait for the GPU", MP_RGB(128, 128, 192)); -GPU::GPU(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& renderer_, bool is_async) - : system{system}, renderer{std::move(renderer_)}, is_async{is_async} { - auto& rasterizer{renderer->Rasterizer()}; - memory_manager = std::make_unique<Tegra::MemoryManager>(system, rasterizer); - dma_pusher = std::make_unique<Tegra::DmaPusher>(*this); - maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, rasterizer, *memory_manager); - fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer); - kepler_compute = std::make_unique<Engines::KeplerCompute>(system, rasterizer, *memory_manager); - maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, *memory_manager); - kepler_memory = std::make_unique<Engines::KeplerMemory>(system, *memory_manager); -} +GPU::GPU(Core::System& system_, bool is_async_, bool use_nvdec_) + : system{system_}, memory_manager{std::make_unique<Tegra::MemoryManager>(system)}, + dma_pusher{std::make_unique<Tegra::DmaPusher>(system, *this)}, + cdma_pusher{std::make_unique<Tegra::CDmaPusher>(*this)}, use_nvdec{use_nvdec_}, + maxwell_3d{std::make_unique<Engines::Maxwell3D>(system, *memory_manager)}, + fermi_2d{std::make_unique<Engines::Fermi2D>()}, + kepler_compute{std::make_unique<Engines::KeplerCompute>(system, *memory_manager)}, + maxwell_dma{std::make_unique<Engines::MaxwellDMA>(system, *memory_manager)}, + kepler_memory{std::make_unique<Engines::KeplerMemory>(system, *memory_manager)}, + shader_notify{std::make_unique<VideoCore::ShaderNotify>()}, is_async{is_async_} {} GPU::~GPU() = default; +void GPU::BindRenderer(std::unique_ptr<VideoCore::RendererBase> renderer_) { + renderer = std::move(renderer_); + + VideoCore::RasterizerInterface& rasterizer = renderer->Rasterizer(); + memory_manager->BindRasterizer(rasterizer); + maxwell_3d->BindRasterizer(rasterizer); + fermi_2d->BindRasterizer(rasterizer); + kepler_compute->BindRasterizer(rasterizer); +} + Engines::Maxwell3D& GPU::Maxwell3D() { return *maxwell_3d; } @@ -65,10 +78,18 @@ DmaPusher& GPU::DmaPusher() { return *dma_pusher; } +Tegra::CDmaPusher& GPU::CDmaPusher() { + return *cdma_pusher; +} + const DmaPusher& GPU::DmaPusher() const { return *dma_pusher; } +const Tegra::CDmaPusher& GPU::CDmaPusher() const { + return *cdma_pusher; +} + void GPU::WaitFence(u32 syncpoint_id, u32 value) { // Synced GPU, is always in sync if (!is_async) { @@ -76,7 +97,7 @@ void GPU::WaitFence(u32 syncpoint_id, u32 value) { } MICROPROFILE_SCOPE(GPU_wait); std::unique_lock lock{sync_mutex}; - sync_cv.wait(lock, [=]() { return syncpoints[syncpoint_id].load() >= value; }); + sync_cv.wait(lock, [=, this] { return syncpoints[syncpoint_id].load() >= value; }); } void GPU::IncrementSyncPoint(const u32 syncpoint_id) { @@ -125,14 +146,38 @@ bool GPU::CancelSyncptInterrupt(const u32 syncpoint_id, const u32 value) { return true; } +u64 GPU::RequestFlush(VAddr addr, std::size_t size) { + std::unique_lock lck{flush_request_mutex}; + const u64 fence = ++last_flush_fence; + flush_requests.emplace_back(fence, addr, size); + return fence; +} + +void GPU::TickWork() { + std::unique_lock lck{flush_request_mutex}; + while (!flush_requests.empty()) { + auto& request = flush_requests.front(); + const u64 fence = request.fence; + const VAddr addr = request.addr; + const std::size_t size = request.size; + flush_requests.pop_front(); + flush_request_mutex.unlock(); + renderer->Rasterizer().FlushRegion(addr, size); + current_flush_fence.store(fence); + flush_request_mutex.lock(); + } +} + u64 GPU::GetTicks() const { // This values were reversed engineered by fincs from NVN // The gpu clock is reported in units of 385/625 nanoseconds constexpr u64 gpu_ticks_num = 384; constexpr u64 gpu_ticks_den = 625; - const u64 cpu_ticks = system.CoreTiming().GetTicks(); - const u64 nanoseconds = Core::Timing::CyclesToNs(cpu_ticks).count(); + u64 nanoseconds = system.CoreTiming().GetGlobalTimeNs().count(); + if (Settings::values.use_fast_gpu_time.GetValue()) { + nanoseconds /= 256; + } const u64 nanoseconds_num = nanoseconds / gpu_ticks_den; const u64 nanoseconds_rem = nanoseconds % gpu_ticks_den; return nanoseconds_num * gpu_ticks_num + (nanoseconds_rem * gpu_ticks_num) / gpu_ticks_den; @@ -142,30 +187,13 @@ void GPU::FlushCommands() { renderer->Rasterizer().FlushCommands(); } -// Note that, traditionally, methods are treated as 4-byte addressable locations, and hence -// their numbers are written down multiplied by 4 in Docs. Here we are not multiply by 4. -// So the values you see in docs might be multiplied by 4. -enum class BufferMethods { - BindObject = 0x0, - Nop = 0x2, - SemaphoreAddressHigh = 0x4, - SemaphoreAddressLow = 0x5, - SemaphoreSequence = 0x6, - SemaphoreTrigger = 0x7, - NotifyIntr = 0x8, - WrcacheFlush = 0x9, - Unk28 = 0xA, - UnkCacheFlush = 0xB, - RefCnt = 0x14, - SemaphoreAcquire = 0x1A, - SemaphoreRelease = 0x1B, - FenceValue = 0x1C, - FenceAction = 0x1D, - Unk78 = 0x1E, - Unk7c = 0x1F, - Yield = 0x20, - NonPullerMethods = 0x40, -}; +void GPU::SyncGuestHost() { + renderer->Rasterizer().SyncGuestHost(); +} + +void GPU::OnCommandListEnd() { + renderer->Rasterizer().ReleaseFences(); +} enum class GpuSemaphoreOperation { AcquireEqual = 0x1, @@ -180,16 +208,32 @@ void GPU::CallMethod(const MethodCall& method_call) { ASSERT(method_call.subchannel < bound_engines.size()); - if (ExecuteMethodOnEngine(method_call)) { + if (ExecuteMethodOnEngine(method_call.method)) { CallEngineMethod(method_call); } else { CallPullerMethod(method_call); } } -bool GPU::ExecuteMethodOnEngine(const MethodCall& method_call) { - const auto method = static_cast<BufferMethods>(method_call.method); - return method >= BufferMethods::NonPullerMethods; +void GPU::CallMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount, + u32 methods_pending) { + LOG_TRACE(HW_GPU, "Processing method {:08X} on subchannel {}", method, subchannel); + + ASSERT(subchannel < bound_engines.size()); + + if (ExecuteMethodOnEngine(method)) { + CallEngineMultiMethod(method, subchannel, base_start, amount, methods_pending); + } else { + for (std::size_t i = 0; i < amount; i++) { + CallPullerMethod( + {method, base_start[i], subchannel, methods_pending - static_cast<u32>(i)}); + } + } +} + +bool GPU::ExecuteMethodOnEngine(u32 method) { + const auto buffer_method = static_cast<BufferMethods>(method); + return buffer_method >= BufferMethods::NonPullerMethods; } void GPU::CallPullerMethod(const MethodCall& method_call) { @@ -209,7 +253,12 @@ void GPU::CallPullerMethod(const MethodCall& method_call) { case BufferMethods::UnkCacheFlush: case BufferMethods::WrcacheFlush: case BufferMethods::FenceValue: + break; case BufferMethods::FenceAction: + ProcessFenceActionMethod(); + break; + case BufferMethods::WaitForInterrupt: + ProcessWaitForInterruptMethod(); break; case BufferMethods::SemaphoreTrigger: { ProcessSemaphoreTriggerMethod(); @@ -250,19 +299,46 @@ void GPU::CallEngineMethod(const MethodCall& method_call) { switch (engine) { case EngineID::FERMI_TWOD_A: - fermi_2d->CallMethod(method_call); + fermi_2d->CallMethod(method_call.method, method_call.argument, method_call.IsLastCall()); + break; + case EngineID::MAXWELL_B: + maxwell_3d->CallMethod(method_call.method, method_call.argument, method_call.IsLastCall()); + break; + case EngineID::KEPLER_COMPUTE_B: + kepler_compute->CallMethod(method_call.method, method_call.argument, + method_call.IsLastCall()); + break; + case EngineID::MAXWELL_DMA_COPY_A: + maxwell_dma->CallMethod(method_call.method, method_call.argument, method_call.IsLastCall()); + break; + case EngineID::KEPLER_INLINE_TO_MEMORY_B: + kepler_memory->CallMethod(method_call.method, method_call.argument, + method_call.IsLastCall()); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented engine"); + } +} + +void GPU::CallEngineMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount, + u32 methods_pending) { + const EngineID engine = bound_engines[subchannel]; + + switch (engine) { + case EngineID::FERMI_TWOD_A: + fermi_2d->CallMultiMethod(method, base_start, amount, methods_pending); break; case EngineID::MAXWELL_B: - maxwell_3d->CallMethod(method_call); + maxwell_3d->CallMultiMethod(method, base_start, amount, methods_pending); break; case EngineID::KEPLER_COMPUTE_B: - kepler_compute->CallMethod(method_call); + kepler_compute->CallMultiMethod(method, base_start, amount, methods_pending); break; case EngineID::MAXWELL_DMA_COPY_A: - maxwell_dma->CallMethod(method_call); + maxwell_dma->CallMultiMethod(method, base_start, amount, methods_pending); break; case EngineID::KEPLER_INLINE_TO_MEMORY_B: - kepler_memory->CallMethod(method_call); + kepler_memory->CallMultiMethod(method, base_start, amount, methods_pending); break; default: UNIMPLEMENTED_MSG("Unimplemented engine"); @@ -273,7 +349,46 @@ void GPU::ProcessBindMethod(const MethodCall& method_call) { // Bind the current subchannel to the desired engine id. LOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", method_call.subchannel, method_call.argument); - bound_engines[method_call.subchannel] = static_cast<EngineID>(method_call.argument); + const auto engine_id = static_cast<EngineID>(method_call.argument); + bound_engines[method_call.subchannel] = static_cast<EngineID>(engine_id); + switch (engine_id) { + case EngineID::FERMI_TWOD_A: + dma_pusher->BindSubchannel(fermi_2d.get(), method_call.subchannel); + break; + case EngineID::MAXWELL_B: + dma_pusher->BindSubchannel(maxwell_3d.get(), method_call.subchannel); + break; + case EngineID::KEPLER_COMPUTE_B: + dma_pusher->BindSubchannel(kepler_compute.get(), method_call.subchannel); + break; + case EngineID::MAXWELL_DMA_COPY_A: + dma_pusher->BindSubchannel(maxwell_dma.get(), method_call.subchannel); + break; + case EngineID::KEPLER_INLINE_TO_MEMORY_B: + dma_pusher->BindSubchannel(kepler_memory.get(), method_call.subchannel); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented engine {:04X}", static_cast<u32>(engine_id)); + } +} + +void GPU::ProcessFenceActionMethod() { + switch (regs.fence_action.op) { + case FenceOperation::Acquire: + WaitFence(regs.fence_action.syncpoint_id, regs.fence_value); + break; + case FenceOperation::Increment: + IncrementSyncPoint(regs.fence_action.syncpoint_id); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented operation {}", + static_cast<u32>(regs.fence_action.op.Value())); + } +} + +void GPU::ProcessWaitForInterruptMethod() { + // TODO(bunnei) ImplementMe + LOG_WARNING(HW_GPU, "(STUBBED) called"); } void GPU::ProcessSemaphoreTriggerMethod() { diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index 1a2d747be..21410e125 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -13,14 +13,15 @@ #include "common/common_types.h" #include "core/hle/service/nvdrv/nvdata.h" #include "core/hle/service/nvflinger/buffer_queue.h" +#include "video_core/cdma_pusher.h" #include "video_core/dma_pusher.h" using CacheAddr = std::uintptr_t; -inline CacheAddr ToCacheAddr(const void* host_ptr) { +[[nodiscard]] inline CacheAddr ToCacheAddr(const void* host_ptr) { return reinterpret_cast<CacheAddr>(host_ptr); } -inline u8* FromCacheAddr(CacheAddr cache_addr) { +[[nodiscard]] inline u8* FromCacheAddr(CacheAddr cache_addr) { return reinterpret_cast<u8*>(cache_addr); } @@ -33,58 +34,68 @@ class System; namespace VideoCore { class RendererBase; +class ShaderNotify; } // namespace VideoCore namespace Tegra { enum class RenderTargetFormat : u32 { NONE = 0x0, - RGBA32_FLOAT = 0xC0, - RGBA32_UINT = 0xC2, - RGBA16_UNORM = 0xC6, - RGBA16_SNORM = 0xC7, - RGBA16_UINT = 0xC9, - RGBA16_FLOAT = 0xCA, - RG32_FLOAT = 0xCB, - RG32_UINT = 0xCD, - RGBX16_FLOAT = 0xCE, - BGRA8_UNORM = 0xCF, - BGRA8_SRGB = 0xD0, - RGB10_A2_UNORM = 0xD1, - RGBA8_UNORM = 0xD5, - RGBA8_SRGB = 0xD6, - RGBA8_SNORM = 0xD7, - RGBA8_UINT = 0xD9, - RG16_UNORM = 0xDA, - RG16_SNORM = 0xDB, - RG16_SINT = 0xDC, - RG16_UINT = 0xDD, - RG16_FLOAT = 0xDE, - R11G11B10_FLOAT = 0xE0, + R32B32G32A32_FLOAT = 0xC0, + R32G32B32A32_SINT = 0xC1, + R32G32B32A32_UINT = 0xC2, + R16G16B16A16_UNORM = 0xC6, + R16G16B16A16_SNORM = 0xC7, + R16G16B16A16_SINT = 0xC8, + R16G16B16A16_UINT = 0xC9, + R16G16B16A16_FLOAT = 0xCA, + R32G32_FLOAT = 0xCB, + R32G32_SINT = 0xCC, + R32G32_UINT = 0xCD, + R16G16B16X16_FLOAT = 0xCE, + B8G8R8A8_UNORM = 0xCF, + B8G8R8A8_SRGB = 0xD0, + A2B10G10R10_UNORM = 0xD1, + A2B10G10R10_UINT = 0xD2, + A8B8G8R8_UNORM = 0xD5, + A8B8G8R8_SRGB = 0xD6, + A8B8G8R8_SNORM = 0xD7, + A8B8G8R8_SINT = 0xD8, + A8B8G8R8_UINT = 0xD9, + R16G16_UNORM = 0xDA, + R16G16_SNORM = 0xDB, + R16G16_SINT = 0xDC, + R16G16_UINT = 0xDD, + R16G16_FLOAT = 0xDE, + B10G11R11_FLOAT = 0xE0, R32_SINT = 0xE3, R32_UINT = 0xE4, R32_FLOAT = 0xE5, - B5G6R5_UNORM = 0xE8, - BGR5A1_UNORM = 0xE9, - RG8_UNORM = 0xEA, - RG8_SNORM = 0xEB, + R5G6B5_UNORM = 0xE8, + A1R5G5B5_UNORM = 0xE9, + R8G8_UNORM = 0xEA, + R8G8_SNORM = 0xEB, + R8G8_SINT = 0xEC, + R8G8_UINT = 0xED, R16_UNORM = 0xEE, R16_SNORM = 0xEF, R16_SINT = 0xF0, R16_UINT = 0xF1, R16_FLOAT = 0xF2, R8_UNORM = 0xF3, + R8_SNORM = 0xF4, + R8_SINT = 0xF5, R8_UINT = 0xF6, }; enum class DepthFormat : u32 { - Z32_FLOAT = 0xA, - Z16_UNORM = 0x13, - S8_Z24_UNORM = 0x14, - Z24_X8_UNORM = 0x15, - Z24_S8_UNORM = 0x16, - Z24_C8_UNORM = 0x18, - Z32_S8_X24_FLOAT = 0x19, + D32_FLOAT = 0xA, + D16_UNORM = 0x13, + S8_UINT_Z24_UNORM = 0x14, + D24X8_UNORM = 0x15, + D24S8_UNORM = 0x16, + D24C8_UNORM = 0x18, + D32_FLOAT_S8X24_UINT = 0x19, }; struct CommandListHeader; @@ -95,9 +106,9 @@ class DebugContext; */ struct FramebufferConfig { enum class PixelFormat : u32 { - ABGR8 = 1, - RGB565 = 4, - BGRA8 = 5, + A8B8G8R8_UNORM = 1, + RGB565_UNORM = 4, + B8G8R8A8_UNORM = 5, }; VAddr address; @@ -132,60 +143,102 @@ class MemoryManager; class GPU { public: - explicit GPU(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& renderer, - bool is_async); - - virtual ~GPU(); - struct MethodCall { u32 method{}; u32 argument{}; u32 subchannel{}; u32 method_count{}; - bool IsLastCall() const { - return method_count <= 1; - } - MethodCall(u32 method, u32 argument, u32 subchannel = 0, u32 method_count = 0) : method(method), argument(argument), subchannel(subchannel), method_count(method_count) {} + + [[nodiscard]] bool IsLastCall() const { + return method_count <= 1; + } }; + explicit GPU(Core::System& system, bool is_async, bool use_nvdec); + virtual ~GPU(); + + /// Binds a renderer to the GPU. + void BindRenderer(std::unique_ptr<VideoCore::RendererBase> renderer); + /// Calls a GPU method. void CallMethod(const MethodCall& method_call); + /// Calls a GPU multivalue method. + void CallMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount, + u32 methods_pending); + + /// Flush all current written commands into the host GPU for execution. void FlushCommands(); + /// Synchronizes CPU writes with Host GPU memory. + void SyncGuestHost(); + /// Signal the ending of command list. + virtual void OnCommandListEnd(); + + /// Request a host GPU memory flush from the CPU. + [[nodiscard]] u64 RequestFlush(VAddr addr, std::size_t size); + + /// Obtains current flush request fence id. + [[nodiscard]] u64 CurrentFlushRequestFence() const { + return current_flush_fence.load(std::memory_order_relaxed); + } + + /// Tick pending requests within the GPU. + void TickWork(); /// Returns a reference to the Maxwell3D GPU engine. - Engines::Maxwell3D& Maxwell3D(); + [[nodiscard]] Engines::Maxwell3D& Maxwell3D(); /// Returns a const reference to the Maxwell3D GPU engine. - const Engines::Maxwell3D& Maxwell3D() const; + [[nodiscard]] const Engines::Maxwell3D& Maxwell3D() const; /// Returns a reference to the KeplerCompute GPU engine. - Engines::KeplerCompute& KeplerCompute(); + [[nodiscard]] Engines::KeplerCompute& KeplerCompute(); /// Returns a reference to the KeplerCompute GPU engine. - const Engines::KeplerCompute& KeplerCompute() const; + [[nodiscard]] const Engines::KeplerCompute& KeplerCompute() const; /// Returns a reference to the GPU memory manager. - Tegra::MemoryManager& MemoryManager(); + [[nodiscard]] Tegra::MemoryManager& MemoryManager(); /// Returns a const reference to the GPU memory manager. - const Tegra::MemoryManager& MemoryManager() const; + [[nodiscard]] const Tegra::MemoryManager& MemoryManager() const; /// Returns a reference to the GPU DMA pusher. - Tegra::DmaPusher& DmaPusher(); + [[nodiscard]] Tegra::DmaPusher& DmaPusher(); - VideoCore::RendererBase& Renderer() { + /// Returns a const reference to the GPU DMA pusher. + [[nodiscard]] const Tegra::DmaPusher& DmaPusher() const; + + /// Returns a reference to the GPU CDMA pusher. + [[nodiscard]] Tegra::CDmaPusher& CDmaPusher(); + + /// Returns a const reference to the GPU CDMA pusher. + [[nodiscard]] const Tegra::CDmaPusher& CDmaPusher() const; + + /// Returns a reference to the underlying renderer. + [[nodiscard]] VideoCore::RendererBase& Renderer() { return *renderer; } - const VideoCore::RendererBase& Renderer() const { + /// Returns a const reference to the underlying renderer. + [[nodiscard]] const VideoCore::RendererBase& Renderer() const { return *renderer; } + /// Returns a reference to the shader notifier. + [[nodiscard]] VideoCore::ShaderNotify& ShaderNotify() { + return *shader_notify; + } + + /// Returns a const reference to the shader notifier. + [[nodiscard]] const VideoCore::ShaderNotify& ShaderNotify() const { + return *shader_notify; + } + // Waits for the GPU to finish working virtual void WaitIdle() const = 0; @@ -194,27 +247,46 @@ public: void IncrementSyncPoint(u32 syncpoint_id); - u32 GetSyncpointValue(u32 syncpoint_id) const; + [[nodiscard]] u32 GetSyncpointValue(u32 syncpoint_id) const; void RegisterSyncptInterrupt(u32 syncpoint_id, u32 value); - bool CancelSyncptInterrupt(u32 syncpoint_id, u32 value); + [[nodiscard]] bool CancelSyncptInterrupt(u32 syncpoint_id, u32 value); - u64 GetTicks() const; + [[nodiscard]] u64 GetTicks() const; - std::unique_lock<std::mutex> LockSync() { + [[nodiscard]] std::unique_lock<std::mutex> LockSync() { return std::unique_lock{sync_mutex}; } - bool IsAsync() const { + [[nodiscard]] bool IsAsync() const { return is_async; } - /// Returns a const reference to the GPU DMA pusher. - const Tegra::DmaPusher& DmaPusher() const; + [[nodiscard]] bool UseNvdec() const { + return use_nvdec; + } + + enum class FenceOperation : u32 { + Acquire = 0, + Increment = 1, + }; + + union FenceAction { + u32 raw; + BitField<0, 1, FenceOperation> op; + BitField<8, 24, u32> syncpoint_id; + + [[nodiscard]] static CommandHeader Build(FenceOperation op, u32 syncpoint_id) { + FenceAction result{}; + result.op.Assign(op); + result.syncpoint_id.Assign(syncpoint_id); + return {result.raw}; + } + }; struct Regs { - static constexpr size_t NUM_REGS = 0x100; + static constexpr size_t NUM_REGS = 0x40; union { struct { @@ -223,7 +295,7 @@ public: u32 address_high; u32 address_low; - GPUVAddr SemaphoreAddress() const { + [[nodiscard]] GPUVAddr SemaphoreAddress() const { return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low); } @@ -233,7 +305,7 @@ public: u32 semaphore_trigger; INSERT_UNION_PADDING_WORDS(0xC); - // The puser and the puller share the reference counter, the pusher only has read + // The pusher and the puller share the reference counter, the pusher only has read // access u32 reference_count; INSERT_UNION_PADDING_WORDS(0x5); @@ -241,10 +313,7 @@ public: u32 semaphore_acquire; u32 semaphore_release; u32 fence_value; - union { - BitField<4, 4, u32> operation; - BitField<8, 8, u32> id; - } fence_action; + FenceAction fence_action; INSERT_UNION_PADDING_WORDS(0xE2); // Puller state @@ -263,9 +332,18 @@ public: /// core timing events. virtual void Start() = 0; + /// Obtain the CPU Context + virtual void ObtainContext() = 0; + + /// Release the CPU Context + virtual void ReleaseContext() = 0; + /// Push GPU command entries to be processed virtual void PushGPUEntries(Tegra::CommandList&& entries) = 0; + /// Push GPU command buffer entries to be processed + virtual void PushCommandBuffer(Tegra::ChCommandHeaderList& entries) = 0; + /// Swap buffers (render frame) virtual void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) = 0; @@ -283,6 +361,8 @@ protected: private: void ProcessBindMethod(const MethodCall& method_call); + void ProcessFenceActionMethod(); + void ProcessWaitForInterruptMethod(); void ProcessSemaphoreTriggerMethod(); void ProcessSemaphoreRelease(); void ProcessSemaphoreAcquire(); @@ -293,17 +373,22 @@ private: /// Calls a GPU engine method. void CallEngineMethod(const MethodCall& method_call); + /// Calls a GPU engine multivalue method. + void CallEngineMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount, + u32 methods_pending); + /// Determines where the method should be executed. - bool ExecuteMethodOnEngine(const MethodCall& method_call); + [[nodiscard]] bool ExecuteMethodOnEngine(u32 method); protected: - std::unique_ptr<Tegra::DmaPusher> dma_pusher; Core::System& system; + std::unique_ptr<Tegra::MemoryManager> memory_manager; + std::unique_ptr<Tegra::DmaPusher> dma_pusher; + std::unique_ptr<Tegra::CDmaPusher> cdma_pusher; std::unique_ptr<VideoCore::RendererBase> renderer; + const bool use_nvdec; private: - std::unique_ptr<Tegra::MemoryManager> memory_manager; - /// Mapping of command subchannels to their bound engine ids std::array<EngineID, 8> bound_engines = {}; /// 3D engine @@ -316,15 +401,31 @@ private: std::unique_ptr<Engines::MaxwellDMA> maxwell_dma; /// Inline memory engine std::unique_ptr<Engines::KeplerMemory> kepler_memory; + /// Shader build notifier + std::unique_ptr<VideoCore::ShaderNotify> shader_notify; std::array<std::atomic<u32>, Service::Nvidia::MaxSyncPoints> syncpoints{}; std::array<std::list<u32>, Service::Nvidia::MaxSyncPoints> syncpt_interrupts; std::mutex sync_mutex; + std::mutex device_mutex; std::condition_variable sync_cv; + struct FlushRequest { + FlushRequest(u64 fence, VAddr addr, std::size_t size) + : fence{fence}, addr{addr}, size{size} {} + u64 fence; + VAddr addr; + std::size_t size; + }; + + std::list<FlushRequest> flush_requests; + std::atomic<u64> current_flush_fence{}; + u64 last_flush_fence{}; + std::mutex flush_request_mutex; + const bool is_async; }; diff --git a/src/video_core/gpu_asynch.cpp b/src/video_core/gpu_asynch.cpp index 20e73a37e..a9baaf7ef 100644 --- a/src/video_core/gpu_asynch.cpp +++ b/src/video_core/gpu_asynch.cpp @@ -10,23 +10,50 @@ namespace VideoCommon { -GPUAsynch::GPUAsynch(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& renderer_, - std::unique_ptr<Core::Frontend::GraphicsContext>&& context) - : GPU(system, std::move(renderer_), true), gpu_thread{system}, - cpu_context(renderer->GetRenderWindow().CreateSharedContext()), - gpu_context(std::move(context)) {} +GPUAsynch::GPUAsynch(Core::System& system, bool use_nvdec) + : GPU{system, true, use_nvdec}, gpu_thread{system} {} GPUAsynch::~GPUAsynch() = default; void GPUAsynch::Start() { + gpu_thread.StartThread(*renderer, renderer->Context(), *dma_pusher, *cdma_pusher); + cpu_context = renderer->GetRenderWindow().CreateSharedContext(); cpu_context->MakeCurrent(); - gpu_thread.StartThread(*renderer, *gpu_context, *dma_pusher); +} + +void GPUAsynch::ObtainContext() { + cpu_context->MakeCurrent(); +} + +void GPUAsynch::ReleaseContext() { + cpu_context->DoneCurrent(); } void GPUAsynch::PushGPUEntries(Tegra::CommandList&& entries) { gpu_thread.SubmitList(std::move(entries)); } +void GPUAsynch::PushCommandBuffer(Tegra::ChCommandHeaderList& entries) { + if (!use_nvdec) { + return; + } + // This condition fires when a video stream ends, clear all intermediary data + if (entries[0].raw == 0xDEADB33F) { + cdma_pusher.reset(); + return; + } + if (!cdma_pusher) { + cdma_pusher = std::make_unique<Tegra::CDmaPusher>(*this); + } + + // SubmitCommandBuffer would make the nvdec operations async, this is not currently working + // TODO(ameerj): RE proper async nvdec operation + // gpu_thread.SubmitCommandBuffer(std::move(entries)); + + cdma_pusher->Push(std::move(entries)); + cdma_pusher->DispatchCalls(); +} + void GPUAsynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { gpu_thread.SwapBuffers(framebuffer); } @@ -52,4 +79,8 @@ void GPUAsynch::WaitIdle() const { gpu_thread.WaitIdle(); } +void GPUAsynch::OnCommandListEnd() { + gpu_thread.OnCommandListEnd(); +} + } // namespace VideoCommon diff --git a/src/video_core/gpu_asynch.h b/src/video_core/gpu_asynch.h index 03fd0eef0..0c0872e73 100644 --- a/src/video_core/gpu_asynch.h +++ b/src/video_core/gpu_asynch.h @@ -20,25 +20,28 @@ namespace VideoCommon { /// Implementation of GPU interface that runs the GPU asynchronously class GPUAsynch final : public Tegra::GPU { public: - explicit GPUAsynch(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& renderer, - std::unique_ptr<Core::Frontend::GraphicsContext>&& context); + explicit GPUAsynch(Core::System& system, bool use_nvdec); ~GPUAsynch() override; void Start() override; + void ObtainContext() override; + void ReleaseContext() override; void PushGPUEntries(Tegra::CommandList&& entries) override; + void PushCommandBuffer(Tegra::ChCommandHeaderList& entries) override; void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override; void FlushRegion(VAddr addr, u64 size) override; void InvalidateRegion(VAddr addr, u64 size) override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; void WaitIdle() const override; + void OnCommandListEnd() override; + protected: void TriggerCpuInterrupt(u32 syncpoint_id, u32 value) const override; private: GPUThread::ThreadManager gpu_thread; std::unique_ptr<Core::Frontend::GraphicsContext> cpu_context; - std::unique_ptr<Core::Frontend::GraphicsContext> gpu_context; }; } // namespace VideoCommon diff --git a/src/video_core/gpu_synch.cpp b/src/video_core/gpu_synch.cpp index 6f38a672a..ecf7bbdf3 100644 --- a/src/video_core/gpu_synch.cpp +++ b/src/video_core/gpu_synch.cpp @@ -7,14 +7,18 @@ namespace VideoCommon { -GPUSynch::GPUSynch(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& renderer, - std::unique_ptr<Core::Frontend::GraphicsContext>&& context) - : GPU(system, std::move(renderer), false), context{std::move(context)} {} +GPUSynch::GPUSynch(Core::System& system, bool use_nvdec) : GPU{system, false, use_nvdec} {} GPUSynch::~GPUSynch() = default; -void GPUSynch::Start() { - context->MakeCurrent(); +void GPUSynch::Start() {} + +void GPUSynch::ObtainContext() { + renderer->Context().MakeCurrent(); +} + +void GPUSynch::ReleaseContext() { + renderer->Context().DoneCurrent(); } void GPUSynch::PushGPUEntries(Tegra::CommandList&& entries) { @@ -22,6 +26,22 @@ void GPUSynch::PushGPUEntries(Tegra::CommandList&& entries) { dma_pusher->DispatchCalls(); } +void GPUSynch::PushCommandBuffer(Tegra::ChCommandHeaderList& entries) { + if (!use_nvdec) { + return; + } + // This condition fires when a video stream ends, clears all intermediary data + if (entries[0].raw == 0xDEADB33F) { + cdma_pusher.reset(); + return; + } + if (!cdma_pusher) { + cdma_pusher = std::make_unique<Tegra::CDmaPusher>(*this); + } + cdma_pusher->Push(std::move(entries)); + cdma_pusher->DispatchCalls(); +} + void GPUSynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { renderer->SwapBuffers(framebuffer); } diff --git a/src/video_core/gpu_synch.h b/src/video_core/gpu_synch.h index 4a6e9a01d..9d778c71a 100644 --- a/src/video_core/gpu_synch.h +++ b/src/video_core/gpu_synch.h @@ -19,12 +19,14 @@ namespace VideoCommon { /// Implementation of GPU interface that runs the GPU synchronously class GPUSynch final : public Tegra::GPU { public: - explicit GPUSynch(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& renderer, - std::unique_ptr<Core::Frontend::GraphicsContext>&& context); + explicit GPUSynch(Core::System& system, bool use_nvdec); ~GPUSynch() override; void Start() override; + void ObtainContext() override; + void ReleaseContext() override; void PushGPUEntries(Tegra::CommandList&& entries) override; + void PushCommandBuffer(Tegra::ChCommandHeaderList& entries) override; void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override; void FlushRegion(VAddr addr, u64 size) override; void InvalidateRegion(VAddr addr, u64 size) override; @@ -34,9 +36,6 @@ public: protected: void TriggerCpuInterrupt([[maybe_unused]] u32 syncpoint_id, [[maybe_unused]] u32 value) const override {} - -private: - std::unique_ptr<Core::Frontend::GraphicsContext> context; }; } // namespace VideoCommon diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp index 10cda686b..4b8f58283 100644 --- a/src/video_core/gpu_thread.cpp +++ b/src/video_core/gpu_thread.cpp @@ -4,8 +4,10 @@ #include "common/assert.h" #include "common/microprofile.h" +#include "common/thread.h" #include "core/core.h" #include "core/frontend/emu_window.h" +#include "core/settings.h" #include "video_core/dma_pusher.h" #include "video_core/gpu.h" #include "video_core/gpu_thread.h" @@ -14,9 +16,14 @@ namespace VideoCommon::GPUThread { /// Runs the GPU thread -static void RunThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, - Tegra::DmaPusher& dma_pusher, SynchState& state) { - MicroProfileOnThreadCreate("GpuThread"); +static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, + Core::Frontend::GraphicsContext& context, Tegra::DmaPusher& dma_pusher, + SynchState& state, Tegra::CDmaPusher& cdma_pusher) { + std::string name = "yuzu:GPU"; + MicroProfileOnThreadCreate(name.c_str()); + Common::SetCurrentThreadName(name.c_str()); + Common::SetCurrentThreadPriority(Common::ThreadPriority::High); + system.RegisterHostThread(); // Wait for first GPU command before acquiring the window context while (state.queue.Empty()) @@ -35,12 +42,20 @@ static void RunThread(VideoCore::RendererBase& renderer, Core::Frontend::Graphic if (const auto submit_list = std::get_if<SubmitListCommand>(&next.data)) { dma_pusher.Push(std::move(submit_list->entries)); dma_pusher.DispatchCalls(); + } else if (const auto command_list = std::get_if<SubmitChCommandEntries>(&next.data)) { + // NVDEC + cdma_pusher.Push(std::move(command_list->entries)); + cdma_pusher.DispatchCalls(); } else if (const auto data = std::get_if<SwapBuffersCommand>(&next.data)) { renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr); + } else if (std::holds_alternative<OnCommandListEndCommand>(next.data)) { + renderer.Rasterizer().ReleaseFences(); + } else if (std::holds_alternative<GPUTickCommand>(next.data)) { + system.GPU().TickWork(); } else if (const auto data = std::get_if<FlushRegionCommand>(&next.data)) { renderer.Rasterizer().FlushRegion(data->addr, data->size); } else if (const auto data = std::get_if<InvalidateRegionCommand>(&next.data)) { - renderer.Rasterizer().InvalidateRegion(data->addr, data->size); + renderer.Rasterizer().OnCPUWrite(data->addr, data->size); } else if (std::holds_alternative<EndProcessingCommand>(next.data)) { return; } else { @@ -64,30 +79,47 @@ ThreadManager::~ThreadManager() { void ThreadManager::StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, - Tegra::DmaPusher& dma_pusher) { - thread = std::thread{RunThread, std::ref(renderer), std::ref(context), std::ref(dma_pusher), - std::ref(state)}; + Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher) { + thread = std::thread(RunThread, std::ref(system), std::ref(renderer), std::ref(context), + std::ref(dma_pusher), std::ref(state), std::ref(cdma_pusher)); } void ThreadManager::SubmitList(Tegra::CommandList&& entries) { PushCommand(SubmitListCommand(std::move(entries))); } +void ThreadManager::SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries) { + PushCommand(SubmitChCommandEntries(std::move(entries))); +} + void ThreadManager::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { PushCommand(SwapBuffersCommand(framebuffer ? std::make_optional(*framebuffer) : std::nullopt)); } void ThreadManager::FlushRegion(VAddr addr, u64 size) { - PushCommand(FlushRegionCommand(addr, size)); + if (!Settings::IsGPULevelHigh()) { + PushCommand(FlushRegionCommand(addr, size)); + return; + } + if (!Settings::IsGPULevelExtreme()) { + return; + } + if (system.Renderer().Rasterizer().MustFlushRegion(addr, size)) { + auto& gpu = system.GPU(); + u64 fence = gpu.RequestFlush(addr, size); + PushCommand(GPUTickCommand()); + while (fence > gpu.CurrentFlushRequestFence()) { + } + } } void ThreadManager::InvalidateRegion(VAddr addr, u64 size) { - system.Renderer().Rasterizer().InvalidateRegion(addr, size); + system.Renderer().Rasterizer().OnCPUWrite(addr, size); } void ThreadManager::FlushAndInvalidateRegion(VAddr addr, u64 size) { // Skip flush on asynch mode, as FlushAndInvalidateRegion is not used for anything too important - InvalidateRegion(addr, size); + system.Renderer().Rasterizer().OnCPUWrite(addr, size); } void ThreadManager::WaitIdle() const { @@ -95,6 +127,10 @@ void ThreadManager::WaitIdle() const { } } +void ThreadManager::OnCommandListEnd() { + PushCommand(OnCommandListEndCommand()); +} + u64 ThreadManager::PushCommand(CommandData&& command_data) { const u64 fence{++state.last_fence}; state.queue.Push(CommandDataContainer(std::move(command_data), fence)); diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index cd74ad330..32a34e3a7 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -37,6 +37,14 @@ struct SubmitListCommand final { Tegra::CommandList entries; }; +/// Command to signal to the GPU thread that a cdma command list is ready for processing +struct SubmitChCommandEntries final { + explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries) + : entries{std::move(entries)} {} + + Tegra::ChCommandHeaderList entries; +}; + /// Command to signal to the GPU thread that a swap buffers is pending struct SwapBuffersCommand final { explicit SwapBuffersCommand(std::optional<const Tegra::FramebufferConfig> framebuffer) @@ -70,9 +78,16 @@ struct FlushAndInvalidateRegionCommand final { u64 size; }; +/// Command called within the gpu, to schedule actions after a command list end +struct OnCommandListEndCommand final {}; + +/// Command to make the gpu look into pending requests +struct GPUTickCommand final {}; + using CommandData = - std::variant<EndProcessingCommand, SubmitListCommand, SwapBuffersCommand, FlushRegionCommand, - InvalidateRegionCommand, FlushAndInvalidateRegionCommand>; + std::variant<EndProcessingCommand, SubmitListCommand, SubmitChCommandEntries, + SwapBuffersCommand, FlushRegionCommand, InvalidateRegionCommand, + FlushAndInvalidateRegionCommand, OnCommandListEndCommand, GPUTickCommand>; struct CommandDataContainer { CommandDataContainer() = default; @@ -102,11 +117,14 @@ public: /// Creates and starts the GPU thread. void StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, - Tegra::DmaPusher& dma_pusher); + Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher); /// Push GPU command entries to be processed void SubmitList(Tegra::CommandList&& entries); + /// Push GPU CDMA command buffer entries to be processed + void SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries); + /// Swap buffers (render frame) void SwapBuffers(const Tegra::FramebufferConfig* framebuffer); @@ -122,6 +140,8 @@ public: // Wait until the gpu thread is idle. void WaitIdle() const; + void OnCommandListEnd(); + private: /// Pushes a command to be executed by the GPU thread u64 PushCommand(CommandData&& command_data); diff --git a/src/video_core/host_shaders/CMakeLists.txt b/src/video_core/host_shaders/CMakeLists.txt new file mode 100644 index 000000000..c157724a9 --- /dev/null +++ b/src/video_core/host_shaders/CMakeLists.txt @@ -0,0 +1,36 @@ +set(SHADER_SOURCES + opengl_present.frag + opengl_present.vert +) + +set(SHADER_INCLUDE ${CMAKE_CURRENT_BINARY_DIR}/include) +set(SHADER_DIR ${SHADER_INCLUDE}/video_core/host_shaders) +set(HOST_SHADERS_INCLUDE ${SHADER_INCLUDE} PARENT_SCOPE) + +set(INPUT_FILE ${CMAKE_CURRENT_SOURCE_DIR}/source_shader.h.in) +set(HEADER_GENERATOR ${CMAKE_CURRENT_SOURCE_DIR}/StringShaderHeader.cmake) + +foreach(FILENAME IN ITEMS ${SHADER_SOURCES}) + string(REPLACE "." "_" SHADER_NAME ${FILENAME}) + set(SOURCE_FILE ${CMAKE_CURRENT_SOURCE_DIR}/${FILENAME}) + set(HEADER_FILE ${SHADER_DIR}/${SHADER_NAME}.h) + add_custom_command( + OUTPUT + ${HEADER_FILE} + COMMAND + ${CMAKE_COMMAND} -P ${HEADER_GENERATOR} ${SOURCE_FILE} ${HEADER_FILE} ${INPUT_FILE} + MAIN_DEPENDENCY + ${SOURCE_FILE} + DEPENDS + ${INPUT_FILE} + # HEADER_GENERATOR should be included here but msbuild seems to assume it's always modified + ) + set(SHADER_HEADERS ${SHADER_HEADERS} ${HEADER_FILE}) +endforeach() + +add_custom_target(host_shaders + DEPENDS + ${SHADER_HEADERS} + SOURCES + ${SHADER_SOURCES} +) diff --git a/src/video_core/host_shaders/StringShaderHeader.cmake b/src/video_core/host_shaders/StringShaderHeader.cmake new file mode 100644 index 000000000..c0fc49768 --- /dev/null +++ b/src/video_core/host_shaders/StringShaderHeader.cmake @@ -0,0 +1,13 @@ +set(SOURCE_FILE ${CMAKE_ARGV3}) +set(HEADER_FILE ${CMAKE_ARGV4}) +set(INPUT_FILE ${CMAKE_ARGV5}) + +get_filename_component(CONTENTS_NAME ${SOURCE_FILE} NAME) +string(REPLACE "." "_" CONTENTS_NAME ${CONTENTS_NAME}) +string(TOUPPER ${CONTENTS_NAME} CONTENTS_NAME) + +file(READ ${SOURCE_FILE} CONTENTS) + +get_filename_component(OUTPUT_DIR ${HEADER_FILE} DIRECTORY) +make_directory(${OUTPUT_DIR}) +configure_file(${INPUT_FILE} ${HEADER_FILE} @ONLY) diff --git a/src/video_core/host_shaders/opengl_present.frag b/src/video_core/host_shaders/opengl_present.frag new file mode 100644 index 000000000..8a4cb024b --- /dev/null +++ b/src/video_core/host_shaders/opengl_present.frag @@ -0,0 +1,10 @@ +#version 430 core + +layout (location = 0) in vec2 frag_tex_coord; +layout (location = 0) out vec4 color; + +layout (binding = 0) uniform sampler2D color_texture; + +void main() { + color = vec4(texture(color_texture, frag_tex_coord).rgb, 1.0f); +} diff --git a/src/video_core/host_shaders/opengl_present.vert b/src/video_core/host_shaders/opengl_present.vert new file mode 100644 index 000000000..2235d31a4 --- /dev/null +++ b/src/video_core/host_shaders/opengl_present.vert @@ -0,0 +1,24 @@ +#version 430 core + +out gl_PerVertex { + vec4 gl_Position; +}; + +layout (location = 0) in vec2 vert_position; +layout (location = 1) in vec2 vert_tex_coord; +layout (location = 0) out vec2 frag_tex_coord; + +// This is a truncated 3x3 matrix for 2D transformations: +// The upper-left 2x2 submatrix performs scaling/rotation/mirroring. +// The third column performs translation. +// The third row could be used for projection, which we don't need in 2D. It hence is assumed to +// implicitly be [0, 0, 1] +layout (location = 0) uniform mat3x2 modelview_matrix; + +void main() { + // Multiply input position by the rotscale part of the matrix and then manually translate by + // the last column. This is equivalent to using a full 3x3 matrix and expanding the vector + // to `vec3(vert_position.xy, 1.0)` + gl_Position = vec4(mat2(modelview_matrix) * vert_position + modelview_matrix[2], 0.0, 1.0); + frag_tex_coord = vert_tex_coord; +} diff --git a/src/video_core/host_shaders/source_shader.h.in b/src/video_core/host_shaders/source_shader.h.in new file mode 100644 index 000000000..ccdb0d2a9 --- /dev/null +++ b/src/video_core/host_shaders/source_shader.h.in @@ -0,0 +1,9 @@ +#pragma once + +#include <string_view> + +namespace HostShaders { + +constexpr std::string_view @CONTENTS_NAME@ = R"(@CONTENTS@)"; + +} // namespace HostShaders diff --git a/src/video_core/macro/macro.cpp b/src/video_core/macro/macro.cpp new file mode 100644 index 000000000..cd21a2112 --- /dev/null +++ b/src/video_core/macro/macro.cpp @@ -0,0 +1,91 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <optional> +#include <boost/container_hash/hash.hpp> +#include "common/assert.h" +#include "common/logging/log.h" +#include "core/settings.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/macro/macro.h" +#include "video_core/macro/macro_hle.h" +#include "video_core/macro/macro_interpreter.h" +#include "video_core/macro/macro_jit_x64.h" + +namespace Tegra { + +MacroEngine::MacroEngine(Engines::Maxwell3D& maxwell3d) + : hle_macros{std::make_unique<Tegra::HLEMacro>(maxwell3d)} {} + +MacroEngine::~MacroEngine() = default; + +void MacroEngine::AddCode(u32 method, u32 data) { + uploaded_macro_code[method].push_back(data); +} + +void MacroEngine::Execute(Engines::Maxwell3D& maxwell3d, u32 method, + const std::vector<u32>& parameters) { + auto compiled_macro = macro_cache.find(method); + if (compiled_macro != macro_cache.end()) { + const auto& cache_info = compiled_macro->second; + if (cache_info.has_hle_program) { + cache_info.hle_program->Execute(parameters, method); + } else { + cache_info.lle_program->Execute(parameters, method); + } + } else { + // Macro not compiled, check if it's uploaded and if so, compile it + std::optional<u32> mid_method; + const auto macro_code = uploaded_macro_code.find(method); + if (macro_code == uploaded_macro_code.end()) { + for (const auto& [method_base, code] : uploaded_macro_code) { + if (method >= method_base && (method - method_base) < code.size()) { + mid_method = method_base; + break; + } + } + if (!mid_method.has_value()) { + UNREACHABLE_MSG("Macro 0x{0:x} was not uploaded", method); + return; + } + } + auto& cache_info = macro_cache[method]; + + if (!mid_method.has_value()) { + cache_info.lle_program = Compile(macro_code->second); + cache_info.hash = boost::hash_value(macro_code->second); + } else { + const auto& macro_cached = uploaded_macro_code[mid_method.value()]; + const auto rebased_method = method - mid_method.value(); + auto& code = uploaded_macro_code[method]; + code.resize(macro_cached.size() - rebased_method); + std::memcpy(code.data(), macro_cached.data() + rebased_method, + code.size() * sizeof(u32)); + cache_info.hash = boost::hash_value(code); + cache_info.lle_program = Compile(code); + } + + auto hle_program = hle_macros->GetHLEProgram(cache_info.hash); + if (hle_program.has_value()) { + cache_info.has_hle_program = true; + cache_info.hle_program = std::move(hle_program.value()); + cache_info.hle_program->Execute(parameters, method); + } else { + cache_info.lle_program->Execute(parameters, method); + } + } +} + +std::unique_ptr<MacroEngine> GetMacroEngine(Engines::Maxwell3D& maxwell3d) { + if (Settings::values.disable_macro_jit) { + return std::make_unique<MacroInterpreter>(maxwell3d); + } +#ifdef ARCHITECTURE_x86_64 + return std::make_unique<MacroJITx64>(maxwell3d); +#else + return std::make_unique<MacroInterpreter>(maxwell3d); +#endif +} + +} // namespace Tegra diff --git a/src/video_core/macro/macro.h b/src/video_core/macro/macro.h new file mode 100644 index 000000000..31ee3440a --- /dev/null +++ b/src/video_core/macro/macro.h @@ -0,0 +1,142 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <unordered_map> +#include <vector> +#include "common/bit_field.h" +#include "common/common_types.h" + +namespace Tegra { + +namespace Engines { +class Maxwell3D; +} + +namespace Macro { +constexpr std::size_t NUM_MACRO_REGISTERS = 8; +enum class Operation : u32 { + ALU = 0, + AddImmediate = 1, + ExtractInsert = 2, + ExtractShiftLeftImmediate = 3, + ExtractShiftLeftRegister = 4, + Read = 5, + Unused = 6, // This operation doesn't seem to be a valid encoding. + Branch = 7, +}; + +enum class ALUOperation : u32 { + Add = 0, + AddWithCarry = 1, + Subtract = 2, + SubtractWithBorrow = 3, + // Operations 4-7 don't seem to be valid encodings. + Xor = 8, + Or = 9, + And = 10, + AndNot = 11, + Nand = 12 +}; + +enum class ResultOperation : u32 { + IgnoreAndFetch = 0, + Move = 1, + MoveAndSetMethod = 2, + FetchAndSend = 3, + MoveAndSend = 4, + FetchAndSetMethod = 5, + MoveAndSetMethodFetchAndSend = 6, + MoveAndSetMethodSend = 7 +}; + +enum class BranchCondition : u32 { + Zero = 0, + NotZero = 1, +}; + +union Opcode { + u32 raw; + BitField<0, 3, Operation> operation; + BitField<4, 3, ResultOperation> result_operation; + BitField<4, 1, BranchCondition> branch_condition; + // If set on a branch, then the branch doesn't have a delay slot. + BitField<5, 1, u32> branch_annul; + BitField<7, 1, u32> is_exit; + BitField<8, 3, u32> dst; + BitField<11, 3, u32> src_a; + BitField<14, 3, u32> src_b; + // The signed immediate overlaps the second source operand and the alu operation. + BitField<14, 18, s32> immediate; + + BitField<17, 5, ALUOperation> alu_operation; + + // Bitfield instructions data + BitField<17, 5, u32> bf_src_bit; + BitField<22, 5, u32> bf_size; + BitField<27, 5, u32> bf_dst_bit; + + u32 GetBitfieldMask() const { + return (1 << bf_size) - 1; + } + + s32 GetBranchTarget() const { + return static_cast<s32>(immediate * sizeof(u32)); + } +}; + +union MethodAddress { + u32 raw; + BitField<0, 12, u32> address; + BitField<12, 6, u32> increment; +}; + +} // namespace Macro + +class HLEMacro; + +class CachedMacro { +public: + virtual ~CachedMacro() = default; + /** + * Executes the macro code with the specified input parameters. + * + * @param parameters The parameters of the macro + * @param method The method to execute + */ + virtual void Execute(const std::vector<u32>& parameters, u32 method) = 0; +}; + +class MacroEngine { +public: + explicit MacroEngine(Engines::Maxwell3D& maxwell3d); + virtual ~MacroEngine(); + + // Store the uploaded macro code to compile them when they're called. + void AddCode(u32 method, u32 data); + + // Compiles the macro if its not in the cache, and executes the compiled macro + void Execute(Engines::Maxwell3D& maxwell3d, u32 method, const std::vector<u32>& parameters); + +protected: + virtual std::unique_ptr<CachedMacro> Compile(const std::vector<u32>& code) = 0; + +private: + struct CacheInfo { + std::unique_ptr<CachedMacro> lle_program{}; + std::unique_ptr<CachedMacro> hle_program{}; + u64 hash{}; + bool has_hle_program{}; + }; + + std::unordered_map<u32, CacheInfo> macro_cache; + std::unordered_map<u32, std::vector<u32>> uploaded_macro_code; + std::unique_ptr<HLEMacro> hle_macros; +}; + +std::unique_ptr<MacroEngine> GetMacroEngine(Engines::Maxwell3D& maxwell3d); + +} // namespace Tegra diff --git a/src/video_core/macro/macro_hle.cpp b/src/video_core/macro/macro_hle.cpp new file mode 100644 index 000000000..df00b57df --- /dev/null +++ b/src/video_core/macro/macro_hle.cpp @@ -0,0 +1,109 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> +#include <vector> +#include "video_core/engines/maxwell_3d.h" +#include "video_core/macro/macro_hle.h" +#include "video_core/rasterizer_interface.h" + +namespace Tegra { + +namespace { +// HLE'd functions +void HLE_771BB18C62444DA0(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) { + const u32 instance_count = parameters[2] & maxwell3d.GetRegisterValue(0xD1B); + + maxwell3d.regs.draw.topology.Assign( + static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0] & 0x3ffffff)); + maxwell3d.regs.vb_base_instance = parameters[5]; + maxwell3d.mme_draw.instance_count = instance_count; + maxwell3d.regs.vb_element_base = parameters[3]; + maxwell3d.regs.index_array.count = parameters[1]; + maxwell3d.regs.index_array.first = parameters[4]; + + if (maxwell3d.ShouldExecute()) { + maxwell3d.Rasterizer().Draw(true, true); + } + maxwell3d.regs.index_array.count = 0; + maxwell3d.mme_draw.instance_count = 0; + maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined; +} + +void HLE_0D61FC9FAAC9FCAD(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) { + const u32 count = (maxwell3d.GetRegisterValue(0xD1B) & parameters[2]); + + maxwell3d.regs.vertex_buffer.first = parameters[3]; + maxwell3d.regs.vertex_buffer.count = parameters[1]; + maxwell3d.regs.vb_base_instance = parameters[4]; + maxwell3d.regs.draw.topology.Assign( + static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0])); + maxwell3d.mme_draw.instance_count = count; + + if (maxwell3d.ShouldExecute()) { + maxwell3d.Rasterizer().Draw(false, true); + } + maxwell3d.regs.vertex_buffer.count = 0; + maxwell3d.mme_draw.instance_count = 0; + maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined; +} + +void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) { + const u32 instance_count = (maxwell3d.GetRegisterValue(0xD1B) & parameters[2]); + const u32 element_base = parameters[4]; + const u32 base_instance = parameters[5]; + maxwell3d.regs.index_array.first = parameters[3]; + maxwell3d.regs.reg_array[0x446] = element_base; // vertex id base? + maxwell3d.regs.index_array.count = parameters[1]; + maxwell3d.regs.vb_element_base = element_base; + maxwell3d.regs.vb_base_instance = base_instance; + maxwell3d.mme_draw.instance_count = instance_count; + maxwell3d.CallMethodFromMME(0x8e3, 0x640); + maxwell3d.CallMethodFromMME(0x8e4, element_base); + maxwell3d.CallMethodFromMME(0x8e5, base_instance); + maxwell3d.regs.draw.topology.Assign( + static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0])); + if (maxwell3d.ShouldExecute()) { + maxwell3d.Rasterizer().Draw(true, true); + } + maxwell3d.regs.reg_array[0x446] = 0x0; // vertex id base? + maxwell3d.regs.index_array.count = 0; + maxwell3d.regs.vb_element_base = 0x0; + maxwell3d.regs.vb_base_instance = 0x0; + maxwell3d.mme_draw.instance_count = 0; + maxwell3d.CallMethodFromMME(0x8e3, 0x640); + maxwell3d.CallMethodFromMME(0x8e4, 0x0); + maxwell3d.CallMethodFromMME(0x8e5, 0x0); + maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined; +} +} // Anonymous namespace + +constexpr std::array<std::pair<u64, HLEFunction>, 3> hle_funcs{{ + {0x771BB18C62444DA0, &HLE_771BB18C62444DA0}, + {0x0D61FC9FAAC9FCAD, &HLE_0D61FC9FAAC9FCAD}, + {0x0217920100488FF7, &HLE_0217920100488FF7}, +}}; + +HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {} +HLEMacro::~HLEMacro() = default; + +std::optional<std::unique_ptr<CachedMacro>> HLEMacro::GetHLEProgram(u64 hash) const { + const auto it = std::find_if(hle_funcs.cbegin(), hle_funcs.cend(), + [hash](const auto& pair) { return pair.first == hash; }); + if (it == hle_funcs.end()) { + return std::nullopt; + } + return std::make_unique<HLEMacroImpl>(maxwell3d, it->second); +} + +HLEMacroImpl::~HLEMacroImpl() = default; + +HLEMacroImpl::HLEMacroImpl(Engines::Maxwell3D& maxwell3d, HLEFunction func) + : maxwell3d(maxwell3d), func(func) {} + +void HLEMacroImpl::Execute(const std::vector<u32>& parameters, u32 method) { + func(maxwell3d, parameters); +} + +} // namespace Tegra diff --git a/src/video_core/macro/macro_hle.h b/src/video_core/macro/macro_hle.h new file mode 100644 index 000000000..37af875a0 --- /dev/null +++ b/src/video_core/macro/macro_hle.h @@ -0,0 +1,44 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <optional> +#include <vector> +#include "common/common_types.h" +#include "video_core/macro/macro.h" + +namespace Tegra { + +namespace Engines { +class Maxwell3D; +} + +using HLEFunction = void (*)(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters); + +class HLEMacro { +public: + explicit HLEMacro(Engines::Maxwell3D& maxwell3d); + ~HLEMacro(); + + std::optional<std::unique_ptr<CachedMacro>> GetHLEProgram(u64 hash) const; + +private: + Engines::Maxwell3D& maxwell3d; +}; + +class HLEMacroImpl : public CachedMacro { +public: + explicit HLEMacroImpl(Engines::Maxwell3D& maxwell3d, HLEFunction func); + ~HLEMacroImpl(); + + void Execute(const std::vector<u32>& parameters, u32 method) override; + +private: + Engines::Maxwell3D& maxwell3d; + HLEFunction func; +}; + +} // namespace Tegra diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro/macro_interpreter.cpp index 42031d80a..bd01fd1f2 100644 --- a/src/video_core/macro_interpreter.cpp +++ b/src/video_core/macro/macro_interpreter.cpp @@ -1,4 +1,4 @@ -// Copyright 2018 yuzu Emulator Project +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -6,109 +6,46 @@ #include "common/logging/log.h" #include "common/microprofile.h" #include "video_core/engines/maxwell_3d.h" -#include "video_core/macro_interpreter.h" +#include "video_core/macro/macro_interpreter.h" MICROPROFILE_DEFINE(MacroInterp, "GPU", "Execute macro interpreter", MP_RGB(128, 128, 192)); namespace Tegra { -namespace { -enum class Operation : u32 { - ALU = 0, - AddImmediate = 1, - ExtractInsert = 2, - ExtractShiftLeftImmediate = 3, - ExtractShiftLeftRegister = 4, - Read = 5, - Unused = 6, // This operation doesn't seem to be a valid encoding. - Branch = 7, -}; -} // Anonymous namespace - -enum class MacroInterpreter::ALUOperation : u32 { - Add = 0, - AddWithCarry = 1, - Subtract = 2, - SubtractWithBorrow = 3, - // Operations 4-7 don't seem to be valid encodings. - Xor = 8, - Or = 9, - And = 10, - AndNot = 11, - Nand = 12 -}; - -enum class MacroInterpreter::ResultOperation : u32 { - IgnoreAndFetch = 0, - Move = 1, - MoveAndSetMethod = 2, - FetchAndSend = 3, - MoveAndSend = 4, - FetchAndSetMethod = 5, - MoveAndSetMethodFetchAndSend = 6, - MoveAndSetMethodSend = 7 -}; - -enum class MacroInterpreter::BranchCondition : u32 { - Zero = 0, - NotZero = 1, -}; - -union MacroInterpreter::Opcode { - u32 raw; - BitField<0, 3, Operation> operation; - BitField<4, 3, ResultOperation> result_operation; - BitField<4, 1, BranchCondition> branch_condition; - // If set on a branch, then the branch doesn't have a delay slot. - BitField<5, 1, u32> branch_annul; - BitField<7, 1, u32> is_exit; - BitField<8, 3, u32> dst; - BitField<11, 3, u32> src_a; - BitField<14, 3, u32> src_b; - // The signed immediate overlaps the second source operand and the alu operation. - BitField<14, 18, s32> immediate; - - BitField<17, 5, ALUOperation> alu_operation; - - // Bitfield instructions data - BitField<17, 5, u32> bf_src_bit; - BitField<22, 5, u32> bf_size; - BitField<27, 5, u32> bf_dst_bit; - - u32 GetBitfieldMask() const { - return (1 << bf_size) - 1; - } +MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) + : MacroEngine::MacroEngine(maxwell3d), maxwell3d(maxwell3d) {} - s32 GetBranchTarget() const { - return static_cast<s32>(immediate * sizeof(u32)); - } -}; +std::unique_ptr<CachedMacro> MacroInterpreter::Compile(const std::vector<u32>& code) { + return std::make_unique<MacroInterpreterImpl>(maxwell3d, code); +} -MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {} +MacroInterpreterImpl::MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d, + const std::vector<u32>& code) + : maxwell3d(maxwell3d), code(code) {} -void MacroInterpreter::Execute(u32 offset, std::size_t num_parameters, const u32* parameters) { +void MacroInterpreterImpl::Execute(const std::vector<u32>& parameters, u32 method) { MICROPROFILE_SCOPE(MacroInterp); Reset(); registers[1] = parameters[0]; + num_parameters = parameters.size(); if (num_parameters > parameters_capacity) { parameters_capacity = num_parameters; this->parameters = std::make_unique<u32[]>(num_parameters); } - std::memcpy(this->parameters.get(), parameters, num_parameters * sizeof(u32)); - this->num_parameters = num_parameters; + std::memcpy(this->parameters.get(), parameters.data(), num_parameters * sizeof(u32)); // Execute the code until we hit an exit condition. bool keep_executing = true; while (keep_executing) { - keep_executing = Step(offset, false); + keep_executing = Step(false); } // Assert the the macro used all the input parameters ASSERT(next_parameter_index == num_parameters); } -void MacroInterpreter::Reset() { +void MacroInterpreterImpl::Reset() { registers = {}; pc = 0; delayed_pc = {}; @@ -120,10 +57,10 @@ void MacroInterpreter::Reset() { carry_flag = false; } -bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { +bool MacroInterpreterImpl::Step(bool is_delay_slot) { u32 base_address = pc; - Opcode opcode = GetOpcode(offset); + Macro::Opcode opcode = GetOpcode(); pc += 4; // Update the program counter if we were delayed @@ -134,18 +71,18 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { } switch (opcode.operation) { - case Operation::ALU: { + case Macro::Operation::ALU: { u32 result = GetALUResult(opcode.alu_operation, GetRegister(opcode.src_a), GetRegister(opcode.src_b)); ProcessResult(opcode.result_operation, opcode.dst, result); break; } - case Operation::AddImmediate: { + case Macro::Operation::AddImmediate: { ProcessResult(opcode.result_operation, opcode.dst, GetRegister(opcode.src_a) + opcode.immediate); break; } - case Operation::ExtractInsert: { + case Macro::Operation::ExtractInsert: { u32 dst = GetRegister(opcode.src_a); u32 src = GetRegister(opcode.src_b); @@ -155,7 +92,7 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { ProcessResult(opcode.result_operation, opcode.dst, dst); break; } - case Operation::ExtractShiftLeftImmediate: { + case Macro::Operation::ExtractShiftLeftImmediate: { u32 dst = GetRegister(opcode.src_a); u32 src = GetRegister(opcode.src_b); @@ -164,7 +101,7 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { ProcessResult(opcode.result_operation, opcode.dst, result); break; } - case Operation::ExtractShiftLeftRegister: { + case Macro::Operation::ExtractShiftLeftRegister: { u32 dst = GetRegister(opcode.src_a); u32 src = GetRegister(opcode.src_b); @@ -173,12 +110,12 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { ProcessResult(opcode.result_operation, opcode.dst, result); break; } - case Operation::Read: { + case Macro::Operation::Read: { u32 result = Read(GetRegister(opcode.src_a) + opcode.immediate); ProcessResult(opcode.result_operation, opcode.dst, result); break; } - case Operation::Branch: { + case Macro::Operation::Branch: { ASSERT_MSG(!is_delay_slot, "Executing a branch in a delay slot is not valid"); u32 value = GetRegister(opcode.src_a); bool taken = EvaluateBranchCondition(opcode.branch_condition, value); @@ -191,7 +128,7 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { delayed_pc = base_address + opcode.GetBranchTarget(); // Execute one more instruction due to the delay slot. - return Step(offset, true); + return Step(true); } break; } @@ -204,51 +141,44 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) { // cause an exit if it's executed inside a delay slot. if (opcode.is_exit && !is_delay_slot) { // Exit has a delay slot, execute the next instruction - Step(offset, true); + Step(true); return false; } return true; } -MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const { - const auto& macro_memory{maxwell3d.GetMacroMemory()}; - ASSERT((pc % sizeof(u32)) == 0); - ASSERT((pc + offset) < macro_memory.size() * sizeof(u32)); - return {macro_memory[offset + pc / sizeof(u32)]}; -} - -u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) { +u32 MacroInterpreterImpl::GetALUResult(Macro::ALUOperation operation, u32 src_a, u32 src_b) { switch (operation) { - case ALUOperation::Add: { + case Macro::ALUOperation::Add: { const u64 result{static_cast<u64>(src_a) + src_b}; carry_flag = result > 0xffffffff; return static_cast<u32>(result); } - case ALUOperation::AddWithCarry: { + case Macro::ALUOperation::AddWithCarry: { const u64 result{static_cast<u64>(src_a) + src_b + (carry_flag ? 1ULL : 0ULL)}; carry_flag = result > 0xffffffff; return static_cast<u32>(result); } - case ALUOperation::Subtract: { + case Macro::ALUOperation::Subtract: { const u64 result{static_cast<u64>(src_a) - src_b}; carry_flag = result < 0x100000000; return static_cast<u32>(result); } - case ALUOperation::SubtractWithBorrow: { + case Macro::ALUOperation::SubtractWithBorrow: { const u64 result{static_cast<u64>(src_a) - src_b - (carry_flag ? 0ULL : 1ULL)}; carry_flag = result < 0x100000000; return static_cast<u32>(result); } - case ALUOperation::Xor: + case Macro::ALUOperation::Xor: return src_a ^ src_b; - case ALUOperation::Or: + case Macro::ALUOperation::Or: return src_a | src_b; - case ALUOperation::And: + case Macro::ALUOperation::And: return src_a & src_b; - case ALUOperation::AndNot: + case Macro::ALUOperation::AndNot: return src_a & ~src_b; - case ALUOperation::Nand: + case Macro::ALUOperation::Nand: return ~(src_a & src_b); default: @@ -257,43 +187,43 @@ u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) } } -void MacroInterpreter::ProcessResult(ResultOperation operation, u32 reg, u32 result) { +void MacroInterpreterImpl::ProcessResult(Macro::ResultOperation operation, u32 reg, u32 result) { switch (operation) { - case ResultOperation::IgnoreAndFetch: + case Macro::ResultOperation::IgnoreAndFetch: // Fetch parameter and ignore result. SetRegister(reg, FetchParameter()); break; - case ResultOperation::Move: + case Macro::ResultOperation::Move: // Move result. SetRegister(reg, result); break; - case ResultOperation::MoveAndSetMethod: + case Macro::ResultOperation::MoveAndSetMethod: // Move result and use as Method Address. SetRegister(reg, result); SetMethodAddress(result); break; - case ResultOperation::FetchAndSend: + case Macro::ResultOperation::FetchAndSend: // Fetch parameter and send result. SetRegister(reg, FetchParameter()); Send(result); break; - case ResultOperation::MoveAndSend: + case Macro::ResultOperation::MoveAndSend: // Move and send result. SetRegister(reg, result); Send(result); break; - case ResultOperation::FetchAndSetMethod: + case Macro::ResultOperation::FetchAndSetMethod: // Fetch parameter and use result as Method Address. SetRegister(reg, FetchParameter()); SetMethodAddress(result); break; - case ResultOperation::MoveAndSetMethodFetchAndSend: + case Macro::ResultOperation::MoveAndSetMethodFetchAndSend: // Move result and use as Method Address, then fetch and send parameter. SetRegister(reg, result); SetMethodAddress(result); Send(FetchParameter()); break; - case ResultOperation::MoveAndSetMethodSend: + case Macro::ResultOperation::MoveAndSetMethodSend: // Move result and use as Method Address, then send bits 12:17 of result. SetRegister(reg, result); SetMethodAddress(result); @@ -304,16 +234,28 @@ void MacroInterpreter::ProcessResult(ResultOperation operation, u32 reg, u32 res } } -u32 MacroInterpreter::FetchParameter() { - ASSERT(next_parameter_index < num_parameters); - return parameters[next_parameter_index++]; +bool MacroInterpreterImpl::EvaluateBranchCondition(Macro::BranchCondition cond, u32 value) const { + switch (cond) { + case Macro::BranchCondition::Zero: + return value == 0; + case Macro::BranchCondition::NotZero: + return value != 0; + } + UNREACHABLE(); + return true; +} + +Macro::Opcode MacroInterpreterImpl::GetOpcode() const { + ASSERT((pc % sizeof(u32)) == 0); + ASSERT(pc < code.size() * sizeof(u32)); + return {code[pc / sizeof(u32)]}; } -u32 MacroInterpreter::GetRegister(u32 register_id) const { +u32 MacroInterpreterImpl::GetRegister(u32 register_id) const { return registers.at(register_id); } -void MacroInterpreter::SetRegister(u32 register_id, u32 value) { +void MacroInterpreterImpl::SetRegister(u32 register_id, u32 value) { // Register 0 is hardwired as the zero register. // Ensure no writes to it actually occur. if (register_id == 0) { @@ -323,30 +265,24 @@ void MacroInterpreter::SetRegister(u32 register_id, u32 value) { registers.at(register_id) = value; } -void MacroInterpreter::SetMethodAddress(u32 address) { +void MacroInterpreterImpl::SetMethodAddress(u32 address) { method_address.raw = address; } -void MacroInterpreter::Send(u32 value) { - maxwell3d.CallMethodFromMME({method_address.address, value}); +void MacroInterpreterImpl::Send(u32 value) { + maxwell3d.CallMethodFromMME(method_address.address, value); // Increment the method address by the method increment. method_address.address.Assign(method_address.address.Value() + method_address.increment.Value()); } -u32 MacroInterpreter::Read(u32 method) const { +u32 MacroInterpreterImpl::Read(u32 method) const { return maxwell3d.GetRegisterValue(method); } -bool MacroInterpreter::EvaluateBranchCondition(BranchCondition cond, u32 value) const { - switch (cond) { - case BranchCondition::Zero: - return value == 0; - case BranchCondition::NotZero: - return value != 0; - } - UNREACHABLE(); - return true; +u32 MacroInterpreterImpl::FetchParameter() { + ASSERT(next_parameter_index < num_parameters); + return parameters[next_parameter_index++]; } } // namespace Tegra diff --git a/src/video_core/macro_interpreter.h b/src/video_core/macro/macro_interpreter.h index 631146d89..90217fc89 100644 --- a/src/video_core/macro_interpreter.h +++ b/src/video_core/macro/macro_interpreter.h @@ -1,44 +1,37 @@ -// Copyright 2018 yuzu Emulator Project +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. #pragma once - #include <array> #include <optional> - +#include <vector> #include "common/bit_field.h" #include "common/common_types.h" +#include "video_core/macro/macro.h" namespace Tegra { namespace Engines { class Maxwell3D; } -class MacroInterpreter final { +class MacroInterpreter final : public MacroEngine { public: explicit MacroInterpreter(Engines::Maxwell3D& maxwell3d); - /** - * Executes the macro code with the specified input parameters. - * @param offset Offset to start execution at. - * @param parameters The parameters of the macro. - */ - void Execute(u32 offset, std::size_t num_parameters, const u32* parameters); +protected: + std::unique_ptr<CachedMacro> Compile(const std::vector<u32>& code) override; private: - enum class ALUOperation : u32; - enum class BranchCondition : u32; - enum class ResultOperation : u32; - - union Opcode; + Engines::Maxwell3D& maxwell3d; +}; - union MethodAddress { - u32 raw; - BitField<0, 12, u32> address; - BitField<12, 6, u32> increment; - }; +class MacroInterpreterImpl : public CachedMacro { +public: + MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& code); + void Execute(const std::vector<u32>& parameters, u32 method) override; +private: /// Resets the execution engine state, zeroing registers, etc. void Reset(); @@ -49,20 +42,20 @@ private: * @param is_delay_slot Whether the current step is being executed due to a delay slot in a * previous instruction. */ - bool Step(u32 offset, bool is_delay_slot); + bool Step(bool is_delay_slot); /// Calculates the result of an ALU operation. src_a OP src_b; - u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b); + u32 GetALUResult(Macro::ALUOperation operation, u32 src_a, u32 src_b); /// Performs the result operation on the input result and stores it in the specified register /// (if necessary). - void ProcessResult(ResultOperation operation, u32 reg, u32 result); + void ProcessResult(Macro::ResultOperation operation, u32 reg, u32 result); /// Evaluates the branch condition and returns whether the branch should be taken or not. - bool EvaluateBranchCondition(BranchCondition cond, u32 value) const; + bool EvaluateBranchCondition(Macro::BranchCondition cond, u32 value) const; /// Reads an opcode at the current program counter location. - Opcode GetOpcode(u32 offset) const; + Macro::Opcode GetOpcode() const; /// Returns the specified register's value. Register 0 is hardcoded to always return 0. u32 GetRegister(u32 register_id) const; @@ -89,13 +82,11 @@ private: /// Program counter to execute at after the delay slot is executed. std::optional<u32> delayed_pc; - static constexpr std::size_t NumMacroRegisters = 8; - /// General purpose macro registers. - std::array<u32, NumMacroRegisters> registers = {}; + std::array<u32, Macro::NUM_MACRO_REGISTERS> registers = {}; /// Method address to use for the next Send instruction. - MethodAddress method_address = {}; + Macro::MethodAddress method_address = {}; /// Input parameters of the current macro. std::unique_ptr<u32[]> parameters; @@ -105,5 +96,7 @@ private: u32 next_parameter_index = 0; bool carry_flag = false; + const std::vector<u32>& code; }; + } // namespace Tegra diff --git a/src/video_core/macro/macro_jit_x64.cpp b/src/video_core/macro/macro_jit_x64.cpp new file mode 100644 index 000000000..954b87515 --- /dev/null +++ b/src/video_core/macro/macro_jit_x64.cpp @@ -0,0 +1,620 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "common/logging/log.h" +#include "common/microprofile.h" +#include "common/x64/xbyak_util.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/macro/macro_interpreter.h" +#include "video_core/macro/macro_jit_x64.h" + +MICROPROFILE_DEFINE(MacroJitCompile, "GPU", "Compile macro JIT", MP_RGB(173, 255, 47)); +MICROPROFILE_DEFINE(MacroJitExecute, "GPU", "Execute macro JIT", MP_RGB(255, 255, 0)); + +namespace Tegra { +constexpr Xbyak::Reg64 STATE = Xbyak::util::rbx; +constexpr Xbyak::Reg32 RESULT = Xbyak::util::ebp; +constexpr Xbyak::Reg64 PARAMETERS = Xbyak::util::r12; +constexpr Xbyak::Reg32 METHOD_ADDRESS = Xbyak::util::r14d; +constexpr Xbyak::Reg64 BRANCH_HOLDER = Xbyak::util::r15; + +static const std::bitset<32> PERSISTENT_REGISTERS = Common::X64::BuildRegSet({ + STATE, + RESULT, + PARAMETERS, + METHOD_ADDRESS, + BRANCH_HOLDER, +}); + +MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d) + : MacroEngine::MacroEngine(maxwell3d), maxwell3d(maxwell3d) {} + +std::unique_ptr<CachedMacro> MacroJITx64::Compile(const std::vector<u32>& code) { + return std::make_unique<MacroJITx64Impl>(maxwell3d, code); +} + +MacroJITx64Impl::MacroJITx64Impl(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& code) + : Xbyak::CodeGenerator(MAX_CODE_SIZE), code(code), maxwell3d(maxwell3d) { + Compile(); +} + +MacroJITx64Impl::~MacroJITx64Impl() = default; + +void MacroJITx64Impl::Execute(const std::vector<u32>& parameters, u32 method) { + MICROPROFILE_SCOPE(MacroJitExecute); + ASSERT_OR_EXECUTE(program != nullptr, { return; }); + JITState state{}; + state.maxwell3d = &maxwell3d; + state.registers = {}; + program(&state, parameters.data()); +} + +void MacroJITx64Impl::Compile_ALU(Macro::Opcode opcode) { + const bool is_a_zero = opcode.src_a == 0; + const bool is_b_zero = opcode.src_b == 0; + const bool valid_operation = !is_a_zero && !is_b_zero; + [[maybe_unused]] const bool is_move_operation = !is_a_zero && is_b_zero; + const bool has_zero_register = is_a_zero || is_b_zero; + const bool no_zero_reg_skip = opcode.alu_operation == Macro::ALUOperation::AddWithCarry || + opcode.alu_operation == Macro::ALUOperation::SubtractWithBorrow; + + Xbyak::Reg32 src_a; + Xbyak::Reg32 src_b; + + if (!optimizer.zero_reg_skip || no_zero_reg_skip) { + src_a = Compile_GetRegister(opcode.src_a, RESULT); + src_b = Compile_GetRegister(opcode.src_b, eax); + } else { + if (!is_a_zero) { + src_a = Compile_GetRegister(opcode.src_a, RESULT); + } + if (!is_b_zero) { + src_b = Compile_GetRegister(opcode.src_b, eax); + } + } + + bool has_emitted = false; + + switch (opcode.alu_operation) { + case Macro::ALUOperation::Add: + if (optimizer.zero_reg_skip) { + if (valid_operation) { + add(src_a, src_b); + } + } else { + add(src_a, src_b); + } + + if (!optimizer.can_skip_carry) { + setc(byte[STATE + offsetof(JITState, carry_flag)]); + } + break; + case Macro::ALUOperation::AddWithCarry: + bt(dword[STATE + offsetof(JITState, carry_flag)], 0); + adc(src_a, src_b); + setc(byte[STATE + offsetof(JITState, carry_flag)]); + break; + case Macro::ALUOperation::Subtract: + if (optimizer.zero_reg_skip) { + if (valid_operation) { + sub(src_a, src_b); + has_emitted = true; + } + } else { + sub(src_a, src_b); + has_emitted = true; + } + if (!optimizer.can_skip_carry && has_emitted) { + setc(byte[STATE + offsetof(JITState, carry_flag)]); + } + break; + case Macro::ALUOperation::SubtractWithBorrow: + bt(dword[STATE + offsetof(JITState, carry_flag)], 0); + sbb(src_a, src_b); + setc(byte[STATE + offsetof(JITState, carry_flag)]); + break; + case Macro::ALUOperation::Xor: + if (optimizer.zero_reg_skip) { + if (valid_operation) { + xor_(src_a, src_b); + } + } else { + xor_(src_a, src_b); + } + break; + case Macro::ALUOperation::Or: + if (optimizer.zero_reg_skip) { + if (valid_operation) { + or_(src_a, src_b); + } + } else { + or_(src_a, src_b); + } + break; + case Macro::ALUOperation::And: + if (optimizer.zero_reg_skip) { + if (!has_zero_register) { + and_(src_a, src_b); + } + } else { + and_(src_a, src_b); + } + break; + case Macro::ALUOperation::AndNot: + if (optimizer.zero_reg_skip) { + if (!is_a_zero) { + not_(src_b); + and_(src_a, src_b); + } + } else { + not_(src_b); + and_(src_a, src_b); + } + break; + case Macro::ALUOperation::Nand: + if (optimizer.zero_reg_skip) { + if (!is_a_zero) { + and_(src_a, src_b); + not_(src_a); + } + } else { + and_(src_a, src_b); + not_(src_a); + } + break; + default: + UNIMPLEMENTED_MSG("Unimplemented ALU operation {}", + static_cast<std::size_t>(opcode.alu_operation.Value())); + break; + } + Compile_ProcessResult(opcode.result_operation, opcode.dst); +} + +void MacroJITx64Impl::Compile_AddImmediate(Macro::Opcode opcode) { + if (optimizer.skip_dummy_addimmediate) { + // Games tend to use this as an exit instruction placeholder. It's to encode an instruction + // without doing anything. In our case we can just not emit anything. + if (opcode.result_operation == Macro::ResultOperation::Move && opcode.dst == 0) { + return; + } + } + // Check for redundant moves + if (optimizer.optimize_for_method_move && + opcode.result_operation == Macro::ResultOperation::MoveAndSetMethod) { + if (next_opcode.has_value()) { + const auto next = *next_opcode; + if (next.result_operation == Macro::ResultOperation::MoveAndSetMethod && + opcode.dst == next.dst) { + return; + } + } + } + if (optimizer.zero_reg_skip && opcode.src_a == 0) { + if (opcode.immediate == 0) { + xor_(RESULT, RESULT); + } else { + mov(RESULT, opcode.immediate); + } + } else { + auto result = Compile_GetRegister(opcode.src_a, RESULT); + if (opcode.immediate > 2) { + add(result, opcode.immediate); + } else if (opcode.immediate == 1) { + inc(result); + } else if (opcode.immediate < 0) { + sub(result, opcode.immediate * -1); + } + } + Compile_ProcessResult(opcode.result_operation, opcode.dst); +} + +void MacroJITx64Impl::Compile_ExtractInsert(Macro::Opcode opcode) { + auto dst = Compile_GetRegister(opcode.src_a, RESULT); + auto src = Compile_GetRegister(opcode.src_b, eax); + + if (opcode.bf_src_bit != 0 && opcode.bf_src_bit != 31) { + shr(src, opcode.bf_src_bit); + } else if (opcode.bf_src_bit == 31) { + xor_(src, src); + } + // Don't bother masking the whole register since we're using a 32 bit register + if (opcode.bf_size != 31 && opcode.bf_size != 0) { + and_(src, opcode.GetBitfieldMask()); + } else if (opcode.bf_size == 0) { + xor_(src, src); + } + if (opcode.bf_dst_bit != 31 && opcode.bf_dst_bit != 0) { + shl(src, opcode.bf_dst_bit); + } else if (opcode.bf_dst_bit == 31) { + xor_(src, src); + } + + const u32 mask = ~(opcode.GetBitfieldMask() << opcode.bf_dst_bit); + if (mask != 0xffffffff) { + and_(dst, mask); + } + or_(dst, src); + Compile_ProcessResult(opcode.result_operation, opcode.dst); +} + +void MacroJITx64Impl::Compile_ExtractShiftLeftImmediate(Macro::Opcode opcode) { + const auto dst = Compile_GetRegister(opcode.src_a, ecx); + const auto src = Compile_GetRegister(opcode.src_b, RESULT); + + shr(src, dst.cvt8()); + if (opcode.bf_size != 0 && opcode.bf_size != 31) { + and_(src, opcode.GetBitfieldMask()); + } else if (opcode.bf_size == 0) { + xor_(src, src); + } + + if (opcode.bf_dst_bit != 0 && opcode.bf_dst_bit != 31) { + shl(src, opcode.bf_dst_bit); + } else if (opcode.bf_dst_bit == 31) { + xor_(src, src); + } + Compile_ProcessResult(opcode.result_operation, opcode.dst); +} + +void MacroJITx64Impl::Compile_ExtractShiftLeftRegister(Macro::Opcode opcode) { + const auto dst = Compile_GetRegister(opcode.src_a, ecx); + const auto src = Compile_GetRegister(opcode.src_b, RESULT); + + if (opcode.bf_src_bit != 0) { + shr(src, opcode.bf_src_bit); + } + + if (opcode.bf_size != 31) { + and_(src, opcode.GetBitfieldMask()); + } + shl(src, dst.cvt8()); + + Compile_ProcessResult(opcode.result_operation, opcode.dst); +} + +void MacroJITx64Impl::Compile_Read(Macro::Opcode opcode) { + if (optimizer.zero_reg_skip && opcode.src_a == 0) { + if (opcode.immediate == 0) { + xor_(RESULT, RESULT); + } else { + mov(RESULT, opcode.immediate); + } + } else { + auto result = Compile_GetRegister(opcode.src_a, RESULT); + if (opcode.immediate > 2) { + add(result, opcode.immediate); + } else if (opcode.immediate == 1) { + inc(result); + } else if (opcode.immediate < 0) { + sub(result, opcode.immediate * -1); + } + } + + // Equivalent to Engines::Maxwell3D::GetRegisterValue: + if (optimizer.enable_asserts) { + Xbyak::Label pass_range_check; + cmp(RESULT, static_cast<u32>(Engines::Maxwell3D::Regs::NUM_REGS)); + jb(pass_range_check); + int3(); + L(pass_range_check); + } + mov(rax, qword[STATE]); + mov(RESULT, + dword[rax + offsetof(Engines::Maxwell3D, regs) + + offsetof(Engines::Maxwell3D::Regs, reg_array) + RESULT.cvt64() * sizeof(u32)]); + + Compile_ProcessResult(opcode.result_operation, opcode.dst); +} + +static void Send(Engines::Maxwell3D* maxwell3d, Macro::MethodAddress method_address, u32 value) { + maxwell3d->CallMethodFromMME(method_address.address, value); +} + +void Tegra::MacroJITx64Impl::Compile_Send(Xbyak::Reg32 value) { + Common::X64::ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + mov(Common::X64::ABI_PARAM1, qword[STATE]); + mov(Common::X64::ABI_PARAM2, METHOD_ADDRESS); + mov(Common::X64::ABI_PARAM3, value); + Common::X64::CallFarFunction(*this, &Send); + Common::X64::ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + + Xbyak::Label dont_process{}; + // Get increment + test(METHOD_ADDRESS, 0x3f000); + // If zero, method address doesn't update + je(dont_process); + + mov(ecx, METHOD_ADDRESS); + and_(METHOD_ADDRESS, 0xfff); + shr(ecx, 12); + and_(ecx, 0x3f); + lea(eax, ptr[rcx + METHOD_ADDRESS.cvt64()]); + sal(ecx, 12); + or_(eax, ecx); + + mov(METHOD_ADDRESS, eax); + + L(dont_process); +} + +void Tegra::MacroJITx64Impl::Compile_Branch(Macro::Opcode opcode) { + ASSERT_MSG(!is_delay_slot, "Executing a branch in a delay slot is not valid"); + const s32 jump_address = + static_cast<s32>(pc) + static_cast<s32>(opcode.GetBranchTarget() / sizeof(s32)); + + Xbyak::Label end; + auto value = Compile_GetRegister(opcode.src_a, eax); + test(value, value); + if (optimizer.has_delayed_pc) { + switch (opcode.branch_condition) { + case Macro::BranchCondition::Zero: + jne(end, T_NEAR); + break; + case Macro::BranchCondition::NotZero: + je(end, T_NEAR); + break; + } + + if (opcode.branch_annul) { + xor_(BRANCH_HOLDER, BRANCH_HOLDER); + jmp(labels[jump_address], T_NEAR); + } else { + Xbyak::Label handle_post_exit{}; + Xbyak::Label skip{}; + jmp(skip, T_NEAR); + if (opcode.is_exit) { + L(handle_post_exit); + // Execute 1 instruction + mov(BRANCH_HOLDER, end_of_code); + // Jump to next instruction to skip delay slot check + jmp(labels[jump_address], T_NEAR); + } else { + L(handle_post_exit); + xor_(BRANCH_HOLDER, BRANCH_HOLDER); + jmp(labels[jump_address], T_NEAR); + } + L(skip); + mov(BRANCH_HOLDER, handle_post_exit); + jmp(delay_skip[pc], T_NEAR); + } + } else { + switch (opcode.branch_condition) { + case Macro::BranchCondition::Zero: + je(labels[jump_address], T_NEAR); + break; + case Macro::BranchCondition::NotZero: + jne(labels[jump_address], T_NEAR); + break; + } + } + + L(end); +} + +void Tegra::MacroJITx64Impl::Optimizer_ScanFlags() { + optimizer.can_skip_carry = true; + optimizer.has_delayed_pc = false; + for (auto raw_op : code) { + Macro::Opcode op{}; + op.raw = raw_op; + + if (op.operation == Macro::Operation::ALU) { + // Scan for any ALU operations which actually use the carry flag, if they don't exist in + // our current code we can skip emitting the carry flag handling operations + if (op.alu_operation == Macro::ALUOperation::AddWithCarry || + op.alu_operation == Macro::ALUOperation::SubtractWithBorrow) { + optimizer.can_skip_carry = false; + } + } + + if (op.operation == Macro::Operation::Branch) { + if (!op.branch_annul) { + optimizer.has_delayed_pc = true; + } + } + } +} + +void MacroJITx64Impl::Compile() { + MICROPROFILE_SCOPE(MacroJitCompile); + labels.fill(Xbyak::Label()); + + Common::X64::ABI_PushRegistersAndAdjustStack(*this, Common::X64::ABI_ALL_CALLEE_SAVED, 8); + // JIT state + mov(STATE, Common::X64::ABI_PARAM1); + mov(PARAMETERS, Common::X64::ABI_PARAM2); + xor_(RESULT, RESULT); + xor_(METHOD_ADDRESS, METHOD_ADDRESS); + xor_(BRANCH_HOLDER, BRANCH_HOLDER); + + mov(dword[STATE + offsetof(JITState, registers) + 4], Compile_FetchParameter()); + + // Track get register for zero registers and mark it as no-op + optimizer.zero_reg_skip = true; + + // AddImmediate tends to be used as a NOP instruction, if we detect this we can + // completely skip the entire code path and no emit anything + optimizer.skip_dummy_addimmediate = true; + + // SMO tends to emit a lot of unnecessary method moves, we can mitigate this by only emitting + // one if our register isn't "dirty" + optimizer.optimize_for_method_move = true; + + // Enable run-time assertions in JITted code + optimizer.enable_asserts = false; + + // Check to see if we can skip emitting certain instructions + Optimizer_ScanFlags(); + + const u32 op_count = static_cast<u32>(code.size()); + for (u32 i = 0; i < op_count; i++) { + if (i < op_count - 1) { + pc = i + 1; + next_opcode = GetOpCode(); + } else { + next_opcode = {}; + } + pc = i; + Compile_NextInstruction(); + } + + L(end_of_code); + + Common::X64::ABI_PopRegistersAndAdjustStack(*this, Common::X64::ABI_ALL_CALLEE_SAVED, 8); + ret(); + ready(); + program = getCode<ProgramType>(); +} + +bool MacroJITx64Impl::Compile_NextInstruction() { + const auto opcode = GetOpCode(); + if (labels[pc].getAddress()) { + return false; + } + + L(labels[pc]); + + switch (opcode.operation) { + case Macro::Operation::ALU: + Compile_ALU(opcode); + break; + case Macro::Operation::AddImmediate: + Compile_AddImmediate(opcode); + break; + case Macro::Operation::ExtractInsert: + Compile_ExtractInsert(opcode); + break; + case Macro::Operation::ExtractShiftLeftImmediate: + Compile_ExtractShiftLeftImmediate(opcode); + break; + case Macro::Operation::ExtractShiftLeftRegister: + Compile_ExtractShiftLeftRegister(opcode); + break; + case Macro::Operation::Read: + Compile_Read(opcode); + break; + case Macro::Operation::Branch: + Compile_Branch(opcode); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented opcode {}", opcode.operation.Value()); + break; + } + + if (optimizer.has_delayed_pc) { + if (opcode.is_exit) { + mov(rax, end_of_code); + test(BRANCH_HOLDER, BRANCH_HOLDER); + cmove(BRANCH_HOLDER, rax); + // Jump to next instruction to skip delay slot check + je(labels[pc + 1], T_NEAR); + } else { + // TODO(ogniK): Optimize delay slot branching + Xbyak::Label no_delay_slot{}; + test(BRANCH_HOLDER, BRANCH_HOLDER); + je(no_delay_slot, T_NEAR); + mov(rax, BRANCH_HOLDER); + xor_(BRANCH_HOLDER, BRANCH_HOLDER); + jmp(rax); + L(no_delay_slot); + } + L(delay_skip[pc]); + if (opcode.is_exit) { + return false; + } + } else { + test(BRANCH_HOLDER, BRANCH_HOLDER); + jne(end_of_code, T_NEAR); + if (opcode.is_exit) { + inc(BRANCH_HOLDER); + return false; + } + } + return true; +} + +Xbyak::Reg32 Tegra::MacroJITx64Impl::Compile_FetchParameter() { + mov(eax, dword[PARAMETERS]); + add(PARAMETERS, sizeof(u32)); + return eax; +} + +Xbyak::Reg32 MacroJITx64Impl::Compile_GetRegister(u32 index, Xbyak::Reg32 dst) { + if (index == 0) { + // Register 0 is always zero + xor_(dst, dst); + } else { + mov(dst, dword[STATE + offsetof(JITState, registers) + index * sizeof(u32)]); + } + + return dst; +} + +void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u32 reg) { + const auto SetRegister = [this](u32 reg, const Xbyak::Reg32& result) { + // Register 0 is supposed to always return 0. NOP is implemented as a store to the zero + // register. + if (reg == 0) { + return; + } + mov(dword[STATE + offsetof(JITState, registers) + reg * sizeof(u32)], result); + }; + const auto SetMethodAddress = [this](const Xbyak::Reg32& reg) { mov(METHOD_ADDRESS, reg); }; + + switch (operation) { + case Macro::ResultOperation::IgnoreAndFetch: + SetRegister(reg, Compile_FetchParameter()); + break; + case Macro::ResultOperation::Move: + SetRegister(reg, RESULT); + break; + case Macro::ResultOperation::MoveAndSetMethod: + SetRegister(reg, RESULT); + SetMethodAddress(RESULT); + break; + case Macro::ResultOperation::FetchAndSend: + // Fetch parameter and send result. + SetRegister(reg, Compile_FetchParameter()); + Compile_Send(RESULT); + break; + case Macro::ResultOperation::MoveAndSend: + // Move and send result. + SetRegister(reg, RESULT); + Compile_Send(RESULT); + break; + case Macro::ResultOperation::FetchAndSetMethod: + // Fetch parameter and use result as Method Address. + SetRegister(reg, Compile_FetchParameter()); + SetMethodAddress(RESULT); + break; + case Macro::ResultOperation::MoveAndSetMethodFetchAndSend: + // Move result and use as Method Address, then fetch and send parameter. + SetRegister(reg, RESULT); + SetMethodAddress(RESULT); + Compile_Send(Compile_FetchParameter()); + break; + case Macro::ResultOperation::MoveAndSetMethodSend: + // Move result and use as Method Address, then send bits 12:17 of result. + SetRegister(reg, RESULT); + SetMethodAddress(RESULT); + shr(RESULT, 12); + and_(RESULT, 0b111111); + Compile_Send(RESULT); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented macro operation {}", static_cast<std::size_t>(operation)); + } +} + +Macro::Opcode MacroJITx64Impl::GetOpCode() const { + ASSERT(pc < code.size()); + return {code[pc]}; +} + +std::bitset<32> MacroJITx64Impl::PersistentCallerSavedRegs() const { + return PERSISTENT_REGISTERS & Common::X64::ABI_ALL_CALLER_SAVED; +} + +} // namespace Tegra diff --git a/src/video_core/macro/macro_jit_x64.h b/src/video_core/macro/macro_jit_x64.h new file mode 100644 index 000000000..a180e7428 --- /dev/null +++ b/src/video_core/macro/macro_jit_x64.h @@ -0,0 +1,98 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <bitset> +#include <xbyak.h> +#include "common/bit_field.h" +#include "common/common_types.h" +#include "common/x64/xbyak_abi.h" +#include "video_core/macro/macro.h" + +namespace Tegra { + +namespace Engines { +class Maxwell3D; +} + +/// MAX_CODE_SIZE is arbitrarily chosen based on current booting games +constexpr size_t MAX_CODE_SIZE = 0x10000; + +class MacroJITx64 final : public MacroEngine { +public: + explicit MacroJITx64(Engines::Maxwell3D& maxwell3d); + +protected: + std::unique_ptr<CachedMacro> Compile(const std::vector<u32>& code) override; + +private: + Engines::Maxwell3D& maxwell3d; +}; + +class MacroJITx64Impl : public Xbyak::CodeGenerator, public CachedMacro { +public: + MacroJITx64Impl(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& code); + ~MacroJITx64Impl(); + + void Execute(const std::vector<u32>& parameters, u32 method) override; + + void Compile_ALU(Macro::Opcode opcode); + void Compile_AddImmediate(Macro::Opcode opcode); + void Compile_ExtractInsert(Macro::Opcode opcode); + void Compile_ExtractShiftLeftImmediate(Macro::Opcode opcode); + void Compile_ExtractShiftLeftRegister(Macro::Opcode opcode); + void Compile_Read(Macro::Opcode opcode); + void Compile_Branch(Macro::Opcode opcode); + +private: + void Optimizer_ScanFlags(); + + void Compile(); + bool Compile_NextInstruction(); + + Xbyak::Reg32 Compile_FetchParameter(); + Xbyak::Reg32 Compile_GetRegister(u32 index, Xbyak::Reg32 dst); + + void Compile_ProcessResult(Macro::ResultOperation operation, u32 reg); + void Compile_Send(Xbyak::Reg32 value); + + Macro::Opcode GetOpCode() const; + std::bitset<32> PersistentCallerSavedRegs() const; + + struct JITState { + Engines::Maxwell3D* maxwell3d{}; + std::array<u32, Macro::NUM_MACRO_REGISTERS> registers{}; + u32 carry_flag{}; + }; + static_assert(offsetof(JITState, maxwell3d) == 0, "Maxwell3D is not at 0x0"); + using ProgramType = void (*)(JITState*, const u32*); + + struct OptimizerState { + bool can_skip_carry{}; + bool has_delayed_pc{}; + bool zero_reg_skip{}; + bool skip_dummy_addimmediate{}; + bool optimize_for_method_move{}; + bool enable_asserts{}; + }; + OptimizerState optimizer{}; + + std::optional<Macro::Opcode> next_opcode{}; + ProgramType program{nullptr}; + + std::array<Xbyak::Label, MAX_CODE_SIZE> labels; + std::array<Xbyak::Label, MAX_CODE_SIZE> delay_skip; + Xbyak::Label end_of_code{}; + + bool is_delay_slot{}; + u32 pc{}; + std::optional<u32> delayed_pc; + + const std::vector<u32>& code; + Engines::Maxwell3D& maxwell3d; +}; + +} // namespace Tegra diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index a3389d0d2..6e70bd362 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -4,186 +4,180 @@ #include "common/alignment.h" #include "common/assert.h" -#include "common/logging/log.h" #include "core/core.h" +#include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/vm_manager.h" #include "core/memory.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" +#include "video_core/renderer_base.h" namespace Tegra { -MemoryManager::MemoryManager(Core::System& system, VideoCore::RasterizerInterface& rasterizer) - : rasterizer{rasterizer}, system{system} { - std::fill(page_table.pointers.begin(), page_table.pointers.end(), nullptr); - std::fill(page_table.attributes.begin(), page_table.attributes.end(), - Common::PageType::Unmapped); - page_table.Resize(address_space_width); +MemoryManager::MemoryManager(Core::System& system_) + : system{system_}, page_table(page_table_size) {} - // Initialize the map with a single free region covering the entire managed space. - VirtualMemoryArea initial_vma; - initial_vma.size = address_space_end; - vma_map.emplace(initial_vma.base, initial_vma); +MemoryManager::~MemoryManager() = default; - UpdatePageTableForVMA(initial_vma); +void MemoryManager::BindRasterizer(VideoCore::RasterizerInterface& rasterizer_) { + rasterizer = &rasterizer_; } -MemoryManager::~MemoryManager() = default; +GPUVAddr MemoryManager::UpdateRange(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size) { + u64 remaining_size{size}; + for (u64 offset{}; offset < size; offset += page_size) { + if (remaining_size < page_size) { + SetPageEntry(gpu_addr + offset, page_entry + offset, remaining_size); + } else { + SetPageEntry(gpu_addr + offset, page_entry + offset); + } + remaining_size -= page_size; + } + return gpu_addr; +} -GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) { - const u64 aligned_size{Common::AlignUp(size, page_size)}; - const GPUVAddr gpu_addr{FindFreeRegion(address_space_base, aligned_size)}; +GPUVAddr MemoryManager::Map(VAddr cpu_addr, GPUVAddr gpu_addr, std::size_t size) { + return UpdateRange(gpu_addr, cpu_addr, size); +} - AllocateMemory(gpu_addr, 0, aligned_size); +GPUVAddr MemoryManager::MapAllocate(VAddr cpu_addr, std::size_t size, std::size_t align) { + return Map(cpu_addr, *FindFreeRange(size, align), size); +} - return gpu_addr; +GPUVAddr MemoryManager::MapAllocate32(VAddr cpu_addr, std::size_t size) { + const std::optional<GPUVAddr> gpu_addr = FindFreeRange(size, 1, true); + ASSERT(gpu_addr); + return Map(cpu_addr, *gpu_addr, size); } -GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) { - const u64 aligned_size{Common::AlignUp(size, page_size)}; +void MemoryManager::Unmap(GPUVAddr gpu_addr, std::size_t size) { + if (!size) { + return; + } - AllocateMemory(gpu_addr, 0, aligned_size); + // Flush and invalidate through the GPU interface, to be asynchronous if possible. + system.GPU().FlushAndInvalidateRegion(*GpuToCpuAddress(gpu_addr), size); - return gpu_addr; + UpdateRange(gpu_addr, PageEntry::State::Unmapped, size); } -GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) { - const u64 aligned_size{Common::AlignUp(size, page_size)}; - const GPUVAddr gpu_addr{FindFreeRegion(address_space_base, aligned_size)}; - - MapBackingMemory(gpu_addr, system.Memory().GetPointer(cpu_addr), aligned_size, cpu_addr); - ASSERT(system.CurrentProcess() - ->VMManager() - .SetMemoryAttribute(cpu_addr, size, Kernel::MemoryAttribute::DeviceMapped, - Kernel::MemoryAttribute::DeviceMapped) - .IsSuccess()); +std::optional<GPUVAddr> MemoryManager::AllocateFixed(GPUVAddr gpu_addr, std::size_t size) { + for (u64 offset{}; offset < size; offset += page_size) { + if (!GetPageEntry(gpu_addr + offset).IsUnmapped()) { + return std::nullopt; + } + } - return gpu_addr; + return UpdateRange(gpu_addr, PageEntry::State::Allocated, size); } -GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size) { - ASSERT((gpu_addr & page_mask) == 0); +GPUVAddr MemoryManager::Allocate(std::size_t size, std::size_t align) { + return *AllocateFixed(*FindFreeRange(size, align), size); +} - const u64 aligned_size{Common::AlignUp(size, page_size)}; +void MemoryManager::TryLockPage(PageEntry page_entry, std::size_t size) { + if (!page_entry.IsValid()) { + return; + } - MapBackingMemory(gpu_addr, system.Memory().GetPointer(cpu_addr), aligned_size, cpu_addr); ASSERT(system.CurrentProcess() - ->VMManager() - .SetMemoryAttribute(cpu_addr, size, Kernel::MemoryAttribute::DeviceMapped, - Kernel::MemoryAttribute::DeviceMapped) + ->PageTable() + .LockForDeviceAddressSpace(page_entry.ToAddress(), size) .IsSuccess()); - return gpu_addr; } -GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) { - ASSERT((gpu_addr & page_mask) == 0); - - const u64 aligned_size{Common::AlignUp(size, page_size)}; - const auto cpu_addr = GpuToCpuAddress(gpu_addr); - ASSERT(cpu_addr); - - // Flush and invalidate through the GPU interface, to be asynchronous if possible. - system.GPU().FlushAndInvalidateRegion(*cpu_addr, aligned_size); +void MemoryManager::TryUnlockPage(PageEntry page_entry, std::size_t size) { + if (!page_entry.IsValid()) { + return; + } - UnmapRange(gpu_addr, aligned_size); ASSERT(system.CurrentProcess() - ->VMManager() - .SetMemoryAttribute(cpu_addr.value(), size, Kernel::MemoryAttribute::DeviceMapped, - Kernel::MemoryAttribute::None) + ->PageTable() + .UnlockForDeviceAddressSpace(page_entry.ToAddress(), size) .IsSuccess()); +} - return gpu_addr; +PageEntry MemoryManager::GetPageEntry(GPUVAddr gpu_addr) const { + return page_table[PageEntryIndex(gpu_addr)]; } -GPUVAddr MemoryManager::FindFreeRegion(GPUVAddr region_start, u64 size) const { - // Find the first Free VMA. - const VMAHandle vma_handle{ - std::find_if(vma_map.begin(), vma_map.end(), [region_start, size](const auto& vma) { - if (vma.second.type != VirtualMemoryArea::Type::Unmapped) { - return false; - } +void MemoryManager::SetPageEntry(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size) { + // TODO(bunnei): We should lock/unlock device regions. This currently causes issues due to + // improper tracking, but should be fixed in the future. - const VAddr vma_end{vma.second.base + vma.second.size}; - return vma_end > region_start && vma_end >= region_start + size; - })}; + //// Unlock the old page + // TryUnlockPage(page_table[PageEntryIndex(gpu_addr)], size); - if (vma_handle == vma_map.end()) { - return {}; - } + //// Lock the new page + // TryLockPage(page_entry, size); - return std::max(region_start, vma_handle->second.base); + page_table[PageEntryIndex(gpu_addr)] = page_entry; } -bool MemoryManager::IsAddressValid(GPUVAddr addr) const { - return (addr >> page_bits) < page_table.pointers.size(); -} +std::optional<GPUVAddr> MemoryManager::FindFreeRange(std::size_t size, std::size_t align, + bool start_32bit_address) const { + if (!align) { + align = page_size; + } else { + align = Common::AlignUp(align, page_size); + } -std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) const { - if (!IsAddressValid(addr)) { - return {}; + u64 available_size{}; + GPUVAddr gpu_addr{start_32bit_address ? address_space_start_low : address_space_start}; + while (gpu_addr + available_size < address_space_size) { + if (GetPageEntry(gpu_addr + available_size).IsUnmapped()) { + available_size += page_size; + + if (available_size >= size) { + return gpu_addr; + } + } else { + gpu_addr += available_size + page_size; + available_size = 0; + + const auto remainder{gpu_addr % align}; + if (remainder) { + gpu_addr = (gpu_addr - remainder) + align; + } + } } - const VAddr cpu_addr{page_table.backing_addr[addr >> page_bits]}; - if (cpu_addr) { - return cpu_addr + (addr & page_mask); + return std::nullopt; +} + +std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) const { + const auto page_entry{GetPageEntry(gpu_addr)}; + if (!page_entry.IsValid()) { + return std::nullopt; } - return {}; + return page_entry.ToAddress() + (gpu_addr & page_mask); } template <typename T> T MemoryManager::Read(GPUVAddr addr) const { - if (!IsAddressValid(addr)) { - return {}; - } - - const u8* page_pointer{GetPointer(addr)}; - if (page_pointer) { + if (auto page_pointer{GetPointer(addr)}; page_pointer) { // NOTE: Avoid adding any extra logic to this fast-path block T value; std::memcpy(&value, page_pointer, sizeof(T)); return value; } - switch (page_table.attributes[addr >> page_bits]) { - case Common::PageType::Unmapped: - LOG_ERROR(HW_GPU, "Unmapped Read{} @ 0x{:08X}", sizeof(T) * 8, addr); - return 0; - case Common::PageType::Memory: - ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", addr); - break; - default: - UNREACHABLE(); - } + UNREACHABLE(); + return {}; } template <typename T> void MemoryManager::Write(GPUVAddr addr, T data) { - if (!IsAddressValid(addr)) { - return; - } - - u8* page_pointer{GetPointer(addr)}; - if (page_pointer) { + if (auto page_pointer{GetPointer(addr)}; page_pointer) { // NOTE: Avoid adding any extra logic to this fast-path block std::memcpy(page_pointer, &data, sizeof(T)); return; } - switch (page_table.attributes[addr >> page_bits]) { - case Common::PageType::Unmapped: - LOG_ERROR(HW_GPU, "Unmapped Write{} 0x{:08X} @ 0x{:016X}", sizeof(data) * 8, - static_cast<u32>(data), addr); - return; - case Common::PageType::Memory: - ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", addr); - break; - default: - UNREACHABLE(); - } + UNREACHABLE(); } template u8 MemoryManager::Read<u8>(GPUVAddr addr) const; @@ -195,71 +189,48 @@ template void MemoryManager::Write<u16>(GPUVAddr addr, u16 data); template void MemoryManager::Write<u32>(GPUVAddr addr, u32 data); template void MemoryManager::Write<u64>(GPUVAddr addr, u64 data); -u8* MemoryManager::GetPointer(GPUVAddr addr) { - if (!IsAddressValid(addr)) { +u8* MemoryManager::GetPointer(GPUVAddr gpu_addr) { + if (!GetPageEntry(gpu_addr).IsValid()) { return {}; } - auto& memory = system.Memory(); - - const VAddr page_addr{page_table.backing_addr[addr >> page_bits]}; - - if (page_addr != 0) { - return memory.GetPointer(page_addr + (addr & page_mask)); + const auto address{GpuToCpuAddress(gpu_addr)}; + if (!address) { + return {}; } - LOG_ERROR(HW_GPU, "Unknown GetPointer @ 0x{:016X}", addr); - return {}; + return system.Memory().GetPointer(*address); } -const u8* MemoryManager::GetPointer(GPUVAddr addr) const { - if (!IsAddressValid(addr)) { +const u8* MemoryManager::GetPointer(GPUVAddr gpu_addr) const { + if (!GetPageEntry(gpu_addr).IsValid()) { return {}; } - const auto& memory = system.Memory(); - - const VAddr page_addr{page_table.backing_addr[addr >> page_bits]}; - - if (page_addr != 0) { - return memory.GetPointer(page_addr + (addr & page_mask)); + const auto address{GpuToCpuAddress(gpu_addr)}; + if (!address) { + return {}; } - LOG_ERROR(HW_GPU, "Unknown GetPointer @ 0x{:016X}", addr); - return {}; -} - -bool MemoryManager::IsBlockContinuous(const GPUVAddr start, const std::size_t size) const { - const std::size_t inner_size = size - 1; - const GPUVAddr end = start + inner_size; - const auto host_ptr_start = reinterpret_cast<std::uintptr_t>(GetPointer(start)); - const auto host_ptr_end = reinterpret_cast<std::uintptr_t>(GetPointer(end)); - const auto range = static_cast<std::size_t>(host_ptr_end - host_ptr_start); - return range == inner_size; + return system.Memory().GetPointer(*address); } -void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, const std::size_t size) const { +void MemoryManager::ReadBlock(GPUVAddr gpu_src_addr, void* dest_buffer, std::size_t size) const { std::size_t remaining_size{size}; - std::size_t page_index{src_addr >> page_bits}; - std::size_t page_offset{src_addr & page_mask}; - - auto& memory = system.Memory(); + std::size_t page_index{gpu_src_addr >> page_bits}; + std::size_t page_offset{gpu_src_addr & page_mask}; while (remaining_size > 0) { const std::size_t copy_amount{ std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; - switch (page_table.attributes[page_index]) { - case Common::PageType::Memory: { - const VAddr src_addr{page_table.backing_addr[page_index] + page_offset}; + if (const auto page_addr{GpuToCpuAddress(page_index << page_bits)}; page_addr) { + const auto src_addr{*page_addr + page_offset}; + // Flush must happen on the rasterizer interface, such that memory is always synchronous // when it is read (even when in asynchronous GPU mode). Fixes Dead Cells title menu. - rasterizer.FlushRegion(src_addr, copy_amount); - memory.ReadBlockUnsafe(src_addr, dest_buffer, copy_amount); - break; - } - default: - UNREACHABLE(); + rasterizer->FlushRegion(src_addr, copy_amount); + system.Memory().ReadBlockUnsafe(src_addr, dest_buffer, copy_amount); } page_index++; @@ -269,24 +240,23 @@ void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, const std::s } } -void MemoryManager::ReadBlockUnsafe(GPUVAddr src_addr, void* dest_buffer, +void MemoryManager::ReadBlockUnsafe(GPUVAddr gpu_src_addr, void* dest_buffer, const std::size_t size) const { std::size_t remaining_size{size}; - std::size_t page_index{src_addr >> page_bits}; - std::size_t page_offset{src_addr & page_mask}; - - auto& memory = system.Memory(); + std::size_t page_index{gpu_src_addr >> page_bits}; + std::size_t page_offset{gpu_src_addr & page_mask}; while (remaining_size > 0) { const std::size_t copy_amount{ std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; - const u8* page_pointer = page_table.pointers[page_index]; - if (page_pointer) { - const VAddr src_addr{page_table.backing_addr[page_index] + page_offset}; - memory.ReadBlockUnsafe(src_addr, dest_buffer, copy_amount); + + if (const auto page_addr{GpuToCpuAddress(page_index << page_bits)}; page_addr) { + const auto src_addr{*page_addr + page_offset}; + system.Memory().ReadBlockUnsafe(src_addr, dest_buffer, copy_amount); } else { std::memset(dest_buffer, 0, copy_amount); } + page_index++; page_offset = 0; dest_buffer = static_cast<u8*>(dest_buffer) + copy_amount; @@ -294,28 +264,22 @@ void MemoryManager::ReadBlockUnsafe(GPUVAddr src_addr, void* dest_buffer, } } -void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, const std::size_t size) { +void MemoryManager::WriteBlock(GPUVAddr gpu_dest_addr, const void* src_buffer, std::size_t size) { std::size_t remaining_size{size}; - std::size_t page_index{dest_addr >> page_bits}; - std::size_t page_offset{dest_addr & page_mask}; - - auto& memory = system.Memory(); + std::size_t page_index{gpu_dest_addr >> page_bits}; + std::size_t page_offset{gpu_dest_addr & page_mask}; while (remaining_size > 0) { const std::size_t copy_amount{ std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; - switch (page_table.attributes[page_index]) { - case Common::PageType::Memory: { - const VAddr dest_addr{page_table.backing_addr[page_index] + page_offset}; + if (const auto page_addr{GpuToCpuAddress(page_index << page_bits)}; page_addr) { + const auto dest_addr{*page_addr + page_offset}; + // Invalidate must happen on the rasterizer interface, such that memory is always // synchronous when it is written (even when in asynchronous GPU mode). - rasterizer.InvalidateRegion(dest_addr, copy_amount); - memory.WriteBlockUnsafe(dest_addr, src_buffer, copy_amount); - break; - } - default: - UNREACHABLE(); + rasterizer->InvalidateRegion(dest_addr, copy_amount); + system.Memory().WriteBlockUnsafe(dest_addr, src_buffer, copy_amount); } page_index++; @@ -325,22 +289,21 @@ void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, const } } -void MemoryManager::WriteBlockUnsafe(GPUVAddr dest_addr, const void* src_buffer, - const std::size_t size) { +void MemoryManager::WriteBlockUnsafe(GPUVAddr gpu_dest_addr, const void* src_buffer, + std::size_t size) { std::size_t remaining_size{size}; - std::size_t page_index{dest_addr >> page_bits}; - std::size_t page_offset{dest_addr & page_mask}; - - auto& memory = system.Memory(); + std::size_t page_index{gpu_dest_addr >> page_bits}; + std::size_t page_offset{gpu_dest_addr & page_mask}; while (remaining_size > 0) { const std::size_t copy_amount{ std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; - u8* page_pointer = page_table.pointers[page_index]; - if (page_pointer) { - const VAddr dest_addr{page_table.backing_addr[page_index] + page_offset}; - memory.WriteBlockUnsafe(dest_addr, src_buffer, copy_amount); + + if (const auto page_addr{GpuToCpuAddress(page_index << page_bits)}; page_addr) { + const auto dest_addr{*page_addr + page_offset}; + system.Memory().WriteBlockUnsafe(dest_addr, src_buffer, copy_amount); } + page_index++; page_offset = 0; src_buffer = static_cast<const u8*>(src_buffer) + copy_amount; @@ -348,270 +311,26 @@ void MemoryManager::WriteBlockUnsafe(GPUVAddr dest_addr, const void* src_buffer, } } -void MemoryManager::CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, const std::size_t size) { +void MemoryManager::CopyBlock(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, std::size_t size) { std::vector<u8> tmp_buffer(size); - ReadBlock(src_addr, tmp_buffer.data(), size); - WriteBlock(dest_addr, tmp_buffer.data(), size); + ReadBlock(gpu_src_addr, tmp_buffer.data(), size); + WriteBlock(gpu_dest_addr, tmp_buffer.data(), size); } -void MemoryManager::CopyBlockUnsafe(GPUVAddr dest_addr, GPUVAddr src_addr, const std::size_t size) { +void MemoryManager::CopyBlockUnsafe(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, + std::size_t size) { std::vector<u8> tmp_buffer(size); - ReadBlockUnsafe(src_addr, tmp_buffer.data(), size); - WriteBlockUnsafe(dest_addr, tmp_buffer.data(), size); -} - -bool MemoryManager::IsGranularRange(GPUVAddr gpu_addr, std::size_t size) { - const VAddr addr = page_table.backing_addr[gpu_addr >> page_bits]; - const std::size_t page = (addr & Memory::PAGE_MASK) + size; - return page <= Memory::PAGE_SIZE; -} - -void MemoryManager::MapPages(GPUVAddr base, u64 size, u8* memory, Common::PageType type, - VAddr backing_addr) { - LOG_DEBUG(HW_GPU, "Mapping {} onto {:016X}-{:016X}", fmt::ptr(memory), base * page_size, - (base + size) * page_size); - - const VAddr end{base + size}; - ASSERT_MSG(end <= page_table.pointers.size(), "out of range mapping at {:016X}", - base + page_table.pointers.size()); - - std::fill(page_table.attributes.begin() + base, page_table.attributes.begin() + end, type); - - if (memory == nullptr) { - std::fill(page_table.pointers.begin() + base, page_table.pointers.begin() + end, memory); - std::fill(page_table.backing_addr.begin() + base, page_table.backing_addr.begin() + end, - backing_addr); - } else { - while (base != end) { - page_table.pointers[base] = memory; - page_table.backing_addr[base] = backing_addr; - - base += 1; - memory += page_size; - backing_addr += page_size; - } - } -} - -void MemoryManager::MapMemoryRegion(GPUVAddr base, u64 size, u8* target, VAddr backing_addr) { - ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: {:016X}", size); - ASSERT_MSG((base & page_mask) == 0, "non-page aligned base: {:016X}", base); - MapPages(base / page_size, size / page_size, target, Common::PageType::Memory, backing_addr); -} - -void MemoryManager::UnmapRegion(GPUVAddr base, u64 size) { - ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: {:016X}", size); - ASSERT_MSG((base & page_mask) == 0, "non-page aligned base: {:016X}", base); - MapPages(base / page_size, size / page_size, nullptr, Common::PageType::Unmapped); -} - -bool VirtualMemoryArea::CanBeMergedWith(const VirtualMemoryArea& next) const { - ASSERT(base + size == next.base); - if (type != next.type) { - return {}; - } - if (type == VirtualMemoryArea::Type::Allocated && (offset + size != next.offset)) { - return {}; - } - if (type == VirtualMemoryArea::Type::Mapped && backing_memory + size != next.backing_memory) { - return {}; - } - return true; -} - -MemoryManager::VMAHandle MemoryManager::FindVMA(GPUVAddr target) const { - if (target >= address_space_end) { - return vma_map.end(); - } else { - return std::prev(vma_map.upper_bound(target)); - } -} - -MemoryManager::VMAIter MemoryManager::Allocate(VMAIter vma_handle) { - VirtualMemoryArea& vma{vma_handle->second}; - - vma.type = VirtualMemoryArea::Type::Allocated; - vma.backing_addr = 0; - vma.backing_memory = {}; - UpdatePageTableForVMA(vma); - - return MergeAdjacent(vma_handle); -} - -MemoryManager::VMAHandle MemoryManager::AllocateMemory(GPUVAddr target, std::size_t offset, - u64 size) { - - // This is the appropriately sized VMA that will turn into our allocation. - VMAIter vma_handle{CarveVMA(target, size)}; - VirtualMemoryArea& vma{vma_handle->second}; - - ASSERT(vma.size == size); - - vma.offset = offset; - - return Allocate(vma_handle); -} - -MemoryManager::VMAHandle MemoryManager::MapBackingMemory(GPUVAddr target, u8* memory, u64 size, - VAddr backing_addr) { - // This is the appropriately sized VMA that will turn into our allocation. - VMAIter vma_handle{CarveVMA(target, size)}; - VirtualMemoryArea& vma{vma_handle->second}; - - ASSERT(vma.size == size); - - vma.type = VirtualMemoryArea::Type::Mapped; - vma.backing_memory = memory; - vma.backing_addr = backing_addr; - UpdatePageTableForVMA(vma); - - return MergeAdjacent(vma_handle); -} - -void MemoryManager::UnmapRange(GPUVAddr target, u64 size) { - VMAIter vma{CarveVMARange(target, size)}; - const VAddr target_end{target + size}; - const VMAIter end{vma_map.end()}; - - // The comparison against the end of the range must be done using addresses since VMAs can be - // merged during this process, causing invalidation of the iterators. - while (vma != end && vma->second.base < target_end) { - // Unmapped ranges return to allocated state and can be reused - // This behavior is used by Super Mario Odyssey, Sonic Forces, and likely other games - vma = std::next(Allocate(vma)); - } - - ASSERT(FindVMA(target)->second.size >= size); -} - -MemoryManager::VMAIter MemoryManager::StripIterConstness(const VMAHandle& iter) { - // This uses a neat C++ trick to convert a const_iterator to a regular iterator, given - // non-const access to its container. - return vma_map.erase(iter, iter); // Erases an empty range of elements -} - -MemoryManager::VMAIter MemoryManager::CarveVMA(GPUVAddr base, u64 size) { - ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: 0x{:016X}", size); - ASSERT_MSG((base & page_mask) == 0, "non-page aligned base: 0x{:016X}", base); - - VMAIter vma_handle{StripIterConstness(FindVMA(base))}; - if (vma_handle == vma_map.end()) { - // Target address is outside the managed range - return {}; - } - - const VirtualMemoryArea& vma{vma_handle->second}; - if (vma.type == VirtualMemoryArea::Type::Mapped) { - // Region is already allocated - return vma_handle; - } - - const VAddr start_in_vma{base - vma.base}; - const VAddr end_in_vma{start_in_vma + size}; - - ASSERT_MSG(end_in_vma <= vma.size, "region size 0x{:016X} is less than required size 0x{:016X}", - vma.size, end_in_vma); - - if (end_in_vma < vma.size) { - // Split VMA at the end of the allocated region - SplitVMA(vma_handle, end_in_vma); - } - if (start_in_vma != 0) { - // Split VMA at the start of the allocated region - vma_handle = SplitVMA(vma_handle, start_in_vma); - } - - return vma_handle; -} - -MemoryManager::VMAIter MemoryManager::CarveVMARange(GPUVAddr target, u64 size) { - ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: 0x{:016X}", size); - ASSERT_MSG((target & page_mask) == 0, "non-page aligned base: 0x{:016X}", target); - - const VAddr target_end{target + size}; - ASSERT(target_end >= target); - ASSERT(size > 0); - - VMAIter begin_vma{StripIterConstness(FindVMA(target))}; - const VMAIter i_end{vma_map.lower_bound(target_end)}; - if (std::any_of(begin_vma, i_end, [](const auto& entry) { - return entry.second.type == VirtualMemoryArea::Type::Unmapped; - })) { - return {}; - } - - if (target != begin_vma->second.base) { - begin_vma = SplitVMA(begin_vma, target - begin_vma->second.base); - } - - VMAIter end_vma{StripIterConstness(FindVMA(target_end))}; - if (end_vma != vma_map.end() && target_end != end_vma->second.base) { - end_vma = SplitVMA(end_vma, target_end - end_vma->second.base); - } - - return begin_vma; -} - -MemoryManager::VMAIter MemoryManager::SplitVMA(VMAIter vma_handle, u64 offset_in_vma) { - VirtualMemoryArea& old_vma{vma_handle->second}; - VirtualMemoryArea new_vma{old_vma}; // Make a copy of the VMA - - // For now, don't allow no-op VMA splits (trying to split at a boundary) because it's probably - // a bug. This restriction might be removed later. - ASSERT(offset_in_vma < old_vma.size); - ASSERT(offset_in_vma > 0); - - old_vma.size = offset_in_vma; - new_vma.base += offset_in_vma; - new_vma.size -= offset_in_vma; - - switch (new_vma.type) { - case VirtualMemoryArea::Type::Unmapped: - break; - case VirtualMemoryArea::Type::Allocated: - new_vma.offset += offset_in_vma; - break; - case VirtualMemoryArea::Type::Mapped: - new_vma.backing_memory += offset_in_vma; - break; - } - - ASSERT(old_vma.CanBeMergedWith(new_vma)); - - return vma_map.emplace_hint(std::next(vma_handle), new_vma.base, new_vma); -} - -MemoryManager::VMAIter MemoryManager::MergeAdjacent(VMAIter iter) { - const VMAIter next_vma{std::next(iter)}; - if (next_vma != vma_map.end() && iter->second.CanBeMergedWith(next_vma->second)) { - iter->second.size += next_vma->second.size; - vma_map.erase(next_vma); - } - - if (iter != vma_map.begin()) { - VMAIter prev_vma{std::prev(iter)}; - if (prev_vma->second.CanBeMergedWith(iter->second)) { - prev_vma->second.size += iter->second.size; - vma_map.erase(iter); - iter = prev_vma; - } - } - - return iter; + ReadBlockUnsafe(gpu_src_addr, tmp_buffer.data(), size); + WriteBlockUnsafe(gpu_dest_addr, tmp_buffer.data(), size); } -void MemoryManager::UpdatePageTableForVMA(const VirtualMemoryArea& vma) { - switch (vma.type) { - case VirtualMemoryArea::Type::Unmapped: - UnmapRegion(vma.base, vma.size); - break; - case VirtualMemoryArea::Type::Allocated: - MapMemoryRegion(vma.base, vma.size, nullptr, vma.backing_addr); - break; - case VirtualMemoryArea::Type::Mapped: - MapMemoryRegion(vma.base, vma.size, vma.backing_memory, vma.backing_addr); - break; +bool MemoryManager::IsGranularRange(GPUVAddr gpu_addr, std::size_t size) const { + const auto cpu_addr{GpuToCpuAddress(gpu_addr)}; + if (!cpu_addr) { + return false; } + const std::size_t page{(*cpu_addr & Core::Memory::PAGE_MASK) + size}; + return page <= Core::Memory::PAGE_SIZE; } } // namespace Tegra diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index 0d9468535..c078193d9 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -6,9 +6,9 @@ #include <map> #include <optional> +#include <vector> #include "common/common_types.h" -#include "common/page_table.h" namespace VideoCore { class RasterizerInterface; @@ -20,58 +20,70 @@ class System; namespace Tegra { -/** - * Represents a VMA in an address space. A VMA is a contiguous region of virtual addressing space - * with homogeneous attributes across its extents. In this particular implementation each VMA is - * also backed by a single host memory allocation. - */ -struct VirtualMemoryArea { - enum class Type : u8 { - Unmapped, - Allocated, - Mapped, +class PageEntry final { +public: + enum class State : u32 { + Unmapped = static_cast<u32>(-1), + Allocated = static_cast<u32>(-2), }; - /// Virtual base address of the region. - GPUVAddr base{}; - /// Size of the region. - u64 size{}; - /// Memory area mapping type. - Type type{Type::Unmapped}; - /// CPU memory mapped address corresponding to this memory area. - VAddr backing_addr{}; - /// Offset into the backing_memory the mapping starts from. - std::size_t offset{}; - /// Pointer backing this VMA. - u8* backing_memory{}; - - /// Tests if this area can be merged to the right with `next`. - bool CanBeMergedWith(const VirtualMemoryArea& next) const; + constexpr PageEntry() = default; + constexpr PageEntry(State state) : state{state} {} + constexpr PageEntry(VAddr addr) : state{static_cast<State>(addr >> ShiftBits)} {} + + [[nodiscard]] constexpr bool IsUnmapped() const { + return state == State::Unmapped; + } + + [[nodiscard]] constexpr bool IsAllocated() const { + return state == State::Allocated; + } + + [[nodiscard]] constexpr bool IsValid() const { + return !IsUnmapped() && !IsAllocated(); + } + + [[nodiscard]] constexpr VAddr ToAddress() const { + if (!IsValid()) { + return {}; + } + + return static_cast<VAddr>(state) << ShiftBits; + } + + [[nodiscard]] constexpr PageEntry operator+(u64 offset) const { + // If this is a reserved value, offsets do not apply + if (!IsValid()) { + return *this; + } + return PageEntry{(static_cast<VAddr>(state) << ShiftBits) + offset}; + } + +private: + static constexpr std::size_t ShiftBits{12}; + + State state{State::Unmapped}; }; +static_assert(sizeof(PageEntry) == 4, "PageEntry is too large"); class MemoryManager final { public: - explicit MemoryManager(Core::System& system, VideoCore::RasterizerInterface& rasterizer); + explicit MemoryManager(Core::System& system); ~MemoryManager(); - GPUVAddr AllocateSpace(u64 size, u64 align); - GPUVAddr AllocateSpace(GPUVAddr addr, u64 size, u64 align); - GPUVAddr MapBufferEx(VAddr cpu_addr, u64 size); - GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr addr, u64 size); - GPUVAddr UnmapBuffer(GPUVAddr addr, u64 size); - std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr) const; + /// Binds a renderer to the memory manager. + void BindRasterizer(VideoCore::RasterizerInterface& rasterizer); + + [[nodiscard]] std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr) const; template <typename T> - T Read(GPUVAddr addr) const; + [[nodiscard]] T Read(GPUVAddr addr) const; template <typename T> void Write(GPUVAddr addr, T data); - u8* GetPointer(GPUVAddr addr); - const u8* GetPointer(GPUVAddr addr) const; - - /// Returns true if the block is continuous in host memory, false otherwise - bool IsBlockContinuous(GPUVAddr start, std::size_t size) const; + [[nodiscard]] u8* GetPointer(GPUVAddr addr); + [[nodiscard]] const u8* GetPointer(GPUVAddr addr) const; /** * ReadBlock and WriteBlock are full read and write operations over virtual @@ -79,9 +91,9 @@ public: * in the Host Memory counterpart. Note: This functions cause Host GPU Memory * Flushes and Invalidations, respectively to each operation. */ - void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const; - void WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size); - void CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size); + void ReadBlock(GPUVAddr gpu_src_addr, void* dest_buffer, std::size_t size) const; + void WriteBlock(GPUVAddr gpu_dest_addr, const void* src_buffer, std::size_t size); + void CopyBlock(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, std::size_t size); /** * ReadBlockUnsafe and WriteBlockUnsafe are special versions of ReadBlock and @@ -93,97 +105,51 @@ public: * WriteBlockUnsafe instead of WriteBlock since it shouldn't invalidate the texture * being flushed. */ - void ReadBlockUnsafe(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const; - void WriteBlockUnsafe(GPUVAddr dest_addr, const void* src_buffer, std::size_t size); - void CopyBlockUnsafe(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size); - - /** - * IsGranularRange checks if a gpu region can be simply read with a pointer - */ - bool IsGranularRange(GPUVAddr gpu_addr, std::size_t size); - -private: - using VMAMap = std::map<GPUVAddr, VirtualMemoryArea>; - using VMAHandle = VMAMap::const_iterator; - using VMAIter = VMAMap::iterator; - - bool IsAddressValid(GPUVAddr addr) const; - void MapPages(GPUVAddr base, u64 size, u8* memory, Common::PageType type, - VAddr backing_addr = 0); - void MapMemoryRegion(GPUVAddr base, u64 size, u8* target, VAddr backing_addr); - void UnmapRegion(GPUVAddr base, u64 size); - - /// Finds the VMA in which the given address is included in, or `vma_map.end()`. - VMAHandle FindVMA(GPUVAddr target) const; - - VMAHandle AllocateMemory(GPUVAddr target, std::size_t offset, u64 size); - - /** - * Maps an unmanaged host memory pointer at a given address. - * - * @param target The guest address to start the mapping at. - * @param memory The memory to be mapped. - * @param size Size of the mapping in bytes. - * @param backing_addr The base address of the range to back this mapping. - */ - VMAHandle MapBackingMemory(GPUVAddr target, u8* memory, u64 size, VAddr backing_addr); - - /// Unmaps a range of addresses, splitting VMAs as necessary. - void UnmapRange(GPUVAddr target, u64 size); - - /// Converts a VMAHandle to a mutable VMAIter. - VMAIter StripIterConstness(const VMAHandle& iter); - - /// Marks as the specified VMA as allocated. - VMAIter Allocate(VMAIter vma); + void ReadBlockUnsafe(GPUVAddr gpu_src_addr, void* dest_buffer, std::size_t size) const; + void WriteBlockUnsafe(GPUVAddr gpu_dest_addr, const void* src_buffer, std::size_t size); + void CopyBlockUnsafe(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, std::size_t size); /** - * Carves a VMA of a specific size at the specified address by splitting Free VMAs while doing - * the appropriate error checking. + * IsGranularRange checks if a gpu region can be simply read with a pointer. */ - VMAIter CarveVMA(GPUVAddr base, u64 size); + [[nodiscard]] bool IsGranularRange(GPUVAddr gpu_addr, std::size_t size) const; - /** - * Splits the edges of the given range of non-Free VMAs so that there is a VMA split at each - * end of the range. - */ - VMAIter CarveVMARange(GPUVAddr base, u64 size); - - /** - * Splits a VMA in two, at the specified offset. - * @returns the right side of the split, with the original iterator becoming the left side. - */ - VMAIter SplitVMA(VMAIter vma, u64 offset_in_vma); - - /** - * Checks for and merges the specified VMA with adjacent ones if possible. - * @returns the merged VMA or the original if no merging was possible. - */ - VMAIter MergeAdjacent(VMAIter vma); - - /// Updates the pages corresponding to this VMA so they match the VMA's attributes. - void UpdatePageTableForVMA(const VirtualMemoryArea& vma); - - /// Finds a free (unmapped region) of the specified size starting at the specified address. - GPUVAddr FindFreeRegion(GPUVAddr region_start, u64 size) const; + [[nodiscard]] GPUVAddr Map(VAddr cpu_addr, GPUVAddr gpu_addr, std::size_t size); + [[nodiscard]] GPUVAddr MapAllocate(VAddr cpu_addr, std::size_t size, std::size_t align); + [[nodiscard]] GPUVAddr MapAllocate32(VAddr cpu_addr, std::size_t size); + [[nodiscard]] std::optional<GPUVAddr> AllocateFixed(GPUVAddr gpu_addr, std::size_t size); + [[nodiscard]] GPUVAddr Allocate(std::size_t size, std::size_t align); + void Unmap(GPUVAddr gpu_addr, std::size_t size); private: + [[nodiscard]] PageEntry GetPageEntry(GPUVAddr gpu_addr) const; + void SetPageEntry(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size = page_size); + GPUVAddr UpdateRange(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size); + [[nodiscard]] std::optional<GPUVAddr> FindFreeRange(std::size_t size, std::size_t align, + bool start_32bit_address = false) const; + + void TryLockPage(PageEntry page_entry, std::size_t size); + void TryUnlockPage(PageEntry page_entry, std::size_t size); + + [[nodiscard]] static constexpr std::size_t PageEntryIndex(GPUVAddr gpu_addr) { + return (gpu_addr >> page_bits) & page_table_mask; + } + + static constexpr u64 address_space_size = 1ULL << 40; + static constexpr u64 address_space_start = 1ULL << 32; + static constexpr u64 address_space_start_low = 1ULL << 16; static constexpr u64 page_bits{16}; static constexpr u64 page_size{1 << page_bits}; static constexpr u64 page_mask{page_size - 1}; + static constexpr u64 page_table_bits{24}; + static constexpr u64 page_table_size{1 << page_table_bits}; + static constexpr u64 page_table_mask{page_table_size - 1}; - /// Address space in bits, according to Tegra X1 TRM - static constexpr u32 address_space_width{40}; - /// Start address for mapping, this is fairly arbitrary but must be non-zero. - static constexpr GPUVAddr address_space_base{0x100000}; - /// End of address space, based on address space in bits. - static constexpr GPUVAddr address_space_end{1ULL << address_space_width}; + Core::System& system; - Common::BackingPageTable page_table{page_bits}; - VMAMap vma_map; - VideoCore::RasterizerInterface& rasterizer; + VideoCore::RasterizerInterface* rasterizer = nullptr; - Core::System& system; + std::vector<PageEntry> page_table; }; } // namespace Tegra diff --git a/src/video_core/morton.cpp b/src/video_core/morton.cpp index 6d522c318..9da9fb4ff 100644 --- a/src/video_core/morton.cpp +++ b/src/video_core/morton.cpp @@ -41,144 +41,168 @@ static void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth } static constexpr ConversionArray morton_to_linear_fns = { - MortonCopy<true, PixelFormat::ABGR8U>, - MortonCopy<true, PixelFormat::ABGR8S>, - MortonCopy<true, PixelFormat::ABGR8UI>, - MortonCopy<true, PixelFormat::B5G6R5U>, - MortonCopy<true, PixelFormat::A2B10G10R10U>, - MortonCopy<true, PixelFormat::A1B5G5R5U>, - MortonCopy<true, PixelFormat::R8U>, - MortonCopy<true, PixelFormat::R8UI>, - MortonCopy<true, PixelFormat::RGBA16F>, - MortonCopy<true, PixelFormat::RGBA16U>, - MortonCopy<true, PixelFormat::RGBA16S>, - MortonCopy<true, PixelFormat::RGBA16UI>, - MortonCopy<true, PixelFormat::R11FG11FB10F>, - MortonCopy<true, PixelFormat::RGBA32UI>, - MortonCopy<true, PixelFormat::DXT1>, - MortonCopy<true, PixelFormat::DXT23>, - MortonCopy<true, PixelFormat::DXT45>, - MortonCopy<true, PixelFormat::DXN1>, - MortonCopy<true, PixelFormat::DXN2UNORM>, - MortonCopy<true, PixelFormat::DXN2SNORM>, - MortonCopy<true, PixelFormat::BC7U>, - MortonCopy<true, PixelFormat::BC6H_UF16>, - MortonCopy<true, PixelFormat::BC6H_SF16>, - MortonCopy<true, PixelFormat::ASTC_2D_4X4>, - MortonCopy<true, PixelFormat::BGRA8>, - MortonCopy<true, PixelFormat::RGBA32F>, - MortonCopy<true, PixelFormat::RG32F>, - MortonCopy<true, PixelFormat::R32F>, - MortonCopy<true, PixelFormat::R16F>, - MortonCopy<true, PixelFormat::R16U>, - MortonCopy<true, PixelFormat::R16S>, - MortonCopy<true, PixelFormat::R16UI>, - MortonCopy<true, PixelFormat::R16I>, - MortonCopy<true, PixelFormat::RG16>, - MortonCopy<true, PixelFormat::RG16F>, - MortonCopy<true, PixelFormat::RG16UI>, - MortonCopy<true, PixelFormat::RG16I>, - MortonCopy<true, PixelFormat::RG16S>, - MortonCopy<true, PixelFormat::RGB32F>, - MortonCopy<true, PixelFormat::RGBA8_SRGB>, - MortonCopy<true, PixelFormat::RG8U>, - MortonCopy<true, PixelFormat::RG8S>, - MortonCopy<true, PixelFormat::RG32UI>, - MortonCopy<true, PixelFormat::RGBX16F>, - MortonCopy<true, PixelFormat::R32UI>, - MortonCopy<true, PixelFormat::R32I>, - MortonCopy<true, PixelFormat::ASTC_2D_8X8>, - MortonCopy<true, PixelFormat::ASTC_2D_8X5>, - MortonCopy<true, PixelFormat::ASTC_2D_5X4>, - MortonCopy<true, PixelFormat::BGRA8_SRGB>, - MortonCopy<true, PixelFormat::DXT1_SRGB>, - MortonCopy<true, PixelFormat::DXT23_SRGB>, - MortonCopy<true, PixelFormat::DXT45_SRGB>, - MortonCopy<true, PixelFormat::BC7U_SRGB>, - MortonCopy<true, PixelFormat::R4G4B4A4U>, + MortonCopy<true, PixelFormat::A8B8G8R8_UNORM>, + MortonCopy<true, PixelFormat::A8B8G8R8_SNORM>, + MortonCopy<true, PixelFormat::A8B8G8R8_SINT>, + MortonCopy<true, PixelFormat::A8B8G8R8_UINT>, + MortonCopy<true, PixelFormat::R5G6B5_UNORM>, + MortonCopy<true, PixelFormat::B5G6R5_UNORM>, + MortonCopy<true, PixelFormat::A1R5G5B5_UNORM>, + MortonCopy<true, PixelFormat::A2B10G10R10_UNORM>, + MortonCopy<true, PixelFormat::A2B10G10R10_UINT>, + MortonCopy<true, PixelFormat::A1B5G5R5_UNORM>, + MortonCopy<true, PixelFormat::R8_UNORM>, + MortonCopy<true, PixelFormat::R8_SNORM>, + MortonCopy<true, PixelFormat::R8_SINT>, + MortonCopy<true, PixelFormat::R8_UINT>, + MortonCopy<true, PixelFormat::R16G16B16A16_FLOAT>, + MortonCopy<true, PixelFormat::R16G16B16A16_UNORM>, + MortonCopy<true, PixelFormat::R16G16B16A16_SNORM>, + MortonCopy<true, PixelFormat::R16G16B16A16_SINT>, + MortonCopy<true, PixelFormat::R16G16B16A16_UINT>, + MortonCopy<true, PixelFormat::B10G11R11_FLOAT>, + MortonCopy<true, PixelFormat::R32G32B32A32_UINT>, + MortonCopy<true, PixelFormat::BC1_RGBA_UNORM>, + MortonCopy<true, PixelFormat::BC2_UNORM>, + MortonCopy<true, PixelFormat::BC3_UNORM>, + MortonCopy<true, PixelFormat::BC4_UNORM>, + MortonCopy<true, PixelFormat::BC4_SNORM>, + MortonCopy<true, PixelFormat::BC5_UNORM>, + MortonCopy<true, PixelFormat::BC5_SNORM>, + MortonCopy<true, PixelFormat::BC7_UNORM>, + MortonCopy<true, PixelFormat::BC6H_UFLOAT>, + MortonCopy<true, PixelFormat::BC6H_SFLOAT>, + MortonCopy<true, PixelFormat::ASTC_2D_4X4_UNORM>, + MortonCopy<true, PixelFormat::B8G8R8A8_UNORM>, + MortonCopy<true, PixelFormat::R32G32B32A32_FLOAT>, + MortonCopy<true, PixelFormat::R32G32B32A32_SINT>, + MortonCopy<true, PixelFormat::R32G32_FLOAT>, + MortonCopy<true, PixelFormat::R32G32_SINT>, + MortonCopy<true, PixelFormat::R32_FLOAT>, + MortonCopy<true, PixelFormat::R16_FLOAT>, + MortonCopy<true, PixelFormat::R16_UNORM>, + MortonCopy<true, PixelFormat::R16_SNORM>, + MortonCopy<true, PixelFormat::R16_UINT>, + MortonCopy<true, PixelFormat::R16_SINT>, + MortonCopy<true, PixelFormat::R16G16_UNORM>, + MortonCopy<true, PixelFormat::R16G16_FLOAT>, + MortonCopy<true, PixelFormat::R16G16_UINT>, + MortonCopy<true, PixelFormat::R16G16_SINT>, + MortonCopy<true, PixelFormat::R16G16_SNORM>, + MortonCopy<true, PixelFormat::R32G32B32_FLOAT>, + MortonCopy<true, PixelFormat::A8B8G8R8_SRGB>, + MortonCopy<true, PixelFormat::R8G8_UNORM>, + MortonCopy<true, PixelFormat::R8G8_SNORM>, + MortonCopy<true, PixelFormat::R8G8_SINT>, + MortonCopy<true, PixelFormat::R8G8_UINT>, + MortonCopy<true, PixelFormat::R32G32_UINT>, + MortonCopy<true, PixelFormat::R16G16B16X16_FLOAT>, + MortonCopy<true, PixelFormat::R32_UINT>, + MortonCopy<true, PixelFormat::R32_SINT>, + MortonCopy<true, PixelFormat::ASTC_2D_8X8_UNORM>, + MortonCopy<true, PixelFormat::ASTC_2D_8X5_UNORM>, + MortonCopy<true, PixelFormat::ASTC_2D_5X4_UNORM>, + MortonCopy<true, PixelFormat::B8G8R8A8_SRGB>, + MortonCopy<true, PixelFormat::BC1_RGBA_SRGB>, + MortonCopy<true, PixelFormat::BC2_SRGB>, + MortonCopy<true, PixelFormat::BC3_SRGB>, + MortonCopy<true, PixelFormat::BC7_SRGB>, + MortonCopy<true, PixelFormat::A4B4G4R4_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>, MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>, MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>, MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_5X5>, + MortonCopy<true, PixelFormat::ASTC_2D_5X5_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_10X8>, + MortonCopy<true, PixelFormat::ASTC_2D_10X8_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_6X6>, + MortonCopy<true, PixelFormat::ASTC_2D_6X6_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_6X6_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_10X10>, + MortonCopy<true, PixelFormat::ASTC_2D_10X10_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_10X10_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_12X12>, + MortonCopy<true, PixelFormat::ASTC_2D_12X12_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_12X12_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_8X6>, + MortonCopy<true, PixelFormat::ASTC_2D_8X6_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_8X6_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_6X5>, + MortonCopy<true, PixelFormat::ASTC_2D_6X5_UNORM>, MortonCopy<true, PixelFormat::ASTC_2D_6X5_SRGB>, - MortonCopy<true, PixelFormat::E5B9G9R9F>, - MortonCopy<true, PixelFormat::Z32F>, - MortonCopy<true, PixelFormat::Z16>, - MortonCopy<true, PixelFormat::Z24S8>, - MortonCopy<true, PixelFormat::S8Z24>, - MortonCopy<true, PixelFormat::Z32FS8>, + MortonCopy<true, PixelFormat::E5B9G9R9_FLOAT>, + MortonCopy<true, PixelFormat::D32_FLOAT>, + MortonCopy<true, PixelFormat::D16_UNORM>, + MortonCopy<true, PixelFormat::D24_UNORM_S8_UINT>, + MortonCopy<true, PixelFormat::S8_UINT_D24_UNORM>, + MortonCopy<true, PixelFormat::D32_FLOAT_S8_UINT>, }; static constexpr ConversionArray linear_to_morton_fns = { - MortonCopy<false, PixelFormat::ABGR8U>, - MortonCopy<false, PixelFormat::ABGR8S>, - MortonCopy<false, PixelFormat::ABGR8UI>, - MortonCopy<false, PixelFormat::B5G6R5U>, - MortonCopy<false, PixelFormat::A2B10G10R10U>, - MortonCopy<false, PixelFormat::A1B5G5R5U>, - MortonCopy<false, PixelFormat::R8U>, - MortonCopy<false, PixelFormat::R8UI>, - MortonCopy<false, PixelFormat::RGBA16F>, - MortonCopy<false, PixelFormat::RGBA16S>, - MortonCopy<false, PixelFormat::RGBA16U>, - MortonCopy<false, PixelFormat::RGBA16UI>, - MortonCopy<false, PixelFormat::R11FG11FB10F>, - MortonCopy<false, PixelFormat::RGBA32UI>, - MortonCopy<false, PixelFormat::DXT1>, - MortonCopy<false, PixelFormat::DXT23>, - MortonCopy<false, PixelFormat::DXT45>, - MortonCopy<false, PixelFormat::DXN1>, - MortonCopy<false, PixelFormat::DXN2UNORM>, - MortonCopy<false, PixelFormat::DXN2SNORM>, - MortonCopy<false, PixelFormat::BC7U>, - MortonCopy<false, PixelFormat::BC6H_UF16>, - MortonCopy<false, PixelFormat::BC6H_SF16>, + MortonCopy<false, PixelFormat::A8B8G8R8_UNORM>, + MortonCopy<false, PixelFormat::A8B8G8R8_SNORM>, + MortonCopy<false, PixelFormat::A8B8G8R8_SINT>, + MortonCopy<false, PixelFormat::A8B8G8R8_UINT>, + MortonCopy<false, PixelFormat::R5G6B5_UNORM>, + MortonCopy<false, PixelFormat::B5G6R5_UNORM>, + MortonCopy<false, PixelFormat::A1R5G5B5_UNORM>, + MortonCopy<false, PixelFormat::A2B10G10R10_UNORM>, + MortonCopy<false, PixelFormat::A2B10G10R10_UINT>, + MortonCopy<false, PixelFormat::A1B5G5R5_UNORM>, + MortonCopy<false, PixelFormat::R8_UNORM>, + MortonCopy<false, PixelFormat::R8_SNORM>, + MortonCopy<false, PixelFormat::R8_SINT>, + MortonCopy<false, PixelFormat::R8_UINT>, + MortonCopy<false, PixelFormat::R16G16B16A16_FLOAT>, + MortonCopy<false, PixelFormat::R16G16B16A16_SNORM>, + MortonCopy<false, PixelFormat::R16G16B16A16_SINT>, + MortonCopy<false, PixelFormat::R16G16B16A16_UNORM>, + MortonCopy<false, PixelFormat::R16G16B16A16_UINT>, + MortonCopy<false, PixelFormat::B10G11R11_FLOAT>, + MortonCopy<false, PixelFormat::R32G32B32A32_UINT>, + MortonCopy<false, PixelFormat::BC1_RGBA_UNORM>, + MortonCopy<false, PixelFormat::BC2_UNORM>, + MortonCopy<false, PixelFormat::BC3_UNORM>, + MortonCopy<false, PixelFormat::BC4_UNORM>, + MortonCopy<false, PixelFormat::BC4_SNORM>, + MortonCopy<false, PixelFormat::BC5_UNORM>, + MortonCopy<false, PixelFormat::BC5_SNORM>, + MortonCopy<false, PixelFormat::BC7_UNORM>, + MortonCopy<false, PixelFormat::BC6H_UFLOAT>, + MortonCopy<false, PixelFormat::BC6H_SFLOAT>, // TODO(Subv): Swizzling ASTC formats are not supported nullptr, - MortonCopy<false, PixelFormat::BGRA8>, - MortonCopy<false, PixelFormat::RGBA32F>, - MortonCopy<false, PixelFormat::RG32F>, - MortonCopy<false, PixelFormat::R32F>, - MortonCopy<false, PixelFormat::R16F>, - MortonCopy<false, PixelFormat::R16U>, - MortonCopy<false, PixelFormat::R16S>, - MortonCopy<false, PixelFormat::R16UI>, - MortonCopy<false, PixelFormat::R16I>, - MortonCopy<false, PixelFormat::RG16>, - MortonCopy<false, PixelFormat::RG16F>, - MortonCopy<false, PixelFormat::RG16UI>, - MortonCopy<false, PixelFormat::RG16I>, - MortonCopy<false, PixelFormat::RG16S>, - MortonCopy<false, PixelFormat::RGB32F>, - MortonCopy<false, PixelFormat::RGBA8_SRGB>, - MortonCopy<false, PixelFormat::RG8U>, - MortonCopy<false, PixelFormat::RG8S>, - MortonCopy<false, PixelFormat::RG32UI>, - MortonCopy<false, PixelFormat::RGBX16F>, - MortonCopy<false, PixelFormat::R32UI>, - MortonCopy<false, PixelFormat::R32I>, + MortonCopy<false, PixelFormat::B8G8R8A8_UNORM>, + MortonCopy<false, PixelFormat::R32G32B32A32_FLOAT>, + MortonCopy<false, PixelFormat::R32G32B32A32_SINT>, + MortonCopy<false, PixelFormat::R32G32_FLOAT>, + MortonCopy<false, PixelFormat::R32G32_SINT>, + MortonCopy<false, PixelFormat::R32_FLOAT>, + MortonCopy<false, PixelFormat::R16_FLOAT>, + MortonCopy<false, PixelFormat::R16_UNORM>, + MortonCopy<false, PixelFormat::R16_SNORM>, + MortonCopy<false, PixelFormat::R16_UINT>, + MortonCopy<false, PixelFormat::R16_SINT>, + MortonCopy<false, PixelFormat::R16G16_UNORM>, + MortonCopy<false, PixelFormat::R16G16_FLOAT>, + MortonCopy<false, PixelFormat::R16G16_UINT>, + MortonCopy<false, PixelFormat::R16G16_SINT>, + MortonCopy<false, PixelFormat::R16G16_SNORM>, + MortonCopy<false, PixelFormat::R32G32B32_FLOAT>, + MortonCopy<false, PixelFormat::A8B8G8R8_SRGB>, + MortonCopy<false, PixelFormat::R8G8_UNORM>, + MortonCopy<false, PixelFormat::R8G8_SNORM>, + MortonCopy<false, PixelFormat::R8G8_SINT>, + MortonCopy<false, PixelFormat::R8G8_UINT>, + MortonCopy<false, PixelFormat::R32G32_UINT>, + MortonCopy<false, PixelFormat::R16G16B16X16_FLOAT>, + MortonCopy<false, PixelFormat::R32_UINT>, + MortonCopy<false, PixelFormat::R32_SINT>, nullptr, nullptr, nullptr, - MortonCopy<false, PixelFormat::BGRA8_SRGB>, - MortonCopy<false, PixelFormat::DXT1_SRGB>, - MortonCopy<false, PixelFormat::DXT23_SRGB>, - MortonCopy<false, PixelFormat::DXT45_SRGB>, - MortonCopy<false, PixelFormat::BC7U_SRGB>, - MortonCopy<false, PixelFormat::R4G4B4A4U>, + MortonCopy<false, PixelFormat::B8G8R8A8_SRGB>, + MortonCopy<false, PixelFormat::BC1_RGBA_SRGB>, + MortonCopy<false, PixelFormat::BC2_SRGB>, + MortonCopy<false, PixelFormat::BC3_SRGB>, + MortonCopy<false, PixelFormat::BC7_SRGB>, + MortonCopy<false, PixelFormat::A4B4G4R4_UNORM>, nullptr, nullptr, nullptr, @@ -197,12 +221,12 @@ static constexpr ConversionArray linear_to_morton_fns = { nullptr, nullptr, nullptr, - MortonCopy<false, PixelFormat::E5B9G9R9F>, - MortonCopy<false, PixelFormat::Z32F>, - MortonCopy<false, PixelFormat::Z16>, - MortonCopy<false, PixelFormat::Z24S8>, - MortonCopy<false, PixelFormat::S8Z24>, - MortonCopy<false, PixelFormat::Z32FS8>, + MortonCopy<false, PixelFormat::E5B9G9R9_FLOAT>, + MortonCopy<false, PixelFormat::D32_FLOAT>, + MortonCopy<false, PixelFormat::D16_UNORM>, + MortonCopy<false, PixelFormat::D24_UNORM_S8_UINT>, + MortonCopy<false, PixelFormat::S8_UINT_D24_UNORM>, + MortonCopy<false, PixelFormat::D32_FLOAT_S8_UINT>, }; static MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFormat format) { diff --git a/src/video_core/query_cache.h b/src/video_core/query_cache.h index 5ea2b01f2..fc54ca0ef 100644 --- a/src/video_core/query_cache.h +++ b/src/video_core/query_cache.h @@ -12,10 +12,12 @@ #include <mutex> #include <optional> #include <unordered_map> +#include <unordered_set> #include <vector> #include "common/assert.h" #include "core/core.h" +#include "core/settings.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" @@ -89,14 +91,15 @@ private: std::shared_ptr<HostCounter> last; }; -template <class QueryCache, class CachedQuery, class CounterStream, class HostCounter, - class QueryPool> +template <class QueryCache, class CachedQuery, class CounterStream, class HostCounter> class QueryCacheBase { public: - explicit QueryCacheBase(Core::System& system, VideoCore::RasterizerInterface& rasterizer) - : system{system}, rasterizer{rasterizer}, streams{{CounterStream{ - static_cast<QueryCache&>(*this), - VideoCore::QueryType::SamplesPassed}}} {} + explicit QueryCacheBase(VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::MemoryManager& gpu_memory_) + : rasterizer{rasterizer_}, maxwell3d{maxwell3d_}, + gpu_memory{gpu_memory_}, streams{{CounterStream{static_cast<QueryCache&>(*this), + VideoCore::QueryType::SamplesPassed}}} {} void InvalidateRegion(VAddr addr, std::size_t size) { std::unique_lock lock{mutex}; @@ -116,26 +119,27 @@ public: */ void Query(GPUVAddr gpu_addr, VideoCore::QueryType type, std::optional<u64> timestamp) { std::unique_lock lock{mutex}; - auto& memory_manager = system.GPU().MemoryManager(); - const std::optional<VAddr> cpu_addr_opt = memory_manager.GpuToCpuAddress(gpu_addr); - ASSERT(cpu_addr_opt); - VAddr cpu_addr = *cpu_addr_opt; + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); + ASSERT(cpu_addr); - CachedQuery* query = TryGet(cpu_addr); + CachedQuery* query = TryGet(*cpu_addr); if (!query) { - ASSERT_OR_EXECUTE(cpu_addr_opt, return;); - const auto host_ptr = memory_manager.GetPointer(gpu_addr); + ASSERT_OR_EXECUTE(cpu_addr, return;); + u8* const host_ptr = gpu_memory.GetPointer(gpu_addr); - query = Register(type, cpu_addr, host_ptr, timestamp.has_value()); + query = Register(type, *cpu_addr, host_ptr, timestamp.has_value()); } query->BindCounter(Stream(type).Current(), timestamp); + if (Settings::values.use_asynchronous_gpu_emulation.GetValue()) { + AsyncFlushQuery(*cpu_addr); + } } /// Updates counters from GPU state. Expected to be called once per draw, clear or dispatch. void UpdateCounters() { std::unique_lock lock{mutex}; - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; Stream(VideoCore::QueryType::SamplesPassed).Update(regs.samplecnt_enable); } @@ -170,8 +174,36 @@ public: return streams[static_cast<std::size_t>(type)]; } -protected: - std::array<QueryPool, VideoCore::NumQueryTypes> query_pools; + void CommitAsyncFlushes() { + committed_flushes.push_back(uncommitted_flushes); + uncommitted_flushes.reset(); + } + + bool HasUncommittedFlushes() const { + return uncommitted_flushes != nullptr; + } + + bool ShouldWaitAsyncFlushes() const { + if (committed_flushes.empty()) { + return false; + } + return committed_flushes.front() != nullptr; + } + + void PopAsyncFlushes() { + if (committed_flushes.empty()) { + return; + } + auto& flush_list = committed_flushes.front(); + if (!flush_list) { + committed_flushes.pop_front(); + return; + } + for (VAddr query_address : *flush_list) { + FlushAndRemoveRegion(query_address, 4); + } + committed_flushes.pop_front(); + } private: /// Flushes a memory range to guest memory and removes it from the cache. @@ -184,8 +216,8 @@ private: return cache_begin < addr_end && addr_begin < cache_end; }; - const u64 page_end = addr_end >> PAGE_SHIFT; - for (u64 page = addr_begin >> PAGE_SHIFT; page <= page_end; ++page) { + const u64 page_end = addr_end >> PAGE_BITS; + for (u64 page = addr_begin >> PAGE_BITS; page <= page_end; ++page) { const auto& it = cached_queries.find(page); if (it == std::end(cached_queries)) { continue; @@ -206,14 +238,14 @@ private: /// Registers the passed parameters as cached and returns a pointer to the stored cached query. CachedQuery* Register(VideoCore::QueryType type, VAddr cpu_addr, u8* host_ptr, bool timestamp) { rasterizer.UpdatePagesCachedCount(cpu_addr, CachedQuery::SizeInBytes(timestamp), 1); - const u64 page = static_cast<u64>(cpu_addr) >> PAGE_SHIFT; + const u64 page = static_cast<u64>(cpu_addr) >> PAGE_BITS; return &cached_queries[page].emplace_back(static_cast<QueryCache&>(*this), type, cpu_addr, host_ptr); } /// Tries to a get a cached query. Returns nullptr on failure. CachedQuery* TryGet(VAddr addr) { - const u64 page = static_cast<u64>(addr) >> PAGE_SHIFT; + const u64 page = static_cast<u64>(addr) >> PAGE_BITS; const auto it = cached_queries.find(page); if (it == std::end(cached_queries)) { return nullptr; @@ -224,17 +256,28 @@ private: return found != std::end(contents) ? &*found : nullptr; } + void AsyncFlushQuery(VAddr addr) { + if (!uncommitted_flushes) { + uncommitted_flushes = std::make_shared<std::unordered_set<VAddr>>(); + } + uncommitted_flushes->insert(addr); + } + static constexpr std::uintptr_t PAGE_SIZE = 4096; - static constexpr unsigned PAGE_SHIFT = 12; + static constexpr unsigned PAGE_BITS = 12; - Core::System& system; VideoCore::RasterizerInterface& rasterizer; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::MemoryManager& gpu_memory; std::recursive_mutex mutex; std::unordered_map<u64, std::vector<CachedQuery>> cached_queries; std::array<CounterStream, VideoCore::NumQueryTypes> streams; + + std::shared_ptr<std::unordered_set<VAddr>> uncommitted_flushes{}; + std::list<std::shared_ptr<std::unordered_set<VAddr>>> committed_flushes; }; template <class QueryCache, class HostCounter> diff --git a/src/video_core/rasterizer_accelerated.cpp b/src/video_core/rasterizer_accelerated.cpp index d01db97da..53622ca05 100644 --- a/src/video_core/rasterizer_accelerated.cpp +++ b/src/video_core/rasterizer_accelerated.cpp @@ -23,15 +23,15 @@ constexpr auto RangeFromInterval(Map& map, const Interval& interval) { } // Anonymous namespace -RasterizerAccelerated::RasterizerAccelerated(Memory::Memory& cpu_memory_) +RasterizerAccelerated::RasterizerAccelerated(Core::Memory::Memory& cpu_memory_) : cpu_memory{cpu_memory_} {} RasterizerAccelerated::~RasterizerAccelerated() = default; void RasterizerAccelerated::UpdatePagesCachedCount(VAddr addr, u64 size, int delta) { std::lock_guard lock{pages_mutex}; - const u64 page_start{addr >> Memory::PAGE_BITS}; - const u64 page_end{(addr + size + Memory::PAGE_SIZE - 1) >> Memory::PAGE_BITS}; + const u64 page_start{addr >> Core::Memory::PAGE_BITS}; + const u64 page_end{(addr + size + Core::Memory::PAGE_SIZE - 1) >> Core::Memory::PAGE_BITS}; // Interval maps will erase segments if count reaches 0, so if delta is negative we have to // subtract after iterating @@ -44,8 +44,8 @@ void RasterizerAccelerated::UpdatePagesCachedCount(VAddr addr, u64 size, int del const auto interval = pair.first & pages_interval; const int count = pair.second; - const VAddr interval_start_addr = boost::icl::first(interval) << Memory::PAGE_BITS; - const VAddr interval_end_addr = boost::icl::last_next(interval) << Memory::PAGE_BITS; + const VAddr interval_start_addr = boost::icl::first(interval) << Core::Memory::PAGE_BITS; + const VAddr interval_end_addr = boost::icl::last_next(interval) << Core::Memory::PAGE_BITS; const u64 interval_size = interval_end_addr - interval_start_addr; if (delta > 0 && count == delta) { diff --git a/src/video_core/rasterizer_accelerated.h b/src/video_core/rasterizer_accelerated.h index 315798e7c..91866d7dd 100644 --- a/src/video_core/rasterizer_accelerated.h +++ b/src/video_core/rasterizer_accelerated.h @@ -11,7 +11,7 @@ #include "common/common_types.h" #include "video_core/rasterizer_interface.h" -namespace Memory { +namespace Core::Memory { class Memory; } @@ -20,7 +20,7 @@ namespace VideoCore { /// Implements the shared part in GPU accelerated rasterizers in RasterizerInterface. class RasterizerAccelerated : public RasterizerInterface { public: - explicit RasterizerAccelerated(Memory::Memory& cpu_memory_); + explicit RasterizerAccelerated(Core::Memory::Memory& cpu_memory_); ~RasterizerAccelerated() override; void UpdatePagesCachedCount(VAddr addr, u64 size, int delta) override; @@ -30,7 +30,7 @@ private: CachedPageMap cached_pages; std::mutex pages_mutex; - Memory::Memory& cpu_memory; + Core::Memory::Memory& cpu_memory; }; } // namespace VideoCore diff --git a/src/video_core/rasterizer_cache.cpp b/src/video_core/rasterizer_cache.cpp deleted file mode 100644 index 093b2cdf4..000000000 --- a/src/video_core/rasterizer_cache.cpp +++ /dev/null @@ -1,7 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "video_core/rasterizer_cache.h" - -RasterizerCacheObject::~RasterizerCacheObject() = default; diff --git a/src/video_core/rasterizer_cache.h b/src/video_core/rasterizer_cache.h deleted file mode 100644 index 22987751e..000000000 --- a/src/video_core/rasterizer_cache.h +++ /dev/null @@ -1,197 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <mutex> -#include <set> -#include <unordered_map> - -#include <boost/icl/interval_map.hpp> -#include <boost/range/iterator_range_core.hpp> - -#include "common/common_types.h" -#include "core/settings.h" -#include "video_core/gpu.h" -#include "video_core/rasterizer_interface.h" - -class RasterizerCacheObject { -public: - explicit RasterizerCacheObject(const VAddr cpu_addr) : cpu_addr{cpu_addr} {} - - virtual ~RasterizerCacheObject(); - - VAddr GetCpuAddr() const { - return cpu_addr; - } - - /// Gets the size of the shader in guest memory, required for cache management - virtual std::size_t GetSizeInBytes() const = 0; - - /// Sets whether the cached object should be considered registered - void SetIsRegistered(bool registered) { - is_registered = registered; - } - - /// Returns true if the cached object is registered - bool IsRegistered() const { - return is_registered; - } - - /// Returns true if the cached object is dirty - bool IsDirty() const { - return is_dirty; - } - - /// Returns ticks from when this cached object was last modified - u64 GetLastModifiedTicks() const { - return last_modified_ticks; - } - - /// Marks an object as recently modified, used to specify whether it is clean or dirty - template <class T> - void MarkAsModified(bool dirty, T& cache) { - is_dirty = dirty; - last_modified_ticks = cache.GetModifiedTicks(); - } - -private: - bool is_registered{}; ///< Whether the object is currently registered with the cache - bool is_dirty{}; ///< Whether the object is dirty (out of sync with guest memory) - u64 last_modified_ticks{}; ///< When the object was last modified, used for in-order flushing - VAddr cpu_addr{}; ///< Cpu address memory, unique from emulated virtual address space -}; - -template <class T> -class RasterizerCache : NonCopyable { - friend class RasterizerCacheObject; - -public: - explicit RasterizerCache(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {} - - /// Write any cached resources overlapping the specified region back to memory - void FlushRegion(VAddr addr, std::size_t size) { - std::lock_guard lock{mutex}; - - const auto& objects{GetSortedObjectsFromRegion(addr, size)}; - for (auto& object : objects) { - FlushObject(object); - } - } - - /// Mark the specified region as being invalidated - void InvalidateRegion(VAddr addr, u64 size) { - std::lock_guard lock{mutex}; - - const auto& objects{GetSortedObjectsFromRegion(addr, size)}; - for (auto& object : objects) { - if (!object->IsRegistered()) { - // Skip duplicates - continue; - } - Unregister(object); - } - } - - /// Invalidates everything in the cache - void InvalidateAll() { - std::lock_guard lock{mutex}; - - while (interval_cache.begin() != interval_cache.end()) { - Unregister(*interval_cache.begin()->second.begin()); - } - } - -protected: - /// Tries to get an object from the cache with the specified cache address - T TryGet(VAddr addr) const { - const auto iter = map_cache.find(addr); - if (iter != map_cache.end()) - return iter->second; - return nullptr; - } - - /// Register an object into the cache - virtual void Register(const T& object) { - std::lock_guard lock{mutex}; - - object->SetIsRegistered(true); - interval_cache.add({GetInterval(object), ObjectSet{object}}); - map_cache.insert({object->GetCpuAddr(), object}); - rasterizer.UpdatePagesCachedCount(object->GetCpuAddr(), object->GetSizeInBytes(), 1); - } - - /// Unregisters an object from the cache - virtual void Unregister(const T& object) { - std::lock_guard lock{mutex}; - - object->SetIsRegistered(false); - rasterizer.UpdatePagesCachedCount(object->GetCpuAddr(), object->GetSizeInBytes(), -1); - const VAddr addr = object->GetCpuAddr(); - interval_cache.subtract({GetInterval(object), ObjectSet{object}}); - map_cache.erase(addr); - } - - /// Returns a ticks counter used for tracking when cached objects were last modified - u64 GetModifiedTicks() { - std::lock_guard lock{mutex}; - - return ++modified_ticks; - } - - virtual void FlushObjectInner(const T& object) = 0; - - /// Flushes the specified object, updating appropriate cache state as needed - void FlushObject(const T& object) { - std::lock_guard lock{mutex}; - - if (!object->IsDirty()) { - return; - } - FlushObjectInner(object); - object->MarkAsModified(false, *this); - } - - std::recursive_mutex mutex; - -private: - /// Returns a list of cached objects from the specified memory region, ordered by access time - std::vector<T> GetSortedObjectsFromRegion(VAddr addr, u64 size) { - if (size == 0) { - return {}; - } - - std::vector<T> objects; - const ObjectInterval interval{addr, addr + size}; - for (auto& pair : boost::make_iterator_range(interval_cache.equal_range(interval))) { - for (auto& cached_object : pair.second) { - if (!cached_object) { - continue; - } - objects.push_back(cached_object); - } - } - - std::sort(objects.begin(), objects.end(), [](const T& a, const T& b) -> bool { - return a->GetLastModifiedTicks() < b->GetLastModifiedTicks(); - }); - - return objects; - } - - using ObjectSet = std::set<T>; - using ObjectCache = std::unordered_map<VAddr, T>; - using IntervalCache = boost::icl::interval_map<VAddr, ObjectSet>; - using ObjectInterval = typename IntervalCache::interval_type; - - static auto GetInterval(const T& object) { - return ObjectInterval::right_open(object->GetCpuAddr(), - object->GetCpuAddr() + object->GetSizeInBytes()); - } - - ObjectCache map_cache; - IntervalCache interval_cache; ///< Cache of objects - u64 modified_ticks{}; ///< Counter of cache state ticks, used for in-order flushing - VideoCore::RasterizerInterface& rasterizer; -}; diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h index 8ae5b9c4e..27ef4c69a 100644 --- a/src/video_core/rasterizer_interface.h +++ b/src/video_core/rasterizer_interface.h @@ -32,7 +32,7 @@ using DiskResourceLoadCallback = std::function<void(LoadCallbackStage, std::size class RasterizerInterface { public: - virtual ~RasterizerInterface() {} + virtual ~RasterizerInterface() = default; /// Dispatches a draw invocation virtual void Draw(bool is_indexed, bool is_instanced) = 0; @@ -49,19 +49,40 @@ public: /// Records a GPU query and caches it virtual void Query(GPUVAddr gpu_addr, QueryType type, std::optional<u64> timestamp) = 0; + /// Signal a GPU based semaphore as a fence + virtual void SignalSemaphore(GPUVAddr addr, u32 value) = 0; + + /// Signal a GPU based syncpoint as a fence + virtual void SignalSyncPoint(u32 value) = 0; + + /// Release all pending fences. + virtual void ReleaseFences() = 0; + /// Notify rasterizer that all caches should be flushed to Switch memory virtual void FlushAll() = 0; /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory virtual void FlushRegion(VAddr addr, u64 size) = 0; + /// Check if the the specified memory area requires flushing to CPU Memory. + virtual bool MustFlushRegion(VAddr addr, u64 size) = 0; + /// Notify rasterizer that any caches of the specified region should be invalidated virtual void InvalidateRegion(VAddr addr, u64 size) = 0; + /// Notify rasterizer that any caches of the specified region are desync with guest + virtual void OnCPUWrite(VAddr addr, u64 size) = 0; + + /// Sync memory between guest and host. + virtual void SyncGuestHost() = 0; + /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory /// and invalidated virtual void FlushAndInvalidateRegion(VAddr addr, u64 size) = 0; + /// Notify the host renderer to wait for previous primitive and compute operations. + virtual void WaitForIdle() = 0; + /// Notify the rasterizer to send all written commands to the host GPU. virtual void FlushCommands() = 0; @@ -69,15 +90,16 @@ public: virtual void TickFrame() = 0; /// Attempt to use a faster method to perform a surface copy - virtual bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, - const Tegra::Engines::Fermi2D::Regs::Surface& dst, - const Tegra::Engines::Fermi2D::Config& copy_config) { + [[nodiscard]] virtual bool AccelerateSurfaceCopy( + const Tegra::Engines::Fermi2D::Regs::Surface& src, + const Tegra::Engines::Fermi2D::Regs::Surface& dst, + const Tegra::Engines::Fermi2D::Config& copy_config) { return false; } /// Attempt to use a faster method to display the framebuffer to screen - virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, - u32 pixel_stride) { + [[nodiscard]] virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config, + VAddr framebuffer_addr, u32 pixel_stride) { return false; } @@ -85,19 +107,16 @@ public: virtual void UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {} /// Initialize disk cached resources for the game being emulated - virtual void LoadDiskResources(const std::atomic_bool& stop_loading = false, - const DiskResourceLoadCallback& callback = {}) {} - - /// Initializes renderer dirty flags - virtual void SetupDirtyFlags() {} + virtual void LoadDiskResources(u64 title_id, const std::atomic_bool& stop_loading, + const DiskResourceLoadCallback& callback) {} /// Grant access to the Guest Driver Profile for recording/obtaining info on the guest driver. - GuestDriverProfile& AccessGuestDriverProfile() { + [[nodiscard]] GuestDriverProfile& AccessGuestDriverProfile() { return guest_driver_profile; } /// Grant access to the Guest Driver Profile for recording/obtaining info on the guest driver. - const GuestDriverProfile& AccessGuestDriverProfile() const { + [[nodiscard]] const GuestDriverProfile& AccessGuestDriverProfile() const { return guest_driver_profile; } diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp index 919d1f2d4..a93a1732c 100644 --- a/src/video_core/renderer_base.cpp +++ b/src/video_core/renderer_base.cpp @@ -9,7 +9,9 @@ namespace VideoCore { -RendererBase::RendererBase(Core::Frontend::EmuWindow& window) : render_window{window} { +RendererBase::RendererBase(Core::Frontend::EmuWindow& window_, + std::unique_ptr<Core::Frontend::GraphicsContext> context_) + : render_window{window_}, context{std::move(context_)} { RefreshBaseSettings(); } @@ -18,7 +20,7 @@ RendererBase::~RendererBase() = default; void RendererBase::RefreshBaseSettings() { UpdateCurrentFramebufferLayout(); - renderer_settings.use_framelimiter = Settings::values.use_frame_limit; + renderer_settings.use_framelimiter = Settings::values.use_frame_limit.GetValue(); renderer_settings.set_background_color = true; } diff --git a/src/video_core/renderer_base.h b/src/video_core/renderer_base.h index 1d85219b6..51dde8eb5 100644 --- a/src/video_core/renderer_base.h +++ b/src/video_core/renderer_base.h @@ -15,7 +15,8 @@ namespace Core::Frontend { class EmuWindow; -} +class GraphicsContext; +} // namespace Core::Frontend namespace VideoCore { @@ -25,18 +26,19 @@ struct RendererSettings { // Screenshot std::atomic<bool> screenshot_requested{false}; - void* screenshot_bits; + void* screenshot_bits{}; std::function<void()> screenshot_complete_callback; Layout::FramebufferLayout screenshot_framebuffer_layout; }; class RendererBase : NonCopyable { public: - explicit RendererBase(Core::Frontend::EmuWindow& window); + explicit RendererBase(Core::Frontend::EmuWindow& window, + std::unique_ptr<Core::Frontend::GraphicsContext> context); virtual ~RendererBase(); /// Initialize the renderer - virtual bool Init() = 0; + [[nodiscard]] virtual bool Init() = 0; /// Shutdown the renderer virtual void ShutDown() = 0; @@ -44,43 +46,46 @@ public: /// Finalize rendering the guest frame and draw into the presentation texture virtual void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) = 0; - /// Draws the latest frame to the window waiting timeout_ms for a frame to arrive (Renderer - /// specific implementation) - /// Returns true if a frame was drawn - virtual bool TryPresent(int timeout_ms) = 0; - // Getter/setter functions: // ------------------------ - f32 GetCurrentFPS() const { + [[nodiscard]] f32 GetCurrentFPS() const { return m_current_fps; } - int GetCurrentFrame() const { + [[nodiscard]] int GetCurrentFrame() const { return m_current_frame; } - RasterizerInterface& Rasterizer() { + [[nodiscard]] RasterizerInterface& Rasterizer() { return *rasterizer; } - const RasterizerInterface& Rasterizer() const { + [[nodiscard]] const RasterizerInterface& Rasterizer() const { return *rasterizer; } - Core::Frontend::EmuWindow& GetRenderWindow() { + [[nodiscard]] Core::Frontend::GraphicsContext& Context() { + return *context; + } + + [[nodiscard]] const Core::Frontend::GraphicsContext& Context() const { + return *context; + } + + [[nodiscard]] Core::Frontend::EmuWindow& GetRenderWindow() { return render_window; } - const Core::Frontend::EmuWindow& GetRenderWindow() const { + [[nodiscard]] const Core::Frontend::EmuWindow& GetRenderWindow() const { return render_window; } - RendererSettings& Settings() { + [[nodiscard]] RendererSettings& Settings() { return renderer_settings; } - const RendererSettings& Settings() const { + [[nodiscard]] const RendererSettings& Settings() const { return renderer_settings; } @@ -94,6 +99,7 @@ public: protected: Core::Frontend::EmuWindow& render_window; ///< Reference to the render window handle. std::unique_ptr<RasterizerInterface> rasterizer; + std::unique_ptr<Core::Frontend::GraphicsContext> context; f32 m_current_fps = 0.0f; ///< Current framerate, should be set by the renderer int m_current_frame = 0; ///< Current frame, should be set by the renderer diff --git a/src/video_core/renderer_opengl/gl_arb_decompiler.cpp b/src/video_core/renderer_opengl/gl_arb_decompiler.cpp new file mode 100644 index 000000000..d6120c23e --- /dev/null +++ b/src/video_core/renderer_opengl/gl_arb_decompiler.cpp @@ -0,0 +1,2126 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <array> +#include <cstddef> +#include <string> +#include <string_view> +#include <utility> +#include <variant> + +#include <fmt/format.h> + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/common_types.h" +#include "video_core/renderer_opengl/gl_arb_decompiler.h" +#include "video_core/renderer_opengl/gl_device.h" +#include "video_core/shader/registry.h" +#include "video_core/shader/shader_ir.h" + +// Predicates in the decompiled code follow the convention that -1 means true and 0 means false. +// GLASM lacks booleans, so they have to be implemented as integers. +// Using -1 for true is useful because both CMP.S and NOT.U can negate it, and CMP.S can be used to +// select between two values, because -1 will be evaluated as true and 0 as false. + +namespace OpenGL { + +namespace { + +using Tegra::Engines::ShaderType; +using Tegra::Shader::Attribute; +using Tegra::Shader::PixelImap; +using Tegra::Shader::Register; +using namespace VideoCommon::Shader; +using Operation = const OperationNode&; + +constexpr std::array INTERNAL_FLAG_NAMES = {"ZERO", "SIGN", "CARRY", "OVERFLOW"}; + +char Swizzle(std::size_t component) { + static constexpr std::string_view SWIZZLE{"xyzw"}; + return SWIZZLE.at(component); +} + +constexpr bool IsGenericAttribute(Attribute::Index index) { + return index >= Attribute::Index::Attribute_0 && index <= Attribute::Index::Attribute_31; +} + +u32 GetGenericAttributeIndex(Attribute::Index index) { + ASSERT(IsGenericAttribute(index)); + return static_cast<u32>(index) - static_cast<u32>(Attribute::Index::Attribute_0); +} + +std::string_view Modifiers(Operation operation) { + const auto meta = std::get_if<MetaArithmetic>(&operation.GetMeta()); + if (meta && meta->precise) { + return ".PREC"; + } + return ""; +} + +std::string_view GetInputFlags(PixelImap attribute) { + switch (attribute) { + case PixelImap::Perspective: + return ""; + case PixelImap::Constant: + return "FLAT "; + case PixelImap::ScreenLinear: + return "NOPERSPECTIVE "; + case PixelImap::Unused: + break; + } + UNIMPLEMENTED_MSG("Unknown attribute usage index={}", static_cast<int>(attribute)); + return {}; +} + +std::string_view ImageType(Tegra::Shader::ImageType image_type) { + switch (image_type) { + case Tegra::Shader::ImageType::Texture1D: + return "1D"; + case Tegra::Shader::ImageType::TextureBuffer: + return "BUFFER"; + case Tegra::Shader::ImageType::Texture1DArray: + return "ARRAY1D"; + case Tegra::Shader::ImageType::Texture2D: + return "2D"; + case Tegra::Shader::ImageType::Texture2DArray: + return "ARRAY2D"; + case Tegra::Shader::ImageType::Texture3D: + return "3D"; + } + UNREACHABLE(); + return {}; +} + +std::string_view StackName(MetaStackClass stack) { + switch (stack) { + case MetaStackClass::Ssy: + return "SSY"; + case MetaStackClass::Pbk: + return "PBK"; + } + UNREACHABLE(); + return ""; +}; + +std::string_view PrimitiveDescription(Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology topology) { + switch (topology) { + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::Points: + return "POINTS"; + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::Lines: + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::LineStrip: + return "LINES"; + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::LinesAdjacency: + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::LineStripAdjacency: + return "LINES_ADJACENCY"; + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::Triangles: + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::TriangleStrip: + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::TriangleFan: + return "TRIANGLES"; + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::TrianglesAdjacency: + case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::TriangleStripAdjacency: + return "TRIANGLES_ADJACENCY"; + default: + UNIMPLEMENTED_MSG("topology={}", static_cast<int>(topology)); + return "POINTS"; + } +} + +std::string_view TopologyName(Tegra::Shader::OutputTopology topology) { + switch (topology) { + case Tegra::Shader::OutputTopology::PointList: + return "POINTS"; + case Tegra::Shader::OutputTopology::LineStrip: + return "LINE_STRIP"; + case Tegra::Shader::OutputTopology::TriangleStrip: + return "TRIANGLE_STRIP"; + default: + UNIMPLEMENTED_MSG("Unknown output topology: {}", static_cast<u32>(topology)); + return "points"; + } +} + +std::string_view StageInputName(ShaderType stage) { + switch (stage) { + case ShaderType::Vertex: + case ShaderType::Geometry: + return "vertex"; + case ShaderType::Fragment: + return "fragment"; + case ShaderType::Compute: + return "invocation"; + default: + UNREACHABLE(); + return ""; + } +} + +std::string TextureType(const MetaTexture& meta) { + if (meta.sampler.is_buffer) { + return "BUFFER"; + } + std::string type; + if (meta.sampler.is_shadow) { + type += "SHADOW"; + } + if (meta.sampler.is_array) { + type += "ARRAY"; + } + type += [&meta] { + switch (meta.sampler.type) { + case Tegra::Shader::TextureType::Texture1D: + return "1D"; + case Tegra::Shader::TextureType::Texture2D: + return "2D"; + case Tegra::Shader::TextureType::Texture3D: + return "3D"; + case Tegra::Shader::TextureType::TextureCube: + return "CUBE"; + } + UNREACHABLE(); + return "2D"; + }(); + return type; +} + +class ARBDecompiler final { +public: + explicit ARBDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, + ShaderType stage, std::string_view identifier); + + std::string Code() const { + return shader_source; + } + +private: + void DefineGlobalMemory(); + + void DeclareHeader(); + void DeclareVertex(); + void DeclareGeometry(); + void DeclareFragment(); + void DeclareCompute(); + void DeclareInputAttributes(); + void DeclareOutputAttributes(); + void DeclareLocalMemory(); + void DeclareGlobalMemory(); + void DeclareConstantBuffers(); + void DeclareRegisters(); + void DeclareTemporaries(); + void DeclarePredicates(); + void DeclareInternalFlags(); + + void InitializeVariables(); + + void DecompileAST(); + void DecompileBranchMode(); + + void VisitAST(const ASTNode& node); + std::string VisitExpression(const Expr& node); + + void VisitBlock(const NodeBlock& bb); + + std::string Visit(const Node& node); + + std::tuple<std::string, std::string, std::size_t> BuildCoords(Operation); + std::string BuildAoffi(Operation); + std::string GlobalMemoryPointer(const GmemNode& gmem); + void Exit(); + + std::string Assign(Operation); + std::string Select(Operation); + std::string FClamp(Operation); + std::string FCastHalf0(Operation); + std::string FCastHalf1(Operation); + std::string FSqrt(Operation); + std::string FSwizzleAdd(Operation); + std::string HAdd2(Operation); + std::string HMul2(Operation); + std::string HFma2(Operation); + std::string HAbsolute(Operation); + std::string HNegate(Operation); + std::string HClamp(Operation); + std::string HCastFloat(Operation); + std::string HUnpack(Operation); + std::string HMergeF32(Operation); + std::string HMergeH0(Operation); + std::string HMergeH1(Operation); + std::string HPack2(Operation); + std::string LogicalAssign(Operation); + std::string LogicalPick2(Operation); + std::string LogicalAnd2(Operation); + std::string FloatOrdered(Operation); + std::string FloatUnordered(Operation); + std::string LogicalAddCarry(Operation); + std::string Texture(Operation); + std::string TextureGather(Operation); + std::string TextureQueryDimensions(Operation); + std::string TextureQueryLod(Operation); + std::string TexelFetch(Operation); + std::string TextureGradient(Operation); + std::string ImageLoad(Operation); + std::string ImageStore(Operation); + std::string Branch(Operation); + std::string BranchIndirect(Operation); + std::string PushFlowStack(Operation); + std::string PopFlowStack(Operation); + std::string Exit(Operation); + std::string Discard(Operation); + std::string EmitVertex(Operation); + std::string EndPrimitive(Operation); + std::string InvocationId(Operation); + std::string YNegate(Operation); + std::string ThreadId(Operation); + std::string ShuffleIndexed(Operation); + std::string Barrier(Operation); + std::string MemoryBarrierGroup(Operation); + std::string MemoryBarrierGlobal(Operation); + + template <const std::string_view& op> + std::string Unary(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("{}{} {}, {};", op, Modifiers(operation), temporary, Visit(operation[0])); + return temporary; + } + + template <const std::string_view& op> + std::string Binary(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("{}{} {}, {}, {};", op, Modifiers(operation), temporary, Visit(operation[0]), + Visit(operation[1])); + return temporary; + } + + template <const std::string_view& op> + std::string Trinary(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("{}{} {}, {}, {}, {};", op, Modifiers(operation), temporary, Visit(operation[0]), + Visit(operation[1]), Visit(operation[2])); + return temporary; + } + + template <const std::string_view& op, bool unordered> + std::string FloatComparison(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("TRUNC.U.CC RC.x, {};", Binary<op>(operation)); + AddLine("MOV.S {}, 0;", temporary); + AddLine("MOV.S {} (NE.x), -1;", temporary); + + const std::string op_a = Visit(operation[0]); + const std::string op_b = Visit(operation[1]); + if constexpr (unordered) { + AddLine("SNE.F RC.x, {}, {};", op_a, op_a); + AddLine("TRUNC.U.CC RC.x, RC.x;"); + AddLine("MOV.S {} (NE.x), -1;", temporary); + AddLine("SNE.F RC.x, {}, {};", op_b, op_b); + AddLine("TRUNC.U.CC RC.x, RC.x;"); + AddLine("MOV.S {} (NE.x), -1;", temporary); + } else if (op == SNE_F) { + AddLine("SNE.F RC.x, {}, {};", op_a, op_a); + AddLine("TRUNC.U.CC RC.x, RC.x;"); + AddLine("MOV.S {} (NE.x), 0;", temporary); + AddLine("SNE.F RC.x, {}, {};", op_b, op_b); + AddLine("TRUNC.U.CC RC.x, RC.x;"); + AddLine("MOV.S {} (NE.x), 0;", temporary); + } + return temporary; + } + + template <const std::string_view& op, bool is_nan> + std::string HalfComparison(Operation operation) { + std::string tmp1 = AllocVectorTemporary(); + const std::string tmp2 = AllocVectorTemporary(); + const std::string op_a = Visit(operation[0]); + const std::string op_b = Visit(operation[1]); + AddLine("UP2H.F {}, {};", tmp1, op_a); + AddLine("UP2H.F {}, {};", tmp2, op_b); + AddLine("{} {}, {}, {};", op, tmp1, tmp1, tmp2); + AddLine("TRUNC.U.CC RC.xy, {};", tmp1); + AddLine("MOV.S {}.xy, {{0, 0, 0, 0}};", tmp1); + AddLine("MOV.S {}.x (NE.x), -1;", tmp1); + AddLine("MOV.S {}.y (NE.y), -1;", tmp1); + if constexpr (is_nan) { + AddLine("MOVC.F RC.x, {};", op_a); + AddLine("MOV.S {}.x (NAN.x), -1;", tmp1); + AddLine("MOVC.F RC.x, {};", op_b); + AddLine("MOV.S {}.y (NAN.x), -1;", tmp1); + } + return tmp1; + } + + template <const std::string_view& op, const std::string_view& type> + std::string AtomicImage(Operation operation) { + const auto& meta = std::get<MetaImage>(operation.GetMeta()); + const u32 image_id = device.GetBaseBindings(stage).image + meta.image.index; + const std::size_t num_coords = operation.GetOperandsCount(); + const std::size_t num_values = meta.values.size(); + + const std::string coord = AllocVectorTemporary(); + const std::string value = AllocVectorTemporary(); + for (std::size_t i = 0; i < num_coords; ++i) { + AddLine("MOV.S {}.{}, {};", coord, Swizzle(i), Visit(operation[i])); + } + for (std::size_t i = 0; i < num_values; ++i) { + AddLine("MOV.F {}.{}, {};", value, Swizzle(i), Visit(meta.values[i])); + } + + AddLine("ATOMIM.{}.{} {}.x, {}, {}, image[{}], {};", op, type, coord, value, coord, + image_id, ImageType(meta.image.type)); + return fmt::format("{}.x", coord); + } + + template <const std::string_view& op, const std::string_view& type> + std::string Atomic(Operation operation) { + std::string temporary = AllocTemporary(); + std::string address; + std::string_view opname; + bool robust = false; + if (const auto gmem = std::get_if<GmemNode>(&*operation[0])) { + address = GlobalMemoryPointer(*gmem); + opname = "ATOM"; + robust = true; + } else if (const auto smem = std::get_if<SmemNode>(&*operation[0])) { + address = fmt::format("shared_mem[{}]", Visit(smem->GetAddress())); + opname = "ATOMS"; + } else { + UNREACHABLE(); + return "{0, 0, 0, 0}"; + } + if (robust) { + AddLine("IF NE.x;"); + } + AddLine("{}.{}.{} {}, {}, {};", opname, op, type, temporary, Visit(operation[1]), address); + if (robust) { + AddLine("ELSE;"); + AddLine("MOV.S {}, 0;", temporary); + AddLine("ENDIF;"); + } + return temporary; + } + + template <char type> + std::string Negate(Operation operation) { + std::string temporary = AllocTemporary(); + if constexpr (type == 'F') { + AddLine("MOV.F32 {}, -{};", temporary, Visit(operation[0])); + } else { + AddLine("MOV.{} {}, -{};", type, temporary, Visit(operation[0])); + } + return temporary; + } + + template <char type> + std::string Absolute(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("MOV.{} {}, |{}|;", type, temporary, Visit(operation[0])); + return temporary; + } + + template <char type> + std::string BitfieldInsert(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("MOV.{} {}.x, {};", type, temporary, Visit(operation[3])); + AddLine("MOV.{} {}.y, {};", type, temporary, Visit(operation[2])); + AddLine("BFI.{} {}.x, {}, {}, {};", type, temporary, temporary, Visit(operation[1]), + Visit(operation[0])); + return fmt::format("{}.x", temporary); + } + + template <char type> + std::string BitfieldExtract(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("MOV.{} {}.x, {};", type, temporary, Visit(operation[2])); + AddLine("MOV.{} {}.y, {};", type, temporary, Visit(operation[1])); + AddLine("BFE.{} {}.x, {}, {};", type, temporary, temporary, Visit(operation[0])); + return fmt::format("{}.x", temporary); + } + + template <char swizzle> + std::string LocalInvocationId(Operation) { + return fmt::format("invocation.localid.{}", swizzle); + } + + template <char swizzle> + std::string WorkGroupId(Operation) { + return fmt::format("invocation.groupid.{}", swizzle); + } + + template <char c1, char c2> + std::string ThreadMask(Operation) { + return fmt::format("{}.thread{}{}mask", StageInputName(stage), c1, c2); + } + + template <typename... Args> + void AddExpression(std::string_view text, Args&&... args) { + shader_source += fmt::format(text, std::forward<Args>(args)...); + } + + template <typename... Args> + void AddLine(std::string_view text, Args&&... args) { + AddExpression(text, std::forward<Args>(args)...); + shader_source += '\n'; + } + + std::string AllocLongVectorTemporary() { + max_long_temporaries = std::max(max_long_temporaries, num_long_temporaries + 1); + return fmt::format("L{}", num_long_temporaries++); + } + + std::string AllocLongTemporary() { + return fmt::format("{}.x", AllocLongVectorTemporary()); + } + + std::string AllocVectorTemporary() { + max_temporaries = std::max(max_temporaries, num_temporaries + 1); + return fmt::format("T{}", num_temporaries++); + } + + std::string AllocTemporary() { + return fmt::format("{}.x", AllocVectorTemporary()); + } + + void ResetTemporaries() noexcept { + num_temporaries = 0; + num_long_temporaries = 0; + } + + const Device& device; + const ShaderIR& ir; + const Registry& registry; + const ShaderType stage; + + std::size_t num_temporaries = 0; + std::size_t max_temporaries = 0; + + std::size_t num_long_temporaries = 0; + std::size_t max_long_temporaries = 0; + + std::map<GlobalMemoryBase, u32> global_memory_names; + + std::string shader_source; + + static constexpr std::string_view ADD_F32 = "ADD.F32"; + static constexpr std::string_view ADD_S = "ADD.S"; + static constexpr std::string_view ADD_U = "ADD.U"; + static constexpr std::string_view MUL_F32 = "MUL.F32"; + static constexpr std::string_view MUL_S = "MUL.S"; + static constexpr std::string_view MUL_U = "MUL.U"; + static constexpr std::string_view DIV_F32 = "DIV.F32"; + static constexpr std::string_view DIV_S = "DIV.S"; + static constexpr std::string_view DIV_U = "DIV.U"; + static constexpr std::string_view MAD_F32 = "MAD.F32"; + static constexpr std::string_view RSQ_F32 = "RSQ.F32"; + static constexpr std::string_view COS_F32 = "COS.F32"; + static constexpr std::string_view SIN_F32 = "SIN.F32"; + static constexpr std::string_view EX2_F32 = "EX2.F32"; + static constexpr std::string_view LG2_F32 = "LG2.F32"; + static constexpr std::string_view SLT_F = "SLT.F32"; + static constexpr std::string_view SLT_S = "SLT.S"; + static constexpr std::string_view SLT_U = "SLT.U"; + static constexpr std::string_view SEQ_F = "SEQ.F32"; + static constexpr std::string_view SEQ_S = "SEQ.S"; + static constexpr std::string_view SEQ_U = "SEQ.U"; + static constexpr std::string_view SLE_F = "SLE.F32"; + static constexpr std::string_view SLE_S = "SLE.S"; + static constexpr std::string_view SLE_U = "SLE.U"; + static constexpr std::string_view SGT_F = "SGT.F32"; + static constexpr std::string_view SGT_S = "SGT.S"; + static constexpr std::string_view SGT_U = "SGT.U"; + static constexpr std::string_view SNE_F = "SNE.F32"; + static constexpr std::string_view SNE_S = "SNE.S"; + static constexpr std::string_view SNE_U = "SNE.U"; + static constexpr std::string_view SGE_F = "SGE.F32"; + static constexpr std::string_view SGE_S = "SGE.S"; + static constexpr std::string_view SGE_U = "SGE.U"; + static constexpr std::string_view AND_S = "AND.S"; + static constexpr std::string_view AND_U = "AND.U"; + static constexpr std::string_view TRUNC_F = "TRUNC.F"; + static constexpr std::string_view TRUNC_S = "TRUNC.S"; + static constexpr std::string_view TRUNC_U = "TRUNC.U"; + static constexpr std::string_view SHL_S = "SHL.S"; + static constexpr std::string_view SHL_U = "SHL.U"; + static constexpr std::string_view SHR_S = "SHR.S"; + static constexpr std::string_view SHR_U = "SHR.U"; + static constexpr std::string_view OR_S = "OR.S"; + static constexpr std::string_view OR_U = "OR.U"; + static constexpr std::string_view XOR_S = "XOR.S"; + static constexpr std::string_view XOR_U = "XOR.U"; + static constexpr std::string_view NOT_S = "NOT.S"; + static constexpr std::string_view NOT_U = "NOT.U"; + static constexpr std::string_view BTC_S = "BTC.S"; + static constexpr std::string_view BTC_U = "BTC.U"; + static constexpr std::string_view BTFM_S = "BTFM.S"; + static constexpr std::string_view BTFM_U = "BTFM.U"; + static constexpr std::string_view ROUND_F = "ROUND.F"; + static constexpr std::string_view CEIL_F = "CEIL.F"; + static constexpr std::string_view FLR_F = "FLR.F"; + static constexpr std::string_view I2F_S = "I2F.S"; + static constexpr std::string_view I2F_U = "I2F.U"; + static constexpr std::string_view MIN_F = "MIN.F"; + static constexpr std::string_view MIN_S = "MIN.S"; + static constexpr std::string_view MIN_U = "MIN.U"; + static constexpr std::string_view MAX_F = "MAX.F"; + static constexpr std::string_view MAX_S = "MAX.S"; + static constexpr std::string_view MAX_U = "MAX.U"; + static constexpr std::string_view MOV_U = "MOV.U"; + static constexpr std::string_view TGBALLOT_U = "TGBALLOT.U"; + static constexpr std::string_view TGALL_U = "TGALL.U"; + static constexpr std::string_view TGANY_U = "TGANY.U"; + static constexpr std::string_view TGEQ_U = "TGEQ.U"; + static constexpr std::string_view EXCH = "EXCH"; + static constexpr std::string_view ADD = "ADD"; + static constexpr std::string_view MIN = "MIN"; + static constexpr std::string_view MAX = "MAX"; + static constexpr std::string_view AND = "AND"; + static constexpr std::string_view OR = "OR"; + static constexpr std::string_view XOR = "XOR"; + static constexpr std::string_view U32 = "U32"; + static constexpr std::string_view S32 = "S32"; + + static constexpr std::size_t NUM_ENTRIES = static_cast<std::size_t>(OperationCode::Amount); + using DecompilerType = std::string (ARBDecompiler::*)(Operation); + static constexpr std::array<DecompilerType, NUM_ENTRIES> OPERATION_DECOMPILERS = { + &ARBDecompiler::Assign, + + &ARBDecompiler::Select, + + &ARBDecompiler::Binary<ADD_F32>, + &ARBDecompiler::Binary<MUL_F32>, + &ARBDecompiler::Binary<DIV_F32>, + &ARBDecompiler::Trinary<MAD_F32>, + &ARBDecompiler::Negate<'F'>, + &ARBDecompiler::Absolute<'F'>, + &ARBDecompiler::FClamp, + &ARBDecompiler::FCastHalf0, + &ARBDecompiler::FCastHalf1, + &ARBDecompiler::Binary<MIN_F>, + &ARBDecompiler::Binary<MAX_F>, + &ARBDecompiler::Unary<COS_F32>, + &ARBDecompiler::Unary<SIN_F32>, + &ARBDecompiler::Unary<EX2_F32>, + &ARBDecompiler::Unary<LG2_F32>, + &ARBDecompiler::Unary<RSQ_F32>, + &ARBDecompiler::FSqrt, + &ARBDecompiler::Unary<ROUND_F>, + &ARBDecompiler::Unary<FLR_F>, + &ARBDecompiler::Unary<CEIL_F>, + &ARBDecompiler::Unary<TRUNC_F>, + &ARBDecompiler::Unary<I2F_S>, + &ARBDecompiler::Unary<I2F_U>, + &ARBDecompiler::FSwizzleAdd, + + &ARBDecompiler::Binary<ADD_S>, + &ARBDecompiler::Binary<MUL_S>, + &ARBDecompiler::Binary<DIV_S>, + &ARBDecompiler::Negate<'S'>, + &ARBDecompiler::Absolute<'S'>, + &ARBDecompiler::Binary<MIN_S>, + &ARBDecompiler::Binary<MAX_S>, + + &ARBDecompiler::Unary<TRUNC_S>, + &ARBDecompiler::Unary<MOV_U>, + &ARBDecompiler::Binary<SHL_S>, + &ARBDecompiler::Binary<SHR_U>, + &ARBDecompiler::Binary<SHR_S>, + &ARBDecompiler::Binary<AND_S>, + &ARBDecompiler::Binary<OR_S>, + &ARBDecompiler::Binary<XOR_S>, + &ARBDecompiler::Unary<NOT_S>, + &ARBDecompiler::BitfieldInsert<'S'>, + &ARBDecompiler::BitfieldExtract<'S'>, + &ARBDecompiler::Unary<BTC_S>, + &ARBDecompiler::Unary<BTFM_S>, + + &ARBDecompiler::Binary<ADD_U>, + &ARBDecompiler::Binary<MUL_U>, + &ARBDecompiler::Binary<DIV_U>, + &ARBDecompiler::Binary<MIN_U>, + &ARBDecompiler::Binary<MAX_U>, + &ARBDecompiler::Unary<TRUNC_U>, + &ARBDecompiler::Unary<MOV_U>, + &ARBDecompiler::Binary<SHL_U>, + &ARBDecompiler::Binary<SHR_U>, + &ARBDecompiler::Binary<SHR_U>, + &ARBDecompiler::Binary<AND_U>, + &ARBDecompiler::Binary<OR_U>, + &ARBDecompiler::Binary<XOR_U>, + &ARBDecompiler::Unary<NOT_U>, + &ARBDecompiler::BitfieldInsert<'U'>, + &ARBDecompiler::BitfieldExtract<'U'>, + &ARBDecompiler::Unary<BTC_U>, + &ARBDecompiler::Unary<BTFM_U>, + + &ARBDecompiler::HAdd2, + &ARBDecompiler::HMul2, + &ARBDecompiler::HFma2, + &ARBDecompiler::HAbsolute, + &ARBDecompiler::HNegate, + &ARBDecompiler::HClamp, + &ARBDecompiler::HCastFloat, + &ARBDecompiler::HUnpack, + &ARBDecompiler::HMergeF32, + &ARBDecompiler::HMergeH0, + &ARBDecompiler::HMergeH1, + &ARBDecompiler::HPack2, + + &ARBDecompiler::LogicalAssign, + &ARBDecompiler::Binary<AND_U>, + &ARBDecompiler::Binary<OR_U>, + &ARBDecompiler::Binary<XOR_U>, + &ARBDecompiler::Unary<NOT_U>, + &ARBDecompiler::LogicalPick2, + &ARBDecompiler::LogicalAnd2, + + &ARBDecompiler::FloatComparison<SLT_F, false>, + &ARBDecompiler::FloatComparison<SEQ_F, false>, + &ARBDecompiler::FloatComparison<SLE_F, false>, + &ARBDecompiler::FloatComparison<SGT_F, false>, + &ARBDecompiler::FloatComparison<SNE_F, false>, + &ARBDecompiler::FloatComparison<SGE_F, false>, + &ARBDecompiler::FloatOrdered, + &ARBDecompiler::FloatUnordered, + &ARBDecompiler::FloatComparison<SLT_F, true>, + &ARBDecompiler::FloatComparison<SEQ_F, true>, + &ARBDecompiler::FloatComparison<SLE_F, true>, + &ARBDecompiler::FloatComparison<SGT_F, true>, + &ARBDecompiler::FloatComparison<SNE_F, true>, + &ARBDecompiler::FloatComparison<SGE_F, true>, + + &ARBDecompiler::Binary<SLT_S>, + &ARBDecompiler::Binary<SEQ_S>, + &ARBDecompiler::Binary<SLE_S>, + &ARBDecompiler::Binary<SGT_S>, + &ARBDecompiler::Binary<SNE_S>, + &ARBDecompiler::Binary<SGE_S>, + + &ARBDecompiler::Binary<SLT_U>, + &ARBDecompiler::Binary<SEQ_U>, + &ARBDecompiler::Binary<SLE_U>, + &ARBDecompiler::Binary<SGT_U>, + &ARBDecompiler::Binary<SNE_U>, + &ARBDecompiler::Binary<SGE_U>, + + &ARBDecompiler::LogicalAddCarry, + + &ARBDecompiler::HalfComparison<SLT_F, false>, + &ARBDecompiler::HalfComparison<SEQ_F, false>, + &ARBDecompiler::HalfComparison<SLE_F, false>, + &ARBDecompiler::HalfComparison<SGT_F, false>, + &ARBDecompiler::HalfComparison<SNE_F, false>, + &ARBDecompiler::HalfComparison<SGE_F, false>, + &ARBDecompiler::HalfComparison<SLT_F, true>, + &ARBDecompiler::HalfComparison<SEQ_F, true>, + &ARBDecompiler::HalfComparison<SLE_F, true>, + &ARBDecompiler::HalfComparison<SGT_F, true>, + &ARBDecompiler::HalfComparison<SNE_F, true>, + &ARBDecompiler::HalfComparison<SGE_F, true>, + + &ARBDecompiler::Texture, + &ARBDecompiler::Texture, + &ARBDecompiler::TextureGather, + &ARBDecompiler::TextureQueryDimensions, + &ARBDecompiler::TextureQueryLod, + &ARBDecompiler::TexelFetch, + &ARBDecompiler::TextureGradient, + + &ARBDecompiler::ImageLoad, + &ARBDecompiler::ImageStore, + + &ARBDecompiler::AtomicImage<ADD, U32>, + &ARBDecompiler::AtomicImage<AND, U32>, + &ARBDecompiler::AtomicImage<OR, U32>, + &ARBDecompiler::AtomicImage<XOR, U32>, + &ARBDecompiler::AtomicImage<EXCH, U32>, + + &ARBDecompiler::Atomic<EXCH, U32>, + &ARBDecompiler::Atomic<ADD, U32>, + &ARBDecompiler::Atomic<MIN, U32>, + &ARBDecompiler::Atomic<MAX, U32>, + &ARBDecompiler::Atomic<AND, U32>, + &ARBDecompiler::Atomic<OR, U32>, + &ARBDecompiler::Atomic<XOR, U32>, + + &ARBDecompiler::Atomic<EXCH, S32>, + &ARBDecompiler::Atomic<ADD, S32>, + &ARBDecompiler::Atomic<MIN, S32>, + &ARBDecompiler::Atomic<MAX, S32>, + &ARBDecompiler::Atomic<AND, S32>, + &ARBDecompiler::Atomic<OR, S32>, + &ARBDecompiler::Atomic<XOR, S32>, + + &ARBDecompiler::Atomic<ADD, U32>, + &ARBDecompiler::Atomic<MIN, U32>, + &ARBDecompiler::Atomic<MAX, U32>, + &ARBDecompiler::Atomic<AND, U32>, + &ARBDecompiler::Atomic<OR, U32>, + &ARBDecompiler::Atomic<XOR, U32>, + + &ARBDecompiler::Atomic<ADD, S32>, + &ARBDecompiler::Atomic<MIN, S32>, + &ARBDecompiler::Atomic<MAX, S32>, + &ARBDecompiler::Atomic<AND, S32>, + &ARBDecompiler::Atomic<OR, S32>, + &ARBDecompiler::Atomic<XOR, S32>, + + &ARBDecompiler::Branch, + &ARBDecompiler::BranchIndirect, + &ARBDecompiler::PushFlowStack, + &ARBDecompiler::PopFlowStack, + &ARBDecompiler::Exit, + &ARBDecompiler::Discard, + + &ARBDecompiler::EmitVertex, + &ARBDecompiler::EndPrimitive, + + &ARBDecompiler::InvocationId, + &ARBDecompiler::YNegate, + &ARBDecompiler::LocalInvocationId<'x'>, + &ARBDecompiler::LocalInvocationId<'y'>, + &ARBDecompiler::LocalInvocationId<'z'>, + &ARBDecompiler::WorkGroupId<'x'>, + &ARBDecompiler::WorkGroupId<'y'>, + &ARBDecompiler::WorkGroupId<'z'>, + + &ARBDecompiler::Unary<TGBALLOT_U>, + &ARBDecompiler::Unary<TGALL_U>, + &ARBDecompiler::Unary<TGANY_U>, + &ARBDecompiler::Unary<TGEQ_U>, + + &ARBDecompiler::ThreadId, + &ARBDecompiler::ThreadMask<'e', 'q'>, + &ARBDecompiler::ThreadMask<'g', 'e'>, + &ARBDecompiler::ThreadMask<'g', 't'>, + &ARBDecompiler::ThreadMask<'l', 'e'>, + &ARBDecompiler::ThreadMask<'l', 't'>, + &ARBDecompiler::ShuffleIndexed, + + &ARBDecompiler::Barrier, + &ARBDecompiler::MemoryBarrierGroup, + &ARBDecompiler::MemoryBarrierGlobal, + }; +}; + +ARBDecompiler::ARBDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, + ShaderType stage, std::string_view identifier) + : device{device}, ir{ir}, registry{registry}, stage{stage} { + DefineGlobalMemory(); + + AddLine("TEMP RC;"); + AddLine("TEMP FSWZA[4];"); + AddLine("TEMP FSWZB[4];"); + if (ir.IsDecompiled()) { + DecompileAST(); + } else { + DecompileBranchMode(); + } + AddLine("END"); + + const std::string code = std::move(shader_source); + DeclareHeader(); + DeclareVertex(); + DeclareGeometry(); + DeclareFragment(); + DeclareCompute(); + DeclareInputAttributes(); + DeclareOutputAttributes(); + DeclareLocalMemory(); + DeclareGlobalMemory(); + DeclareConstantBuffers(); + DeclareRegisters(); + DeclareTemporaries(); + DeclarePredicates(); + DeclareInternalFlags(); + + shader_source += code; +} + +std::string_view HeaderStageName(ShaderType stage) { + switch (stage) { + case ShaderType::Vertex: + return "vp"; + case ShaderType::Geometry: + return "gp"; + case ShaderType::Fragment: + return "fp"; + case ShaderType::Compute: + return "cp"; + default: + UNREACHABLE(); + return ""; + } +} + +void ARBDecompiler::DefineGlobalMemory() { + u32 binding = 0; + for (const auto& pair : ir.GetGlobalMemory()) { + const GlobalMemoryBase base = pair.first; + global_memory_names.emplace(base, binding); + ++binding; + } +} + +void ARBDecompiler::DeclareHeader() { + AddLine("!!NV{}5.0", HeaderStageName(stage)); + // Enabling this allows us to cheat on some instructions like TXL with SHADOWARRAY2D + AddLine("OPTION NV_internal;"); + AddLine("OPTION NV_gpu_program_fp64;"); + AddLine("OPTION NV_shader_thread_group;"); + if (ir.UsesWarps() && device.HasWarpIntrinsics()) { + AddLine("OPTION NV_shader_thread_shuffle;"); + } + if (stage == ShaderType::Vertex) { + if (device.HasNvViewportArray2()) { + AddLine("OPTION NV_viewport_array2;"); + } + } + if (stage == ShaderType::Fragment) { + AddLine("OPTION ARB_draw_buffers;"); + } + if (device.HasImageLoadFormatted()) { + AddLine("OPTION EXT_shader_image_load_formatted;"); + } +} + +void ARBDecompiler::DeclareVertex() { + if (stage != ShaderType::Vertex) { + return; + } + AddLine("OUTPUT result_clip[] = {{ result.clip[0..7] }};"); +} + +void ARBDecompiler::DeclareGeometry() { + if (stage != ShaderType::Geometry) { + return; + } + const auto& info = registry.GetGraphicsInfo(); + const auto& header = ir.GetHeader(); + AddLine("PRIMITIVE_IN {};", PrimitiveDescription(info.primitive_topology)); + AddLine("PRIMITIVE_OUT {};", TopologyName(header.common3.output_topology)); + AddLine("VERTICES_OUT {};", header.common4.max_output_vertices.Value()); + AddLine("ATTRIB vertex_position = vertex.position;"); +} + +void ARBDecompiler::DeclareFragment() { + if (stage != ShaderType::Fragment) { + return; + } + AddLine("OUTPUT result_color7 = result.color[7];"); + AddLine("OUTPUT result_color6 = result.color[6];"); + AddLine("OUTPUT result_color5 = result.color[5];"); + AddLine("OUTPUT result_color4 = result.color[4];"); + AddLine("OUTPUT result_color3 = result.color[3];"); + AddLine("OUTPUT result_color2 = result.color[2];"); + AddLine("OUTPUT result_color1 = result.color[1];"); + AddLine("OUTPUT result_color0 = result.color;"); +} + +void ARBDecompiler::DeclareCompute() { + if (stage != ShaderType::Compute) { + return; + } + const ComputeInfo& info = registry.GetComputeInfo(); + AddLine("GROUP_SIZE {} {} {};", info.workgroup_size[0], info.workgroup_size[1], + info.workgroup_size[2]); + if (info.shared_memory_size_in_words == 0) { + return; + } + const u32 limit = device.GetMaxComputeSharedMemorySize(); + u32 size_in_bytes = info.shared_memory_size_in_words * 4; + if (size_in_bytes > limit) { + LOG_ERROR(Render_OpenGL, "Shared memory size {} is clamped to host's limit {}", + size_in_bytes, limit); + size_in_bytes = limit; + } + + AddLine("SHARED_MEMORY {};", size_in_bytes); + AddLine("SHARED shared_mem[] = {{program.sharedmem}};"); +} + +void ARBDecompiler::DeclareInputAttributes() { + if (stage == ShaderType::Compute) { + return; + } + const std::string_view stage_name = StageInputName(stage); + for (const auto attribute : ir.GetInputAttributes()) { + if (!IsGenericAttribute(attribute)) { + continue; + } + const u32 index = GetGenericAttributeIndex(attribute); + + std::string_view suffix; + if (stage == ShaderType::Fragment) { + const auto input_mode{ir.GetHeader().ps.GetPixelImap(index)}; + if (input_mode == PixelImap::Unused) { + return; + } + suffix = GetInputFlags(input_mode); + } + AddLine("{}ATTRIB in_attr{}[] = {{ {}.attrib[{}..{}] }};", suffix, index, stage_name, index, + index); + } +} + +void ARBDecompiler::DeclareOutputAttributes() { + if (stage == ShaderType::Compute) { + return; + } + for (const auto attribute : ir.GetOutputAttributes()) { + if (!IsGenericAttribute(attribute)) { + continue; + } + const u32 index = GetGenericAttributeIndex(attribute); + AddLine("OUTPUT out_attr{}[] = {{ result.attrib[{}..{}] }};", index, index, index); + } +} + +void ARBDecompiler::DeclareLocalMemory() { + u64 size = 0; + if (stage == ShaderType::Compute) { + size = registry.GetComputeInfo().local_memory_size_in_words * 4ULL; + } else { + size = ir.GetHeader().GetLocalMemorySize(); + } + if (size == 0) { + return; + } + const u64 element_count = Common::AlignUp(size, 4) / 4; + AddLine("TEMP lmem[{}];", element_count); +} + +void ARBDecompiler::DeclareGlobalMemory() { + const size_t num_entries = ir.GetGlobalMemory().size(); + if (num_entries > 0) { + AddLine("PARAM c[{}] = {{ program.local[0..{}] }};", num_entries, num_entries - 1); + } +} + +void ARBDecompiler::DeclareConstantBuffers() { + u32 binding = 0; + for (const auto& cbuf : ir.GetConstantBuffers()) { + AddLine("CBUFFER cbuf{}[] = {{ program.buffer[{}] }};", cbuf.first, binding); + ++binding; + } +} + +void ARBDecompiler::DeclareRegisters() { + for (const u32 gpr : ir.GetRegisters()) { + AddLine("TEMP R{};", gpr); + } +} + +void ARBDecompiler::DeclareTemporaries() { + for (std::size_t i = 0; i < max_temporaries; ++i) { + AddLine("TEMP T{};", i); + } + for (std::size_t i = 0; i < max_long_temporaries; ++i) { + AddLine("LONG TEMP L{};", i); + } +} + +void ARBDecompiler::DeclarePredicates() { + for (const Tegra::Shader::Pred pred : ir.GetPredicates()) { + AddLine("TEMP P{};", static_cast<u64>(pred)); + } +} + +void ARBDecompiler::DeclareInternalFlags() { + for (const char* name : INTERNAL_FLAG_NAMES) { + AddLine("TEMP {};", name); + } +} + +void ARBDecompiler::InitializeVariables() { + AddLine("MOV.F32 FSWZA[0], -1;"); + AddLine("MOV.F32 FSWZA[1], 1;"); + AddLine("MOV.F32 FSWZA[2], -1;"); + AddLine("MOV.F32 FSWZA[3], 0;"); + AddLine("MOV.F32 FSWZB[0], -1;"); + AddLine("MOV.F32 FSWZB[1], -1;"); + AddLine("MOV.F32 FSWZB[2], 1;"); + AddLine("MOV.F32 FSWZB[3], -1;"); + + if (stage == ShaderType::Vertex || stage == ShaderType::Geometry) { + AddLine("MOV.F result.position, {{0, 0, 0, 1}};"); + } + for (const auto attribute : ir.GetOutputAttributes()) { + if (!IsGenericAttribute(attribute)) { + continue; + } + const u32 index = GetGenericAttributeIndex(attribute); + AddLine("MOV.F result.attrib[{}], {{0, 0, 0, 1}};", index); + } + for (const u32 gpr : ir.GetRegisters()) { + AddLine("MOV.F R{}, {{0, 0, 0, 0}};", gpr); + } + for (const Tegra::Shader::Pred pred : ir.GetPredicates()) { + AddLine("MOV.U P{}, {{0, 0, 0, 0}};", static_cast<u64>(pred)); + } +} + +void ARBDecompiler::DecompileAST() { + const u32 num_flow_variables = ir.GetASTNumVariables(); + for (u32 i = 0; i < num_flow_variables; ++i) { + AddLine("TEMP F{};", i); + } + for (u32 i = 0; i < num_flow_variables; ++i) { + AddLine("MOV.U F{}, {{0, 0, 0, 0}};", i); + } + + InitializeVariables(); + + VisitAST(ir.GetASTProgram()); +} + +void ARBDecompiler::DecompileBranchMode() { + static constexpr u32 FLOW_STACK_SIZE = 20; + if (!ir.IsFlowStackDisabled()) { + AddLine("TEMP SSY[{}];", FLOW_STACK_SIZE); + AddLine("TEMP PBK[{}];", FLOW_STACK_SIZE); + AddLine("TEMP SSY_TOP;"); + AddLine("TEMP PBK_TOP;"); + } + + AddLine("TEMP PC;"); + + if (!ir.IsFlowStackDisabled()) { + AddLine("MOV.U SSY_TOP.x, 0;"); + AddLine("MOV.U PBK_TOP.x, 0;"); + } + + InitializeVariables(); + + const auto basic_block_end = ir.GetBasicBlocks().end(); + auto basic_block_it = ir.GetBasicBlocks().begin(); + const u32 first_address = basic_block_it->first; + AddLine("MOV.U PC.x, {};", first_address); + + AddLine("REP;"); + + std::size_t num_blocks = 0; + while (basic_block_it != basic_block_end) { + const auto& [address, bb] = *basic_block_it; + ++num_blocks; + + AddLine("SEQ.S.CC RC.x, PC.x, {};", address); + AddLine("IF NE.x;"); + + VisitBlock(bb); + + ++basic_block_it; + + if (basic_block_it != basic_block_end) { + const auto op = std::get_if<OperationNode>(&*bb[bb.size() - 1]); + if (!op || op->GetCode() != OperationCode::Branch) { + const u32 next_address = basic_block_it->first; + AddLine("MOV.U PC.x, {};", next_address); + AddLine("CONT;"); + } + } + + AddLine("ELSE;"); + } + AddLine("RET;"); + while (num_blocks--) { + AddLine("ENDIF;"); + } + + AddLine("ENDREP;"); +} + +void ARBDecompiler::VisitAST(const ASTNode& node) { + if (const auto ast = std::get_if<ASTProgram>(&*node->GetInnerData())) { + for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + VisitAST(current); + } + } else if (const auto ast = std::get_if<ASTIfThen>(&*node->GetInnerData())) { + const std::string condition = VisitExpression(ast->condition); + ResetTemporaries(); + + AddLine("MOVC.U RC.x, {};", condition); + AddLine("IF NE.x;"); + for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + VisitAST(current); + } + AddLine("ENDIF;"); + } else if (const auto ast = std::get_if<ASTIfElse>(&*node->GetInnerData())) { + AddLine("ELSE;"); + for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + VisitAST(current); + } + } else if (const auto ast = std::get_if<ASTBlockDecoded>(&*node->GetInnerData())) { + VisitBlock(ast->nodes); + } else if (const auto ast = std::get_if<ASTVarSet>(&*node->GetInnerData())) { + AddLine("MOV.U F{}, {};", ast->index, VisitExpression(ast->condition)); + ResetTemporaries(); + } else if (const auto ast = std::get_if<ASTDoWhile>(&*node->GetInnerData())) { + const std::string condition = VisitExpression(ast->condition); + ResetTemporaries(); + AddLine("REP;"); + for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + VisitAST(current); + } + AddLine("MOVC.U RC.x, {};", condition); + AddLine("BRK (NE.x);"); + AddLine("ENDREP;"); + } else if (const auto ast = std::get_if<ASTReturn>(&*node->GetInnerData())) { + const bool is_true = ExprIsTrue(ast->condition); + if (!is_true) { + AddLine("MOVC.U RC.x, {};", VisitExpression(ast->condition)); + AddLine("IF NE.x;"); + ResetTemporaries(); + } + if (ast->kills) { + AddLine("KIL TR;"); + } else { + Exit(); + } + if (!is_true) { + AddLine("ENDIF;"); + } + } else if (const auto ast = std::get_if<ASTBreak>(&*node->GetInnerData())) { + if (ExprIsTrue(ast->condition)) { + AddLine("BRK;"); + } else { + AddLine("MOVC.U RC.x, {};", VisitExpression(ast->condition)); + AddLine("BRK (NE.x);"); + ResetTemporaries(); + } + } else if (std::holds_alternative<ASTLabel>(*node->GetInnerData())) { + // Nothing to do + } else { + UNREACHABLE(); + } +} + +std::string ARBDecompiler::VisitExpression(const Expr& node) { + if (const auto expr = std::get_if<ExprAnd>(&*node)) { + std::string result = AllocTemporary(); + AddLine("AND.U {}, {}, {};", result, VisitExpression(expr->operand1), + VisitExpression(expr->operand2)); + return result; + } + if (const auto expr = std::get_if<ExprOr>(&*node)) { + std::string result = AllocTemporary(); + AddLine("OR.U {}, {}, {};", result, VisitExpression(expr->operand1), + VisitExpression(expr->operand2)); + return result; + } + if (const auto expr = std::get_if<ExprNot>(&*node)) { + std::string result = AllocTemporary(); + AddLine("CMP.S {}, {}, 0, -1;", result, VisitExpression(expr->operand1)); + return result; + } + if (const auto expr = std::get_if<ExprPredicate>(&*node)) { + return fmt::format("P{}.x", static_cast<u64>(expr->predicate)); + } + if (const auto expr = std::get_if<ExprCondCode>(&*node)) { + return Visit(ir.GetConditionCode(expr->cc)); + } + if (const auto expr = std::get_if<ExprVar>(&*node)) { + return fmt::format("F{}.x", expr->var_index); + } + if (const auto expr = std::get_if<ExprBoolean>(&*node)) { + return expr->value ? "0xffffffff" : "0"; + } + if (const auto expr = std::get_if<ExprGprEqual>(&*node)) { + std::string result = AllocTemporary(); + AddLine("SEQ.U {}, R{}.x, {};", result, expr->gpr, expr->value); + return result; + } + UNREACHABLE(); + return "0"; +} + +void ARBDecompiler::VisitBlock(const NodeBlock& bb) { + for (const auto& node : bb) { + Visit(node); + } +} + +std::string ARBDecompiler::Visit(const Node& node) { + if (const auto operation = std::get_if<OperationNode>(&*node)) { + if (const auto amend_index = operation->GetAmendIndex()) { + Visit(ir.GetAmendNode(*amend_index)); + } + const std::size_t index = static_cast<std::size_t>(operation->GetCode()); + if (index >= OPERATION_DECOMPILERS.size()) { + UNREACHABLE_MSG("Out of bounds operation: {}", index); + return {}; + } + const auto decompiler = OPERATION_DECOMPILERS[index]; + if (decompiler == nullptr) { + UNREACHABLE_MSG("Undefined operation: {}", index); + return {}; + } + return (this->*decompiler)(*operation); + } + + if (const auto gpr = std::get_if<GprNode>(&*node)) { + const u32 index = gpr->GetIndex(); + if (index == Register::ZeroIndex) { + return "{0, 0, 0, 0}.x"; + } + return fmt::format("R{}.x", index); + } + + if (const auto cv = std::get_if<CustomVarNode>(&*node)) { + return fmt::format("CV{}.x", cv->GetIndex()); + } + + if (const auto immediate = std::get_if<ImmediateNode>(&*node)) { + std::string temporary = AllocTemporary(); + AddLine("MOV.U {}, {};", temporary, immediate->GetValue()); + return temporary; + } + + if (const auto predicate = std::get_if<PredicateNode>(&*node)) { + std::string temporary = AllocTemporary(); + switch (const auto index = predicate->GetIndex(); index) { + case Tegra::Shader::Pred::UnusedIndex: + AddLine("MOV.S {}, -1;", temporary); + break; + case Tegra::Shader::Pred::NeverExecute: + AddLine("MOV.S {}, 0;", temporary); + break; + default: + AddLine("MOV.S {}, P{}.x;", temporary, static_cast<u64>(index)); + break; + } + if (predicate->IsNegated()) { + AddLine("CMP.S {}, {}, 0, -1;", temporary, temporary); + } + return temporary; + } + + if (const auto abuf = std::get_if<AbufNode>(&*node)) { + if (abuf->IsPhysicalBuffer()) { + UNIMPLEMENTED_MSG("Physical buffers are not implemented"); + return "{0, 0, 0, 0}.x"; + } + + const Attribute::Index index = abuf->GetIndex(); + const u32 element = abuf->GetElement(); + const char swizzle = Swizzle(element); + switch (index) { + case Attribute::Index::Position: { + if (stage == ShaderType::Geometry) { + return fmt::format("{}_position[{}].{}", StageInputName(stage), + Visit(abuf->GetBuffer()), swizzle); + } else { + return fmt::format("{}.position.{}", StageInputName(stage), swizzle); + } + } + case Attribute::Index::TessCoordInstanceIDVertexID: + ASSERT(stage == ShaderType::Vertex); + switch (element) { + case 2: + return "vertex.instance"; + case 3: + return "vertex.id"; + } + UNIMPLEMENTED_MSG("Unmanaged TessCoordInstanceIDVertexID element={}", element); + break; + case Attribute::Index::PointCoord: + switch (element) { + case 0: + return "fragment.pointcoord.x"; + case 1: + return "fragment.pointcoord.y"; + } + UNIMPLEMENTED(); + break; + case Attribute::Index::FrontFacing: { + ASSERT(stage == ShaderType::Fragment); + ASSERT(element == 3); + const std::string temporary = AllocVectorTemporary(); + AddLine("SGT.S RC.x, fragment.facing, {{0, 0, 0, 0}};"); + AddLine("MOV.U.CC RC.x, -RC;"); + AddLine("MOV.S {}.x, 0;", temporary); + AddLine("MOV.S {}.x (NE.x), -1;", temporary); + return fmt::format("{}.x", temporary); + } + default: + if (IsGenericAttribute(index)) { + if (stage == ShaderType::Geometry) { + return fmt::format("in_attr{}[{}][0].{}", GetGenericAttributeIndex(index), + Visit(abuf->GetBuffer()), swizzle); + } else { + return fmt::format("{}.attrib[{}].{}", StageInputName(stage), + GetGenericAttributeIndex(index), swizzle); + } + } + UNIMPLEMENTED_MSG("Unimplemented input attribute={}", static_cast<int>(index)); + break; + } + return "{0, 0, 0, 0}.x"; + } + + if (const auto cbuf = std::get_if<CbufNode>(&*node)) { + std::string offset_string; + const auto& offset = cbuf->GetOffset(); + if (const auto imm = std::get_if<ImmediateNode>(&*offset)) { + offset_string = std::to_string(imm->GetValue()); + } else { + offset_string = Visit(offset); + } + std::string temporary = AllocTemporary(); + AddLine("LDC.F32 {}, cbuf{}[{}];", temporary, cbuf->GetIndex(), offset_string); + return temporary; + } + + if (const auto gmem = std::get_if<GmemNode>(&*node)) { + std::string temporary = AllocTemporary(); + AddLine("MOV {}, 0;", temporary); + AddLine("LOAD.U32 {} (NE.x), {};", temporary, GlobalMemoryPointer(*gmem)); + return temporary; + } + + if (const auto lmem = std::get_if<LmemNode>(&*node)) { + std::string temporary = Visit(lmem->GetAddress()); + AddLine("SHR.U {}, {}, 2;", temporary, temporary); + AddLine("MOV.U {}, lmem[{}].x;", temporary, temporary); + return temporary; + } + + if (const auto smem = std::get_if<SmemNode>(&*node)) { + std::string temporary = Visit(smem->GetAddress()); + AddLine("LDS.U32 {}, shared_mem[{}];", temporary, temporary); + return temporary; + } + + if (const auto internal_flag = std::get_if<InternalFlagNode>(&*node)) { + const std::size_t index = static_cast<std::size_t>(internal_flag->GetFlag()); + return fmt::format("{}.x", INTERNAL_FLAG_NAMES[index]); + } + + if (const auto conditional = std::get_if<ConditionalNode>(&*node)) { + if (const auto amend_index = conditional->GetAmendIndex()) { + Visit(ir.GetAmendNode(*amend_index)); + } + AddLine("MOVC.U RC.x, {};", Visit(conditional->GetCondition())); + AddLine("IF NE.x;"); + VisitBlock(conditional->GetCode()); + AddLine("ENDIF;"); + return {}; + } + + if ([[maybe_unused]] const auto cmt = std::get_if<CommentNode>(&*node)) { + // Uncommenting this will generate invalid code. GLASM lacks comments. + // AddLine("// {}", cmt->GetText()); + return {}; + } + + UNIMPLEMENTED(); + return {}; +} + +std::tuple<std::string, std::string, std::size_t> ARBDecompiler::BuildCoords(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + UNIMPLEMENTED_IF(meta.sampler.is_indexed); + + const bool is_extended = meta.sampler.is_shadow && meta.sampler.is_array && + meta.sampler.type == Tegra::Shader::TextureType::TextureCube; + const std::size_t count = operation.GetOperandsCount(); + std::string temporary = AllocVectorTemporary(); + std::size_t i = 0; + for (; i < count; ++i) { + AddLine("MOV.F {}.{}, {};", temporary, Swizzle(i), Visit(operation[i])); + } + if (meta.sampler.is_array) { + AddLine("I2F.S {}.{}, {};", temporary, Swizzle(i), Visit(meta.array)); + ++i; + } + if (meta.sampler.is_shadow) { + std::string compare = Visit(meta.depth_compare); + if (is_extended) { + ASSERT(i == 4); + std::string extra_coord = AllocVectorTemporary(); + AddLine("MOV.F {}.x, {};", extra_coord, compare); + return {fmt::format("{}, {}", temporary, extra_coord), extra_coord, 0}; + } + AddLine("MOV.F {}.{}, {};", temporary, Swizzle(i), compare); + ++i; + } + return {temporary, temporary, i}; +} + +std::string ARBDecompiler::BuildAoffi(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + if (meta.aoffi.empty()) { + return {}; + } + const std::string temporary = AllocVectorTemporary(); + std::size_t i = 0; + for (auto& node : meta.aoffi) { + AddLine("MOV.S {}.{}, {};", temporary, Swizzle(i++), Visit(node)); + } + return fmt::format(", offset({})", temporary); +} + +std::string ARBDecompiler::GlobalMemoryPointer(const GmemNode& gmem) { + // Read a bindless SSBO, return its address and set CC accordingly + // address = c[binding].xy + // length = c[binding].z + const u32 binding = global_memory_names.at(gmem.GetDescriptor()); + + const std::string pointer = AllocLongVectorTemporary(); + std::string temporary = AllocTemporary(); + + AddLine("PK64.U {}, c[{}];", pointer, binding); + AddLine("SUB.U {}, {}, {};", temporary, Visit(gmem.GetRealAddress()), + Visit(gmem.GetBaseAddress())); + AddLine("CVT.U64.U32 {}.z, {};", pointer, temporary); + AddLine("ADD.U64 {}.x, {}.x, {}.z;", pointer, pointer, pointer); + // Compare offset to length and set CC + AddLine("SLT.U.CC RC.x, {}, c[{}].z;", temporary, binding); + return fmt::format("{}.x", pointer); +} + +void ARBDecompiler::Exit() { + if (stage != ShaderType::Fragment) { + AddLine("RET;"); + return; + } + + const auto safe_get_register = [this](u32 reg) -> std::string { + // TODO(Rodrigo): Replace with contains once C++20 releases + const auto& used_registers = ir.GetRegisters(); + if (used_registers.find(reg) != used_registers.end()) { + return fmt::format("R{}.x", reg); + } + return "{0, 0, 0, 0}.x"; + }; + + const auto& header = ir.GetHeader(); + u32 current_reg = 0; + for (u32 rt = 0; rt < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; ++rt) { + for (u32 component = 0; component < 4; ++component) { + if (!header.ps.IsColorComponentOutputEnabled(rt, component)) { + continue; + } + AddLine("MOV.F result_color{}.{}, {};", rt, Swizzle(component), + safe_get_register(current_reg)); + ++current_reg; + } + } + if (header.ps.omap.depth) { + AddLine("MOV.F result.depth.z, {};", safe_get_register(current_reg + 1)); + } + + AddLine("RET;"); +} + +std::string ARBDecompiler::Assign(Operation operation) { + const Node& dest = operation[0]; + const Node& src = operation[1]; + + std::string dest_name; + if (const auto gpr = std::get_if<GprNode>(&*dest)) { + if (gpr->GetIndex() == Register::ZeroIndex) { + // Writing to Register::ZeroIndex is a no op + return {}; + } + dest_name = fmt::format("R{}.x", gpr->GetIndex()); + } else if (const auto abuf = std::get_if<AbufNode>(&*dest)) { + const u32 element = abuf->GetElement(); + const char swizzle = Swizzle(element); + switch (const Attribute::Index index = abuf->GetIndex()) { + case Attribute::Index::Position: + dest_name = fmt::format("result.position.{}", swizzle); + break; + case Attribute::Index::LayerViewportPointSize: + switch (element) { + case 0: + UNIMPLEMENTED(); + return {}; + case 1: + case 2: + if (!device.HasNvViewportArray2()) { + LOG_ERROR( + Render_OpenGL, + "NV_viewport_array2 is missing. Maxwell gen 2 or better is required."); + return {}; + } + dest_name = element == 1 ? "result.layer.x" : "result.viewport.x"; + break; + case 3: + dest_name = "result.pointsize.x"; + break; + } + break; + case Attribute::Index::ClipDistances0123: + dest_name = fmt::format("result.clip[{}].x", element); + break; + case Attribute::Index::ClipDistances4567: + dest_name = fmt::format("result.clip[{}].x", element + 4); + break; + default: + if (!IsGenericAttribute(index)) { + UNREACHABLE(); + return {}; + } + dest_name = + fmt::format("result.attrib[{}].{}", GetGenericAttributeIndex(index), swizzle); + break; + } + } else if (const auto lmem = std::get_if<LmemNode>(&*dest)) { + const std::string address = Visit(lmem->GetAddress()); + AddLine("SHR.U {}, {}, 2;", address, address); + dest_name = fmt::format("lmem[{}].x", address); + } else if (const auto smem = std::get_if<SmemNode>(&*dest)) { + AddLine("STS.U32 {}, shared_mem[{}];", Visit(src), Visit(smem->GetAddress())); + ResetTemporaries(); + return {}; + } else if (const auto gmem = std::get_if<GmemNode>(&*dest)) { + AddLine("IF NE.x;"); + AddLine("STORE.U32 {}, {};", Visit(src), GlobalMemoryPointer(*gmem)); + AddLine("ENDIF;"); + ResetTemporaries(); + return {}; + } else { + UNREACHABLE(); + ResetTemporaries(); + return {}; + } + + AddLine("MOV.U {}, {};", dest_name, Visit(src)); + ResetTemporaries(); + return {}; +} + +std::string ARBDecompiler::Select(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("CMP.S {}, {}, {}, {};", temporary, Visit(operation[0]), Visit(operation[1]), + Visit(operation[2])); + return temporary; +} + +std::string ARBDecompiler::FClamp(Operation operation) { + // 1.0f in hex, replace with std::bit_cast on C++20 + static constexpr u32 POSITIVE_ONE = 0x3f800000; + + std::string temporary = AllocTemporary(); + const Node& value = operation[0]; + const Node& low = operation[1]; + const Node& high = operation[2]; + const auto* const imm_low = std::get_if<ImmediateNode>(&*low); + const auto* const imm_high = std::get_if<ImmediateNode>(&*high); + if (imm_low && imm_high && imm_low->GetValue() == 0 && imm_high->GetValue() == POSITIVE_ONE) { + AddLine("MOV.F32.SAT {}, {};", temporary, Visit(value)); + } else { + AddLine("MIN.F {}, {}, {};", temporary, Visit(value), Visit(high)); + AddLine("MAX.F {}, {}, {};", temporary, temporary, Visit(low)); + } + return temporary; +} + +std::string ARBDecompiler::FCastHalf0(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.x, {};", temporary, Visit(operation[0])); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::FCastHalf1(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.y, {};", temporary, Visit(operation[0])); + AddLine("MOV {}.x, {}.y;", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::FSqrt(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("RSQ.F32 {}, {};", temporary, Visit(operation[0])); + AddLine("RCP.F32 {}, {};", temporary, temporary); + return temporary; +} + +std::string ARBDecompiler::FSwizzleAdd(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + if (!device.HasWarpIntrinsics()) { + LOG_ERROR(Render_OpenGL, + "NV_shader_thread_shuffle is missing. Kepler or better is required."); + AddLine("ADD.F {}.x, {}, {};", temporary, Visit(operation[0]), Visit(operation[1])); + return fmt::format("{}.x", temporary); + } + + AddLine("AND.U {}.z, {}.threadid, 3;", temporary, StageInputName(stage)); + AddLine("SHL.U {}.z, {}.z, 1;", temporary, temporary); + AddLine("SHR.U {}.z, {}, {}.z;", temporary, Visit(operation[2]), temporary); + AddLine("AND.U {}.z, {}.z, 3;", temporary, temporary); + AddLine("MUL.F32 {}.x, {}, FSWZA[{}.z];", temporary, Visit(operation[0]), temporary); + AddLine("MUL.F32 {}.y, {}, FSWZB[{}.z];", temporary, Visit(operation[1]), temporary); + AddLine("ADD.F32 {}.x, {}.x, {}.y;", temporary, temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HAdd2(Operation operation) { + const std::string tmp1 = AllocVectorTemporary(); + const std::string tmp2 = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", tmp1, Visit(operation[0])); + AddLine("UP2H.F {}.xy, {};", tmp2, Visit(operation[1])); + AddLine("ADD.F16 {}, {}, {};", tmp1, tmp1, tmp2); + AddLine("PK2H.F {}.x, {};", tmp1, tmp1); + return fmt::format("{}.x", tmp1); +} + +std::string ARBDecompiler::HMul2(Operation operation) { + const std::string tmp1 = AllocVectorTemporary(); + const std::string tmp2 = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", tmp1, Visit(operation[0])); + AddLine("UP2H.F {}.xy, {};", tmp2, Visit(operation[1])); + AddLine("MUL.F16 {}, {}, {};", tmp1, tmp1, tmp2); + AddLine("PK2H.F {}.x, {};", tmp1, tmp1); + return fmt::format("{}.x", tmp1); +} + +std::string ARBDecompiler::HFma2(Operation operation) { + const std::string tmp1 = AllocVectorTemporary(); + const std::string tmp2 = AllocVectorTemporary(); + const std::string tmp3 = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", tmp1, Visit(operation[0])); + AddLine("UP2H.F {}.xy, {};", tmp2, Visit(operation[1])); + AddLine("UP2H.F {}.xy, {};", tmp3, Visit(operation[2])); + AddLine("MAD.F16 {}, {}, {}, {};", tmp1, tmp1, tmp2, tmp3); + AddLine("PK2H.F {}.x, {};", tmp1, tmp1); + return fmt::format("{}.x", tmp1); +} + +std::string ARBDecompiler::HAbsolute(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, Visit(operation[0])); + AddLine("PK2H.F {}.x, |{}|;", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HNegate(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, Visit(operation[0])); + AddLine("MOVC.S RC.x, {};", Visit(operation[1])); + AddLine("MOV.F {}.x (NE.x), -{}.x;", temporary, temporary); + AddLine("MOVC.S RC.x, {};", Visit(operation[2])); + AddLine("MOV.F {}.y (NE.x), -{}.y;", temporary, temporary); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HClamp(Operation operation) { + const std::string tmp1 = AllocVectorTemporary(); + const std::string tmp2 = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", tmp1, Visit(operation[0])); + AddLine("MOV.U {}.x, {};", tmp2, Visit(operation[1])); + AddLine("MOV.U {}.y, {}.x;", tmp2, tmp2); + AddLine("MAX.F {}, {}, {};", tmp1, tmp1, tmp2); + AddLine("MOV.U {}.x, {};", tmp2, Visit(operation[2])); + AddLine("MOV.U {}.y, {}.x;", tmp2, tmp2); + AddLine("MIN.F {}, {}, {};", tmp1, tmp1, tmp2); + AddLine("PK2H.F {}.x, {};", tmp1, tmp1); + return fmt::format("{}.x", tmp1); +} + +std::string ARBDecompiler::HCastFloat(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("MOV.F {}.y, {{0, 0, 0, 0}};", temporary); + AddLine("MOV.F {}.x, {};", temporary, Visit(operation[0])); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HUnpack(Operation operation) { + std::string operand = Visit(operation[0]); + switch (std::get<Tegra::Shader::HalfType>(operation.GetMeta())) { + case Tegra::Shader::HalfType::H0_H1: + return operand; + case Tegra::Shader::HalfType::F32: { + const std::string temporary = AllocVectorTemporary(); + AddLine("MOV.U {}.x, {};", temporary, operand); + AddLine("MOV.U {}.y, {}.x;", temporary, temporary); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); + } + case Tegra::Shader::HalfType::H0_H0: { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, operand); + AddLine("MOV.U {}.y, {}.x;", temporary, temporary); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); + } + case Tegra::Shader::HalfType::H1_H1: { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, operand); + AddLine("MOV.U {}.x, {}.y;", temporary, temporary); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); + } + } + UNREACHABLE(); + return "{0, 0, 0, 0}.x"; +} + +std::string ARBDecompiler::HMergeF32(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, Visit(operation[0])); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HMergeH0(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, Visit(operation[0])); + AddLine("UP2H.F {}.zw, {};", temporary, Visit(operation[1])); + AddLine("MOV.U {}.x, {}.z;", temporary, temporary); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HMergeH1(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("UP2H.F {}.xy, {};", temporary, Visit(operation[0])); + AddLine("UP2H.F {}.zw, {};", temporary, Visit(operation[1])); + AddLine("MOV.U {}.y, {}.w;", temporary, temporary); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::HPack2(Operation operation) { + const std::string temporary = AllocVectorTemporary(); + AddLine("MOV.U {}.x, {};", temporary, Visit(operation[0])); + AddLine("MOV.U {}.y, {};", temporary, Visit(operation[1])); + AddLine("PK2H.F {}.x, {};", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::LogicalAssign(Operation operation) { + const Node& dest = operation[0]; + const Node& src = operation[1]; + + std::string target; + + if (const auto pred = std::get_if<PredicateNode>(&*dest)) { + ASSERT_MSG(!pred->IsNegated(), "Negating logical assignment"); + + const Tegra::Shader::Pred index = pred->GetIndex(); + switch (index) { + case Tegra::Shader::Pred::NeverExecute: + case Tegra::Shader::Pred::UnusedIndex: + // Writing to these predicates is a no-op + return {}; + } + target = fmt::format("P{}.x", static_cast<u64>(index)); + } else if (const auto internal_flag = std::get_if<InternalFlagNode>(&*dest)) { + const std::size_t index = static_cast<std::size_t>(internal_flag->GetFlag()); + target = fmt::format("{}.x", INTERNAL_FLAG_NAMES[index]); + } else { + UNREACHABLE(); + ResetTemporaries(); + return {}; + } + + AddLine("MOV.U {}, {};", target, Visit(src)); + ResetTemporaries(); + return {}; +} + +std::string ARBDecompiler::LogicalPick2(Operation operation) { + std::string temporary = AllocTemporary(); + const u32 index = std::get<ImmediateNode>(*operation[1]).GetValue(); + AddLine("MOV.U {}, {}.{};", temporary, Visit(operation[0]), Swizzle(index)); + return temporary; +} + +std::string ARBDecompiler::LogicalAnd2(Operation operation) { + std::string temporary = AllocTemporary(); + const std::string op = Visit(operation[0]); + AddLine("AND.U {}, {}.x, {}.y;", temporary, op, op); + return temporary; +} + +std::string ARBDecompiler::FloatOrdered(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("MOVC.F32 RC.x, {};", Visit(operation[0])); + AddLine("MOVC.F32 RC.y, {};", Visit(operation[1])); + AddLine("MOV.S {}, -1;", temporary); + AddLine("MOV.S {} (NAN.x), 0;", temporary); + AddLine("MOV.S {} (NAN.y), 0;", temporary); + return temporary; +} + +std::string ARBDecompiler::FloatUnordered(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("MOVC.F32 RC.x, {};", Visit(operation[0])); + AddLine("MOVC.F32 RC.y, {};", Visit(operation[1])); + AddLine("MOV.S {}, 0;", temporary); + AddLine("MOV.S {} (NAN.x), -1;", temporary); + AddLine("MOV.S {} (NAN.y), -1;", temporary); + return temporary; +} + +std::string ARBDecompiler::LogicalAddCarry(Operation operation) { + std::string temporary = AllocTemporary(); + AddLine("ADDC.U RC, {}, {};", Visit(operation[0]), Visit(operation[1])); + AddLine("MOV.S {}, 0;", temporary); + AddLine("IF CF.x;"); + AddLine("MOV.S {}, -1;", temporary); + AddLine("ENDIF;"); + return temporary; +} + +std::string ARBDecompiler::Texture(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; + const auto [coords, temporary, swizzle] = BuildCoords(operation); + + std::string_view opcode = "TEX"; + std::string extra; + if (meta.bias) { + ASSERT(!meta.lod); + opcode = "TXB"; + + if (swizzle < 4) { + AddLine("MOV.F {}.w, {};", temporary, Visit(meta.bias)); + } else { + const std::string bias = AllocTemporary(); + AddLine("MOV.F {}, {};", bias, Visit(meta.bias)); + extra = fmt::format(" {},", bias); + } + } + if (meta.lod) { + ASSERT(!meta.bias); + opcode = "TXL"; + + if (swizzle < 4) { + AddLine("MOV.F {}.w, {};", temporary, Visit(meta.lod)); + } else { + const std::string lod = AllocTemporary(); + AddLine("MOV.F {}, {};", lod, Visit(meta.lod)); + extra = fmt::format(" {},", lod); + } + } + + AddLine("{}.F {}, {},{} texture[{}], {}{};", opcode, temporary, coords, extra, sampler_id, + TextureType(meta), BuildAoffi(operation)); + AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::TextureGather(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; + const auto [coords, temporary, swizzle] = BuildCoords(operation); + + std::string comp; + if (!meta.sampler.is_shadow) { + const auto& immediate = std::get<ImmediateNode>(*meta.component); + comp = fmt::format(".{}", Swizzle(immediate.GetValue())); + } + + AddLine("TXG.F {}, {}, texture[{}]{}, {}{};", temporary, temporary, sampler_id, comp, + TextureType(meta), BuildAoffi(operation)); + AddLine("MOV.U {}.x, {}.{};", temporary, coords, Swizzle(meta.element)); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::TextureQueryDimensions(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + const std::string temporary = AllocVectorTemporary(); + const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; + + ASSERT(!meta.sampler.is_array); + + const std::string lod = operation.GetOperandsCount() > 0 ? Visit(operation[0]) : "0"; + AddLine("TXQ {}, {}, texture[{}], {};", temporary, lod, sampler_id, TextureType(meta)); + AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::TextureQueryLod(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + const std::string temporary = AllocVectorTemporary(); + const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; + + ASSERT(!meta.sampler.is_array); + + const std::size_t count = operation.GetOperandsCount(); + for (std::size_t i = 0; i < count; ++i) { + AddLine("MOV.F {}.{}, {};", temporary, Swizzle(i), Visit(operation[i])); + } + AddLine("LOD.F {}, {}, texture[{}], {};", temporary, temporary, sampler_id, TextureType(meta)); + AddLine("MUL.F32 {}, {}, {{256, 256, 0, 0}};", temporary, temporary); + AddLine("TRUNC.S {}, {};", temporary, temporary); + AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::TexelFetch(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; + const auto [coords, temporary, swizzle] = BuildCoords(operation); + + if (!meta.sampler.is_buffer) { + ASSERT(swizzle < 4); + AddLine("MOV.F {}.w, {};", temporary, Visit(meta.lod)); + } + AddLine("TXF.F {}, {}, texture[{}], {}{};", temporary, coords, sampler_id, TextureType(meta), + BuildAoffi(operation)); + AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::TextureGradient(Operation operation) { + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); + const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; + const std::string ddx = AllocVectorTemporary(); + const std::string ddy = AllocVectorTemporary(); + const std::string coord = std::get<1>(BuildCoords(operation)); + + const std::size_t num_components = meta.derivates.size() / 2; + for (std::size_t index = 0; index < num_components; ++index) { + const char swizzle = Swizzle(index); + AddLine("MOV.F {}.{}, {};", ddx, swizzle, Visit(meta.derivates[index * 2])); + AddLine("MOV.F {}.{}, {};", ddy, swizzle, Visit(meta.derivates[index * 2 + 1])); + } + + const std::string_view result = coord; + AddLine("TXD.F {}, {}, {}, {}, texture[{}], {}{};", result, coord, ddx, ddy, sampler_id, + TextureType(meta), BuildAoffi(operation)); + AddLine("MOV.F {}.x, {}.{};", result, result, Swizzle(meta.element)); + return fmt::format("{}.x", result); +} + +std::string ARBDecompiler::ImageLoad(Operation operation) { + const auto& meta = std::get<MetaImage>(operation.GetMeta()); + const u32 image_id = device.GetBaseBindings(stage).image + meta.image.index; + const std::size_t count = operation.GetOperandsCount(); + const std::string_view type = ImageType(meta.image.type); + + const std::string temporary = AllocVectorTemporary(); + for (std::size_t i = 0; i < count; ++i) { + AddLine("MOV.S {}.{}, {};", temporary, Swizzle(i), Visit(operation[i])); + } + AddLine("LOADIM.F {}, {}, image[{}], {};", temporary, temporary, image_id, type); + AddLine("MOV.F {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::ImageStore(Operation operation) { + const auto& meta = std::get<MetaImage>(operation.GetMeta()); + const u32 image_id = device.GetBaseBindings(stage).image + meta.image.index; + const std::size_t num_coords = operation.GetOperandsCount(); + const std::size_t num_values = meta.values.size(); + const std::string_view type = ImageType(meta.image.type); + + const std::string coord = AllocVectorTemporary(); + const std::string value = AllocVectorTemporary(); + for (std::size_t i = 0; i < num_coords; ++i) { + AddLine("MOV.S {}.{}, {};", coord, Swizzle(i), Visit(operation[i])); + } + for (std::size_t i = 0; i < num_values; ++i) { + AddLine("MOV.F {}.{}, {};", value, Swizzle(i), Visit(meta.values[i])); + } + AddLine("STOREIM.F image[{}], {}, {}, {};", image_id, value, coord, type); + return {}; +} + +std::string ARBDecompiler::Branch(Operation operation) { + const auto target = std::get<ImmediateNode>(*operation[0]); + AddLine("MOV.U PC.x, {};", target.GetValue()); + AddLine("CONT;"); + return {}; +} + +std::string ARBDecompiler::BranchIndirect(Operation operation) { + AddLine("MOV.U PC.x, {};", Visit(operation[0])); + AddLine("CONT;"); + return {}; +} + +std::string ARBDecompiler::PushFlowStack(Operation operation) { + const auto stack = std::get<MetaStackClass>(operation.GetMeta()); + const u32 target = std::get<ImmediateNode>(*operation[0]).GetValue(); + const std::string_view stack_name = StackName(stack); + AddLine("MOV.U {}[{}_TOP.x].x, {};", stack_name, stack_name, target); + AddLine("ADD.S {}_TOP.x, {}_TOP.x, 1;", stack_name, stack_name); + return {}; +} + +std::string ARBDecompiler::PopFlowStack(Operation operation) { + const auto stack = std::get<MetaStackClass>(operation.GetMeta()); + const std::string_view stack_name = StackName(stack); + AddLine("SUB.S {}_TOP.x, {}_TOP.x, 1;", stack_name, stack_name); + AddLine("MOV.U PC.x, {}[{}_TOP.x].x;", stack_name, stack_name); + AddLine("CONT;"); + return {}; +} + +std::string ARBDecompiler::Exit(Operation) { + Exit(); + return {}; +} + +std::string ARBDecompiler::Discard(Operation) { + AddLine("KIL TR;"); + return {}; +} + +std::string ARBDecompiler::EmitVertex(Operation) { + AddLine("EMIT;"); + return {}; +} + +std::string ARBDecompiler::EndPrimitive(Operation) { + AddLine("ENDPRIM;"); + return {}; +} + +std::string ARBDecompiler::InvocationId(Operation) { + return "primitive.invocation"; +} + +std::string ARBDecompiler::YNegate(Operation) { + LOG_WARNING(Render_OpenGL, "(STUBBED)"); + std::string temporary = AllocTemporary(); + AddLine("MOV.F {}, 1;", temporary); + return temporary; +} + +std::string ARBDecompiler::ThreadId(Operation) { + return fmt::format("{}.threadid", StageInputName(stage)); +} + +std::string ARBDecompiler::ShuffleIndexed(Operation operation) { + if (!device.HasWarpIntrinsics()) { + LOG_ERROR(Render_OpenGL, + "NV_shader_thread_shuffle is missing. Kepler or better is required."); + return Visit(operation[0]); + } + const std::string temporary = AllocVectorTemporary(); + AddLine("SHFIDX.U {}, {}, {}, {{31, 0, 0, 0}};", temporary, Visit(operation[0]), + Visit(operation[1])); + AddLine("MOV.U {}.x, {}.y;", temporary, temporary); + return fmt::format("{}.x", temporary); +} + +std::string ARBDecompiler::Barrier(Operation) { + AddLine("BAR;"); + return {}; +} + +std::string ARBDecompiler::MemoryBarrierGroup(Operation) { + AddLine("MEMBAR.CTA;"); + return {}; +} + +std::string ARBDecompiler::MemoryBarrierGlobal(Operation) { + AddLine("MEMBAR;"); + return {}; +} + +} // Anonymous namespace + +std::string DecompileAssemblyShader(const Device& device, const VideoCommon::Shader::ShaderIR& ir, + const VideoCommon::Shader::Registry& registry, + Tegra::Engines::ShaderType stage, std::string_view identifier) { + return ARBDecompiler(device, ir, registry, stage, identifier).Code(); +} + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_arb_decompiler.h b/src/video_core/renderer_opengl/gl_arb_decompiler.h new file mode 100644 index 000000000..6afc87220 --- /dev/null +++ b/src/video_core/renderer_opengl/gl_arb_decompiler.h @@ -0,0 +1,29 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <string> +#include <string_view> + +#include "common/common_types.h" + +namespace Tegra::Engines { +enum class ShaderType : u32; +} + +namespace VideoCommon::Shader { +class ShaderIR; +class Registry; +} // namespace VideoCommon::Shader + +namespace OpenGL { + +class Device; + +std::string DecompileAssemblyShader(const Device& device, const VideoCommon::Shader::ShaderIR& ir, + const VideoCommon::Shader::Registry& registry, + Tegra::Engines::ShaderType stage, std::string_view identifier); + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index 4eb37a96c..b1c4cd62f 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -8,6 +8,7 @@ #include "common/assert.h" #include "common/microprofile.h" +#include "video_core/buffer_cache/buffer_cache.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" @@ -21,22 +22,54 @@ using Maxwell = Tegra::Engines::Maxwell3D::Regs; MICROPROFILE_DEFINE(OpenGL_Buffer_Download, "OpenGL", "Buffer Download", MP_RGB(192, 192, 128)); -CachedBufferBlock::CachedBufferBlock(VAddr cpu_addr, const std::size_t size) +Buffer::Buffer(const Device& device, VAddr cpu_addr, std::size_t size) : VideoCommon::BufferBlock{cpu_addr, size} { gl_buffer.Create(); glNamedBufferData(gl_buffer.handle, static_cast<GLsizeiptr>(size), nullptr, GL_DYNAMIC_DRAW); + if (device.UseAssemblyShaders() || device.HasVertexBufferUnifiedMemory()) { + glMakeNamedBufferResidentNV(gl_buffer.handle, GL_READ_WRITE); + glGetNamedBufferParameterui64vNV(gl_buffer.handle, GL_BUFFER_GPU_ADDRESS_NV, &gpu_address); + } } -CachedBufferBlock::~CachedBufferBlock() = default; +Buffer::~Buffer() = default; + +void Buffer::Upload(std::size_t offset, std::size_t size, const u8* data) { + glNamedBufferSubData(Handle(), static_cast<GLintptr>(offset), static_cast<GLsizeiptr>(size), + data); +} -OGLBufferCache::OGLBufferCache(RasterizerOpenGL& rasterizer, Core::System& system, - const Device& device, std::size_t stream_size) - : GenericBufferCache{rasterizer, system, std::make_unique<OGLStreamBuffer>(stream_size, true)} { +void Buffer::Download(std::size_t offset, std::size_t size, u8* data) { + MICROPROFILE_SCOPE(OpenGL_Buffer_Download); + const GLsizeiptr gl_size = static_cast<GLsizeiptr>(size); + const GLintptr gl_offset = static_cast<GLintptr>(offset); + if (read_buffer.handle == 0) { + read_buffer.Create(); + glNamedBufferData(read_buffer.handle, static_cast<GLsizeiptr>(Size()), nullptr, + GL_STREAM_READ); + } + glMemoryBarrier(GL_BUFFER_UPDATE_BARRIER_BIT); + glCopyNamedBufferSubData(gl_buffer.handle, read_buffer.handle, gl_offset, gl_offset, gl_size); + glGetNamedBufferSubData(read_buffer.handle, gl_offset, gl_size, data); +} + +void Buffer::CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, + std::size_t size) { + glCopyNamedBufferSubData(src.Handle(), Handle(), static_cast<GLintptr>(src_offset), + static_cast<GLintptr>(dst_offset), static_cast<GLsizeiptr>(size)); +} + +OGLBufferCache::OGLBufferCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, + const Device& device_, std::size_t stream_size) + : GenericBufferCache{rasterizer, gpu_memory, cpu_memory, + std::make_unique<OGLStreamBuffer>(device_, stream_size, true)}, + device{device_} { if (!device.HasFastBufferSubData()) { return; } - static constexpr auto size = static_cast<GLsizeiptr>(Maxwell::MaxConstBufferSize); + static constexpr GLsizeiptr size = static_cast<GLsizeiptr>(Maxwell::MaxConstBufferSize); glCreateBuffers(static_cast<GLsizei>(std::size(cbufs)), std::data(cbufs)); for (const GLuint cbuf : cbufs) { glNamedBufferData(cbuf, size, nullptr, GL_STREAM_DRAW); @@ -47,49 +80,21 @@ OGLBufferCache::~OGLBufferCache() { glDeleteBuffers(static_cast<GLsizei>(std::size(cbufs)), std::data(cbufs)); } -Buffer OGLBufferCache::CreateBlock(VAddr cpu_addr, std::size_t size) { - return std::make_shared<CachedBufferBlock>(cpu_addr, size); -} - -void OGLBufferCache::WriteBarrier() { - glMemoryBarrier(GL_ALL_BARRIER_BITS); +std::shared_ptr<Buffer> OGLBufferCache::CreateBlock(VAddr cpu_addr, std::size_t size) { + return std::make_shared<Buffer>(device, cpu_addr, size); } -const GLuint* OGLBufferCache::ToHandle(const Buffer& buffer) { - return buffer->GetHandle(); -} - -const GLuint* OGLBufferCache::GetEmptyBuffer(std::size_t) { - static const GLuint null_buffer = 0; - return &null_buffer; -} - -void OGLBufferCache::UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - const u8* data) { - glNamedBufferSubData(*buffer->GetHandle(), static_cast<GLintptr>(offset), - static_cast<GLsizeiptr>(size), data); -} - -void OGLBufferCache::DownloadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - u8* data) { - MICROPROFILE_SCOPE(OpenGL_Buffer_Download); - glGetNamedBufferSubData(*buffer->GetHandle(), static_cast<GLintptr>(offset), - static_cast<GLsizeiptr>(size), data); -} - -void OGLBufferCache::CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset, - std::size_t dst_offset, std::size_t size) { - glCopyNamedBufferSubData(*src->GetHandle(), *dst->GetHandle(), - static_cast<GLintptr>(src_offset), static_cast<GLintptr>(dst_offset), - static_cast<GLsizeiptr>(size)); +OGLBufferCache::BufferInfo OGLBufferCache::GetEmptyBuffer(std::size_t) { + return {0, 0, 0}; } OGLBufferCache::BufferInfo OGLBufferCache::ConstBufferUpload(const void* raw_pointer, std::size_t size) { DEBUG_ASSERT(cbuf_cursor < std::size(cbufs)); - const GLuint& cbuf = cbufs[cbuf_cursor++]; + const GLuint cbuf = cbufs[cbuf_cursor++]; + glNamedBufferSubData(cbuf, 0, static_cast<GLsizeiptr>(size), raw_pointer); - return {&cbuf, 0}; + return {cbuf, 0, 0}; } } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h index d94a11252..f75b32e31 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.h +++ b/src/video_core/renderer_opengl/gl_buffer_cache.h @@ -10,7 +10,6 @@ #include "common/common_types.h" #include "video_core/buffer_cache/buffer_cache.h" #include "video_core/engines/maxwell_3d.h" -#include "video_core/rasterizer_cache.h" #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_stream_buffer.h" @@ -24,59 +23,59 @@ class Device; class OGLStreamBuffer; class RasterizerOpenGL; -class CachedBufferBlock; +class Buffer : public VideoCommon::BufferBlock { +public: + explicit Buffer(const Device& device, VAddr cpu_addr, std::size_t size); + ~Buffer(); -using Buffer = std::shared_ptr<CachedBufferBlock>; -using GenericBufferCache = VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer>; + void Upload(std::size_t offset, std::size_t size, const u8* data); -class CachedBufferBlock : public VideoCommon::BufferBlock { -public: - explicit CachedBufferBlock(VAddr cpu_addr, const std::size_t size); - ~CachedBufferBlock(); + void Download(std::size_t offset, std::size_t size, u8* data); + + void CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, + std::size_t size); - const GLuint* GetHandle() const { - return &gl_buffer.handle; + GLuint Handle() const noexcept { + return gl_buffer.handle; + } + + u64 Address() const noexcept { + return gpu_address; } private: - OGLBuffer gl_buffer{}; + OGLBuffer gl_buffer; + OGLBuffer read_buffer; + u64 gpu_address = 0; }; +using GenericBufferCache = VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer>; class OGLBufferCache final : public GenericBufferCache { public: - explicit OGLBufferCache(RasterizerOpenGL& rasterizer, Core::System& system, + explicit OGLBufferCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, const Device& device, std::size_t stream_size); ~OGLBufferCache(); - const GLuint* GetEmptyBuffer(std::size_t) override; + BufferInfo GetEmptyBuffer(std::size_t) override; void Acquire() noexcept { cbuf_cursor = 0; } protected: - Buffer CreateBlock(VAddr cpu_addr, std::size_t size) override; - - void WriteBarrier() override; - - const GLuint* ToHandle(const Buffer& buffer) override; - - void UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - const u8* data) override; - - void DownloadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - u8* data) override; - - void CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset, - std::size_t dst_offset, std::size_t size) override; + std::shared_ptr<Buffer> CreateBlock(VAddr cpu_addr, std::size_t size) override; BufferInfo ConstBufferUpload(const void* raw_pointer, std::size_t size) override; private: + static constexpr std::size_t NUM_CBUFS = Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers * + Tegra::Engines::Maxwell3D::Regs::MaxShaderProgram; + + const Device& device; + std::size_t cbuf_cursor = 0; - std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers * - Tegra::Engines::Maxwell3D::Regs::MaxShaderProgram> - cbufs; + std::array<GLuint, NUM_CBUFS> cbufs{}; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp index c286502ba..a94e4f72e 100644 --- a/src/video_core/renderer_opengl/gl_device.cpp +++ b/src/video_core/renderer_opengl/gl_device.cpp @@ -6,6 +6,7 @@ #include <array> #include <cstddef> #include <cstring> +#include <limits> #include <optional> #include <vector> @@ -13,6 +14,7 @@ #include "common/logging/log.h" #include "common/scope_exit.h" +#include "core/settings.h" #include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_resource_manager.h" @@ -25,24 +27,27 @@ constexpr u32 ReservedUniformBlocks = 1; constexpr u32 NumStages = 5; -constexpr std::array LimitUBOs = {GL_MAX_VERTEX_UNIFORM_BLOCKS, GL_MAX_TESS_CONTROL_UNIFORM_BLOCKS, - GL_MAX_TESS_EVALUATION_UNIFORM_BLOCKS, - GL_MAX_GEOMETRY_UNIFORM_BLOCKS, GL_MAX_FRAGMENT_UNIFORM_BLOCKS}; +constexpr std::array LimitUBOs = { + GL_MAX_VERTEX_UNIFORM_BLOCKS, GL_MAX_TESS_CONTROL_UNIFORM_BLOCKS, + GL_MAX_TESS_EVALUATION_UNIFORM_BLOCKS, GL_MAX_GEOMETRY_UNIFORM_BLOCKS, + GL_MAX_FRAGMENT_UNIFORM_BLOCKS, GL_MAX_COMPUTE_UNIFORM_BLOCKS}; constexpr std::array LimitSSBOs = { - GL_MAX_VERTEX_SHADER_STORAGE_BLOCKS, GL_MAX_TESS_CONTROL_SHADER_STORAGE_BLOCKS, + GL_MAX_VERTEX_SHADER_STORAGE_BLOCKS, GL_MAX_TESS_CONTROL_SHADER_STORAGE_BLOCKS, GL_MAX_TESS_EVALUATION_SHADER_STORAGE_BLOCKS, GL_MAX_GEOMETRY_SHADER_STORAGE_BLOCKS, - GL_MAX_FRAGMENT_SHADER_STORAGE_BLOCKS}; + GL_MAX_FRAGMENT_SHADER_STORAGE_BLOCKS, GL_MAX_COMPUTE_SHADER_STORAGE_BLOCKS}; -constexpr std::array LimitSamplers = { - GL_MAX_VERTEX_TEXTURE_IMAGE_UNITS, GL_MAX_TESS_CONTROL_TEXTURE_IMAGE_UNITS, - GL_MAX_TESS_EVALUATION_TEXTURE_IMAGE_UNITS, GL_MAX_GEOMETRY_TEXTURE_IMAGE_UNITS, - GL_MAX_TEXTURE_IMAGE_UNITS}; +constexpr std::array LimitSamplers = {GL_MAX_VERTEX_TEXTURE_IMAGE_UNITS, + GL_MAX_TESS_CONTROL_TEXTURE_IMAGE_UNITS, + GL_MAX_TESS_EVALUATION_TEXTURE_IMAGE_UNITS, + GL_MAX_GEOMETRY_TEXTURE_IMAGE_UNITS, + GL_MAX_TEXTURE_IMAGE_UNITS, + GL_MAX_COMPUTE_TEXTURE_IMAGE_UNITS}; -constexpr std::array LimitImages = {GL_MAX_VERTEX_IMAGE_UNIFORMS, - GL_MAX_TESS_CONTROL_IMAGE_UNIFORMS, - GL_MAX_TESS_EVALUATION_IMAGE_UNIFORMS, - GL_MAX_GEOMETRY_IMAGE_UNIFORMS, GL_MAX_FRAGMENT_IMAGE_UNIFORMS}; +constexpr std::array LimitImages = { + GL_MAX_VERTEX_IMAGE_UNIFORMS, GL_MAX_TESS_CONTROL_IMAGE_UNIFORMS, + GL_MAX_TESS_EVALUATION_IMAGE_UNIFORMS, GL_MAX_GEOMETRY_IMAGE_UNIFORMS, + GL_MAX_FRAGMENT_IMAGE_UNIFORMS, GL_MAX_COMPUTE_IMAGE_UNIFORMS}; template <typename T> T GetInteger(GLenum pname) { @@ -84,10 +89,17 @@ u32 Extract(u32& base, u32& num, u32 amount, std::optional<GLenum> limit = {}) { return std::exchange(base, base + amount); } +std::array<u32, Tegra::Engines::MaxShaderTypes> BuildMaxUniformBuffers() noexcept { + std::array<u32, Tegra::Engines::MaxShaderTypes> max; + std::transform(LimitUBOs.begin(), LimitUBOs.end(), max.begin(), + [](GLenum pname) { return GetInteger<u32>(pname); }); + return max; +} + std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindings() noexcept { std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> bindings; - static std::array<std::size_t, 5> stage_swizzle = {0, 1, 2, 3, 4}; + static constexpr std::array<std::size_t, 5> stage_swizzle{0, 1, 2, 3, 4}; const u32 total_ubos = GetInteger<u32>(GL_MAX_UNIFORM_BUFFER_BINDINGS); const u32 total_ssbos = GetInteger<u32>(GL_MAX_SHADER_STORAGE_BUFFER_BINDINGS); const u32 total_samplers = GetInteger<u32>(GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS); @@ -111,16 +123,24 @@ std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindin u32 num_images = GetInteger<u32>(GL_MAX_IMAGE_UNITS); u32 base_images = 0; - // Reserve more image bindings on fragment and vertex stages. + // GL_MAX_IMAGE_UNITS is guaranteed by the spec to have a minimum value of 8. + // Due to the limitation of GL_MAX_IMAGE_UNITS, reserve at least 4 image bindings on the + // fragment stage, and at least 1 for the rest of the stages. + // So far games are observed to use 1 image binding on vertex and 4 on fragment stages. + + // Reserve at least 4 image bindings on the fragment stage. bindings[4].image = - Extract(base_images, num_images, num_images / NumStages + 2, LimitImages[4]); - bindings[0].image = - Extract(base_images, num_images, num_images / NumStages + 1, LimitImages[0]); + Extract(base_images, num_images, std::max(4U, num_images / NumStages), LimitImages[4]); + + // This is guaranteed to be at least 1. + const u32 total_extracted_images = num_images / (NumStages - 1); // Reserve the other image bindings. - const u32 total_extracted_images = num_images / (NumStages - 2); - for (std::size_t i = 2; i < NumStages; ++i) { + for (std::size_t i = 0; i < NumStages; ++i) { const std::size_t stage = stage_swizzle[i]; + if (stage == 4) { + continue; + } bindings[stage].image = Extract(base_images, num_images, total_extracted_images, LimitImages[stage]); } @@ -132,6 +152,7 @@ std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindin } bool IsASTCSupported() { + static constexpr std::array targets = {GL_TEXTURE_2D, GL_TEXTURE_2D_ARRAY}; static constexpr std::array formats = { GL_COMPRESSED_RGBA_ASTC_4x4_KHR, GL_COMPRESSED_RGBA_ASTC_5x4_KHR, GL_COMPRESSED_RGBA_ASTC_5x5_KHR, GL_COMPRESSED_RGBA_ASTC_6x5_KHR, @@ -148,59 +169,94 @@ bool IsASTCSupported() { GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x8_KHR, GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR, GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x10_KHR, GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR, }; - return std::find_if_not(formats.begin(), formats.end(), [](GLenum format) { - GLint supported; - glGetInternalformativ(GL_TEXTURE_2D, format, GL_INTERNALFORMAT_SUPPORTED, 1, - &supported); - return supported == GL_TRUE; - }) == formats.end(); + static constexpr std::array required_support = { + GL_VERTEX_TEXTURE, GL_TESS_CONTROL_TEXTURE, GL_TESS_EVALUATION_TEXTURE, + GL_GEOMETRY_TEXTURE, GL_FRAGMENT_TEXTURE, GL_COMPUTE_TEXTURE, + }; + + for (const GLenum target : targets) { + for (const GLenum format : formats) { + for (const GLenum support : required_support) { + GLint value; + glGetInternalformativ(target, format, support, 1, &value); + if (value != GL_FULL_SUPPORT) { + return false; + } + } + } + } + return true; } } // Anonymous namespace -Device::Device() : base_bindings{BuildBaseBindings()} { +Device::Device() + : max_uniform_buffers{BuildMaxUniformBuffers()}, base_bindings{BuildBaseBindings()} { const std::string_view vendor = reinterpret_cast<const char*>(glGetString(GL_VENDOR)); - const auto renderer = reinterpret_cast<const char*>(glGetString(GL_RENDERER)); + const std::string_view version = reinterpret_cast<const char*>(glGetString(GL_VERSION)); const std::vector extensions = GetExtensions(); const bool is_nvidia = vendor == "NVIDIA Corporation"; const bool is_amd = vendor == "ATI Technologies Inc."; - const bool is_intel = vendor == "Intel"; - const bool is_intel_proprietary = is_intel && std::strstr(renderer, "Mesa") == nullptr; + + bool disable_fast_buffer_sub_data = false; + if (is_nvidia && version == "4.6.0 NVIDIA 443.24") { + LOG_WARNING( + Render_OpenGL, + "Beta driver 443.24 is known to have issues. There might be performance issues."); + disable_fast_buffer_sub_data = true; + } uniform_buffer_alignment = GetInteger<std::size_t>(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT); shader_storage_alignment = GetInteger<std::size_t>(GL_SHADER_STORAGE_BUFFER_OFFSET_ALIGNMENT); max_vertex_attributes = GetInteger<u32>(GL_MAX_VERTEX_ATTRIBS); max_varyings = GetInteger<u32>(GL_MAX_VARYING_VECTORS); + max_compute_shared_memory_size = GetInteger<u32>(GL_MAX_COMPUTE_SHARED_MEMORY_SIZE); has_warp_intrinsics = GLAD_GL_NV_gpu_shader5 && GLAD_GL_NV_shader_thread_group && GLAD_GL_NV_shader_thread_shuffle; has_shader_ballot = GLAD_GL_ARB_shader_ballot; has_vertex_viewport_layer = GLAD_GL_ARB_shader_viewport_layer_array; has_image_load_formatted = HasExtension(extensions, "GL_EXT_shader_image_load_formatted"); + has_texture_shadow_lod = HasExtension(extensions, "GL_EXT_texture_shadow_lod"); has_astc = IsASTCSupported(); has_variable_aoffi = TestVariableAoffi(); has_component_indexing_bug = is_amd; has_precise_bug = TestPreciseBug(); - has_broken_compute = is_intel_proprietary; - has_fast_buffer_sub_data = is_nvidia; + has_nv_viewport_array2 = GLAD_GL_NV_viewport_array2; + has_vertex_buffer_unified_memory = GLAD_GL_NV_vertex_buffer_unified_memory; + + // At the moment of writing this, only Nvidia's driver optimizes BufferSubData on exclusive + // uniform buffers as "push constants" + has_fast_buffer_sub_data = is_nvidia && !disable_fast_buffer_sub_data; + + use_assembly_shaders = Settings::values.use_assembly_shaders.GetValue() && + GLAD_GL_NV_gpu_program5 && GLAD_GL_NV_compute_program5 && + GLAD_GL_NV_transform_feedback && GLAD_GL_NV_transform_feedback2; + + use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue(); LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi); LOG_INFO(Render_OpenGL, "Renderer_ComponentIndexingBug: {}", has_component_indexing_bug); LOG_INFO(Render_OpenGL, "Renderer_PreciseBug: {}", has_precise_bug); + + if (Settings::values.use_assembly_shaders.GetValue() && !use_assembly_shaders) { + LOG_ERROR(Render_OpenGL, "Assembly shaders enabled but not supported"); + } } Device::Device(std::nullptr_t) { - uniform_buffer_alignment = 0; + max_uniform_buffers.fill(std::numeric_limits<u32>::max()); + uniform_buffer_alignment = 4; + shader_storage_alignment = 4; max_vertex_attributes = 16; max_varyings = 15; + max_compute_shared_memory_size = 0x10000; has_warp_intrinsics = true; has_shader_ballot = true; has_vertex_viewport_layer = true; has_image_load_formatted = true; + has_texture_shadow_lod = true; has_variable_aoffi = true; - has_component_indexing_bug = false; - has_broken_compute = false; - has_precise_bug = false; } bool Device::TestVariableAoffi() { diff --git a/src/video_core/renderer_opengl/gl_device.h b/src/video_core/renderer_opengl/gl_device.h index a55050cb5..8a4b6b9fc 100644 --- a/src/video_core/renderer_opengl/gl_device.h +++ b/src/video_core/renderer_opengl/gl_device.h @@ -24,6 +24,10 @@ public: explicit Device(); explicit Device(std::nullptr_t); + u32 GetMaxUniformBuffers(Tegra::Engines::ShaderType shader_type) const noexcept { + return max_uniform_buffers[static_cast<std::size_t>(shader_type)]; + } + const BaseBindings& GetBaseBindings(std::size_t stage_index) const noexcept { return base_bindings[stage_index]; } @@ -48,6 +52,10 @@ public: return max_varyings; } + u32 GetMaxComputeSharedMemorySize() const { + return max_compute_shared_memory_size; + } + bool HasWarpIntrinsics() const { return has_warp_intrinsics; } @@ -64,6 +72,14 @@ public: return has_image_load_formatted; } + bool HasTextureShadowLod() const { + return has_texture_shadow_lod; + } + + bool HasVertexBufferUnifiedMemory() const { + return has_vertex_buffer_unified_memory; + } + bool HasASTC() const { return has_astc; } @@ -80,33 +96,47 @@ public: return has_precise_bug; } - bool HasBrokenCompute() const { - return has_broken_compute; - } - bool HasFastBufferSubData() const { return has_fast_buffer_sub_data; } + bool HasNvViewportArray2() const { + return has_nv_viewport_array2; + } + + bool UseAssemblyShaders() const { + return use_assembly_shaders; + } + + bool UseAsynchronousShaders() const { + return use_asynchronous_shaders; + } + private: static bool TestVariableAoffi(); static bool TestPreciseBug(); - std::array<BaseBindings, Tegra::Engines::MaxShaderTypes> base_bindings; + std::array<u32, Tegra::Engines::MaxShaderTypes> max_uniform_buffers{}; + std::array<BaseBindings, Tegra::Engines::MaxShaderTypes> base_bindings{}; std::size_t uniform_buffer_alignment{}; std::size_t shader_storage_alignment{}; u32 max_vertex_attributes{}; u32 max_varyings{}; + u32 max_compute_shared_memory_size{}; bool has_warp_intrinsics{}; bool has_shader_ballot{}; bool has_vertex_viewport_layer{}; bool has_image_load_formatted{}; + bool has_texture_shadow_lod{}; + bool has_vertex_buffer_unified_memory{}; bool has_astc{}; bool has_variable_aoffi{}; bool has_component_indexing_bug{}; bool has_precise_bug{}; - bool has_broken_compute{}; bool has_fast_buffer_sub_data{}; + bool has_nv_viewport_array2{}; + bool use_assembly_shaders{}; + bool use_asynchronous_shaders{}; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_fence_manager.cpp b/src/video_core/renderer_opengl/gl_fence_manager.cpp new file mode 100644 index 000000000..b532fdcc2 --- /dev/null +++ b/src/video_core/renderer_opengl/gl_fence_manager.cpp @@ -0,0 +1,73 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" + +#include <glad/glad.h> + +#include "video_core/renderer_opengl/gl_buffer_cache.h" +#include "video_core/renderer_opengl/gl_fence_manager.h" + +namespace OpenGL { + +GLInnerFence::GLInnerFence(u32 payload, bool is_stubbed) : FenceBase(payload, is_stubbed) {} + +GLInnerFence::GLInnerFence(GPUVAddr address, u32 payload, bool is_stubbed) + : FenceBase(address, payload, is_stubbed) {} + +GLInnerFence::~GLInnerFence() = default; + +void GLInnerFence::Queue() { + if (is_stubbed) { + return; + } + ASSERT(sync_object.handle == 0); + sync_object.Create(); +} + +bool GLInnerFence::IsSignaled() const { + if (is_stubbed) { + return true; + } + ASSERT(sync_object.handle != 0); + GLsizei length; + GLint sync_status; + glGetSynciv(sync_object.handle, GL_SYNC_STATUS, sizeof(GLint), &length, &sync_status); + return sync_status == GL_SIGNALED; +} + +void GLInnerFence::Wait() { + if (is_stubbed) { + return; + } + ASSERT(sync_object.handle != 0); + glClientWaitSync(sync_object.handle, 0, GL_TIMEOUT_IGNORED); +} + +FenceManagerOpenGL::FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, + TextureCacheOpenGL& texture_cache, + OGLBufferCache& buffer_cache, QueryCache& query_cache) + : GenericFenceManager{rasterizer, gpu, texture_cache, buffer_cache, query_cache} {} + +Fence FenceManagerOpenGL::CreateFence(u32 value, bool is_stubbed) { + return std::make_shared<GLInnerFence>(value, is_stubbed); +} + +Fence FenceManagerOpenGL::CreateFence(GPUVAddr addr, u32 value, bool is_stubbed) { + return std::make_shared<GLInnerFence>(addr, value, is_stubbed); +} + +void FenceManagerOpenGL::QueueFence(Fence& fence) { + fence->Queue(); +} + +bool FenceManagerOpenGL::IsFenceSignaled(Fence& fence) const { + return fence->IsSignaled(); +} + +void FenceManagerOpenGL::WaitFence(Fence& fence) { + fence->Wait(); +} + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_fence_manager.h b/src/video_core/renderer_opengl/gl_fence_manager.h new file mode 100644 index 000000000..da1dcdace --- /dev/null +++ b/src/video_core/renderer_opengl/gl_fence_manager.h @@ -0,0 +1,52 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> + +#include "common/common_types.h" +#include "video_core/fence_manager.h" +#include "video_core/renderer_opengl/gl_buffer_cache.h" +#include "video_core/renderer_opengl/gl_query_cache.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/renderer_opengl/gl_texture_cache.h" + +namespace OpenGL { + +class GLInnerFence : public VideoCommon::FenceBase { +public: + GLInnerFence(u32 payload, bool is_stubbed); + GLInnerFence(GPUVAddr address, u32 payload, bool is_stubbed); + ~GLInnerFence(); + + void Queue(); + + bool IsSignaled() const; + + void Wait(); + +private: + OGLSync sync_object; +}; + +using Fence = std::shared_ptr<GLInnerFence>; +using GenericFenceManager = + VideoCommon::FenceManager<Fence, TextureCacheOpenGL, OGLBufferCache, QueryCache>; + +class FenceManagerOpenGL final : public GenericFenceManager { +public: + explicit FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, + TextureCacheOpenGL& texture_cache, OGLBufferCache& buffer_cache, + QueryCache& query_cache); + +protected: + Fence CreateFence(u32 value, bool is_stubbed) override; + Fence CreateFence(GPUVAddr addr, u32 value, bool is_stubbed) override; + void QueueFence(Fence& fence) override; + bool IsFenceSignaled(Fence& fence) const override; + void WaitFence(Fence& fence) override; +}; + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_query_cache.cpp b/src/video_core/renderer_opengl/gl_query_cache.cpp index f12e9f55f..1a3d9720e 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.cpp +++ b/src/video_core/renderer_opengl/gl_query_cache.cpp @@ -30,12 +30,11 @@ constexpr GLenum GetTarget(VideoCore::QueryType type) { } // Anonymous namespace -QueryCache::QueryCache(Core::System& system, RasterizerOpenGL& gl_rasterizer) - : VideoCommon::QueryCacheBase< - QueryCache, CachedQuery, CounterStream, HostCounter, - std::vector<OGLQuery>>{system, - static_cast<VideoCore::RasterizerInterface&>(gl_rasterizer)}, - gl_rasterizer{gl_rasterizer} {} +QueryCache::QueryCache(RasterizerOpenGL& rasterizer, Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::MemoryManager& gpu_memory) + : VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter>( + rasterizer, maxwell3d, gpu_memory), + gl_rasterizer{rasterizer} {} QueryCache::~QueryCache() = default; @@ -90,13 +89,15 @@ u64 HostCounter::BlockingQuery() const { CachedQuery::CachedQuery(QueryCache& cache, VideoCore::QueryType type, VAddr cpu_addr, u8* host_ptr) : VideoCommon::CachedQueryBase<HostCounter>{cpu_addr, host_ptr}, cache{&cache}, type{type} {} +CachedQuery::~CachedQuery() = default; + CachedQuery::CachedQuery(CachedQuery&& rhs) noexcept : VideoCommon::CachedQueryBase<HostCounter>(std::move(rhs)), cache{rhs.cache}, type{rhs.type} {} CachedQuery& CachedQuery::operator=(CachedQuery&& rhs) noexcept { - VideoCommon::CachedQueryBase<HostCounter>::operator=(std::move(rhs)); cache = rhs.cache; type = rhs.type; + CachedQueryBase<HostCounter>::operator=(std::move(rhs)); return *this; } diff --git a/src/video_core/renderer_opengl/gl_query_cache.h b/src/video_core/renderer_opengl/gl_query_cache.h index d8e7052a1..82cac51ee 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.h +++ b/src/video_core/renderer_opengl/gl_query_cache.h @@ -26,10 +26,11 @@ class RasterizerOpenGL; using CounterStream = VideoCommon::CounterStreamBase<QueryCache, HostCounter>; -class QueryCache final : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, - HostCounter, std::vector<OGLQuery>> { +class QueryCache final + : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter> { public: - explicit QueryCache(Core::System& system, RasterizerOpenGL& rasterizer); + explicit QueryCache(RasterizerOpenGL& rasterizer, Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::MemoryManager& gpu_memory); ~QueryCache(); OGLQuery AllocateQuery(VideoCore::QueryType type); @@ -40,6 +41,7 @@ public: private: RasterizerOpenGL& gl_rasterizer; + std::array<std::vector<OGLQuery>, VideoCore::NumQueryTypes> query_pools; }; class HostCounter final : public VideoCommon::HostCounterBase<QueryCache, HostCounter> { @@ -62,10 +64,12 @@ class CachedQuery final : public VideoCommon::CachedQueryBase<HostCounter> { public: explicit CachedQuery(QueryCache& cache, VideoCore::QueryType type, VAddr cpu_addr, u8* host_ptr); - CachedQuery(CachedQuery&& rhs) noexcept; - CachedQuery(const CachedQuery&) = delete; + ~CachedQuery() override; + CachedQuery(CachedQuery&& rhs) noexcept; CachedQuery& operator=(CachedQuery&& rhs) noexcept; + + CachedQuery(const CachedQuery&) = delete; CachedQuery& operator=(const CachedQuery&) = delete; void Flush() override; diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index f4598fbf7..cfddbde5d 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -30,6 +30,7 @@ #include "video_core/renderer_opengl/gl_shader_cache.h" #include "video_core/renderer_opengl/maxwell_to_gl.h" #include "video_core/renderer_opengl/renderer_opengl.h" +#include "video_core/shader_cache.h" namespace OpenGL { @@ -54,19 +55,36 @@ MICROPROFILE_DEFINE(OpenGL_PrimitiveAssembly, "OpenGL", "Prim Asmbl", MP_RGB(255 namespace { -constexpr std::size_t NumSupportedVertexAttributes = 16; +constexpr std::size_t NUM_CONST_BUFFERS_PER_STAGE = 18; +constexpr std::size_t NUM_CONST_BUFFERS_BYTES_PER_STAGE = + NUM_CONST_BUFFERS_PER_STAGE * Maxwell::MaxConstBufferSize; +constexpr std::size_t TOTAL_CONST_BUFFER_BYTES = + NUM_CONST_BUFFERS_BYTES_PER_STAGE * Maxwell::MaxShaderStage; + +constexpr std::size_t NUM_SUPPORTED_VERTEX_ATTRIBUTES = 16; +constexpr std::size_t NUM_SUPPORTED_VERTEX_BINDINGS = 16; template <typename Engine, typename Entry> Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry, ShaderType shader_type, std::size_t index = 0) { - if (entry.IsBindless()) { - const Tegra::Texture::TextureHandle tex_handle = - engine.AccessConstBuffer32(shader_type, entry.GetBuffer(), entry.GetOffset()); - return engine.GetTextureInfo(tex_handle); + if constexpr (std::is_same_v<Entry, SamplerEntry>) { + if (entry.is_separated) { + const u32 buffer_1 = entry.buffer; + const u32 buffer_2 = entry.secondary_buffer; + const u32 offset_1 = entry.offset; + const u32 offset_2 = entry.secondary_offset; + const u32 handle_1 = engine.AccessConstBuffer32(shader_type, buffer_1, offset_1); + const u32 handle_2 = engine.AccessConstBuffer32(shader_type, buffer_2, offset_2); + return engine.GetTextureInfo(handle_1 | handle_2); + } + } + if (entry.is_bindless) { + const u32 handle = engine.AccessConstBuffer32(shader_type, entry.buffer, entry.offset); + return engine.GetTextureInfo(handle); } + const auto& gpu_profile = engine.AccessGuestDriverProfile(); - const u32 offset = - entry.GetOffset() + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); + const u32 offset = entry.offset + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) { return engine.GetStageTexture(shader_type, offset); } else { @@ -89,23 +107,84 @@ std::size_t GetConstBufferSize(const Tegra::Engines::ConstBufferInfo& buffer, return buffer.size; } +/// Translates hardware transform feedback indices +/// @param location Hardware location +/// @return Pair of ARB_transform_feedback3 token stream first and third arguments +/// @note Read https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_transform_feedback3.txt +std::pair<GLint, GLint> TransformFeedbackEnum(u8 location) { + const u8 index = location / 4; + if (index >= 8 && index <= 39) { + return {GL_GENERIC_ATTRIB_NV, index - 8}; + } + if (index >= 48 && index <= 55) { + return {GL_TEXTURE_COORD_NV, index - 48}; + } + switch (index) { + case 7: + return {GL_POSITION, 0}; + case 40: + return {GL_PRIMARY_COLOR_NV, 0}; + case 41: + return {GL_SECONDARY_COLOR_NV, 0}; + case 42: + return {GL_BACK_PRIMARY_COLOR_NV, 0}; + case 43: + return {GL_BACK_SECONDARY_COLOR_NV, 0}; + } + UNIMPLEMENTED_MSG("index={}", static_cast<int>(index)); + return {GL_POSITION, 0}; +} + void oglEnable(GLenum cap, bool state) { (state ? glEnable : glDisable)(cap); } +void UpdateBindlessSSBOs(GLenum target, const BindlessSSBO* ssbos, size_t num_ssbos) { + if (num_ssbos == 0) { + return; + } + glProgramLocalParametersI4uivNV(target, 0, static_cast<GLsizei>(num_ssbos), + reinterpret_cast<const GLuint*>(ssbos)); +} + } // Anonymous namespace -RasterizerOpenGL::RasterizerOpenGL(Core::System& system, Core::Frontend::EmuWindow& emu_window, - ScreenInfo& info, GLShader::ProgramManager& program_manager, - StateTracker& state_tracker) - : RasterizerAccelerated{system.Memory()}, texture_cache{system, *this, device, state_tracker}, - shader_cache{*this, system, emu_window, device}, query_cache{system, *this}, system{system}, - screen_info{info}, program_manager{program_manager}, state_tracker{state_tracker}, - buffer_cache{*this, system, device, STREAM_BUFFER_SIZE} { +RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu_, + Core::Memory::Memory& cpu_memory, const Device& device_, + ScreenInfo& screen_info_, ProgramManager& program_manager_, + StateTracker& state_tracker_) + : RasterizerAccelerated{cpu_memory}, gpu(gpu_), maxwell3d(gpu.Maxwell3D()), + kepler_compute(gpu.KeplerCompute()), gpu_memory(gpu.MemoryManager()), device(device_), + screen_info(screen_info_), program_manager(program_manager_), state_tracker(state_tracker_), + texture_cache(*this, maxwell3d, gpu_memory, device, state_tracker), + shader_cache(*this, emu_window, gpu, maxwell3d, kepler_compute, gpu_memory, device), + query_cache(*this, maxwell3d, gpu_memory), + buffer_cache(*this, gpu_memory, cpu_memory, device, STREAM_BUFFER_SIZE), + fence_manager(*this, gpu, texture_cache, buffer_cache, query_cache), + async_shaders(emu_window) { CheckExtensions(); + + unified_uniform_buffer.Create(); + glNamedBufferStorage(unified_uniform_buffer.handle, TOTAL_CONST_BUFFER_BYTES, nullptr, 0); + + if (device.UseAssemblyShaders()) { + glCreateBuffers(static_cast<GLsizei>(staging_cbufs.size()), staging_cbufs.data()); + for (const GLuint cbuf : staging_cbufs) { + glNamedBufferStorage(cbuf, static_cast<GLsizeiptr>(Maxwell::MaxConstBufferSize), + nullptr, 0); + } + } + + if (device.UseAsynchronousShaders()) { + async_shaders.AllocateWorkers(); + } } -RasterizerOpenGL::~RasterizerOpenGL() {} +RasterizerOpenGL::~RasterizerOpenGL() { + if (device.UseAssemblyShaders()) { + glDeleteBuffers(static_cast<GLsizei>(staging_cbufs.size()), staging_cbufs.data()); + } +} void RasterizerOpenGL::CheckExtensions() { if (!GLAD_GL_ARB_texture_filter_anisotropic && !GLAD_GL_EXT_texture_filter_anisotropic) { @@ -116,8 +195,7 @@ void RasterizerOpenGL::CheckExtensions() { } void RasterizerOpenGL::SetupVertexFormat() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::VertexFormats]) { return; } @@ -131,13 +209,13 @@ void RasterizerOpenGL::SetupVertexFormat() { // avoid OpenGL errors. // TODO(Subv): Analyze the shader to identify which attributes are actually used and don't // assume every shader uses them all. - for (std::size_t index = 0; index < NumSupportedVertexAttributes; ++index) { + for (std::size_t index = 0; index < NUM_SUPPORTED_VERTEX_ATTRIBUTES; ++index) { if (!flags[Dirty::VertexFormat0 + index]) { continue; } flags[Dirty::VertexFormat0 + index] = false; - const auto attrib = gpu.regs.vertex_attrib_format[index]; + const auto attrib = maxwell3d.regs.vertex_attrib_format[index]; const auto gl_index = static_cast<GLuint>(index); // Disable constant attributes. @@ -150,9 +228,10 @@ void RasterizerOpenGL::SetupVertexFormat() { if (attrib.type == Maxwell::VertexAttribute::Type::SignedInt || attrib.type == Maxwell::VertexAttribute::Type::UnsignedInt) { glVertexAttribIFormat(gl_index, attrib.ComponentCount(), - MaxwellToGL::VertexType(attrib), attrib.offset); + MaxwellToGL::VertexFormat(attrib), attrib.offset); } else { - glVertexAttribFormat(gl_index, attrib.ComponentCount(), MaxwellToGL::VertexType(attrib), + glVertexAttribFormat(gl_index, attrib.ComponentCount(), + MaxwellToGL::VertexFormat(attrib), attrib.IsNormalized() ? GL_TRUE : GL_FALSE, attrib.offset); } glVertexAttribBinding(gl_index, attrib.buffer); @@ -160,8 +239,7 @@ void RasterizerOpenGL::SetupVertexFormat() { } void RasterizerOpenGL::SetupVertexBuffer() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::VertexBuffers]) { return; } @@ -169,9 +247,11 @@ void RasterizerOpenGL::SetupVertexBuffer() { MICROPROFILE_SCOPE(OpenGL_VB); + const bool use_unified_memory = device.HasVertexBufferUnifiedMemory(); + // Upload all guest vertex arrays sequentially to our buffer - const auto& regs = gpu.regs; - for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { + const auto& regs = maxwell3d.regs; + for (std::size_t index = 0; index < NUM_SUPPORTED_VERTEX_BINDINGS; ++index) { if (!flags[Dirty::VertexBuffer0 + index]) { continue; } @@ -184,27 +264,37 @@ void RasterizerOpenGL::SetupVertexBuffer() { const GPUVAddr start = vertex_array.StartAddress(); const GPUVAddr end = regs.vertex_array_limit[index].LimitAddress(); - - ASSERT(end > start); - const u64 size = end - start + 1; - const auto [vertex_buffer, vertex_buffer_offset] = buffer_cache.UploadMemory(start, size); - - // Bind the vertex array to the buffer at the current offset. - vertex_array_pushbuffer.SetVertexBuffer(static_cast<GLuint>(index), vertex_buffer, - vertex_buffer_offset, vertex_array.stride); + ASSERT(end >= start); + + const GLuint gl_index = static_cast<GLuint>(index); + const u64 size = end - start; + if (size == 0) { + glBindVertexBuffer(gl_index, 0, 0, vertex_array.stride); + if (use_unified_memory) { + glBufferAddressRangeNV(GL_VERTEX_ATTRIB_ARRAY_ADDRESS_NV, gl_index, 0, 0); + } + continue; + } + const auto info = buffer_cache.UploadMemory(start, size); + if (use_unified_memory) { + glBindVertexBuffer(gl_index, 0, 0, vertex_array.stride); + glBufferAddressRangeNV(GL_VERTEX_ATTRIB_ARRAY_ADDRESS_NV, gl_index, + info.address + info.offset, size); + } else { + glBindVertexBuffer(gl_index, info.handle, info.offset, vertex_array.stride); + } } } void RasterizerOpenGL::SetupVertexInstances() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::VertexInstances]) { return; } flags[Dirty::VertexInstances] = false; - const auto& regs = gpu.regs; - for (std::size_t index = 0; index < NumSupportedVertexAttributes; ++index) { + const auto& regs = maxwell3d.regs; + for (std::size_t index = 0; index < NUM_SUPPORTED_VERTEX_ATTRIBUTES; ++index) { if (!flags[Dirty::VertexInstance0 + index]) { continue; } @@ -219,24 +309,23 @@ void RasterizerOpenGL::SetupVertexInstances() { GLintptr RasterizerOpenGL::SetupIndexBuffer() { MICROPROFILE_SCOPE(OpenGL_Index); - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; const std::size_t size = CalculateIndexBufferSize(); - const auto [buffer, offset] = buffer_cache.UploadMemory(regs.index_array.IndexStart(), size); - vertex_array_pushbuffer.SetIndexBuffer(buffer); - return offset; + const auto info = buffer_cache.UploadMemory(regs.index_array.IndexStart(), size); + glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, info.handle); + return info.offset; } void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { MICROPROFILE_SCOPE(OpenGL_Shader); - auto& gpu = system.GPU().Maxwell3D(); u32 clip_distances = 0; for (std::size_t index = 0; index < Maxwell::MaxShaderProgram; ++index) { - const auto& shader_config = gpu.regs.shader_config[index]; + const auto& shader_config = maxwell3d.regs.shader_config[index]; const auto program{static_cast<Maxwell::ShaderProgram>(index)}; // Skip stages that are not enabled - if (!gpu.regs.IsShaderConfigEnabled(index)) { + if (!maxwell3d.regs.IsShaderConfigEnabled(index)) { switch (program) { case Maxwell::ShaderProgram::Geometry: program_manager.UseGeometryShader(0); @@ -251,23 +340,15 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { } // Currently this stages are not supported in the OpenGL backend. - // Todo(Blinkhawk): Port tesselation shaders from Vulkan to OpenGL - if (program == Maxwell::ShaderProgram::TesselationControl) { - continue; - } else if (program == Maxwell::ShaderProgram::TesselationEval) { + // TODO(Blinkhawk): Port tesselation shaders from Vulkan to OpenGL + if (program == Maxwell::ShaderProgram::TesselationControl || + program == Maxwell::ShaderProgram::TesselationEval) { continue; } - Shader shader{shader_cache.GetStageProgram(program)}; + Shader* const shader = shader_cache.GetStageProgram(program, async_shaders); - // Stage indices are 0 - 5 - const std::size_t stage = index == 0 ? 0 : index - 1; - SetupDrawConstBuffers(stage, shader); - SetupDrawGlobalMemory(stage, shader); - SetupDrawTextures(stage, shader); - SetupDrawImages(stage, shader); - - const GLuint program_handle = shader->GetHandle(); + const GLuint program_handle = shader->IsBuilt() ? shader->GetHandle() : 0; switch (program) { case Maxwell::ShaderProgram::VertexA: case Maxwell::ShaderProgram::VertexB: @@ -284,6 +365,13 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { shader_config.enable.Value(), shader_config.offset); } + // Stage indices are 0 - 5 + const std::size_t stage = index == 0 ? 0 : index - 1; + SetupDrawConstBuffers(stage, shader); + SetupDrawGlobalMemory(stage, shader); + SetupDrawTextures(stage, shader); + SetupDrawImages(stage, shader); + // Workaround for Intel drivers. // When a clip distance is enabled but not set in the shader it crops parts of the screen // (sometimes it's half the screen, sometimes three quarters). To avoid this, enable the @@ -298,11 +386,11 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { } SyncClipEnabled(clip_distances); - gpu.dirty.flags[Dirty::Shaders] = false; + maxwell3d.dirty.flags[Dirty::Shaders] = false; } std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const { - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; std::size_t size = 0; for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) { @@ -312,49 +400,42 @@ std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const { const GPUVAddr start = regs.vertex_array[index].StartAddress(); const GPUVAddr end = regs.vertex_array_limit[index].LimitAddress(); - ASSERT(end > start); - size += end - start + 1; + size += end - start; + ASSERT(end >= start); } return size; } std::size_t RasterizerOpenGL::CalculateIndexBufferSize() const { - const auto& regs = system.GPU().Maxwell3D().regs; - - return static_cast<std::size_t>(regs.index_array.count) * - static_cast<std::size_t>(regs.index_array.FormatSizeInBytes()); + return static_cast<std::size_t>(maxwell3d.regs.index_array.count) * + static_cast<std::size_t>(maxwell3d.regs.index_array.FormatSizeInBytes()); } -void RasterizerOpenGL::LoadDiskResources(const std::atomic_bool& stop_loading, +void RasterizerOpenGL::LoadDiskResources(u64 title_id, const std::atomic_bool& stop_loading, const VideoCore::DiskResourceLoadCallback& callback) { - shader_cache.LoadDiskCache(stop_loading, callback); -} - -void RasterizerOpenGL::SetupDirtyFlags() { - state_tracker.Initialize(); + shader_cache.LoadDiskCache(title_id, stop_loading, callback); } void RasterizerOpenGL::ConfigureFramebuffers() { MICROPROFILE_SCOPE(OpenGL_Framebuffer); - auto& gpu = system.GPU().Maxwell3D(); - if (!gpu.dirty.flags[VideoCommon::Dirty::RenderTargets]) { + if (!maxwell3d.dirty.flags[VideoCommon::Dirty::RenderTargets]) { return; } - gpu.dirty.flags[VideoCommon::Dirty::RenderTargets] = false; + maxwell3d.dirty.flags[VideoCommon::Dirty::RenderTargets] = false; texture_cache.GuardRenderTargets(true); - View depth_surface = texture_cache.GetDepthBufferSurface(); + View depth_surface = texture_cache.GetDepthBufferSurface(true); - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; UNIMPLEMENTED_IF(regs.rt_separate_frag_data == 0); // Bind the framebuffer surfaces FramebufferCacheKey key; const auto colors_count = static_cast<std::size_t>(regs.rt_control.count); for (std::size_t index = 0; index < colors_count; ++index) { - View color_surface{texture_cache.GetColorBufferSurface(index)}; + View color_surface{texture_cache.GetColorBufferSurface(index, true)}; if (!color_surface) { continue; } @@ -378,40 +459,62 @@ void RasterizerOpenGL::ConfigureFramebuffers() { glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer_cache.GetFramebuffer(key)); } -void RasterizerOpenGL::ConfigureClearFramebuffer(bool using_color_fb, bool using_depth_fb, - bool using_stencil_fb) { - auto& gpu = system.GPU().Maxwell3D(); - const auto& regs = gpu.regs; +void RasterizerOpenGL::ConfigureClearFramebuffer(bool using_color, bool using_depth_stencil) { + const auto& regs = maxwell3d.regs; texture_cache.GuardRenderTargets(true); View color_surface; - if (using_color_fb) { + + if (using_color) { + // Determine if we have to preserve the contents. + // First we have to make sure all clear masks are enabled. + bool preserve_contents = !regs.clear_buffers.R || !regs.clear_buffers.G || + !regs.clear_buffers.B || !regs.clear_buffers.A; const std::size_t index = regs.clear_buffers.RT; - color_surface = texture_cache.GetColorBufferSurface(index); + if (regs.clear_flags.scissor) { + // Then we have to confirm scissor testing clears the whole image. + const auto& scissor = regs.scissor_test[0]; + preserve_contents |= scissor.min_x > 0; + preserve_contents |= scissor.min_y > 0; + preserve_contents |= scissor.max_x < regs.rt[index].width; + preserve_contents |= scissor.max_y < regs.rt[index].height; + } + + color_surface = texture_cache.GetColorBufferSurface(index, preserve_contents); texture_cache.MarkColorBufferInUse(index); } + View depth_surface; - if (using_depth_fb || using_stencil_fb) { - depth_surface = texture_cache.GetDepthBufferSurface(); + if (using_depth_stencil) { + bool preserve_contents = false; + if (regs.clear_flags.scissor) { + // For depth stencil clears we only have to confirm scissor test covers the whole image. + const auto& scissor = regs.scissor_test[0]; + preserve_contents |= scissor.min_x > 0; + preserve_contents |= scissor.min_y > 0; + preserve_contents |= scissor.max_x < regs.zeta_width; + preserve_contents |= scissor.max_y < regs.zeta_height; + } + + depth_surface = texture_cache.GetDepthBufferSurface(preserve_contents); texture_cache.MarkDepthBufferInUse(); } texture_cache.GuardRenderTargets(false); FramebufferCacheKey key; - key.colors[0] = color_surface; - key.zeta = depth_surface; + key.colors[0] = std::move(color_surface); + key.zeta = std::move(depth_surface); state_tracker.NotifyFramebuffer(); glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer_cache.GetFramebuffer(key)); } void RasterizerOpenGL::Clear() { - const auto& gpu = system.GPU().Maxwell3D(); - if (!gpu.ShouldExecute()) { + if (!maxwell3d.ShouldExecute()) { return; } - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; bool use_color{}; bool use_depth{}; bool use_stencil{}; @@ -419,8 +522,7 @@ void RasterizerOpenGL::Clear() { if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B || regs.clear_buffers.A) { use_color = true; - } - if (use_color) { + state_tracker.NotifyColorMask0(); glColorMaski(0, regs.clear_buffers.R != 0, regs.clear_buffers.G != 0, regs.clear_buffers.B != 0, regs.clear_buffers.A != 0); @@ -458,7 +560,7 @@ void RasterizerOpenGL::Clear() { UNIMPLEMENTED_IF(regs.clear_flags.viewport); - ConfigureClearFramebuffer(use_color, use_depth, use_stencil); + ConfigureClearFramebuffer(use_color, use_depth || use_stencil); if (use_color) { glClearBufferfv(GL_COLOR, 0, regs.clear_color); @@ -477,7 +579,6 @@ void RasterizerOpenGL::Clear() { void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { MICROPROFILE_SCOPE(OpenGL_Drawing); - auto& gpu = system.GPU().Maxwell3D(); query_cache.UpdateCounters(); @@ -502,6 +603,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { SyncFramebufferSRGB(); buffer_cache.Acquire(); + current_cbuf = 0; std::size_t buffer_size = CalculateVertexArraysSize(); @@ -511,20 +613,28 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { } // Uniform space for the 5 shader stages - buffer_size = Common::AlignUp<std::size_t>(buffer_size, 4) + - (sizeof(GLShader::MaxwellUniformData) + device.GetUniformBufferAlignment()) * - Maxwell::MaxShaderStage; + buffer_size = + Common::AlignUp<std::size_t>(buffer_size, 4) + + (sizeof(MaxwellUniformData) + device.GetUniformBufferAlignment()) * Maxwell::MaxShaderStage; // Add space for at least 18 constant buffers buffer_size += Maxwell::MaxConstBuffers * (Maxwell::MaxConstBufferSize + device.GetUniformBufferAlignment()); // Prepare the vertex array. - buffer_cache.Map(buffer_size); + const bool invalidated = buffer_cache.Map(buffer_size); + + if (invalidated) { + // When the stream buffer has been invalidated, we have to consider vertex buffers as dirty + auto& dirty = maxwell3d.dirty.flags; + dirty[Dirty::VertexBuffers] = true; + for (int index = Dirty::VertexBuffer0; index <= Dirty::VertexBuffer31; ++index) { + dirty[index] = true; + } + } // Prepare vertex array format. SetupVertexFormat(); - vertex_array_pushbuffer.Setup(); // Upload vertex and index data. SetupVertexBuffer(); @@ -534,21 +644,19 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { index_buffer_offset = SetupIndexBuffer(); } - // Prepare packed bindings. - bind_ubo_pushbuffer.Setup(); - bind_ssbo_pushbuffer.Setup(); - // Setup emulation uniform buffer. - GLShader::MaxwellUniformData ubo; - ubo.SetFromRegs(gpu); - const auto [buffer, offset] = - buffer_cache.UploadHostMemory(&ubo, sizeof(ubo), device.GetUniformBufferAlignment()); - bind_ubo_pushbuffer.Push(EmulationUniformBlockBinding, buffer, offset, - static_cast<GLsizeiptr>(sizeof(ubo))); + if (!device.UseAssemblyShaders()) { + MaxwellUniformData ubo; + ubo.SetFromRegs(maxwell3d); + const auto info = + buffer_cache.UploadHostMemory(&ubo, sizeof(ubo), device.GetUniformBufferAlignment()); + glBindBufferRange(GL_UNIFORM_BUFFER, EmulationUniformBlockBinding, info.handle, info.offset, + static_cast<GLsizeiptr>(sizeof(ubo))); + } // Setup shaders and their used resources. texture_cache.GuardSamplers(true); - const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(gpu.regs.draw.topology); + const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(maxwell3d.regs.draw.topology); SetupShaders(primitive_mode); texture_cache.GuardSamplers(false); @@ -557,11 +665,6 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { // Signal the buffer cache that we are not going to upload more things. buffer_cache.Unmap(); - // Now that we are no longer uploading data, we can safely bind the buffers to OpenGL. - vertex_array_pushbuffer.Bind(); - bind_ubo_pushbuffer.Bind(); - bind_ssbo_pushbuffer.Bind(); - program_manager.BindGraphicsPipeline(); if (texture_cache.TextureBarrier()) { @@ -570,14 +673,14 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { BeginTransformFeedback(primitive_mode); - const GLuint base_instance = static_cast<GLuint>(gpu.regs.vb_base_instance); + const GLuint base_instance = static_cast<GLuint>(maxwell3d.regs.vb_base_instance); const GLsizei num_instances = - static_cast<GLsizei>(is_instanced ? gpu.mme_draw.instance_count : 1); + static_cast<GLsizei>(is_instanced ? maxwell3d.mme_draw.instance_count : 1); if (is_indexed) { - const GLint base_vertex = static_cast<GLint>(gpu.regs.vb_element_base); - const GLsizei num_vertices = static_cast<GLsizei>(gpu.regs.index_array.count); + const GLint base_vertex = static_cast<GLint>(maxwell3d.regs.vb_element_base); + const GLsizei num_vertices = static_cast<GLsizei>(maxwell3d.regs.index_array.count); const GLvoid* offset = reinterpret_cast<const GLvoid*>(index_buffer_offset); - const GLenum format = MaxwellToGL::IndexFormat(gpu.regs.index_array.format); + const GLenum format = MaxwellToGL::IndexFormat(maxwell3d.regs.index_array.format); if (num_instances == 1 && base_instance == 0 && base_vertex == 0) { glDrawElements(primitive_mode, num_vertices, format, offset); } else if (num_instances == 1 && base_instance == 0) { @@ -596,8 +699,8 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { base_instance); } } else { - const GLint base_vertex = static_cast<GLint>(gpu.regs.vertex_buffer.first); - const GLsizei num_vertices = static_cast<GLsizei>(gpu.regs.vertex_buffer.count); + const GLint base_vertex = static_cast<GLint>(maxwell3d.regs.vertex_buffer.first); + const GLsizei num_vertices = static_cast<GLsizei>(maxwell3d.regs.vertex_buffer.count); if (num_instances == 1 && base_instance == 0) { glDrawArrays(primitive_mode, base_vertex, num_vertices); } else if (base_instance == 0) { @@ -611,37 +714,32 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { EndTransformFeedback(); ++num_queued_commands; + + gpu.TickWork(); } void RasterizerOpenGL::DispatchCompute(GPUVAddr code_addr) { - if (device.HasBrokenCompute()) { - return; - } - buffer_cache.Acquire(); + current_cbuf = 0; auto kernel = shader_cache.GetComputeKernel(code_addr); + program_manager.BindCompute(kernel->GetHandle()); + SetupComputeTextures(kernel); SetupComputeImages(kernel); - program_manager.BindComputeShader(kernel->GetHandle()); const std::size_t buffer_size = Tegra::Engines::KeplerCompute::NumConstBuffers * (Maxwell::MaxConstBufferSize + device.GetUniformBufferAlignment()); buffer_cache.Map(buffer_size); - bind_ubo_pushbuffer.Setup(); - bind_ssbo_pushbuffer.Setup(); - SetupComputeConstBuffers(kernel); SetupComputeGlobalMemory(kernel); buffer_cache.Unmap(); - bind_ubo_pushbuffer.Bind(); - bind_ssbo_pushbuffer.Bind(); - - const auto& launch_desc = system.GPU().KeplerCompute().launch_description; + const auto& launch_desc = kepler_compute.launch_description; + program_manager.BindCompute(kernel->GetHandle()); glDispatchCompute(launch_desc.grid_dim_x, launch_desc.grid_dim_y, launch_desc.grid_dim_z); ++num_queued_commands; } @@ -667,6 +765,13 @@ void RasterizerOpenGL::FlushRegion(VAddr addr, u64 size) { query_cache.FlushRegion(addr, size); } +bool RasterizerOpenGL::MustFlushRegion(VAddr addr, u64 size) { + if (!Settings::IsGPULevelHigh()) { + return buffer_cache.MustFlushRegion(addr, size); + } + return texture_cache.MustFlushRegion(addr, size) || buffer_cache.MustFlushRegion(addr, size); +} + void RasterizerOpenGL::InvalidateRegion(VAddr addr, u64 size) { MICROPROFILE_SCOPE(OpenGL_CacheManagement); if (addr == 0 || size == 0) { @@ -678,13 +783,64 @@ void RasterizerOpenGL::InvalidateRegion(VAddr addr, u64 size) { query_cache.InvalidateRegion(addr, size); } +void RasterizerOpenGL::OnCPUWrite(VAddr addr, u64 size) { + MICROPROFILE_SCOPE(OpenGL_CacheManagement); + if (addr == 0 || size == 0) { + return; + } + texture_cache.OnCPUWrite(addr, size); + shader_cache.OnCPUWrite(addr, size); + buffer_cache.OnCPUWrite(addr, size); +} + +void RasterizerOpenGL::SyncGuestHost() { + MICROPROFILE_SCOPE(OpenGL_CacheManagement); + texture_cache.SyncGuestHost(); + buffer_cache.SyncGuestHost(); + shader_cache.SyncGuestHost(); +} + +void RasterizerOpenGL::SignalSemaphore(GPUVAddr addr, u32 value) { + if (!gpu.IsAsync()) { + gpu_memory.Write<u32>(addr, value); + return; + } + fence_manager.SignalSemaphore(addr, value); +} + +void RasterizerOpenGL::SignalSyncPoint(u32 value) { + if (!gpu.IsAsync()) { + gpu.IncrementSyncPoint(value); + return; + } + fence_manager.SignalSyncPoint(value); +} + +void RasterizerOpenGL::ReleaseFences() { + if (!gpu.IsAsync()) { + return; + } + fence_manager.WaitPendingFences(); +} + void RasterizerOpenGL::FlushAndInvalidateRegion(VAddr addr, u64 size) { - if (Settings::values.use_accurate_gpu_emulation) { + if (Settings::IsGPULevelExtreme()) { FlushRegion(addr, size); } InvalidateRegion(addr, size); } +void RasterizerOpenGL::WaitForIdle() { + // Place a barrier on everything that is not framebuffer related. + // This is related to another flag that is not currently implemented. + glMemoryBarrier(GL_VERTEX_ATTRIB_ARRAY_BARRIER_BIT | GL_ELEMENT_ARRAY_BARRIER_BIT | + GL_UNIFORM_BARRIER_BIT | GL_TEXTURE_FETCH_BARRIER_BIT | + GL_SHADER_IMAGE_ACCESS_BARRIER_BIT | GL_COMMAND_BARRIER_BIT | + GL_PIXEL_BUFFER_BARRIER_BIT | GL_TEXTURE_UPDATE_BARRIER_BIT | + GL_BUFFER_UPDATE_BARRIER_BIT | GL_TRANSFORM_FEEDBACK_BARRIER_BIT | + GL_SHADER_STORAGE_BARRIER_BIT | GL_QUERY_BUFFER_BARRIER_BIT); +} + void RasterizerOpenGL::FlushCommands() { // Only flush when we have commands queued to OpenGL. if (num_queued_commands == 0) { @@ -739,40 +895,72 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config, return true; } -void RasterizerOpenGL::SetupDrawConstBuffers(std::size_t stage_index, const Shader& shader) { +void RasterizerOpenGL::SetupDrawConstBuffers(std::size_t stage_index, Shader* shader) { + static constexpr std::array PARAMETER_LUT{ + GL_VERTEX_PROGRAM_PARAMETER_BUFFER_NV, GL_TESS_CONTROL_PROGRAM_PARAMETER_BUFFER_NV, + GL_TESS_EVALUATION_PROGRAM_PARAMETER_BUFFER_NV, GL_GEOMETRY_PROGRAM_PARAMETER_BUFFER_NV, + GL_FRAGMENT_PROGRAM_PARAMETER_BUFFER_NV, + }; MICROPROFILE_SCOPE(OpenGL_UBO); - const auto& stages = system.GPU().Maxwell3D().state.shader_stages; + const auto& stages = maxwell3d.state.shader_stages; const auto& shader_stage = stages[stage_index]; - - u32 binding = device.GetBaseBindings(stage_index).uniform_buffer; - for (const auto& entry : shader->GetEntries().const_buffers) { - const auto& buffer = shader_stage.const_buffers[entry.GetIndex()]; - SetupConstBuffer(binding++, buffer, entry); + const auto& entries = shader->GetEntries(); + const bool use_unified = entries.use_unified_uniforms; + const std::size_t base_unified_offset = stage_index * NUM_CONST_BUFFERS_BYTES_PER_STAGE; + + const auto base_bindings = device.GetBaseBindings(stage_index); + u32 binding = device.UseAssemblyShaders() ? 0 : base_bindings.uniform_buffer; + for (const auto& entry : entries.const_buffers) { + const u32 index = entry.GetIndex(); + const auto& buffer = shader_stage.const_buffers[index]; + SetupConstBuffer(PARAMETER_LUT[stage_index], binding, buffer, entry, use_unified, + base_unified_offset + index * Maxwell::MaxConstBufferSize); + ++binding; + } + if (use_unified) { + const u32 index = static_cast<u32>(base_bindings.shader_storage_buffer + + entries.global_memory_entries.size()); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, index, unified_uniform_buffer.handle, + base_unified_offset, NUM_CONST_BUFFERS_BYTES_PER_STAGE); } } -void RasterizerOpenGL::SetupComputeConstBuffers(const Shader& kernel) { +void RasterizerOpenGL::SetupComputeConstBuffers(Shader* kernel) { MICROPROFILE_SCOPE(OpenGL_UBO); - const auto& launch_desc = system.GPU().KeplerCompute().launch_description; + const auto& launch_desc = kepler_compute.launch_description; + const auto& entries = kernel->GetEntries(); + const bool use_unified = entries.use_unified_uniforms; u32 binding = 0; - for (const auto& entry : kernel->GetEntries().const_buffers) { + for (const auto& entry : entries.const_buffers) { const auto& config = launch_desc.const_buffer_config[entry.GetIndex()]; const std::bitset<8> mask = launch_desc.const_buffer_enable_mask.Value(); Tegra::Engines::ConstBufferInfo buffer; buffer.address = config.Address(); buffer.size = config.size; buffer.enabled = mask[entry.GetIndex()]; - SetupConstBuffer(binding++, buffer, entry); + SetupConstBuffer(GL_COMPUTE_PROGRAM_PARAMETER_BUFFER_NV, binding, buffer, entry, + use_unified, entry.GetIndex() * Maxwell::MaxConstBufferSize); + ++binding; + } + if (use_unified) { + const GLuint index = static_cast<GLuint>(entries.global_memory_entries.size()); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, index, unified_uniform_buffer.handle, 0, + NUM_CONST_BUFFERS_BYTES_PER_STAGE); } } -void RasterizerOpenGL::SetupConstBuffer(u32 binding, const Tegra::Engines::ConstBufferInfo& buffer, - const ConstBufferEntry& entry) { +void RasterizerOpenGL::SetupConstBuffer(GLenum stage, u32 binding, + const Tegra::Engines::ConstBufferInfo& buffer, + const ConstBufferEntry& entry, bool use_unified, + std::size_t unified_offset) { if (!buffer.enabled) { // Set values to zero to unbind buffers - bind_ubo_pushbuffer.Push(binding, buffer_cache.GetEmptyBuffer(sizeof(float)), 0, - sizeof(float)); + if (device.UseAssemblyShaders()) { + glBindBufferRangeNV(stage, entry.GetIndex(), 0, 0, 0); + } else { + glBindBufferRange(GL_UNIFORM_BUFFER, binding, 0, 0, sizeof(float)); + } return; } @@ -780,68 +968,112 @@ void RasterizerOpenGL::SetupConstBuffer(u32 binding, const Tegra::Engines::Const // UBO alignment requirements. const std::size_t size = Common::AlignUp(GetConstBufferSize(buffer, entry), sizeof(GLvec4)); - const auto alignment = device.GetUniformBufferAlignment(); - const auto [cbuf, offset] = buffer_cache.UploadMemory(buffer.address, size, alignment, false, - device.HasFastBufferSubData()); - bind_ubo_pushbuffer.Push(binding, cbuf, offset, size); + const bool fast_upload = !use_unified && device.HasFastBufferSubData(); + + const std::size_t alignment = use_unified ? 4 : device.GetUniformBufferAlignment(); + const GPUVAddr gpu_addr = buffer.address; + auto info = buffer_cache.UploadMemory(gpu_addr, size, alignment, false, fast_upload); + + if (device.UseAssemblyShaders()) { + UNIMPLEMENTED_IF(use_unified); + if (info.offset != 0) { + const GLuint staging_cbuf = staging_cbufs[current_cbuf++]; + glCopyNamedBufferSubData(info.handle, staging_cbuf, info.offset, 0, size); + info.handle = staging_cbuf; + info.offset = 0; + } + glBindBufferRangeNV(stage, binding, info.handle, info.offset, size); + return; + } + + if (use_unified) { + glCopyNamedBufferSubData(info.handle, unified_uniform_buffer.handle, info.offset, + unified_offset, size); + } else { + glBindBufferRange(GL_UNIFORM_BUFFER, binding, info.handle, info.offset, size); + } } -void RasterizerOpenGL::SetupDrawGlobalMemory(std::size_t stage_index, const Shader& shader) { - auto& gpu{system.GPU()}; - auto& memory_manager{gpu.MemoryManager()}; - const auto cbufs{gpu.Maxwell3D().state.shader_stages[stage_index]}; +void RasterizerOpenGL::SetupDrawGlobalMemory(std::size_t stage_index, Shader* shader) { + static constexpr std::array TARGET_LUT = { + GL_VERTEX_PROGRAM_NV, GL_TESS_CONTROL_PROGRAM_NV, GL_TESS_EVALUATION_PROGRAM_NV, + GL_GEOMETRY_PROGRAM_NV, GL_FRAGMENT_PROGRAM_NV, + }; + + const auto& cbufs{maxwell3d.state.shader_stages[stage_index]}; + const auto& entries{shader->GetEntries().global_memory_entries}; - u32 binding = device.GetBaseBindings(stage_index).shader_storage_buffer; - for (const auto& entry : shader->GetEntries().global_memory_entries) { - const auto addr{cbufs.const_buffers[entry.GetCbufIndex()].address + entry.GetCbufOffset()}; - const auto gpu_addr{memory_manager.Read<u64>(addr)}; - const auto size{memory_manager.Read<u32>(addr + 8)}; - SetupGlobalMemory(binding++, entry, gpu_addr, size); + std::array<BindlessSSBO, 32> ssbos; + ASSERT(entries.size() < ssbos.size()); + + const bool assembly_shaders = device.UseAssemblyShaders(); + u32 binding = assembly_shaders ? 0 : device.GetBaseBindings(stage_index).shader_storage_buffer; + for (const auto& entry : entries) { + const GPUVAddr addr{cbufs.const_buffers[entry.cbuf_index].address + entry.cbuf_offset}; + const GPUVAddr gpu_addr{gpu_memory.Read<u64>(addr)}; + const u32 size{gpu_memory.Read<u32>(addr + 8)}; + SetupGlobalMemory(binding, entry, gpu_addr, size, &ssbos[binding]); + ++binding; + } + if (assembly_shaders) { + UpdateBindlessSSBOs(TARGET_LUT[stage_index], ssbos.data(), entries.size()); } } -void RasterizerOpenGL::SetupComputeGlobalMemory(const Shader& kernel) { - auto& gpu{system.GPU()}; - auto& memory_manager{gpu.MemoryManager()}; - const auto cbufs{gpu.KeplerCompute().launch_description.const_buffer_config}; +void RasterizerOpenGL::SetupComputeGlobalMemory(Shader* kernel) { + const auto& cbufs{kepler_compute.launch_description.const_buffer_config}; + const auto& entries{kernel->GetEntries().global_memory_entries}; + + std::array<BindlessSSBO, 32> ssbos; + ASSERT(entries.size() < ssbos.size()); u32 binding = 0; - for (const auto& entry : kernel->GetEntries().global_memory_entries) { - const auto addr{cbufs[entry.GetCbufIndex()].Address() + entry.GetCbufOffset()}; - const auto gpu_addr{memory_manager.Read<u64>(addr)}; - const auto size{memory_manager.Read<u32>(addr + 8)}; - SetupGlobalMemory(binding++, entry, gpu_addr, size); + for (const auto& entry : entries) { + const GPUVAddr addr{cbufs[entry.cbuf_index].Address() + entry.cbuf_offset}; + const GPUVAddr gpu_addr{gpu_memory.Read<u64>(addr)}; + const u32 size{gpu_memory.Read<u32>(addr + 8)}; + SetupGlobalMemory(binding, entry, gpu_addr, size, &ssbos[binding]); + ++binding; + } + if (device.UseAssemblyShaders()) { + UpdateBindlessSSBOs(GL_COMPUTE_PROGRAM_NV, ssbos.data(), ssbos.size()); } } void RasterizerOpenGL::SetupGlobalMemory(u32 binding, const GlobalMemoryEntry& entry, - GPUVAddr gpu_addr, std::size_t size) { - const auto alignment{device.GetShaderStorageBufferAlignment()}; - const auto [ssbo, buffer_offset] = - buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.IsWritten()); - bind_ssbo_pushbuffer.Push(binding, ssbo, buffer_offset, static_cast<GLsizeiptr>(size)); + GPUVAddr gpu_addr, size_t size, BindlessSSBO* ssbo) { + const size_t alignment{device.GetShaderStorageBufferAlignment()}; + const auto info = buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.is_written); + if (device.UseAssemblyShaders()) { + *ssbo = BindlessSSBO{ + .address = static_cast<GLuint64EXT>(info.address + info.offset), + .length = static_cast<GLsizei>(size), + .padding = 0, + }; + } else { + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, binding, info.handle, info.offset, + static_cast<GLsizeiptr>(size)); + } } -void RasterizerOpenGL::SetupDrawTextures(std::size_t stage_index, const Shader& shader) { +void RasterizerOpenGL::SetupDrawTextures(std::size_t stage_index, Shader* shader) { MICROPROFILE_SCOPE(OpenGL_Texture); - const auto& maxwell3d = system.GPU().Maxwell3D(); u32 binding = device.GetBaseBindings(stage_index).sampler; for (const auto& entry : shader->GetEntries().samplers) { const auto shader_type = static_cast<ShaderType>(stage_index); - for (std::size_t i = 0; i < entry.Size(); ++i) { + for (std::size_t i = 0; i < entry.size; ++i) { const auto texture = GetTextureInfo(maxwell3d, entry, shader_type, i); SetupTexture(binding++, texture, entry); } } } -void RasterizerOpenGL::SetupComputeTextures(const Shader& kernel) { +void RasterizerOpenGL::SetupComputeTextures(Shader* kernel) { MICROPROFILE_SCOPE(OpenGL_Texture); - const auto& compute = system.GPU().KeplerCompute(); u32 binding = 0; for (const auto& entry : kernel->GetEntries().samplers) { - for (std::size_t i = 0; i < entry.Size(); ++i) { - const auto texture = GetTextureInfo(compute, entry, ShaderType::Compute, i); + for (std::size_t i = 0; i < entry.size; ++i) { + const auto texture = GetTextureInfo(kepler_compute, entry, ShaderType::Compute, i); SetupTexture(binding++, texture, entry); } } @@ -856,33 +1088,27 @@ void RasterizerOpenGL::SetupTexture(u32 binding, const Tegra::Texture::FullTextu glBindTextureUnit(binding, 0); return; } - glBindTextureUnit(binding, view->GetTexture()); - - if (view->GetSurfaceParams().IsBuffer()) { - return; + const GLuint handle = view->GetTexture(texture.tic.x_source, texture.tic.y_source, + texture.tic.z_source, texture.tic.w_source); + glBindTextureUnit(binding, handle); + if (!view->GetSurfaceParams().IsBuffer()) { + glBindSampler(binding, sampler_cache.GetSampler(texture.tsc)); } - // Apply swizzle to textures that are not buffers. - view->ApplySwizzle(texture.tic.x_source, texture.tic.y_source, texture.tic.z_source, - texture.tic.w_source); - - glBindSampler(binding, sampler_cache.GetSampler(texture.tsc)); } -void RasterizerOpenGL::SetupDrawImages(std::size_t stage_index, const Shader& shader) { - const auto& maxwell3d = system.GPU().Maxwell3D(); +void RasterizerOpenGL::SetupDrawImages(std::size_t stage_index, Shader* shader) { u32 binding = device.GetBaseBindings(stage_index).image; for (const auto& entry : shader->GetEntries().images) { - const auto shader_type = static_cast<Tegra::Engines::ShaderType>(stage_index); + const auto shader_type = static_cast<ShaderType>(stage_index); const auto tic = GetTextureInfo(maxwell3d, entry, shader_type).tic; SetupImage(binding++, tic, entry); } } -void RasterizerOpenGL::SetupComputeImages(const Shader& shader) { - const auto& compute = system.GPU().KeplerCompute(); +void RasterizerOpenGL::SetupComputeImages(Shader* shader) { u32 binding = 0; for (const auto& entry : shader->GetEntries().images) { - const auto tic = GetTextureInfo(compute, entry, Tegra::Engines::ShaderType::Compute).tic; + const auto tic = GetTextureInfo(kepler_compute, entry, ShaderType::Compute).tic; SetupImage(binding++, tic, entry); } } @@ -894,27 +1120,43 @@ void RasterizerOpenGL::SetupImage(u32 binding, const Tegra::Texture::TICEntry& t glBindImageTexture(binding, 0, 0, GL_FALSE, 0, GL_READ_ONLY, GL_R8); return; } - if (!tic.IsBuffer()) { - view->ApplySwizzle(tic.x_source, tic.y_source, tic.z_source, tic.w_source); - } - if (entry.IsWritten()) { + if (entry.is_written) { view->MarkAsModified(texture_cache.Tick()); } - glBindImageTexture(binding, view->GetTexture(), 0, GL_TRUE, 0, GL_READ_WRITE, - view->GetFormat()); + const GLuint handle = view->GetTexture(tic.x_source, tic.y_source, tic.z_source, tic.w_source); + glBindImageTexture(binding, handle, 0, GL_TRUE, 0, GL_READ_WRITE, view->GetFormat()); } void RasterizerOpenGL::SyncViewport() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; - const auto& regs = gpu.regs; + auto& flags = maxwell3d.dirty.flags; + const auto& regs = maxwell3d.regs; const bool dirty_viewport = flags[Dirty::Viewports]; + const bool dirty_clip_control = flags[Dirty::ClipControl]; + + if (dirty_clip_control || flags[Dirty::FrontFace]) { + flags[Dirty::FrontFace] = false; + + GLenum mode = MaxwellToGL::FrontFace(regs.front_face); + if (regs.screen_y_control.triangle_rast_flip != 0 && + regs.viewport_transform[0].scale_y < 0.0f) { + switch (mode) { + case GL_CW: + mode = GL_CCW; + break; + case GL_CCW: + mode = GL_CW; + break; + } + } + glFrontFace(mode); + } + if (dirty_viewport || flags[Dirty::ClipControl]) { flags[Dirty::ClipControl] = false; bool flip_y = false; - if (regs.viewport_transform[0].scale_y < 0.0) { + if (regs.viewport_transform[0].scale_y < 0.0f) { flip_y = !flip_y; } if (regs.screen_y_control.y_negate != 0) { @@ -946,34 +1188,36 @@ void RasterizerOpenGL::SyncViewport() { const GLdouble near_depth = src.translate_z - src.scale_z * reduce_z; const GLdouble far_depth = src.translate_z + src.scale_z; glDepthRangeIndexed(static_cast<GLuint>(i), near_depth, far_depth); + + if (!GLAD_GL_NV_viewport_swizzle) { + continue; + } + glViewportSwizzleNV(static_cast<GLuint>(i), MaxwellToGL::ViewportSwizzle(src.swizzle.x), + MaxwellToGL::ViewportSwizzle(src.swizzle.y), + MaxwellToGL::ViewportSwizzle(src.swizzle.z), + MaxwellToGL::ViewportSwizzle(src.swizzle.w)); } } } void RasterizerOpenGL::SyncDepthClamp() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::DepthClampEnabled]) { return; } flags[Dirty::DepthClampEnabled] = false; - const auto& state = gpu.regs.view_volume_clip_control; - UNIMPLEMENTED_IF_MSG(state.depth_clamp_far != state.depth_clamp_near, - "Unimplemented depth clamp separation!"); - - oglEnable(GL_DEPTH_CLAMP, state.depth_clamp_far || state.depth_clamp_near); + oglEnable(GL_DEPTH_CLAMP, maxwell3d.regs.view_volume_clip_control.depth_clamp_disabled == 0); } void RasterizerOpenGL::SyncClipEnabled(u32 clip_mask) { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::ClipDistances] && !flags[Dirty::Shaders]) { return; } flags[Dirty::ClipDistances] = false; - clip_mask &= gpu.regs.clip_distance_enabled; + clip_mask &= maxwell3d.regs.clip_distance_enabled; if (clip_mask == last_clip_distance_mask) { return; } @@ -989,9 +1233,8 @@ void RasterizerOpenGL::SyncClipCoef() { } void RasterizerOpenGL::SyncCullMode() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; - const auto& regs = gpu.regs; + auto& flags = maxwell3d.dirty.flags; + const auto& regs = maxwell3d.regs; if (flags[Dirty::CullTest]) { flags[Dirty::CullTest] = false; @@ -1003,34 +1246,27 @@ void RasterizerOpenGL::SyncCullMode() { glDisable(GL_CULL_FACE); } } - - if (flags[Dirty::FrontFace]) { - flags[Dirty::FrontFace] = false; - glFrontFace(MaxwellToGL::FrontFace(regs.front_face)); - } } void RasterizerOpenGL::SyncPrimitiveRestart() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::PrimitiveRestart]) { return; } flags[Dirty::PrimitiveRestart] = false; - if (gpu.regs.primitive_restart.enabled) { + if (maxwell3d.regs.primitive_restart.enabled) { glEnable(GL_PRIMITIVE_RESTART); - glPrimitiveRestartIndex(gpu.regs.primitive_restart.index); + glPrimitiveRestartIndex(maxwell3d.regs.primitive_restart.index); } else { glDisable(GL_PRIMITIVE_RESTART); } } void RasterizerOpenGL::SyncDepthTestState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; + const auto& regs = maxwell3d.regs; - const auto& regs = gpu.regs; if (flags[Dirty::DepthMask]) { flags[Dirty::DepthMask] = false; glDepthMask(regs.depth_write_enabled ? GL_TRUE : GL_FALSE); @@ -1048,14 +1284,13 @@ void RasterizerOpenGL::SyncDepthTestState() { } void RasterizerOpenGL::SyncStencilTestState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::StencilTest]) { return; } flags[Dirty::StencilTest] = false; - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; oglEnable(GL_STENCIL_TEST, regs.stencil_enable); glStencilFuncSeparate(GL_FRONT, MaxwellToGL::ComparisonOp(regs.stencil_front_func_func), @@ -1080,25 +1315,24 @@ void RasterizerOpenGL::SyncStencilTestState() { } void RasterizerOpenGL::SyncRasterizeEnable() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::RasterizeEnable]) { return; } flags[Dirty::RasterizeEnable] = false; - oglEnable(GL_RASTERIZER_DISCARD, gpu.regs.rasterize_enable == 0); + oglEnable(GL_RASTERIZER_DISCARD, maxwell3d.regs.rasterize_enable == 0); } void RasterizerOpenGL::SyncPolygonModes() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::PolygonModes]) { return; } flags[Dirty::PolygonModes] = false; - if (gpu.regs.fill_rectangle) { + const auto& regs = maxwell3d.regs; + if (regs.fill_rectangle) { if (!GLAD_GL_NV_fill_rectangle) { LOG_ERROR(Render_OpenGL, "GL_NV_fill_rectangle used and not supported"); glPolygonMode(GL_FRONT_AND_BACK, GL_FILL); @@ -1111,27 +1345,26 @@ void RasterizerOpenGL::SyncPolygonModes() { return; } - if (gpu.regs.polygon_mode_front == gpu.regs.polygon_mode_back) { + if (regs.polygon_mode_front == regs.polygon_mode_back) { flags[Dirty::PolygonModeFront] = false; flags[Dirty::PolygonModeBack] = false; - glPolygonMode(GL_FRONT_AND_BACK, MaxwellToGL::PolygonMode(gpu.regs.polygon_mode_front)); + glPolygonMode(GL_FRONT_AND_BACK, MaxwellToGL::PolygonMode(regs.polygon_mode_front)); return; } if (flags[Dirty::PolygonModeFront]) { flags[Dirty::PolygonModeFront] = false; - glPolygonMode(GL_FRONT, MaxwellToGL::PolygonMode(gpu.regs.polygon_mode_front)); + glPolygonMode(GL_FRONT, MaxwellToGL::PolygonMode(regs.polygon_mode_front)); } if (flags[Dirty::PolygonModeBack]) { flags[Dirty::PolygonModeBack] = false; - glPolygonMode(GL_BACK, MaxwellToGL::PolygonMode(gpu.regs.polygon_mode_back)); + glPolygonMode(GL_BACK, MaxwellToGL::PolygonMode(regs.polygon_mode_back)); } } void RasterizerOpenGL::SyncColorMask() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::ColorMasks]) { return; } @@ -1140,7 +1373,7 @@ void RasterizerOpenGL::SyncColorMask() { const bool force = flags[Dirty::ColorMaskCommon]; flags[Dirty::ColorMaskCommon] = false; - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; if (regs.color_mask_common) { if (!force && !flags[Dirty::ColorMask0]) { return; @@ -1165,33 +1398,30 @@ void RasterizerOpenGL::SyncColorMask() { } void RasterizerOpenGL::SyncMultiSampleState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::MultisampleControl]) { return; } flags[Dirty::MultisampleControl] = false; - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; oglEnable(GL_SAMPLE_ALPHA_TO_COVERAGE, regs.multisample_control.alpha_to_coverage); oglEnable(GL_SAMPLE_ALPHA_TO_ONE, regs.multisample_control.alpha_to_one); } void RasterizerOpenGL::SyncFragmentColorClampState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::FragmentClampColor]) { return; } flags[Dirty::FragmentClampColor] = false; - glClampColor(GL_CLAMP_FRAGMENT_COLOR, gpu.regs.frag_color_clamp ? GL_TRUE : GL_FALSE); + glClampColor(GL_CLAMP_FRAGMENT_COLOR, maxwell3d.regs.frag_color_clamp ? GL_TRUE : GL_FALSE); } void RasterizerOpenGL::SyncBlendState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; - const auto& regs = gpu.regs; + auto& flags = maxwell3d.dirty.flags; + const auto& regs = maxwell3d.regs; if (flags[Dirty::BlendColor]) { flags[Dirty::BlendColor] = false; @@ -1248,14 +1478,13 @@ void RasterizerOpenGL::SyncBlendState() { } void RasterizerOpenGL::SyncLogicOpState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::LogicOp]) { return; } flags[Dirty::LogicOp] = false; - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; if (regs.logic_op.enable) { glEnable(GL_COLOR_LOGIC_OP); glLogicOp(MaxwellToGL::LogicOp(regs.logic_op.operation)); @@ -1265,14 +1494,13 @@ void RasterizerOpenGL::SyncLogicOpState() { } void RasterizerOpenGL::SyncScissorTest() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::Scissors]) { return; } flags[Dirty::Scissors] = false; - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; for (std::size_t index = 0; index < Maxwell::NumViewports; ++index) { if (!flags[Dirty::Scissor0 + index]) { continue; @@ -1291,16 +1519,15 @@ void RasterizerOpenGL::SyncScissorTest() { } void RasterizerOpenGL::SyncPointState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::PointSize]) { return; } flags[Dirty::PointSize] = false; - oglEnable(GL_POINT_SPRITE, gpu.regs.point_sprite_enable); + oglEnable(GL_POINT_SPRITE, maxwell3d.regs.point_sprite_enable); - if (gpu.regs.vp_point_size.enable) { + if (maxwell3d.regs.vp_point_size.enable) { // By definition of GL_POINT_SIZE, it only matters if GL_PROGRAM_POINT_SIZE is disabled. glEnable(GL_PROGRAM_POINT_SIZE); return; @@ -1308,32 +1535,30 @@ void RasterizerOpenGL::SyncPointState() { // Limit the point size to 1 since nouveau sometimes sets a point size of 0 (and that's invalid // in OpenGL). - glPointSize(std::max(1.0f, gpu.regs.point_size)); + glPointSize(std::max(1.0f, maxwell3d.regs.point_size)); glDisable(GL_PROGRAM_POINT_SIZE); } void RasterizerOpenGL::SyncLineState() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::LineWidth]) { return; } flags[Dirty::LineWidth] = false; - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; oglEnable(GL_LINE_SMOOTH, regs.line_smooth_enable); glLineWidth(regs.line_smooth_enable ? regs.line_width_smooth : regs.line_width_aliased); } void RasterizerOpenGL::SyncPolygonOffset() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::PolygonOffset]) { return; } flags[Dirty::PolygonOffset] = false; - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; oglEnable(GL_POLYGON_OFFSET_FILL, regs.polygon_offset_fill_enable); oglEnable(GL_POLYGON_OFFSET_LINE, regs.polygon_offset_line_enable); oglEnable(GL_POLYGON_OFFSET_POINT, regs.polygon_offset_point_enable); @@ -1347,18 +1572,13 @@ void RasterizerOpenGL::SyncPolygonOffset() { } void RasterizerOpenGL::SyncAlphaTest() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::AlphaTest]) { return; } flags[Dirty::AlphaTest] = false; - const auto& regs = gpu.regs; - if (regs.alpha_test_enabled && regs.rt_control.count > 1) { - LOG_WARNING(Render_OpenGL, "Alpha testing with more than one render target is not tested"); - } - + const auto& regs = maxwell3d.regs; if (regs.alpha_test_enabled) { glEnable(GL_ALPHA_TEST); glAlphaFunc(MaxwellToGL::ComparisonOp(regs.alpha_test_func), regs.alpha_test_ref); @@ -1368,22 +1588,79 @@ void RasterizerOpenGL::SyncAlphaTest() { } void RasterizerOpenGL::SyncFramebufferSRGB() { - auto& gpu = system.GPU().Maxwell3D(); - auto& flags = gpu.dirty.flags; + auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::FramebufferSRGB]) { return; } flags[Dirty::FramebufferSRGB] = false; - oglEnable(GL_FRAMEBUFFER_SRGB, gpu.regs.framebuffer_srgb); + oglEnable(GL_FRAMEBUFFER_SRGB, maxwell3d.regs.framebuffer_srgb); +} + +void RasterizerOpenGL::SyncTransformFeedback() { + // TODO(Rodrigo): Inject SKIP_COMPONENTS*_NV when required. An unimplemented message will signal + // when this is required. + const auto& regs = maxwell3d.regs; + + static constexpr std::size_t STRIDE = 3; + std::array<GLint, 128 * STRIDE * Maxwell::NumTransformFeedbackBuffers> attribs; + std::array<GLint, Maxwell::NumTransformFeedbackBuffers> streams; + + GLint* cursor = attribs.data(); + GLint* current_stream = streams.data(); + + for (std::size_t feedback = 0; feedback < Maxwell::NumTransformFeedbackBuffers; ++feedback) { + const auto& layout = regs.tfb_layouts[feedback]; + UNIMPLEMENTED_IF_MSG(layout.stride != layout.varying_count * 4, "Stride padding"); + if (layout.varying_count == 0) { + continue; + } + + *current_stream = static_cast<GLint>(feedback); + if (current_stream != streams.data()) { + // When stepping one stream, push the expected token + cursor[0] = GL_NEXT_BUFFER_NV; + cursor[1] = 0; + cursor[2] = 0; + cursor += STRIDE; + } + ++current_stream; + + const auto& locations = regs.tfb_varying_locs[feedback]; + std::optional<u8> current_index; + for (u32 offset = 0; offset < layout.varying_count; ++offset) { + const u8 location = locations[offset]; + const u8 index = location / 4; + + if (current_index == index) { + // Increase number of components of the previous attachment + ++cursor[-2]; + continue; + } + current_index = index; + + std::tie(cursor[0], cursor[2]) = TransformFeedbackEnum(location); + cursor[1] = 1; + cursor += STRIDE; + } + } + + const GLsizei num_attribs = static_cast<GLsizei>((cursor - attribs.data()) / STRIDE); + const GLsizei num_strides = static_cast<GLsizei>(current_stream - streams.data()); + glTransformFeedbackStreamAttribsNV(num_attribs, attribs.data(), num_strides, streams.data(), + GL_INTERLEAVED_ATTRIBS); } void RasterizerOpenGL::BeginTransformFeedback(GLenum primitive_mode) { - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; if (regs.tfb_enabled == 0) { return; } + if (device.UseAssemblyShaders()) { + SyncTransformFeedback(); + } + UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationControl) || regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationEval) || regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::Geometry)); @@ -1410,11 +1687,15 @@ void RasterizerOpenGL::BeginTransformFeedback(GLenum primitive_mode) { static_cast<GLsizeiptr>(size)); } + // We may have to call BeginTransformFeedbackNV here since they seem to call different + // implementations on Nvidia's driver (the pointer is different) but we are using + // ARB_transform_feedback3 features with NV_transform_feedback interactions and the ARB + // extension doesn't define BeginTransformFeedback (without NV) interactions. It just works. glBeginTransformFeedback(GL_POINTS); } void RasterizerOpenGL::EndTransformFeedback() { - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; if (regs.tfb_enabled == 0) { return; } @@ -1431,8 +1712,9 @@ void RasterizerOpenGL::EndTransformFeedback() { const GLuint handle = transform_feedback_buffers[index].handle; const GPUVAddr gpu_addr = binding.Address(); const std::size_t size = binding.buffer_size; - const auto [dest_buffer, offset] = buffer_cache.UploadMemory(gpu_addr, size, 4, true); - glCopyNamedBufferSubData(handle, *dest_buffer, 0, offset, static_cast<GLsizeiptr>(size)); + const auto info = buffer_cache.UploadMemory(gpu_addr, size, 4, true); + glCopyNamedBufferSubData(handle, info.handle, 0, info.offset, + static_cast<GLsizeiptr>(size)); } } diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 435da4425..1d0f585fa 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -19,10 +19,10 @@ #include "video_core/engines/const_buffer_info.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/rasterizer_accelerated.h" -#include "video_core/rasterizer_cache.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" #include "video_core/renderer_opengl/gl_device.h" +#include "video_core/renderer_opengl/gl_fence_manager.h" #include "video_core/renderer_opengl/gl_framebuffer_cache.h" #include "video_core/renderer_opengl/gl_query_cache.h" #include "video_core/renderer_opengl/gl_resource_manager.h" @@ -33,10 +33,11 @@ #include "video_core/renderer_opengl/gl_state_tracker.h" #include "video_core/renderer_opengl/gl_texture_cache.h" #include "video_core/renderer_opengl/utils.h" +#include "video_core/shader/async_shaders.h" #include "video_core/textures/texture.h" -namespace Core { -class System; +namespace Core::Memory { +class Memory; } namespace Core::Frontend { @@ -52,10 +53,18 @@ namespace OpenGL { struct ScreenInfo; struct DrawParameters; +struct BindlessSSBO { + GLuint64EXT address; + GLsizei length; + GLsizei padding; +}; +static_assert(sizeof(BindlessSSBO) * CHAR_BIT == 128); + class RasterizerOpenGL : public VideoCore::RasterizerAccelerated { public: - explicit RasterizerOpenGL(Core::System& system, Core::Frontend::EmuWindow& emu_window, - ScreenInfo& info, GLShader::ProgramManager& program_manager, + explicit RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu, + Core::Memory::Memory& cpu_memory, const Device& device, + ScreenInfo& screen_info, ProgramManager& program_manager, StateTracker& state_tracker); ~RasterizerOpenGL() override; @@ -66,8 +75,15 @@ public: void Query(GPUVAddr gpu_addr, VideoCore::QueryType type, std::optional<u64> timestamp) override; void FlushAll() override; void FlushRegion(VAddr addr, u64 size) override; + bool MustFlushRegion(VAddr addr, u64 size) override; void InvalidateRegion(VAddr addr, u64 size) override; + void OnCPUWrite(VAddr addr, u64 size) override; + void SyncGuestHost() override; + void SignalSemaphore(GPUVAddr addr, u32 value) override; + void SignalSyncPoint(u32 value) override; + void ReleaseFences() override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; + void WaitForIdle() override; void FlushCommands() override; void TickFrame() override; bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, @@ -75,56 +91,65 @@ public: const Tegra::Engines::Fermi2D::Config& copy_config) override; bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, u32 pixel_stride) override; - void LoadDiskResources(const std::atomic_bool& stop_loading, + void LoadDiskResources(u64 title_id, const std::atomic_bool& stop_loading, const VideoCore::DiskResourceLoadCallback& callback) override; - void SetupDirtyFlags() override; /// Returns true when there are commands queued to the OpenGL server. bool AnyCommandQueued() const { return num_queued_commands > 0; } + VideoCommon::Shader::AsyncShaders& GetAsyncShaders() { + return async_shaders; + } + + const VideoCommon::Shader::AsyncShaders& GetAsyncShaders() const { + return async_shaders; + } + private: /// Configures the color and depth framebuffer states. void ConfigureFramebuffers(); - void ConfigureClearFramebuffer(bool using_color_fb, bool using_depth_fb, bool using_stencil_fb); + /// Configures the color and depth framebuffer for clearing. + void ConfigureClearFramebuffer(bool using_color, bool using_depth_stencil); /// Configures the current constbuffers to use for the draw command. - void SetupDrawConstBuffers(std::size_t stage_index, const Shader& shader); + void SetupDrawConstBuffers(std::size_t stage_index, Shader* shader); /// Configures the current constbuffers to use for the kernel invocation. - void SetupComputeConstBuffers(const Shader& kernel); + void SetupComputeConstBuffers(Shader* kernel); /// Configures a constant buffer. - void SetupConstBuffer(u32 binding, const Tegra::Engines::ConstBufferInfo& buffer, - const ConstBufferEntry& entry); + void SetupConstBuffer(GLenum stage, u32 binding, const Tegra::Engines::ConstBufferInfo& buffer, + const ConstBufferEntry& entry, bool use_unified, + std::size_t unified_offset); /// Configures the current global memory entries to use for the draw command. - void SetupDrawGlobalMemory(std::size_t stage_index, const Shader& shader); + void SetupDrawGlobalMemory(std::size_t stage_index, Shader* shader); /// Configures the current global memory entries to use for the kernel invocation. - void SetupComputeGlobalMemory(const Shader& kernel); + void SetupComputeGlobalMemory(Shader* kernel); - /// Configures a constant buffer. + /// Configures a global memory buffer. void SetupGlobalMemory(u32 binding, const GlobalMemoryEntry& entry, GPUVAddr gpu_addr, - std::size_t size); + size_t size, BindlessSSBO* ssbo); /// Configures the current textures to use for the draw command. - void SetupDrawTextures(std::size_t stage_index, const Shader& shader); + void SetupDrawTextures(std::size_t stage_index, Shader* shader); /// Configures the textures used in a compute shader. - void SetupComputeTextures(const Shader& kernel); + void SetupComputeTextures(Shader* kernel); /// Configures a texture. void SetupTexture(u32 binding, const Tegra::Texture::FullTextureInfo& texture, const SamplerEntry& entry); /// Configures images in a graphics shader. - void SetupDrawImages(std::size_t stage_index, const Shader& shader); + void SetupDrawImages(std::size_t stage_index, Shader* shader); /// Configures images in a compute shader. - void SetupComputeImages(const Shader& shader); + void SetupComputeImages(Shader* shader); /// Configures an image. void SetupImage(u32 binding, const Tegra::Texture::TICEntry& tic, const ImageEntry& entry); @@ -192,6 +217,10 @@ private: /// Syncs the framebuffer sRGB state to match the guest state void SyncFramebufferSRGB(); + /// Syncs transform feedback state to match guest state + /// @note Only valid on assembly shaders + void SyncTransformFeedback(); + /// Begin a transform feedback void BeginTransformFeedback(GLenum primitive_mode); @@ -215,31 +244,42 @@ private: void SetupShaders(GLenum primitive_mode); - const Device device; + Tegra::GPU& gpu; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::Engines::KeplerCompute& kepler_compute; + Tegra::MemoryManager& gpu_memory; + + const Device& device; + ScreenInfo& screen_info; + ProgramManager& program_manager; + StateTracker& state_tracker; TextureCacheOpenGL texture_cache; ShaderCacheOpenGL shader_cache; SamplerCacheOpenGL sampler_cache; FramebufferCacheOpenGL framebuffer_cache; QueryCache query_cache; + OGLBufferCache buffer_cache; + FenceManagerOpenGL fence_manager; - Core::System& system; - ScreenInfo& screen_info; - GLShader::ProgramManager& program_manager; - StateTracker& state_tracker; + VideoCommon::Shader::AsyncShaders async_shaders; static constexpr std::size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024; - OGLBufferCache buffer_cache; - VertexArrayPushBuffer vertex_array_pushbuffer{state_tracker}; - BindBuffersRangePushBuffer bind_ubo_pushbuffer{GL_UNIFORM_BUFFER}; - BindBuffersRangePushBuffer bind_ssbo_pushbuffer{GL_SHADER_STORAGE_BUFFER}; + GLint vertex_binding = 0; std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::NumTransformFeedbackBuffers> transform_feedback_buffers; std::bitset<Tegra::Engines::Maxwell3D::Regs::NumTransformFeedbackBuffers> enabled_transform_feedback_buffers; + static constexpr std::size_t NUM_CONSTANT_BUFFERS = + Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers * + Tegra::Engines::Maxwell3D::Regs::MaxShaderProgram; + std::array<GLuint, NUM_CONSTANT_BUFFERS> staging_cbufs{}; + std::size_t current_cbuf = 0; + OGLBuffer unified_uniform_buffer; + /// Number of commands queued to the OpenGL driver. Reseted on flush. std::size_t num_queued_commands = 0; diff --git a/src/video_core/renderer_opengl/gl_resource_manager.cpp b/src/video_core/renderer_opengl/gl_resource_manager.cpp index 97803d480..0ebcec427 100644 --- a/src/video_core/renderer_opengl/gl_resource_manager.cpp +++ b/src/video_core/renderer_opengl/gl_resource_manager.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <string_view> #include <utility> #include <glad/glad.h> #include "common/common_types.h" @@ -82,11 +83,13 @@ void OGLSampler::Release() { handle = 0; } -void OGLShader::Create(const char* source, GLenum type) { - if (handle != 0) +void OGLShader::Create(std::string_view source, GLenum type) { + if (handle != 0) { return; - if (source == nullptr) + } + if (source.empty()) { return; + } MICROPROFILE_SCOPE(OpenGL_ResourceCreation); handle = GLShader::LoadShader(source, type); @@ -125,6 +128,15 @@ void OGLProgram::Release() { handle = 0; } +void OGLAssemblyProgram::Release() { + if (handle == 0) { + return; + } + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteProgramsARB(1, &handle); + handle = 0; +} + void OGLPipeline::Create() { if (handle != 0) return; diff --git a/src/video_core/renderer_opengl/gl_resource_manager.h b/src/video_core/renderer_opengl/gl_resource_manager.h index de93f4212..f48398669 100644 --- a/src/video_core/renderer_opengl/gl_resource_manager.h +++ b/src/video_core/renderer_opengl/gl_resource_manager.h @@ -4,6 +4,7 @@ #pragma once +#include <string_view> #include <utility> #include <glad/glad.h> #include "common/common_types.h" @@ -127,7 +128,7 @@ public: return *this; } - void Create(const char* source, GLenum type); + void Create(std::string_view source, GLenum type); void Release(); @@ -167,6 +168,28 @@ public: GLuint handle = 0; }; +class OGLAssemblyProgram : private NonCopyable { +public: + OGLAssemblyProgram() = default; + + OGLAssemblyProgram(OGLAssemblyProgram&& o) noexcept : handle(std::exchange(o.handle, 0)) {} + + ~OGLAssemblyProgram() { + Release(); + } + + OGLAssemblyProgram& operator=(OGLAssemblyProgram&& o) noexcept { + Release(); + handle = std::exchange(o.handle, 0); + return *this; + } + + /// Deletes the internal OpenGL resource + void Release(); + + GLuint handle = 0; +}; + class OGLPipeline : private NonCopyable { public: OGLPipeline() = default; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 12c6dcfde..bd56bed0c 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -10,8 +10,6 @@ #include <thread> #include <unordered_set> -#include <boost/functional/hash.hpp> - #include "common/alignment.h" #include "common/assert.h" #include "common/logging/log.h" @@ -22,83 +20,35 @@ #include "video_core/engines/maxwell_3d.h" #include "video_core/engines/shader_type.h" #include "video_core/memory_manager.h" +#include "video_core/renderer_opengl/gl_arb_decompiler.h" #include "video_core/renderer_opengl/gl_rasterizer.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_shader_cache.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" #include "video_core/renderer_opengl/gl_shader_disk_cache.h" #include "video_core/renderer_opengl/gl_state_tracker.h" #include "video_core/renderer_opengl/utils.h" +#include "video_core/shader/memory_util.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" +#include "video_core/shader_cache.h" +#include "video_core/shader_notify.h" namespace OpenGL { using Tegra::Engines::ShaderType; -using VideoCommon::Shader::CompileDepth; -using VideoCommon::Shader::CompilerSettings; +using VideoCommon::Shader::GetShaderAddress; +using VideoCommon::Shader::GetShaderCode; +using VideoCommon::Shader::GetUniqueIdentifier; +using VideoCommon::Shader::KERNEL_MAIN_OFFSET; using VideoCommon::Shader::ProgramCode; using VideoCommon::Shader::Registry; using VideoCommon::Shader::ShaderIR; +using VideoCommon::Shader::STAGE_MAIN_OFFSET; namespace { -constexpr u32 STAGE_MAIN_OFFSET = 10; -constexpr u32 KERNEL_MAIN_OFFSET = 0; - -constexpr CompilerSettings COMPILER_SETTINGS{CompileDepth::FullDecompile}; - -/// Gets the address for the specified shader stage program -GPUVAddr GetShaderAddress(Core::System& system, Maxwell::ShaderProgram program) { - const auto& gpu{system.GPU().Maxwell3D()}; - const auto& shader_config{gpu.regs.shader_config[static_cast<std::size_t>(program)]}; - return gpu.regs.code_address.CodeAddress() + shader_config.offset; -} - -/// Gets if the current instruction offset is a scheduler instruction -constexpr bool IsSchedInstruction(std::size_t offset, std::size_t main_offset) { - // Sched instructions appear once every 4 instructions. - constexpr std::size_t SchedPeriod = 4; - const std::size_t absolute_offset = offset - main_offset; - return (absolute_offset % SchedPeriod) == 0; -} - -/// Calculates the size of a program stream -std::size_t CalculateProgramSize(const ProgramCode& program) { - constexpr std::size_t start_offset = 10; - // This is the encoded version of BRA that jumps to itself. All Nvidia - // shaders end with one. - constexpr u64 self_jumping_branch = 0xE2400FFFFF07000FULL; - constexpr u64 mask = 0xFFFFFFFFFF7FFFFFULL; - std::size_t offset = start_offset; - while (offset < program.size()) { - const u64 instruction = program[offset]; - if (!IsSchedInstruction(offset, start_offset)) { - if ((instruction & mask) == self_jumping_branch) { - // End on Maxwell's "nop" instruction - break; - } - if (instruction == 0) { - break; - } - } - offset++; - } - // The last instruction is included in the program size - return std::min(offset + 1, program.size()); -} - -/// Gets the shader program code from memory for the specified address -ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, const GPUVAddr gpu_addr, - const u8* host_ptr) { - ProgramCode code(VideoCommon::Shader::MAX_PROGRAM_LENGTH); - ASSERT_OR_EXECUTE(host_ptr != nullptr, { - std::fill(code.begin(), code.end(), 0); - return code; - }); - memory_manager.ReadBlockUnsafe(gpu_addr, code.data(), code.size() * sizeof(u64)); - code.resize(CalculateProgramSize(code)); - return code; -} +constexpr VideoCommon::Shader::CompilerSettings COMPILER_SETTINGS{}; /// Gets the shader type from a Maxwell program type constexpr GLenum GetGLShaderType(ShaderType shader_type) { @@ -116,17 +66,6 @@ constexpr GLenum GetGLShaderType(ShaderType shader_type) { } } -/// Hashes one (or two) program streams -u64 GetUniqueIdentifier(ShaderType shader_type, bool is_a, const ProgramCode& code, - const ProgramCode& code_b = {}) { - u64 unique_identifier = boost::hash_value(code); - if (is_a) { - // VertexA programs include two programs - boost::hash_combine(unique_identifier, boost::hash_value(code_b)); - } - return unique_identifier; -} - constexpr const char* GetShaderTypeName(ShaderType shader_type) { switch (shader_type) { case ShaderType::Vertex: @@ -162,6 +101,24 @@ constexpr ShaderType GetShaderType(Maxwell::ShaderProgram program_type) { return {}; } +constexpr GLenum AssemblyEnum(ShaderType shader_type) { + switch (shader_type) { + case ShaderType::Vertex: + return GL_VERTEX_PROGRAM_NV; + case ShaderType::TesselationControl: + return GL_TESS_CONTROL_PROGRAM_NV; + case ShaderType::TesselationEval: + return GL_TESS_EVALUATION_PROGRAM_NV; + case ShaderType::Geometry: + return GL_GEOMETRY_PROGRAM_NV; + case ShaderType::Fragment: + return GL_FRAGMENT_PROGRAM_NV; + case ShaderType::Compute: + return GL_COMPUTE_PROGRAM_NV; + } + return {}; +} + std::string MakeShaderID(u64 unique_identifier, ShaderType shader_type) { return fmt::format("{}{:016X}", GetShaderTypeName(shader_type), unique_identifier); } @@ -170,7 +127,7 @@ std::shared_ptr<Registry> MakeRegistry(const ShaderDiskCacheEntry& entry) { const VideoCore::GuestDriverProfile guest_profile{entry.texture_handler_size}; const VideoCommon::Shader::SerializedRegistryInfo info{guest_profile, entry.bound_buffer, entry.graphics_info, entry.compute_info}; - const auto registry = std::make_shared<Registry>(entry.type, info); + auto registry = std::make_shared<Registry>(entry.type, info); for (const auto& [address, value] : entry.keys) { const auto [buffer, offset] = address; registry->InsertKey(buffer, offset, value); @@ -185,21 +142,6 @@ std::shared_ptr<Registry> MakeRegistry(const ShaderDiskCacheEntry& entry) { return registry; } -std::shared_ptr<OGLProgram> BuildShader(const Device& device, ShaderType shader_type, - u64 unique_identifier, const ShaderIR& ir, - const Registry& registry, bool hint_retrievable = false) { - const std::string shader_id = MakeShaderID(unique_identifier, shader_type); - LOG_INFO(Render_OpenGL, "{}", shader_id); - - const std::string glsl = DecompileShader(device, ir, registry, shader_type, shader_id); - OGLShader shader; - shader.Create(glsl.c_str(), GetGLShaderType(shader_type)); - - auto program = std::make_shared<OGLProgram>(); - program->Create(true, hint_retrievable, shader.handle); - return program; -} - std::unordered_set<GLenum> GetSupportedFormats() { GLint num_formats; glGetIntegerv(GL_NUM_PROGRAM_BINARY_FORMATS, &num_formats); @@ -216,55 +158,138 @@ std::unordered_set<GLenum> GetSupportedFormats() { } // Anonymous namespace -CachedShader::CachedShader(VAddr cpu_addr, std::size_t size_in_bytes, - std::shared_ptr<VideoCommon::Shader::Registry> registry, - ShaderEntries entries, std::shared_ptr<OGLProgram> program) - : RasterizerCacheObject{cpu_addr}, registry{std::move(registry)}, entries{std::move(entries)}, - size_in_bytes{size_in_bytes}, program{std::move(program)} {} +ProgramSharedPtr BuildShader(const Device& device, ShaderType shader_type, u64 unique_identifier, + const ShaderIR& ir, const Registry& registry, bool hint_retrievable) { + const std::string shader_id = MakeShaderID(unique_identifier, shader_type); + LOG_INFO(Render_OpenGL, "{}", shader_id); + + auto program = std::make_shared<ProgramHandle>(); + + if (device.UseAssemblyShaders()) { + const std::string arb = + DecompileAssemblyShader(device, ir, registry, shader_type, shader_id); + + GLuint& arb_prog = program->assembly_program.handle; + +// Commented out functions signal OpenGL errors but are compatible with apitrace. +// Use them only to capture and replay on apitrace. +#if 0 + glGenProgramsNV(1, &arb_prog); + glLoadProgramNV(AssemblyEnum(shader_type), arb_prog, static_cast<GLsizei>(arb.size()), + reinterpret_cast<const GLubyte*>(arb.data())); +#else + glGenProgramsARB(1, &arb_prog); + glNamedProgramStringEXT(arb_prog, AssemblyEnum(shader_type), GL_PROGRAM_FORMAT_ASCII_ARB, + static_cast<GLsizei>(arb.size()), arb.data()); +#endif + const auto err = reinterpret_cast<const char*>(glGetString(GL_PROGRAM_ERROR_STRING_NV)); + if (err && *err) { + LOG_CRITICAL(Render_OpenGL, "{}", err); + LOG_INFO(Render_OpenGL, "\n{}", arb); + } + } else { + const std::string glsl = DecompileShader(device, ir, registry, shader_type, shader_id); + OGLShader shader; + shader.Create(glsl.c_str(), GetGLShaderType(shader_type)); + + program->source_program.Create(true, hint_retrievable, shader.handle); + } -CachedShader::~CachedShader() = default; + return program; +} -GLuint CachedShader::GetHandle() const { +Shader::Shader(std::shared_ptr<VideoCommon::Shader::Registry> registry_, ShaderEntries entries_, + ProgramSharedPtr program_, bool is_built) + : registry{std::move(registry_)}, entries{std::move(entries_)}, program{std::move(program_)}, + is_built(is_built) { + handle = program->assembly_program.handle; + if (handle == 0) { + handle = program->source_program.handle; + } + if (is_built) { + ASSERT(handle != 0); + } +} + +Shader::~Shader() = default; + +GLuint Shader::GetHandle() const { DEBUG_ASSERT(registry->IsConsistent()); - return program->handle; + return handle; } -Shader CachedShader::CreateStageFromMemory(const ShaderParameters& params, - Maxwell::ShaderProgram program_type, ProgramCode code, - ProgramCode code_b) { +bool Shader::IsBuilt() const { + return is_built; +} + +void Shader::AsyncOpenGLBuilt(OGLProgram new_program) { + program->source_program = std::move(new_program); + handle = program->source_program.handle; + is_built = true; +} + +void Shader::AsyncGLASMBuilt(OGLAssemblyProgram new_program) { + program->assembly_program = std::move(new_program); + handle = program->assembly_program.handle; + is_built = true; +} + +std::unique_ptr<Shader> Shader::CreateStageFromMemory( + const ShaderParameters& params, Maxwell::ShaderProgram program_type, ProgramCode code, + ProgramCode code_b, VideoCommon::Shader::AsyncShaders& async_shaders, VAddr cpu_addr) { const auto shader_type = GetShaderType(program_type); - const std::size_t size_in_bytes = code.size() * sizeof(u64); - auto registry = std::make_shared<Registry>(shader_type, params.system.GPU().Maxwell3D()); - const ShaderIR ir(code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, *registry); - // TODO(Rodrigo): Handle VertexA shaders - // std::optional<ShaderIR> ir_b; - // if (!code_b.empty()) { - // ir_b.emplace(code_b, STAGE_MAIN_OFFSET); - // } - auto program = BuildShader(params.device, shader_type, params.unique_identifier, ir, *registry); + auto& gpu = params.gpu; + gpu.ShaderNotify().MarkSharderBuilding(); + + auto registry = std::make_shared<Registry>(shader_type, gpu.Maxwell3D()); + if (!async_shaders.IsShaderAsync(gpu) || !params.device.UseAsynchronousShaders()) { + const ShaderIR ir(code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, *registry); + // TODO(Rodrigo): Handle VertexA shaders + // std::optional<ShaderIR> ir_b; + // if (!code_b.empty()) { + // ir_b.emplace(code_b, STAGE_MAIN_OFFSET); + // } + auto program = + BuildShader(params.device, shader_type, params.unique_identifier, ir, *registry); + ShaderDiskCacheEntry entry; + entry.type = shader_type; + entry.code = std::move(code); + entry.code_b = std::move(code_b); + entry.unique_identifier = params.unique_identifier; + entry.bound_buffer = registry->GetBoundBuffer(); + entry.graphics_info = registry->GetGraphicsInfo(); + entry.keys = registry->GetKeys(); + entry.bound_samplers = registry->GetBoundSamplers(); + entry.bindless_samplers = registry->GetBindlessSamplers(); + params.disk_cache.SaveEntry(std::move(entry)); + + gpu.ShaderNotify().MarkShaderComplete(); + + return std::unique_ptr<Shader>(new Shader(std::move(registry), + MakeEntries(params.device, ir, shader_type), + std::move(program), true)); + } else { + // Required for entries + const ShaderIR ir(code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, *registry); + auto entries = MakeEntries(params.device, ir, shader_type); - ShaderDiskCacheEntry entry; - entry.type = shader_type; - entry.code = std::move(code); - entry.code_b = std::move(code_b); - entry.unique_identifier = params.unique_identifier; - entry.bound_buffer = registry->GetBoundBuffer(); - entry.graphics_info = registry->GetGraphicsInfo(); - entry.keys = registry->GetKeys(); - entry.bound_samplers = registry->GetBoundSamplers(); - entry.bindless_samplers = registry->GetBindlessSamplers(); - params.disk_cache.SaveEntry(std::move(entry)); + async_shaders.QueueOpenGLShader(params.device, shader_type, params.unique_identifier, + std::move(code), std::move(code_b), STAGE_MAIN_OFFSET, + COMPILER_SETTINGS, *registry, cpu_addr); - return std::shared_ptr<CachedShader>(new CachedShader( - params.cpu_addr, size_in_bytes, std::move(registry), MakeEntries(ir), std::move(program))); + auto program = std::make_shared<ProgramHandle>(); + return std::unique_ptr<Shader>( + new Shader(std::move(registry), std::move(entries), std::move(program), false)); + } } -Shader CachedShader::CreateKernelFromMemory(const ShaderParameters& params, ProgramCode code) { - const std::size_t size_in_bytes = code.size() * sizeof(u64); +std::unique_ptr<Shader> Shader::CreateKernelFromMemory(const ShaderParameters& params, + ProgramCode code) { + auto& gpu = params.gpu; + gpu.ShaderNotify().MarkSharderBuilding(); - auto& engine = params.system.GPU().KeplerCompute(); - auto registry = std::make_shared<Registry>(ShaderType::Compute, engine); + auto registry = std::make_shared<Registry>(ShaderType::Compute, params.engine); const ShaderIR ir(code, KERNEL_MAIN_OFFSET, COMPILER_SETTINGS, *registry); const u64 uid = params.unique_identifier; auto program = BuildShader(params.device, ShaderType::Compute, uid, ir, *registry); @@ -280,31 +305,43 @@ Shader CachedShader::CreateKernelFromMemory(const ShaderParameters& params, Prog entry.bindless_samplers = registry->GetBindlessSamplers(); params.disk_cache.SaveEntry(std::move(entry)); - return std::shared_ptr<CachedShader>(new CachedShader( - params.cpu_addr, size_in_bytes, std::move(registry), MakeEntries(ir), std::move(program))); + gpu.ShaderNotify().MarkShaderComplete(); + + return std::unique_ptr<Shader>(new Shader(std::move(registry), + MakeEntries(params.device, ir, ShaderType::Compute), + std::move(program))); } -Shader CachedShader::CreateFromCache(const ShaderParameters& params, - const PrecompiledShader& precompiled_shader, - std::size_t size_in_bytes) { - return std::shared_ptr<CachedShader>( - new CachedShader(params.cpu_addr, size_in_bytes, precompiled_shader.registry, - precompiled_shader.entries, precompiled_shader.program)); +std::unique_ptr<Shader> Shader::CreateFromCache(const ShaderParameters& params, + const PrecompiledShader& precompiled_shader) { + return std::unique_ptr<Shader>(new Shader( + precompiled_shader.registry, precompiled_shader.entries, precompiled_shader.program)); } -ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer, Core::System& system, - Core::Frontend::EmuWindow& emu_window, const Device& device) - : RasterizerCache{rasterizer}, system{system}, emu_window{emu_window}, device{device}, - disk_cache{system} {} +ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer, + Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_, + Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::Engines::KeplerCompute& kepler_compute_, + Tegra::MemoryManager& gpu_memory_, const Device& device_) + : VideoCommon::ShaderCache<Shader>{rasterizer}, emu_window{emu_window_}, gpu{gpu_}, + gpu_memory{gpu_memory_}, maxwell3d{maxwell3d_}, + kepler_compute{kepler_compute_}, device{device_} {} -void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, +ShaderCacheOpenGL::~ShaderCacheOpenGL() = default; + +void ShaderCacheOpenGL::LoadDiskCache(u64 title_id, const std::atomic_bool& stop_loading, const VideoCore::DiskResourceLoadCallback& callback) { + disk_cache.BindTitleID(title_id); const std::optional transferable = disk_cache.LoadTransferable(); if (!transferable) { return; } - const std::vector gl_cache = disk_cache.LoadPrecompiled(); + std::vector<ShaderDiskCachePrecompiled> gl_cache; + if (!device.UseAssemblyShaders()) { + // Only load precompiled cache when we are not using assembly shaders + gl_cache = disk_cache.LoadPrecompiled(); + } const auto supported_formats = GetSupportedFormats(); // Track if precompiled cache was altered during loading to know if we have to @@ -343,7 +380,7 @@ void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, auto registry = MakeRegistry(entry); const ShaderIR ir(entry.code, main_offset, COMPILER_SETTINGS, *registry); - std::shared_ptr<OGLProgram> program; + ProgramSharedPtr program; if (precompiled_entry) { // If the shader is precompiled, attempt to load it with program = GeneratePrecompiledProgram(entry, *precompiled_entry, supported_formats); @@ -359,7 +396,7 @@ void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, PrecompiledShader shader; shader.program = std::move(program); shader.registry = std::move(registry); - shader.entries = MakeEntries(ir); + shader.entries = MakeEntries(device, ir, entry.type); std::scoped_lock lock{mutex}; if (callback) { @@ -370,7 +407,7 @@ void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, } }; - const auto num_workers{static_cast<std::size_t>(std::thread::hardware_concurrency() + 1ULL)}; + const std::size_t num_workers{std::max(1U, std::thread::hardware_concurrency())}; const std::size_t bucket_size{transferable->size() / num_workers}; std::vector<std::unique_ptr<Core::Frontend::GraphicsContext>> contexts(num_workers); std::vector<std::thread> threads(num_workers); @@ -397,6 +434,11 @@ void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, return; } + if (device.UseAssemblyShaders()) { + // Don't store precompiled binaries for assembly shaders. + return; + } + // TODO(Rodrigo): Do state tracking for transferable shaders and do a dummy draw // before precompiling them @@ -404,7 +446,7 @@ void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, const u64 id = (*transferable)[i].unique_identifier; const auto it = find_precompiled(id); if (it == gl_cache.end()) { - const GLuint program = runtime_cache.at(id).program->handle; + const GLuint program = runtime_cache.at(id).program->source_program.handle; disk_cache.SavePrecompiled(id, program); precompiled_cache_altered = true; } @@ -415,7 +457,7 @@ void ShaderCacheOpenGL::LoadDiskCache(const std::atomic_bool& stop_loading, } } -std::shared_ptr<OGLProgram> ShaderCacheOpenGL::GeneratePrecompiledProgram( +ProgramSharedPtr ShaderCacheOpenGL::GeneratePrecompiledProgram( const ShaderDiskCacheEntry& entry, const ShaderDiskCachePrecompiled& precompiled_entry, const std::unordered_set<GLenum>& supported_formats) { if (supported_formats.find(precompiled_entry.binary_format) == supported_formats.end()) { @@ -423,15 +465,15 @@ std::shared_ptr<OGLProgram> ShaderCacheOpenGL::GeneratePrecompiledProgram( return {}; } - auto program = std::make_shared<OGLProgram>(); - program->handle = glCreateProgram(); - glProgramParameteri(program->handle, GL_PROGRAM_SEPARABLE, GL_TRUE); - glProgramBinary(program->handle, precompiled_entry.binary_format, - precompiled_entry.binary.data(), + auto program = std::make_shared<ProgramHandle>(); + GLuint& handle = program->source_program.handle; + handle = glCreateProgram(); + glProgramParameteri(handle, GL_PROGRAM_SEPARABLE, GL_TRUE); + glProgramBinary(handle, precompiled_entry.binary_format, precompiled_entry.binary.data(), static_cast<GLsizei>(precompiled_entry.binary.size())); GLint link_status; - glGetProgramiv(program->handle, GL_LINK_STATUS, &link_status); + glGetProgramiv(handle, GL_LINK_STATUS, &link_status); if (link_status == GL_FALSE) { LOG_INFO(Render_OpenGL, "Precompiled cache rejected by the driver, removing"); return {}; @@ -440,77 +482,122 @@ std::shared_ptr<OGLProgram> ShaderCacheOpenGL::GeneratePrecompiledProgram( return program; } -Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) { - if (!system.GPU().Maxwell3D().dirty.flags[Dirty::Shaders]) { - return last_shaders[static_cast<std::size_t>(program)]; +Shader* ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program, + VideoCommon::Shader::AsyncShaders& async_shaders) { + if (!maxwell3d.dirty.flags[Dirty::Shaders]) { + auto* last_shader = last_shaders[static_cast<std::size_t>(program)]; + if (last_shader->IsBuilt()) { + return last_shader; + } } - auto& memory_manager{system.GPU().MemoryManager()}; - const GPUVAddr address{GetShaderAddress(system, program)}; + const GPUVAddr address{GetShaderAddress(maxwell3d, program)}; + + if (device.UseAsynchronousShaders() && async_shaders.HasCompletedWork()) { + auto completed_work = async_shaders.GetCompletedWork(); + for (auto& work : completed_work) { + Shader* shader = TryGet(work.cpu_address); + gpu.ShaderNotify().MarkShaderComplete(); + if (shader == nullptr) { + continue; + } + using namespace VideoCommon::Shader; + if (work.backend == AsyncShaders::Backend::OpenGL) { + shader->AsyncOpenGLBuilt(std::move(work.program.opengl)); + } else if (work.backend == AsyncShaders::Backend::GLASM) { + shader->AsyncGLASMBuilt(std::move(work.program.glasm)); + } + + auto& registry = shader->GetRegistry(); + + ShaderDiskCacheEntry entry; + entry.type = work.shader_type; + entry.code = std::move(work.code); + entry.code_b = std::move(work.code_b); + entry.unique_identifier = work.uid; + entry.bound_buffer = registry.GetBoundBuffer(); + entry.graphics_info = registry.GetGraphicsInfo(); + entry.keys = registry.GetKeys(); + entry.bound_samplers = registry.GetBoundSamplers(); + entry.bindless_samplers = registry.GetBindlessSamplers(); + disk_cache.SaveEntry(std::move(entry)); + } + } // Look up shader in the cache based on address - const auto cpu_addr{memory_manager.GpuToCpuAddress(address)}; - Shader shader{cpu_addr ? TryGet(*cpu_addr) : nullptr}; - if (shader) { + const std::optional<VAddr> cpu_addr{gpu_memory.GpuToCpuAddress(address)}; + if (Shader* const shader{cpu_addr ? TryGet(*cpu_addr) : null_shader.get()}) { return last_shaders[static_cast<std::size_t>(program)] = shader; } - const auto host_ptr{memory_manager.GetPointer(address)}; + const u8* const host_ptr{gpu_memory.GetPointer(address)}; // No shader found - create a new one - ProgramCode code{GetShaderCode(memory_manager, address, host_ptr)}; + ProgramCode code{GetShaderCode(gpu_memory, address, host_ptr, false)}; ProgramCode code_b; if (program == Maxwell::ShaderProgram::VertexA) { - const GPUVAddr address_b{GetShaderAddress(system, Maxwell::ShaderProgram::VertexB)}; - code_b = GetShaderCode(memory_manager, address_b, memory_manager.GetPointer(address_b)); + const GPUVAddr address_b{GetShaderAddress(maxwell3d, Maxwell::ShaderProgram::VertexB)}; + const u8* host_ptr_b = gpu_memory.GetPointer(address_b); + code_b = GetShaderCode(gpu_memory, address_b, host_ptr_b, false); } + const std::size_t code_size = code.size() * sizeof(u64); - const auto unique_identifier = GetUniqueIdentifier( + const u64 unique_identifier = GetUniqueIdentifier( GetShaderType(program), program == Maxwell::ShaderProgram::VertexA, code, code_b); - const ShaderParameters params{system, disk_cache, device, - *cpu_addr, host_ptr, unique_identifier}; + const ShaderParameters params{gpu, maxwell3d, disk_cache, device, + *cpu_addr, host_ptr, unique_identifier}; + std::unique_ptr<Shader> shader; const auto found = runtime_cache.find(unique_identifier); if (found == runtime_cache.end()) { - shader = CachedShader::CreateStageFromMemory(params, program, std::move(code), - std::move(code_b)); + shader = Shader::CreateStageFromMemory(params, program, std::move(code), std::move(code_b), + async_shaders, cpu_addr.value_or(0)); } else { - const std::size_t size_in_bytes = code.size() * sizeof(u64); - shader = CachedShader::CreateFromCache(params, found->second, size_in_bytes); + shader = Shader::CreateFromCache(params, found->second); } - Register(shader); - return last_shaders[static_cast<std::size_t>(program)] = shader; + Shader* const result = shader.get(); + if (cpu_addr) { + Register(std::move(shader), *cpu_addr, code_size); + } else { + null_shader = std::move(shader); + } + + return last_shaders[static_cast<std::size_t>(program)] = result; } -Shader ShaderCacheOpenGL::GetComputeKernel(GPUVAddr code_addr) { - auto& memory_manager{system.GPU().MemoryManager()}; - const auto cpu_addr{memory_manager.GpuToCpuAddress(code_addr)}; +Shader* ShaderCacheOpenGL::GetComputeKernel(GPUVAddr code_addr) { + const std::optional<VAddr> cpu_addr{gpu_memory.GpuToCpuAddress(code_addr)}; - auto kernel = cpu_addr ? TryGet(*cpu_addr) : nullptr; - if (kernel) { + if (Shader* const kernel = cpu_addr ? TryGet(*cpu_addr) : null_kernel.get()) { return kernel; } - const auto host_ptr{memory_manager.GetPointer(code_addr)}; // No kernel found, create a new one - auto code{GetShaderCode(memory_manager, code_addr, host_ptr)}; - const auto unique_identifier{GetUniqueIdentifier(ShaderType::Compute, false, code)}; + const u8* host_ptr{gpu_memory.GetPointer(code_addr)}; + ProgramCode code{GetShaderCode(gpu_memory, code_addr, host_ptr, true)}; + const std::size_t code_size{code.size() * sizeof(u64)}; + const u64 unique_identifier{GetUniqueIdentifier(ShaderType::Compute, false, code)}; - const ShaderParameters params{system, disk_cache, device, - *cpu_addr, host_ptr, unique_identifier}; + const ShaderParameters params{gpu, kepler_compute, disk_cache, device, + *cpu_addr, host_ptr, unique_identifier}; + std::unique_ptr<Shader> kernel; const auto found = runtime_cache.find(unique_identifier); if (found == runtime_cache.end()) { - kernel = CachedShader::CreateKernelFromMemory(params, std::move(code)); + kernel = Shader::CreateKernelFromMemory(params, std::move(code)); } else { - const std::size_t size_in_bytes = code.size() * sizeof(u64); - kernel = CachedShader::CreateFromCache(params, found->second, size_in_bytes); + kernel = Shader::CreateFromCache(params, found->second); } - Register(kernel); - return kernel; + Shader* const result = kernel.get(); + if (cpu_addr) { + Register(std::move(kernel), *cpu_addr, code_size); + } else { + null_kernel = std::move(kernel); + } + return result; } } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_cache.h b/src/video_core/renderer_opengl/gl_shader_cache.h index c836df5bd..1708af06a 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.h +++ b/src/video_core/renderer_opengl/gl_shader_cache.h @@ -18,114 +18,143 @@ #include "common/common_types.h" #include "video_core/engines/shader_type.h" -#include "video_core/rasterizer_cache.h" #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" #include "video_core/renderer_opengl/gl_shader_disk_cache.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" +#include "video_core/shader_cache.h" -namespace Core { -class System; +namespace Tegra { +class MemoryManager; } namespace Core::Frontend { class EmuWindow; } +namespace VideoCommon::Shader { +class AsyncShaders; +} + namespace OpenGL { -class CachedShader; class Device; class RasterizerOpenGL; -struct UnspecializedShader; -using Shader = std::shared_ptr<CachedShader>; using Maxwell = Tegra::Engines::Maxwell3D::Regs; +struct ProgramHandle { + OGLProgram source_program; + OGLAssemblyProgram assembly_program; +}; +using ProgramSharedPtr = std::shared_ptr<ProgramHandle>; + struct PrecompiledShader { - std::shared_ptr<OGLProgram> program; + ProgramSharedPtr program; std::shared_ptr<VideoCommon::Shader::Registry> registry; ShaderEntries entries; }; struct ShaderParameters { - Core::System& system; + Tegra::GPU& gpu; + Tegra::Engines::ConstBufferEngineInterface& engine; ShaderDiskCacheOpenGL& disk_cache; const Device& device; VAddr cpu_addr; - u8* host_ptr; + const u8* host_ptr; u64 unique_identifier; }; -class CachedShader final : public RasterizerCacheObject { +ProgramSharedPtr BuildShader(const Device& device, Tegra::Engines::ShaderType shader_type, + u64 unique_identifier, const VideoCommon::Shader::ShaderIR& ir, + const VideoCommon::Shader::Registry& registry, + bool hint_retrievable = false); + +class Shader final { public: - ~CachedShader(); + ~Shader(); /// Gets the GL program handle for the shader GLuint GetHandle() const; - /// Returns the size in bytes of the shader - std::size_t GetSizeInBytes() const override { - return size_in_bytes; - } + bool IsBuilt() const; /// Gets the shader entries for the shader const ShaderEntries& GetEntries() const { return entries; } - static Shader CreateStageFromMemory(const ShaderParameters& params, - Maxwell::ShaderProgram program_type, - ProgramCode program_code, ProgramCode program_code_b); - static Shader CreateKernelFromMemory(const ShaderParameters& params, ProgramCode code); + const VideoCommon::Shader::Registry& GetRegistry() const { + return *registry; + } + + /// Mark a OpenGL shader as built + void AsyncOpenGLBuilt(OGLProgram new_program); + + /// Mark a GLASM shader as built + void AsyncGLASMBuilt(OGLAssemblyProgram new_program); + + static std::unique_ptr<Shader> CreateStageFromMemory( + const ShaderParameters& params, Maxwell::ShaderProgram program_type, + ProgramCode program_code, ProgramCode program_code_b, + VideoCommon::Shader::AsyncShaders& async_shaders, VAddr cpu_addr); + + static std::unique_ptr<Shader> CreateKernelFromMemory(const ShaderParameters& params, + ProgramCode code); - static Shader CreateFromCache(const ShaderParameters& params, - const PrecompiledShader& precompiled_shader, - std::size_t size_in_bytes); + static std::unique_ptr<Shader> CreateFromCache(const ShaderParameters& params, + const PrecompiledShader& precompiled_shader); private: - explicit CachedShader(VAddr cpu_addr, std::size_t size_in_bytes, - std::shared_ptr<VideoCommon::Shader::Registry> registry, - ShaderEntries entries, std::shared_ptr<OGLProgram> program); + explicit Shader(std::shared_ptr<VideoCommon::Shader::Registry> registry, ShaderEntries entries, + ProgramSharedPtr program, bool is_built = true); std::shared_ptr<VideoCommon::Shader::Registry> registry; ShaderEntries entries; - std::size_t size_in_bytes = 0; - std::shared_ptr<OGLProgram> program; + ProgramSharedPtr program; + GLuint handle = 0; + bool is_built{}; }; -class ShaderCacheOpenGL final : public RasterizerCache<Shader> { +class ShaderCacheOpenGL final : public VideoCommon::ShaderCache<Shader> { public: - explicit ShaderCacheOpenGL(RasterizerOpenGL& rasterizer, Core::System& system, - Core::Frontend::EmuWindow& emu_window, const Device& device); + explicit ShaderCacheOpenGL(RasterizerOpenGL& rasterizer, Core::Frontend::EmuWindow& emu_window, + Tegra::GPU& gpu, Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::Engines::KeplerCompute& kepler_compute, + Tegra::MemoryManager& gpu_memory, const Device& device); + ~ShaderCacheOpenGL() override; /// Loads disk cache for the current game - void LoadDiskCache(const std::atomic_bool& stop_loading, + void LoadDiskCache(u64 title_id, const std::atomic_bool& stop_loading, const VideoCore::DiskResourceLoadCallback& callback); /// Gets the current specified shader stage program - Shader GetStageProgram(Maxwell::ShaderProgram program); + Shader* GetStageProgram(Maxwell::ShaderProgram program, + VideoCommon::Shader::AsyncShaders& async_shaders); /// Gets a compute kernel in the passed address - Shader GetComputeKernel(GPUVAddr code_addr); - -protected: - // We do not have to flush this cache as things in it are never modified by us. - void FlushObjectInner(const Shader& object) override {} + Shader* GetComputeKernel(GPUVAddr code_addr); private: - std::shared_ptr<OGLProgram> GeneratePrecompiledProgram( + ProgramSharedPtr GeneratePrecompiledProgram( const ShaderDiskCacheEntry& entry, const ShaderDiskCachePrecompiled& precompiled_entry, const std::unordered_set<GLenum>& supported_formats); - Core::System& system; Core::Frontend::EmuWindow& emu_window; + Tegra::GPU& gpu; + Tegra::MemoryManager& gpu_memory; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::Engines::KeplerCompute& kepler_compute; const Device& device; + ShaderDiskCacheOpenGL disk_cache; std::unordered_map<u64, PrecompiledShader> runtime_cache; - std::array<Shader, Maxwell::MaxShaderProgram> last_shaders; + std::unique_ptr<Shader> null_shader; + std::unique_ptr<Shader> null_kernel; + + std::array<Shader*, Maxwell::MaxShaderProgram> last_shaders{}; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index b1804e9ea..95ca96c8e 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -37,6 +37,7 @@ using Tegra::Shader::IpaMode; using Tegra::Shader::IpaSampleMode; using Tegra::Shader::PixelImap; using Tegra::Shader::Register; +using Tegra::Shader::TextureType; using VideoCommon::Shader::BuildTransformFeedback; using VideoCommon::Shader::Registry; @@ -61,8 +62,8 @@ struct TextureDerivates {}; using TextureArgument = std::pair<Type, Node>; using TextureIR = std::variant<TextureOffset, TextureDerivates, TextureArgument>; -constexpr u32 MAX_CONSTBUFFER_ELEMENTS = - static_cast<u32>(Maxwell::MaxConstBufferSize) / (4 * sizeof(float)); +constexpr u32 MAX_CONSTBUFFER_SCALARS = static_cast<u32>(Maxwell::MaxConstBufferSize) / sizeof(u32); +constexpr u32 MAX_CONSTBUFFER_ELEMENTS = MAX_CONSTBUFFER_SCALARS / sizeof(u32); constexpr std::string_view CommonDeclarations = R"(#define ftoi floatBitsToInt #define ftou floatBitsToUint @@ -402,6 +403,13 @@ std::string FlowStackTopName(MetaStackClass stack) { return fmt::format("{}_flow_stack_top", GetFlowStackPrefix(stack)); } +bool UseUnifiedUniforms(const Device& device, const ShaderIR& ir, ShaderType stage) { + const u32 num_ubos = static_cast<u32>(ir.GetConstantBuffers().size()); + // We waste one UBO for emulation + const u32 num_available_ubos = device.GetMaxUniformBuffers(stage) - 1; + return num_ubos > num_available_ubos; +} + struct GenericVaryingDescription { std::string name; u8 first_element = 0; @@ -412,8 +420,9 @@ class GLSLDecompiler final { public: explicit GLSLDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, ShaderType stage, std::string_view identifier, std::string_view suffix) - : device{device}, ir{ir}, registry{registry}, stage{stage}, - identifier{identifier}, suffix{suffix}, header{ir.GetHeader()} { + : device{device}, ir{ir}, registry{registry}, stage{stage}, identifier{identifier}, + suffix{suffix}, header{ir.GetHeader()}, use_unified_uniforms{ + UseUnifiedUniforms(device, ir, stage)} { if (stage != ShaderType::Compute) { transform_feedback = BuildTransformFeedback(registry.GetGraphicsInfo()); } @@ -484,7 +493,7 @@ private: code.AddLine("switch (jmp_to) {{"); for (const auto& pair : ir.GetBasicBlocks()) { - const auto [address, bb] = pair; + const auto& [address, bb] = pair; code.AddLine("case 0x{:X}U: {{", address); ++code.scope; @@ -518,6 +527,9 @@ private: if (device.HasImageLoadFormatted()) { code.AddLine("#extension GL_EXT_shader_image_load_formatted : require"); } + if (device.HasTextureShadowLod()) { + code.AddLine("#extension GL_EXT_texture_shadow_lod : require"); + } if (device.HasWarpIntrinsics()) { code.AddLine("#extension GL_NV_gpu_shader5 : require"); code.AddLine("#extension GL_NV_shader_thread_group : require"); @@ -590,8 +602,15 @@ private: return; } const auto& info = registry.GetComputeInfo(); - if (const u32 size = info.shared_memory_size_in_words; size > 0) { - code.AddLine("shared uint smem[{}];", size); + if (u32 size = info.shared_memory_size_in_words * 4; size > 0) { + const u32 limit = device.GetMaxComputeSharedMemorySize(); + if (size > limit) { + LOG_ERROR(Render_OpenGL, "Shared memory size {} is clamped to host's limit {}", + size, limit); + size = limit; + } + + code.AddLine("shared uint smem[{}];", size / 4); code.AddNewLine(); } code.AddLine("layout (local_size_x = {}, local_size_y = {}, local_size_z = {}) in;", @@ -618,7 +637,9 @@ private: break; } } - if (stage != ShaderType::Vertex || device.HasVertexViewportLayer()) { + + if (stage != ShaderType::Geometry && + (stage != ShaderType::Vertex || device.HasVertexViewportLayer())) { if (ir.UsesLayer()) { code.AddLine("int gl_Layer;"); } @@ -647,6 +668,16 @@ private: --code.scope; code.AddLine("}};"); code.AddNewLine(); + + if (stage == ShaderType::Geometry) { + if (ir.UsesLayer()) { + code.AddLine("out int gl_Layer;"); + } + if (ir.UsesViewportIndex()) { + code.AddLine("out int gl_ViewportIndex;"); + } + } + code.AddNewLine(); } void DeclareRegisters() { @@ -782,7 +813,7 @@ private: const u8 location = static_cast<u8>(static_cast<u32>(index) * 4 + element); const auto it = transform_feedback.find(location); if (it == transform_feedback.end()) { - return {}; + return std::nullopt; } return it->second.components; } @@ -834,11 +865,24 @@ private: } void DeclareConstantBuffers() { + if (use_unified_uniforms) { + const u32 binding = device.GetBaseBindings(stage).shader_storage_buffer + + static_cast<u32>(ir.GetGlobalMemory().size()); + code.AddLine("layout (std430, binding = {}) readonly buffer UnifiedUniforms {{", + binding); + code.AddLine(" uint cbufs[];"); + code.AddLine("}};"); + code.AddNewLine(); + return; + } + u32 binding = device.GetBaseBindings(stage).uniform_buffer; - for (const auto& [index, cbuf] : ir.GetConstantBuffers()) { + for (const auto [index, info] : ir.GetConstantBuffers()) { + const u32 num_elements = Common::AlignUp(info.GetSize(), 4) / 4; + const u32 size = info.IsIndirect() ? MAX_CONSTBUFFER_ELEMENTS : num_elements; code.AddLine("layout (std140, binding = {}) uniform {} {{", binding++, GetConstBufferBlock(index)); - code.AddLine(" uvec4 {}[{}];", GetConstBuffer(index), MAX_CONSTBUFFER_ELEMENTS); + code.AddLine(" uvec4 {}[{}];", GetConstBuffer(index), size); code.AddLine("}};"); code.AddNewLine(); } @@ -869,37 +913,37 @@ private: for (const auto& sampler : ir.GetSamplers()) { const std::string name = GetSampler(sampler); const std::string description = fmt::format("layout (binding = {}) uniform", binding); - binding += sampler.IsIndexed() ? sampler.Size() : 1; + binding += sampler.is_indexed ? sampler.size : 1; std::string sampler_type = [&]() { - if (sampler.IsBuffer()) { + if (sampler.is_buffer) { return "samplerBuffer"; } - switch (sampler.GetType()) { - case Tegra::Shader::TextureType::Texture1D: + switch (sampler.type) { + case TextureType::Texture1D: return "sampler1D"; - case Tegra::Shader::TextureType::Texture2D: + case TextureType::Texture2D: return "sampler2D"; - case Tegra::Shader::TextureType::Texture3D: + case TextureType::Texture3D: return "sampler3D"; - case Tegra::Shader::TextureType::TextureCube: + case TextureType::TextureCube: return "samplerCube"; default: UNREACHABLE(); return "sampler2D"; } }(); - if (sampler.IsArray()) { + if (sampler.is_array) { sampler_type += "Array"; } - if (sampler.IsShadow()) { + if (sampler.is_shadow) { sampler_type += "Shadow"; } - if (!sampler.IsIndexed()) { + if (!sampler.is_indexed) { code.AddLine("{} {} {};", description, sampler_type, name); } else { - code.AddLine("{} {} {}[{}];", description, sampler_type, name, sampler.Size()); + code.AddLine("{} {} {}[{}];", description, sampler_type, name, sampler.size); } } if (!ir.GetSamplers().empty()) { @@ -945,14 +989,14 @@ private: u32 binding = device.GetBaseBindings(stage).image; for (const auto& image : ir.GetImages()) { std::string qualifier = "coherent volatile"; - if (image.IsRead() && !image.IsWritten()) { + if (image.is_read && !image.is_written) { qualifier += " readonly"; - } else if (image.IsWritten() && !image.IsRead()) { + } else if (image.is_written && !image.is_read) { qualifier += " writeonly"; } - const char* format = image.IsAtomic() ? "r32ui, " : ""; - const char* type_declaration = GetImageTypeDeclaration(image.GetType()); + const char* format = image.is_atomic ? "r32ui, " : ""; + const char* type_declaration = GetImageTypeDeclaration(image.type); code.AddLine("layout ({}binding = {}) {} uniform uimage{} {};", format, binding++, qualifier, type_declaration, GetImage(image)); } @@ -1037,42 +1081,51 @@ private: if (const auto cbuf = std::get_if<CbufNode>(&*node)) { const Node offset = cbuf->GetOffset(); + const u32 base_unified_offset = cbuf->GetIndex() * MAX_CONSTBUFFER_SCALARS; + if (const auto immediate = std::get_if<ImmediateNode>(&*offset)) { // Direct access const u32 offset_imm = immediate->GetValue(); ASSERT_MSG(offset_imm % 4 == 0, "Unaligned cbuf direct access"); - return {fmt::format("{}[{}][{}]", GetConstBuffer(cbuf->GetIndex()), - offset_imm / (4 * 4), (offset_imm / 4) % 4), - Type::Uint}; + if (use_unified_uniforms) { + return {fmt::format("cbufs[{}]", base_unified_offset + offset_imm / 4), + Type::Uint}; + } else { + return {fmt::format("{}[{}][{}]", GetConstBuffer(cbuf->GetIndex()), + offset_imm / (4 * 4), (offset_imm / 4) % 4), + Type::Uint}; + } } - if (std::holds_alternative<OperationNode>(*offset)) { - // Indirect access - const std::string final_offset = code.GenerateTemporary(); - code.AddLine("uint {} = {} >> 2;", final_offset, Visit(offset).AsUint()); + // Indirect access + if (use_unified_uniforms) { + return {fmt::format("cbufs[{} + ({} >> 2)]", base_unified_offset, + Visit(offset).AsUint()), + Type::Uint}; + } - if (!device.HasComponentIndexingBug()) { - return {fmt::format("{}[{} >> 2][{} & 3]", GetConstBuffer(cbuf->GetIndex()), - final_offset, final_offset), - Type::Uint}; - } + const std::string final_offset = code.GenerateTemporary(); + code.AddLine("uint {} = {} >> 2;", final_offset, Visit(offset).AsUint()); - // AMD's proprietary GLSL compiler emits ill code for variable component access. - // To bypass this driver bug generate 4 ifs, one per each component. - const std::string pack = code.GenerateTemporary(); - code.AddLine("uvec4 {} = {}[{} >> 2];", pack, GetConstBuffer(cbuf->GetIndex()), - final_offset); - - const std::string result = code.GenerateTemporary(); - code.AddLine("uint {};", result); - for (u32 swizzle = 0; swizzle < 4; ++swizzle) { - code.AddLine("if (({} & 3) == {}) {} = {}{};", final_offset, swizzle, result, - pack, GetSwizzle(swizzle)); - } - return {result, Type::Uint}; + if (!device.HasComponentIndexingBug()) { + return {fmt::format("{}[{} >> 2][{} & 3]", GetConstBuffer(cbuf->GetIndex()), + final_offset, final_offset), + Type::Uint}; } - UNREACHABLE_MSG("Unmanaged offset node type"); + // AMD's proprietary GLSL compiler emits ill code for variable component access. + // To bypass this driver bug generate 4 ifs, one per each component. + const std::string pack = code.GenerateTemporary(); + code.AddLine("uvec4 {} = {}[{} >> 2];", pack, GetConstBuffer(cbuf->GetIndex()), + final_offset); + + const std::string result = code.GenerateTemporary(); + code.AddLine("uint {};", result); + for (u32 swizzle = 0; swizzle < 4; ++swizzle) { + code.AddLine("if (({} & 3) == {}) {} = {}{};", final_offset, swizzle, result, pack, + GetSwizzle(swizzle)); + } + return {result, Type::Uint}; } if (const auto gmem = std::get_if<GmemNode>(&*node)) { @@ -1144,6 +1197,7 @@ private: return {"gl_FragCoord"s + GetSwizzle(element), Type::Float}; default: UNREACHABLE(); + return {"0", Type::Int}; } case Attribute::Index::FrontColor: return {"gl_Color"s + GetSwizzle(element), Type::Float}; @@ -1241,21 +1295,21 @@ private: switch (element) { case 0: UNIMPLEMENTED(); - return {}; + return std::nullopt; case 1: if (stage == ShaderType::Vertex && !device.HasVertexViewportLayer()) { - return {}; + return std::nullopt; } return {{"gl_Layer", Type::Int}}; case 2: if (stage == ShaderType::Vertex && !device.HasVertexViewportLayer()) { - return {}; + return std::nullopt; } return {{"gl_ViewportIndex", Type::Int}}; case 3: return {{"gl_PointSize", Type::Float}}; } - return {}; + return std::nullopt; case Attribute::Index::FrontColor: return {{"gl_FrontColor"s + GetSwizzle(element), Type::Float}}; case Attribute::Index::FrontSecondaryColor: @@ -1278,7 +1332,7 @@ private: Type::Float}}; } UNIMPLEMENTED_MSG("Unhandled output attribute: {}", static_cast<u32>(attribute)); - return {}; + return std::nullopt; } } @@ -1335,16 +1389,27 @@ private: ASSERT(meta); const std::size_t count = operation.GetOperandsCount(); - const bool has_array = meta->sampler.IsArray(); - const bool has_shadow = meta->sampler.IsShadow(); + const bool has_array = meta->sampler.is_array; + const bool has_shadow = meta->sampler.is_shadow; + const bool workaround_lod_array_shadow_as_grad = + !device.HasTextureShadowLod() && function_suffix == "Lod" && meta->sampler.is_shadow && + ((meta->sampler.type == TextureType::Texture2D && meta->sampler.is_array) || + meta->sampler.type == TextureType::TextureCube); + + std::string expr = "texture"; + + if (workaround_lod_array_shadow_as_grad) { + expr += "Grad"; + } else { + expr += function_suffix; + } - std::string expr = "texture" + function_suffix; if (!meta->aoffi.empty()) { expr += "Offset"; } else if (!meta->ptp.empty()) { expr += "Offsets"; } - if (!meta->sampler.IsIndexed()) { + if (!meta->sampler.is_indexed) { expr += '(' + GetSampler(meta->sampler) + ", "; } else { expr += '(' + GetSampler(meta->sampler) + '[' + Visit(meta->index).AsUint() + "], "; @@ -1372,6 +1437,18 @@ private: expr += ')'; } + if (workaround_lod_array_shadow_as_grad) { + switch (meta->sampler.type) { + case TextureType::Texture2D: + return expr + ", vec2(0.0), vec2(0.0))"; + case TextureType::TextureCube: + return expr + ", vec3(0.0), vec3(0.0))"; + default: + UNREACHABLE(); + break; + } + } + for (const auto& variant : extras) { if (const auto argument = std::get_if<TextureArgument>(&variant)) { expr += GenerateTextureArgument(*argument); @@ -1482,8 +1559,8 @@ private: dy += '('; for (std::size_t index = 0; index < components; ++index) { - const auto operand_x{derivates.at(index * 2)}; - const auto operand_y{derivates.at(index * 2 + 1)}; + const auto& operand_x{derivates.at(index * 2)}; + const auto& operand_y{derivates.at(index * 2 + 1)}; dx += Visit(operand_x).AsFloat(); dy += Visit(operand_y).AsFloat(); @@ -1536,7 +1613,9 @@ private: Expression target; if (const auto gpr = std::get_if<GprNode>(&*dest)) { if (gpr->GetIndex() == Register::ZeroIndex) { - // Writing to Register::ZeroIndex is a no op + // Writing to Register::ZeroIndex is a no op but we still have to visit the source + // as it might have side effects. + code.AddLine("{};", Visit(src).GetCode()); return {}; } target = {GetRegister(gpr->GetIndex()), Type::Float}; @@ -1838,38 +1917,48 @@ private: Type::HalfFloat}; } - template <Type type> - Expression LogicalLessThan(Operation operation) { - return GenerateBinaryInfix(operation, "<", Type::Bool, type, type); - } - - template <Type type> - Expression LogicalEqual(Operation operation) { - return GenerateBinaryInfix(operation, "==", Type::Bool, type, type); - } + template <const std::string_view& op, Type type, bool unordered = false> + Expression Comparison(Operation operation) { + static_assert(!unordered || type == Type::Float); - template <Type type> - Expression LogicalLessEqual(Operation operation) { - return GenerateBinaryInfix(operation, "<=", Type::Bool, type, type); - } + Expression expr = GenerateBinaryInfix(operation, op, Type::Bool, type, type); - template <Type type> - Expression LogicalGreaterThan(Operation operation) { - return GenerateBinaryInfix(operation, ">", Type::Bool, type, type); + if constexpr (op.compare("!=") == 0 && type == Type::Float && !unordered) { + // GLSL's operator!=(float, float) doesn't seem be ordered. This happens on both AMD's + // and Nvidia's proprietary stacks. Manually force an ordered comparison. + return {fmt::format("({} && !isnan({}) && !isnan({}))", expr.AsBool(), + VisitOperand(operation, 0).AsFloat(), + VisitOperand(operation, 1).AsFloat()), + Type::Bool}; + } + if constexpr (!unordered) { + return expr; + } + // Unordered comparisons are always true for NaN operands. + return {fmt::format("({} || isnan({}) || isnan({}))", expr.AsBool(), + VisitOperand(operation, 0).AsFloat(), + VisitOperand(operation, 1).AsFloat()), + Type::Bool}; } - template <Type type> - Expression LogicalNotEqual(Operation operation) { - return GenerateBinaryInfix(operation, "!=", Type::Bool, type, type); + Expression FOrdered(Operation operation) { + return {fmt::format("(!isnan({}) && !isnan({}))", VisitOperand(operation, 0).AsFloat(), + VisitOperand(operation, 1).AsFloat()), + Type::Bool}; } - template <Type type> - Expression LogicalGreaterEqual(Operation operation) { - return GenerateBinaryInfix(operation, ">=", Type::Bool, type, type); + Expression FUnordered(Operation operation) { + return {fmt::format("(isnan({}) || isnan({}))", VisitOperand(operation, 0).AsFloat(), + VisitOperand(operation, 1).AsFloat()), + Type::Bool}; } - Expression LogicalFIsNan(Operation operation) { - return GenerateUnary(operation, "isnan", Type::Bool, Type::Float); + Expression LogicalAddCarry(Operation operation) { + const std::string carry = code.GenerateTemporary(); + code.AddLine("uint {};", carry); + code.AddLine("uaddCarry({}, {}, {});", VisitOperand(operation, 0).AsUint(), + VisitOperand(operation, 1).AsUint(), carry); + return {fmt::format("({} != 0)", carry), Type::Bool}; } Expression LogicalAssign(Operation operation) { @@ -1967,24 +2056,39 @@ private: } Expression Texture(Operation operation) { - const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); - ASSERT(meta); - - std::string expr = GenerateTexture( - operation, "", {TextureOffset{}, TextureArgument{Type::Float, meta->bias}}); - if (meta->sampler.IsShadow()) { - expr = "vec4(" + expr + ')'; + const auto meta = std::get<MetaTexture>(operation.GetMeta()); + const bool separate_dc = meta.sampler.type == TextureType::TextureCube && + meta.sampler.is_array && meta.sampler.is_shadow; + // TODO: Replace this with an array and make GenerateTexture use C++20 std::span + const std::vector<TextureIR> extras{ + TextureOffset{}, + TextureArgument{Type::Float, meta.bias}, + }; + std::string expr = GenerateTexture(operation, "", extras, separate_dc); + if (meta.sampler.is_shadow) { + expr = fmt::format("vec4({})", expr); } - return {expr + GetSwizzle(meta->element), Type::Float}; + return {expr + GetSwizzle(meta.element), Type::Float}; } Expression TextureLod(Operation operation) { const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); ASSERT(meta); - std::string expr = GenerateTexture( - operation, "Lod", {TextureArgument{Type::Float, meta->lod}, TextureOffset{}}); - if (meta->sampler.IsShadow()) { + std::string expr{}; + + if (!device.HasTextureShadowLod() && meta->sampler.is_shadow && + ((meta->sampler.type == TextureType::Texture2D && meta->sampler.is_array) || + meta->sampler.type == TextureType::TextureCube)) { + LOG_ERROR(Render_OpenGL, + "Device lacks GL_EXT_texture_shadow_lod, using textureGrad as a workaround"); + expr = GenerateTexture(operation, "Lod", {}); + } else { + expr = GenerateTexture(operation, "Lod", + {TextureArgument{Type::Float, meta->lod}, TextureOffset{}}); + } + + if (meta->sampler.is_shadow) { expr = "vec4(" + expr + ')'; } return {expr + GetSwizzle(meta->element), Type::Float}; @@ -1993,11 +2097,11 @@ private: Expression TextureGather(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); - const auto type = meta.sampler.IsShadow() ? Type::Float : Type::Int; - const bool separate_dc = meta.sampler.IsShadow(); + const auto type = meta.sampler.is_shadow ? Type::Float : Type::Int; + const bool separate_dc = meta.sampler.is_shadow; std::vector<TextureIR> ir; - if (meta.sampler.IsShadow()) { + if (meta.sampler.is_shadow) { ir = {TextureOffset{}}; } else { ir = {TextureOffset{}, TextureArgument{type, meta.component}}; @@ -2042,7 +2146,7 @@ private: constexpr std::array constructors = {"int", "ivec2", "ivec3", "ivec4"}; const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); ASSERT(meta); - UNIMPLEMENTED_IF(meta->sampler.IsArray()); + UNIMPLEMENTED_IF(meta->sampler.is_array); const std::size_t count = operation.GetOperandsCount(); std::string expr = "texelFetch("; @@ -2063,7 +2167,7 @@ private: } expr += ')'; - if (meta->lod && !meta->sampler.IsBuffer()) { + if (meta->lod && !meta->sampler.is_buffer) { expr += ", "; expr += Visit(meta->lod).AsInt(); } @@ -2074,12 +2178,10 @@ private: } Expression TextureGradient(Operation operation) { - const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); - ASSERT(meta); - + const auto& meta = std::get<MetaTexture>(operation.GetMeta()); std::string expr = GenerateTexture(operation, "Grad", {TextureDerivates{}, TextureOffset{}}); - return {std::move(expr) + GetSwizzle(meta->element), Type::Float}; + return {std::move(expr) + GetSwizzle(meta.element), Type::Float}; } Expression ImageLoad(Operation operation) { @@ -2295,6 +2397,18 @@ private: return {"gl_SubGroupInvocationARB", Type::Uint}; } + template <const std::string_view& comparison> + Expression ThreadMask(Operation) { + if (device.HasWarpIntrinsics()) { + return {fmt::format("gl_Thread{}MaskNV", comparison), Type::Uint}; + } + if (device.HasShaderBallot()) { + return {fmt::format("uint(gl_SubGroup{}MaskARB)", comparison), Type::Uint}; + } + LOG_ERROR(Render_OpenGL, "Thread mask intrinsics are required by the shader"); + return {"0U", Type::Uint}; + } + Expression ShuffleIndexed(Operation operation) { std::string value = VisitOperand(operation, 0).AsFloat(); @@ -2307,7 +2421,21 @@ private: return {fmt::format("readInvocationARB({}, {})", value, index), Type::Float}; } - Expression MemoryBarrierGL(Operation) { + Expression Barrier(Operation) { + if (!ir.IsDecompiled()) { + LOG_ERROR(Render_OpenGL, "barrier() used but shader is not decompiled"); + return {}; + } + code.AddLine("barrier();"); + return {}; + } + + Expression MemoryBarrierGroup(Operation) { + code.AddLine("groupMemoryBarrier();"); + return {}; + } + + Expression MemoryBarrierGlobal(Operation) { code.AddLine("memoryBarrier();"); return {}; } @@ -2316,6 +2444,19 @@ private: Func() = delete; ~Func() = delete; + static constexpr std::string_view LessThan = "<"; + static constexpr std::string_view Equal = "=="; + static constexpr std::string_view LessEqual = "<="; + static constexpr std::string_view GreaterThan = ">"; + static constexpr std::string_view NotEqual = "!="; + static constexpr std::string_view GreaterEqual = ">="; + + static constexpr std::string_view Eq = "Eq"; + static constexpr std::string_view Ge = "Ge"; + static constexpr std::string_view Gt = "Gt"; + static constexpr std::string_view Le = "Le"; + static constexpr std::string_view Lt = "Lt"; + static constexpr std::string_view Add = "Add"; static constexpr std::string_view Min = "Min"; static constexpr std::string_view Max = "Max"; @@ -2417,27 +2558,36 @@ private: &GLSLDecompiler::LogicalPick2, &GLSLDecompiler::LogicalAnd2, - &GLSLDecompiler::LogicalLessThan<Type::Float>, - &GLSLDecompiler::LogicalEqual<Type::Float>, - &GLSLDecompiler::LogicalLessEqual<Type::Float>, - &GLSLDecompiler::LogicalGreaterThan<Type::Float>, - &GLSLDecompiler::LogicalNotEqual<Type::Float>, - &GLSLDecompiler::LogicalGreaterEqual<Type::Float>, - &GLSLDecompiler::LogicalFIsNan, - - &GLSLDecompiler::LogicalLessThan<Type::Int>, - &GLSLDecompiler::LogicalEqual<Type::Int>, - &GLSLDecompiler::LogicalLessEqual<Type::Int>, - &GLSLDecompiler::LogicalGreaterThan<Type::Int>, - &GLSLDecompiler::LogicalNotEqual<Type::Int>, - &GLSLDecompiler::LogicalGreaterEqual<Type::Int>, - - &GLSLDecompiler::LogicalLessThan<Type::Uint>, - &GLSLDecompiler::LogicalEqual<Type::Uint>, - &GLSLDecompiler::LogicalLessEqual<Type::Uint>, - &GLSLDecompiler::LogicalGreaterThan<Type::Uint>, - &GLSLDecompiler::LogicalNotEqual<Type::Uint>, - &GLSLDecompiler::LogicalGreaterEqual<Type::Uint>, + &GLSLDecompiler::Comparison<Func::LessThan, Type::Float, false>, + &GLSLDecompiler::Comparison<Func::Equal, Type::Float, false>, + &GLSLDecompiler::Comparison<Func::LessEqual, Type::Float, false>, + &GLSLDecompiler::Comparison<Func::GreaterThan, Type::Float, false>, + &GLSLDecompiler::Comparison<Func::NotEqual, Type::Float, false>, + &GLSLDecompiler::Comparison<Func::GreaterEqual, Type::Float, false>, + &GLSLDecompiler::FOrdered, + &GLSLDecompiler::FUnordered, + &GLSLDecompiler::Comparison<Func::LessThan, Type::Float, true>, + &GLSLDecompiler::Comparison<Func::Equal, Type::Float, true>, + &GLSLDecompiler::Comparison<Func::LessEqual, Type::Float, true>, + &GLSLDecompiler::Comparison<Func::GreaterThan, Type::Float, true>, + &GLSLDecompiler::Comparison<Func::NotEqual, Type::Float, true>, + &GLSLDecompiler::Comparison<Func::GreaterEqual, Type::Float, true>, + + &GLSLDecompiler::Comparison<Func::LessThan, Type::Int>, + &GLSLDecompiler::Comparison<Func::Equal, Type::Int>, + &GLSLDecompiler::Comparison<Func::LessEqual, Type::Int>, + &GLSLDecompiler::Comparison<Func::GreaterThan, Type::Int>, + &GLSLDecompiler::Comparison<Func::NotEqual, Type::Int>, + &GLSLDecompiler::Comparison<Func::GreaterEqual, Type::Int>, + + &GLSLDecompiler::Comparison<Func::LessThan, Type::Uint>, + &GLSLDecompiler::Comparison<Func::Equal, Type::Uint>, + &GLSLDecompiler::Comparison<Func::LessEqual, Type::Uint>, + &GLSLDecompiler::Comparison<Func::GreaterThan, Type::Uint>, + &GLSLDecompiler::Comparison<Func::NotEqual, Type::Uint>, + &GLSLDecompiler::Comparison<Func::GreaterEqual, Type::Uint>, + + &GLSLDecompiler::LogicalAddCarry, &GLSLDecompiler::Logical2HLessThan<false>, &GLSLDecompiler::Logical2HEqual<false>, @@ -2524,9 +2674,16 @@ private: &GLSLDecompiler::VoteEqual, &GLSLDecompiler::ThreadId, + &GLSLDecompiler::ThreadMask<Func::Eq>, + &GLSLDecompiler::ThreadMask<Func::Ge>, + &GLSLDecompiler::ThreadMask<Func::Gt>, + &GLSLDecompiler::ThreadMask<Func::Le>, + &GLSLDecompiler::ThreadMask<Func::Lt>, &GLSLDecompiler::ShuffleIndexed, - &GLSLDecompiler::MemoryBarrierGL, + &GLSLDecompiler::Barrier, + &GLSLDecompiler::MemoryBarrierGroup, + &GLSLDecompiler::MemoryBarrierGlobal, }; static_assert(operation_decompilers.size() == static_cast<std::size_t>(OperationCode::Amount)); @@ -2596,11 +2753,11 @@ private: } std::string GetSampler(const Sampler& sampler) const { - return AppendSuffix(static_cast<u32>(sampler.GetIndex()), "sampler"); + return AppendSuffix(sampler.index, "sampler"); } std::string GetImage(const Image& image) const { - return AppendSuffix(static_cast<u32>(image.GetIndex()), "image"); + return AppendSuffix(image.index, "image"); } std::string AppendSuffix(u32 index, std::string_view name) const { @@ -2623,15 +2780,6 @@ private: return std::min<u32>(device.GetMaxVaryings(), Maxwell::NumVaryings); } - bool IsRenderTargetEnabled(u32 render_target) const { - for (u32 component = 0; component < 4; ++component) { - if (header.ps.IsColorComponentOutputEnabled(render_target, component)) { - return true; - } - } - return false; - } - const Device& device; const ShaderIR& ir; const Registry& registry; @@ -2639,6 +2787,7 @@ private: const std::string_view identifier; const std::string_view suffix; const Header header; + const bool use_unified_uniforms; std::unordered_map<u8, VaryingTFB> transform_feedback; ShaderWriter code; @@ -2834,7 +2983,7 @@ void GLSLDecompiler::DecompileAST() { } // Anonymous namespace -ShaderEntries MakeEntries(const VideoCommon::Shader::ShaderIR& ir) { +ShaderEntries MakeEntries(const Device& device, const ShaderIR& ir, ShaderType stage) { ShaderEntries entries; for (const auto& cbuf : ir.GetConstantBuffers()) { entries.const_buffers.emplace_back(cbuf.second.GetMaxOffset(), cbuf.second.IsIndirect(), @@ -2855,6 +3004,7 @@ ShaderEntries MakeEntries(const VideoCommon::Shader::ShaderIR& ir) { entries.clip_distances = (clip_distances[i] ? 1U : 0U) << i; } entries.shader_length = ir.GetLength(); + entries.use_unified_uniforms = UseUnifiedUniforms(device, ir, stage); return entries; } diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h index e7dbd810c..451c9689a 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.h +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h @@ -33,36 +33,19 @@ public: } private: - u32 index{}; + u32 index = 0; }; -class GlobalMemoryEntry { -public: - explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, bool is_written) +struct GlobalMemoryEntry { + constexpr explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, + bool is_written) : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_read{is_read}, is_written{ is_written} {} - u32 GetCbufIndex() const { - return cbuf_index; - } - - u32 GetCbufOffset() const { - return cbuf_offset; - } - - bool IsRead() const { - return is_read; - } - - bool IsWritten() const { - return is_written; - } - -private: - u32 cbuf_index{}; - u32 cbuf_offset{}; - bool is_read{}; - bool is_written{}; + u32 cbuf_index = 0; + u32 cbuf_offset = 0; + bool is_read = false; + bool is_written = false; }; struct ShaderEntries { @@ -70,11 +53,13 @@ struct ShaderEntries { std::vector<GlobalMemoryEntry> global_memory_entries; std::vector<SamplerEntry> samplers; std::vector<ImageEntry> images; - u32 clip_distances{}; std::size_t shader_length{}; + u32 clip_distances{}; + bool use_unified_uniforms{}; }; -ShaderEntries MakeEntries(const VideoCommon::Shader::ShaderIR& ir); +ShaderEntries MakeEntries(const Device& device, const VideoCommon::Shader::ShaderIR& ir, + Tegra::Engines::ShaderType stage); std::string DecompileShader(const Device& device, const VideoCommon::Shader::ShaderIR& ir, const VideoCommon::Shader::Registry& registry, diff --git a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp index 9e95a122b..70dd0c3c6 100644 --- a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp @@ -29,6 +29,8 @@ using VideoCommon::Shader::KeyMap; namespace { +using VideoCommon::Shader::SeparateSamplerKey; + using ShaderCacheVersionHash = std::array<u8, 64>; struct ConstBufferKey { @@ -37,18 +39,26 @@ struct ConstBufferKey { u32 value = 0; }; -struct BoundSamplerKey { +struct BoundSamplerEntry { u32 offset = 0; Tegra::Engines::SamplerDescriptor sampler; }; -struct BindlessSamplerKey { +struct SeparateSamplerEntry { + u32 cbuf1 = 0; + u32 cbuf2 = 0; + u32 offset1 = 0; + u32 offset2 = 0; + Tegra::Engines::SamplerDescriptor sampler; +}; + +struct BindlessSamplerEntry { u32 cbuf = 0; u32 offset = 0; Tegra::Engines::SamplerDescriptor sampler; }; -constexpr u32 NativeVersion = 20; +constexpr u32 NativeVersion = 21; ShaderCacheVersionHash GetShaderCacheVersionHash() { ShaderCacheVersionHash hash{}; @@ -63,7 +73,7 @@ ShaderDiskCacheEntry::ShaderDiskCacheEntry() = default; ShaderDiskCacheEntry::~ShaderDiskCacheEntry() = default; -bool ShaderDiskCacheEntry::Load(FileUtil::IOFile& file) { +bool ShaderDiskCacheEntry::Load(Common::FS::IOFile& file) { if (file.ReadBytes(&type, sizeof(u32)) != sizeof(u32)) { return false; } @@ -87,12 +97,14 @@ bool ShaderDiskCacheEntry::Load(FileUtil::IOFile& file) { u32 texture_handler_size_value; u32 num_keys; u32 num_bound_samplers; + u32 num_separate_samplers; u32 num_bindless_samplers; if (file.ReadArray(&unique_identifier, 1) != 1 || file.ReadArray(&bound_buffer, 1) != 1 || file.ReadArray(&is_texture_handler_size_known, 1) != 1 || file.ReadArray(&texture_handler_size_value, 1) != 1 || file.ReadArray(&graphics_info, 1) != 1 || file.ReadArray(&compute_info, 1) != 1 || file.ReadArray(&num_keys, 1) != 1 || file.ReadArray(&num_bound_samplers, 1) != 1 || + file.ReadArray(&num_separate_samplers, 1) != 1 || file.ReadArray(&num_bindless_samplers, 1) != 1) { return false; } @@ -101,29 +113,38 @@ bool ShaderDiskCacheEntry::Load(FileUtil::IOFile& file) { } std::vector<ConstBufferKey> flat_keys(num_keys); - std::vector<BoundSamplerKey> flat_bound_samplers(num_bound_samplers); - std::vector<BindlessSamplerKey> flat_bindless_samplers(num_bindless_samplers); + std::vector<BoundSamplerEntry> flat_bound_samplers(num_bound_samplers); + std::vector<SeparateSamplerEntry> flat_separate_samplers(num_separate_samplers); + std::vector<BindlessSamplerEntry> flat_bindless_samplers(num_bindless_samplers); if (file.ReadArray(flat_keys.data(), flat_keys.size()) != flat_keys.size() || file.ReadArray(flat_bound_samplers.data(), flat_bound_samplers.size()) != flat_bound_samplers.size() || + file.ReadArray(flat_separate_samplers.data(), flat_separate_samplers.size()) != + flat_separate_samplers.size() || file.ReadArray(flat_bindless_samplers.data(), flat_bindless_samplers.size()) != flat_bindless_samplers.size()) { return false; } - for (const auto& key : flat_keys) { - keys.insert({{key.cbuf, key.offset}, key.value}); + for (const auto& entry : flat_keys) { + keys.insert({{entry.cbuf, entry.offset}, entry.value}); + } + for (const auto& entry : flat_bound_samplers) { + bound_samplers.emplace(entry.offset, entry.sampler); } - for (const auto& key : flat_bound_samplers) { - bound_samplers.emplace(key.offset, key.sampler); + for (const auto& entry : flat_separate_samplers) { + SeparateSamplerKey key; + key.buffers = {entry.cbuf1, entry.cbuf2}; + key.offsets = {entry.offset1, entry.offset2}; + separate_samplers.emplace(key, entry.sampler); } - for (const auto& key : flat_bindless_samplers) { - bindless_samplers.insert({{key.cbuf, key.offset}, key.sampler}); + for (const auto& entry : flat_bindless_samplers) { + bindless_samplers.insert({{entry.cbuf, entry.offset}, entry.sampler}); } return true; } -bool ShaderDiskCacheEntry::Save(FileUtil::IOFile& file) const { +bool ShaderDiskCacheEntry::Save(Common::FS::IOFile& file) const { if (file.WriteObject(static_cast<u32>(type)) != 1 || file.WriteObject(static_cast<u32>(code.size())) != 1 || file.WriteObject(static_cast<u32>(code_b.size())) != 1) { @@ -142,6 +163,7 @@ bool ShaderDiskCacheEntry::Save(FileUtil::IOFile& file) const { file.WriteObject(graphics_info) != 1 || file.WriteObject(compute_info) != 1 || file.WriteObject(static_cast<u32>(keys.size())) != 1 || file.WriteObject(static_cast<u32>(bound_samplers.size())) != 1 || + file.WriteObject(static_cast<u32>(separate_samplers.size())) != 1 || file.WriteObject(static_cast<u32>(bindless_samplers.size())) != 1) { return false; } @@ -152,48 +174,64 @@ bool ShaderDiskCacheEntry::Save(FileUtil::IOFile& file) const { flat_keys.push_back(ConstBufferKey{address.first, address.second, value}); } - std::vector<BoundSamplerKey> flat_bound_samplers; + std::vector<BoundSamplerEntry> flat_bound_samplers; flat_bound_samplers.reserve(bound_samplers.size()); for (const auto& [address, sampler] : bound_samplers) { - flat_bound_samplers.push_back(BoundSamplerKey{address, sampler}); + flat_bound_samplers.push_back(BoundSamplerEntry{address, sampler}); } - std::vector<BindlessSamplerKey> flat_bindless_samplers; + std::vector<SeparateSamplerEntry> flat_separate_samplers; + flat_separate_samplers.reserve(separate_samplers.size()); + for (const auto& [key, sampler] : separate_samplers) { + SeparateSamplerEntry entry; + std::tie(entry.cbuf1, entry.cbuf2) = key.buffers; + std::tie(entry.offset1, entry.offset2) = key.offsets; + entry.sampler = sampler; + flat_separate_samplers.push_back(entry); + } + + std::vector<BindlessSamplerEntry> flat_bindless_samplers; flat_bindless_samplers.reserve(bindless_samplers.size()); for (const auto& [address, sampler] : bindless_samplers) { flat_bindless_samplers.push_back( - BindlessSamplerKey{address.first, address.second, sampler}); + BindlessSamplerEntry{address.first, address.second, sampler}); } return file.WriteArray(flat_keys.data(), flat_keys.size()) == flat_keys.size() && file.WriteArray(flat_bound_samplers.data(), flat_bound_samplers.size()) == flat_bound_samplers.size() && + file.WriteArray(flat_separate_samplers.data(), flat_separate_samplers.size()) == + flat_separate_samplers.size() && file.WriteArray(flat_bindless_samplers.data(), flat_bindless_samplers.size()) == flat_bindless_samplers.size(); } -ShaderDiskCacheOpenGL::ShaderDiskCacheOpenGL(Core::System& system) : system{system} {} +ShaderDiskCacheOpenGL::ShaderDiskCacheOpenGL() = default; ShaderDiskCacheOpenGL::~ShaderDiskCacheOpenGL() = default; +void ShaderDiskCacheOpenGL::BindTitleID(u64 title_id_) { + title_id = title_id_; +} + std::optional<std::vector<ShaderDiskCacheEntry>> ShaderDiskCacheOpenGL::LoadTransferable() { // Skip games without title id - const bool has_title_id = system.CurrentProcess()->GetTitleID() != 0; - if (!Settings::values.use_disk_shader_cache || !has_title_id) { - return {}; + const bool has_title_id = title_id != 0; + if (!Settings::values.use_disk_shader_cache.GetValue() || !has_title_id) { + return std::nullopt; } - FileUtil::IOFile file(GetTransferablePath(), "rb"); + Common::FS::IOFile file(GetTransferablePath(), "rb"); if (!file.IsOpen()) { LOG_INFO(Render_OpenGL, "No transferable shader cache found"); is_usable = true; - return {}; + return std::nullopt; } u32 version{}; if (file.ReadBytes(&version, sizeof(version)) != sizeof(version)) { LOG_ERROR(Render_OpenGL, "Failed to get transferable cache version, skipping it"); - return {}; + return std::nullopt; } if (version < NativeVersion) { @@ -201,12 +239,12 @@ std::optional<std::vector<ShaderDiskCacheEntry>> ShaderDiskCacheOpenGL::LoadTran file.Close(); InvalidateTransferable(); is_usable = true; - return {}; + return std::nullopt; } if (version > NativeVersion) { LOG_WARNING(Render_OpenGL, "Transferable shader cache was generated with a newer version " "of the emulator, skipping"); - return {}; + return std::nullopt; } // Version is valid, load the shaders @@ -215,7 +253,7 @@ std::optional<std::vector<ShaderDiskCacheEntry>> ShaderDiskCacheOpenGL::LoadTran ShaderDiskCacheEntry& entry = entries.emplace_back(); if (!entry.Load(file)) { LOG_ERROR(Render_OpenGL, "Failed to load transferable raw entry, skipping"); - return {}; + return std::nullopt; } } @@ -228,7 +266,7 @@ std::vector<ShaderDiskCachePrecompiled> ShaderDiskCacheOpenGL::LoadPrecompiled() return {}; } - FileUtil::IOFile file(GetPrecompiledPath(), "rb"); + Common::FS::IOFile file(GetPrecompiledPath(), "rb"); if (!file.IsOpen()) { LOG_INFO(Render_OpenGL, "No precompiled shader cache found"); return {}; @@ -245,7 +283,7 @@ std::vector<ShaderDiskCachePrecompiled> ShaderDiskCacheOpenGL::LoadPrecompiled() } std::optional<std::vector<ShaderDiskCachePrecompiled>> ShaderDiskCacheOpenGL::LoadPrecompiledFile( - FileUtil::IOFile& file) { + Common::FS::IOFile& file) { // Read compressed file from disk and decompress to virtual precompiled cache file std::vector<u8> compressed(file.GetSize()); file.ReadBytes(compressed.data(), compressed.size()); @@ -256,12 +294,12 @@ std::optional<std::vector<ShaderDiskCachePrecompiled>> ShaderDiskCacheOpenGL::Lo ShaderCacheVersionHash file_hash{}; if (!LoadArrayFromPrecompiled(file_hash.data(), file_hash.size())) { precompiled_cache_virtual_file_offset = 0; - return {}; + return std::nullopt; } if (GetShaderCacheVersionHash() != file_hash) { LOG_INFO(Render_OpenGL, "Precompiled cache is from another version of the emulator"); precompiled_cache_virtual_file_offset = 0; - return {}; + return std::nullopt; } std::vector<ShaderDiskCachePrecompiled> entries; @@ -271,19 +309,19 @@ std::optional<std::vector<ShaderDiskCachePrecompiled>> ShaderDiskCacheOpenGL::Lo if (!LoadObjectFromPrecompiled(entry.unique_identifier) || !LoadObjectFromPrecompiled(entry.binary_format) || !LoadObjectFromPrecompiled(binary_size)) { - return {}; + return std::nullopt; } entry.binary.resize(binary_size); if (!LoadArrayFromPrecompiled(entry.binary.data(), entry.binary.size())) { - return {}; + return std::nullopt; } } return entries; } void ShaderDiskCacheOpenGL::InvalidateTransferable() { - if (!FileUtil::Delete(GetTransferablePath())) { + if (!Common::FS::Delete(GetTransferablePath())) { LOG_ERROR(Render_OpenGL, "Failed to invalidate transferable file={}", GetTransferablePath()); } @@ -294,7 +332,7 @@ void ShaderDiskCacheOpenGL::InvalidatePrecompiled() { // Clear virtaul precompiled cache file precompiled_cache_virtual_file.Resize(0); - if (!FileUtil::Delete(GetPrecompiledPath())) { + if (!Common::FS::Delete(GetPrecompiledPath())) { LOG_ERROR(Render_OpenGL, "Failed to invalidate precompiled file={}", GetPrecompiledPath()); } } @@ -310,7 +348,7 @@ void ShaderDiskCacheOpenGL::SaveEntry(const ShaderDiskCacheEntry& entry) { return; } - FileUtil::IOFile file = AppendTransferableFile(); + Common::FS::IOFile file = AppendTransferableFile(); if (!file.IsOpen()) { return; } @@ -352,15 +390,15 @@ void ShaderDiskCacheOpenGL::SavePrecompiled(u64 unique_identifier, GLuint progra } } -FileUtil::IOFile ShaderDiskCacheOpenGL::AppendTransferableFile() const { +Common::FS::IOFile ShaderDiskCacheOpenGL::AppendTransferableFile() const { if (!EnsureDirectories()) { return {}; } const auto transferable_path{GetTransferablePath()}; - const bool existed = FileUtil::Exists(transferable_path); + const bool existed = Common::FS::Exists(transferable_path); - FileUtil::IOFile file(transferable_path, "ab"); + Common::FS::IOFile file(transferable_path, "ab"); if (!file.IsOpen()) { LOG_ERROR(Render_OpenGL, "Failed to open transferable cache in path={}", transferable_path); return {}; @@ -392,7 +430,7 @@ void ShaderDiskCacheOpenGL::SaveVirtualPrecompiledFile() { Common::Compression::CompressDataZSTDDefault(uncompressed.data(), uncompressed.size()); const auto precompiled_path{GetPrecompiledPath()}; - FileUtil::IOFile file(precompiled_path, "wb"); + Common::FS::IOFile file(precompiled_path, "wb"); if (!file.IsOpen()) { LOG_ERROR(Render_OpenGL, "Failed to open precompiled cache in path={}", precompiled_path); @@ -406,24 +444,24 @@ void ShaderDiskCacheOpenGL::SaveVirtualPrecompiledFile() { bool ShaderDiskCacheOpenGL::EnsureDirectories() const { const auto CreateDir = [](const std::string& dir) { - if (!FileUtil::CreateDir(dir)) { + if (!Common::FS::CreateDir(dir)) { LOG_ERROR(Render_OpenGL, "Failed to create directory={}", dir); return false; } return true; }; - return CreateDir(FileUtil::GetUserPath(FileUtil::UserPath::ShaderDir)) && + return CreateDir(Common::FS::GetUserPath(Common::FS::UserPath::ShaderDir)) && CreateDir(GetBaseDir()) && CreateDir(GetTransferableDir()) && CreateDir(GetPrecompiledDir()); } std::string ShaderDiskCacheOpenGL::GetTransferablePath() const { - return FileUtil::SanitizePath(GetTransferableDir() + DIR_SEP_CHR + GetTitleID() + ".bin"); + return Common::FS::SanitizePath(GetTransferableDir() + DIR_SEP_CHR + GetTitleID() + ".bin"); } std::string ShaderDiskCacheOpenGL::GetPrecompiledPath() const { - return FileUtil::SanitizePath(GetPrecompiledDir() + DIR_SEP_CHR + GetTitleID() + ".bin"); + return Common::FS::SanitizePath(GetPrecompiledDir() + DIR_SEP_CHR + GetTitleID() + ".bin"); } std::string ShaderDiskCacheOpenGL::GetTransferableDir() const { @@ -435,11 +473,11 @@ std::string ShaderDiskCacheOpenGL::GetPrecompiledDir() const { } std::string ShaderDiskCacheOpenGL::GetBaseDir() const { - return FileUtil::GetUserPath(FileUtil::UserPath::ShaderDir) + DIR_SEP "opengl"; + return Common::FS::GetUserPath(Common::FS::UserPath::ShaderDir) + DIR_SEP "opengl"; } std::string ShaderDiskCacheOpenGL::GetTitleID() const { - return fmt::format("{:016X}", system.CurrentProcess()->GetTitleID()); + return fmt::format("{:016X}", title_id); } } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_disk_cache.h b/src/video_core/renderer_opengl/gl_shader_disk_cache.h index d5be52e40..aef841c1d 100644 --- a/src/video_core/renderer_opengl/gl_shader_disk_cache.h +++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.h @@ -21,11 +21,7 @@ #include "video_core/engines/shader_type.h" #include "video_core/shader/registry.h" -namespace Core { -class System; -} - -namespace FileUtil { +namespace Common::FS { class IOFile; } @@ -38,9 +34,9 @@ struct ShaderDiskCacheEntry { ShaderDiskCacheEntry(); ~ShaderDiskCacheEntry(); - bool Load(FileUtil::IOFile& file); + bool Load(Common::FS::IOFile& file); - bool Save(FileUtil::IOFile& file) const; + bool Save(Common::FS::IOFile& file) const; bool HasProgramA() const { return !code.empty() && !code_b.empty(); @@ -57,6 +53,7 @@ struct ShaderDiskCacheEntry { VideoCommon::Shader::ComputeInfo compute_info; VideoCommon::Shader::KeyMap keys; VideoCommon::Shader::BoundSamplerMap bound_samplers; + VideoCommon::Shader::SeparateSamplerMap separate_samplers; VideoCommon::Shader::BindlessSamplerMap bindless_samplers; }; @@ -69,9 +66,12 @@ struct ShaderDiskCachePrecompiled { class ShaderDiskCacheOpenGL { public: - explicit ShaderDiskCacheOpenGL(Core::System& system); + explicit ShaderDiskCacheOpenGL(); ~ShaderDiskCacheOpenGL(); + /// Binds a title ID for all future operations. + void BindTitleID(u64 title_id); + /// Loads transferable cache. If file has a old version or on failure, it deletes the file. std::optional<std::vector<ShaderDiskCacheEntry>> LoadTransferable(); @@ -96,10 +96,10 @@ public: private: /// Loads the transferable cache. Returns empty on failure. std::optional<std::vector<ShaderDiskCachePrecompiled>> LoadPrecompiledFile( - FileUtil::IOFile& file); + Common::FS::IOFile& file); /// Opens current game's transferable file and write it's header if it doesn't exist - FileUtil::IOFile AppendTransferableFile() const; + Common::FS::IOFile AppendTransferableFile() const; /// Save precompiled header to precompiled_cache_in_memory void SavePrecompiledHeaderToVirtualPrecompiledCache(); @@ -156,8 +156,6 @@ private: return LoadArrayFromPrecompiled(&object, 1); } - Core::System& system; - // Stores whole precompiled cache which will be read from or saved to the precompiled chache // file FileSys::VectorVfsFile precompiled_cache_virtual_file; @@ -167,8 +165,11 @@ private: // Stored transferable shaders std::unordered_set<u64> stored_transferable; + /// Title ID to operate on + u64 title_id = 0; + // The cache has been loaded at boot - bool is_usable{}; + bool is_usable = false; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_manager.cpp b/src/video_core/renderer_opengl/gl_shader_manager.cpp index 9c7b0adbd..691c6c79b 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.cpp +++ b/src/video_core/renderer_opengl/gl_shader_manager.cpp @@ -6,47 +6,124 @@ #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_shader_manager.h" -namespace OpenGL::GLShader { +namespace OpenGL { -ProgramManager::ProgramManager() = default; +namespace { + +void BindProgram(GLenum stage, GLuint current, GLuint old, bool& enabled) { + if (current == old) { + return; + } + if (current == 0) { + if (enabled) { + enabled = false; + glDisable(stage); + } + return; + } + if (!enabled) { + enabled = true; + glEnable(stage); + } + glBindProgramARB(stage, current); +} + +} // Anonymous namespace + +ProgramManager::ProgramManager(const Device& device) + : use_assembly_programs{device.UseAssemblyShaders()} { + if (use_assembly_programs) { + glEnable(GL_COMPUTE_PROGRAM_NV); + } else { + graphics_pipeline.Create(); + glBindProgramPipeline(graphics_pipeline.handle); + } +} ProgramManager::~ProgramManager() = default; -void ProgramManager::Create() { - graphics_pipeline.Create(); - glBindProgramPipeline(graphics_pipeline.handle); +void ProgramManager::BindCompute(GLuint program) { + if (use_assembly_programs) { + glBindProgramARB(GL_COMPUTE_PROGRAM_NV, program); + } else { + is_graphics_bound = false; + glUseProgram(program); + } } void ProgramManager::BindGraphicsPipeline() { + if (!use_assembly_programs) { + UpdateSourcePrograms(); + } +} + +void ProgramManager::BindHostPipeline(GLuint pipeline) { + if (use_assembly_programs) { + if (geometry_enabled) { + geometry_enabled = false; + old_state.geometry = 0; + glDisable(GL_GEOMETRY_PROGRAM_NV); + } + } else { + if (!is_graphics_bound) { + glUseProgram(0); + } + } + glBindProgramPipeline(pipeline); +} + +void ProgramManager::RestoreGuestPipeline() { + if (use_assembly_programs) { + glBindProgramPipeline(0); + } else { + glBindProgramPipeline(graphics_pipeline.handle); + } +} + +void ProgramManager::UseVertexShader(GLuint program) { + if (use_assembly_programs) { + BindProgram(GL_VERTEX_PROGRAM_NV, program, current_state.vertex, vertex_enabled); + } + current_state.vertex = program; +} + +void ProgramManager::UseGeometryShader(GLuint program) { + if (use_assembly_programs) { + BindProgram(GL_GEOMETRY_PROGRAM_NV, program, current_state.vertex, geometry_enabled); + } + current_state.geometry = program; +} + +void ProgramManager::UseFragmentShader(GLuint program) { + if (use_assembly_programs) { + BindProgram(GL_FRAGMENT_PROGRAM_NV, program, current_state.vertex, fragment_enabled); + } + current_state.fragment = program; +} + +void ProgramManager::UpdateSourcePrograms() { if (!is_graphics_bound) { is_graphics_bound = true; glUseProgram(0); } - // Avoid updating the pipeline when values have no changed - if (old_state == current_state) { - return; - } - - // Workaround for AMD bug - static constexpr GLenum all_used_stages{GL_VERTEX_SHADER_BIT | GL_GEOMETRY_SHADER_BIT | - GL_FRAGMENT_SHADER_BIT}; const GLuint handle = graphics_pipeline.handle; - glUseProgramStages(handle, all_used_stages, 0); - glUseProgramStages(handle, GL_VERTEX_SHADER_BIT, current_state.vertex_shader); - glUseProgramStages(handle, GL_GEOMETRY_SHADER_BIT, current_state.geometry_shader); - glUseProgramStages(handle, GL_FRAGMENT_SHADER_BIT, current_state.fragment_shader); + const auto update_state = [handle](GLenum stage, GLuint current, GLuint old) { + if (current == old) { + return; + } + glUseProgramStages(handle, stage, current); + }; + update_state(GL_VERTEX_SHADER_BIT, current_state.vertex, old_state.vertex); + update_state(GL_GEOMETRY_SHADER_BIT, current_state.geometry, old_state.geometry); + update_state(GL_FRAGMENT_SHADER_BIT, current_state.fragment, old_state.fragment); old_state = current_state; } -void ProgramManager::BindComputeShader(GLuint program) { - is_graphics_bound = false; - glUseProgram(program); -} - void MaxwellUniformData::SetFromRegs(const Tegra::Engines::Maxwell3D& maxwell) { const auto& regs = maxwell.regs; @@ -54,4 +131,4 @@ void MaxwellUniformData::SetFromRegs(const Tegra::Engines::Maxwell3D& maxwell) { y_direction = regs.screen_y_control.y_negate == 0 ? 1.0f : -1.0f; } -} // namespace OpenGL::GLShader +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h index d2e47f2a9..950e0dfcb 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.h +++ b/src/video_core/renderer_opengl/gl_shader_manager.h @@ -11,7 +11,9 @@ #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/maxwell_to_gl.h" -namespace OpenGL::GLShader { +namespace OpenGL { + +class Device; /// Uniform structure for the Uniform Buffer Object, all vectors must be 16-byte aligned /// @note Always keep a vec4 at the end. The GL spec is not clear whether the alignment at @@ -28,50 +30,47 @@ static_assert(sizeof(MaxwellUniformData) < 16384, class ProgramManager { public: - explicit ProgramManager(); + explicit ProgramManager(const Device& device); ~ProgramManager(); - void Create(); + /// Binds a compute program + void BindCompute(GLuint program); - /// Updates the graphics pipeline and binds it. + /// Updates bound programs. void BindGraphicsPipeline(); - /// Binds a compute shader. - void BindComputeShader(GLuint program); - - void UseVertexShader(GLuint program) { - current_state.vertex_shader = program; - } + /// Binds an OpenGL pipeline object unsynchronized with the guest state. + void BindHostPipeline(GLuint pipeline); - void UseGeometryShader(GLuint program) { - current_state.geometry_shader = program; - } + /// Rewinds BindHostPipeline state changes. + void RestoreGuestPipeline(); - void UseFragmentShader(GLuint program) { - current_state.fragment_shader = program; - } + void UseVertexShader(GLuint program); + void UseGeometryShader(GLuint program); + void UseFragmentShader(GLuint program); private: struct PipelineState { - bool operator==(const PipelineState& rhs) const noexcept { - return vertex_shader == rhs.vertex_shader && fragment_shader == rhs.fragment_shader && - geometry_shader == rhs.geometry_shader; - } - - bool operator!=(const PipelineState& rhs) const noexcept { - return !operator==(rhs); - } - - GLuint vertex_shader = 0; - GLuint fragment_shader = 0; - GLuint geometry_shader = 0; + GLuint vertex = 0; + GLuint geometry = 0; + GLuint fragment = 0; }; + /// Update GLSL programs. + void UpdateSourcePrograms(); + OGLPipeline graphics_pipeline; - OGLPipeline compute_pipeline; + PipelineState current_state; PipelineState old_state; + + bool use_assembly_programs = false; + bool is_graphics_bound = true; + + bool vertex_enabled = false; + bool geometry_enabled = false; + bool fragment_enabled = false; }; -} // namespace OpenGL::GLShader +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_util.cpp b/src/video_core/renderer_opengl/gl_shader_util.cpp index 9e74eda0d..4bf0d6090 100644 --- a/src/video_core/renderer_opengl/gl_shader_util.cpp +++ b/src/video_core/renderer_opengl/gl_shader_util.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <string_view> #include <vector> #include <glad/glad.h> #include "common/assert.h" @@ -11,7 +12,8 @@ namespace OpenGL::GLShader { namespace { -const char* GetStageDebugName(GLenum type) { + +std::string_view StageDebugName(GLenum type) { switch (type) { case GL_VERTEX_SHADER: return "vertex"; @@ -25,12 +27,17 @@ const char* GetStageDebugName(GLenum type) { UNIMPLEMENTED(); return "unknown"; } + } // Anonymous namespace -GLuint LoadShader(const char* source, GLenum type) { - const char* debug_type = GetStageDebugName(type); +GLuint LoadShader(std::string_view source, GLenum type) { + const std::string_view debug_type = StageDebugName(type); const GLuint shader_id = glCreateShader(type); - glShaderSource(shader_id, 1, &source, nullptr); + + const GLchar* source_string = source.data(); + const GLint source_length = static_cast<GLint>(source.size()); + + glShaderSource(shader_id, 1, &source_string, &source_length); LOG_DEBUG(Render_OpenGL, "Compiling {} shader...", debug_type); glCompileShader(shader_id); diff --git a/src/video_core/renderer_opengl/gl_shader_util.h b/src/video_core/renderer_opengl/gl_shader_util.h index 03b7548c2..1b770532e 100644 --- a/src/video_core/renderer_opengl/gl_shader_util.h +++ b/src/video_core/renderer_opengl/gl_shader_util.h @@ -38,7 +38,7 @@ void LogShaderSource(T... shaders) { * @param source String of the GLSL shader program * @param type Type of the shader (GL_VERTEX_SHADER, GL_GEOMETRY_SHADER or GL_FRAGMENT_SHADER) */ -GLuint LoadShader(const char* source, GLenum type); +GLuint LoadShader(std::string_view source, GLenum type); /** * Utility function to create and compile an OpenGL GLSL shader program (vertex + fragment shader) diff --git a/src/video_core/renderer_opengl/gl_state_tracker.cpp b/src/video_core/renderer_opengl/gl_state_tracker.cpp index d24fad3de..6bcf831f2 100644 --- a/src/video_core/renderer_opengl/gl_state_tracker.cpp +++ b/src/video_core/renderer_opengl/gl_state_tracker.cpp @@ -214,10 +214,8 @@ void SetupDirtyMisc(Tables& tables) { } // Anonymous namespace -StateTracker::StateTracker(Core::System& system) : system{system} {} - -void StateTracker::Initialize() { - auto& dirty = system.GPU().Maxwell3D().dirty; +StateTracker::StateTracker(Tegra::GPU& gpu) : flags{gpu.Maxwell3D().dirty.flags} { + auto& dirty = gpu.Maxwell3D().dirty; auto& tables = dirty.tables; SetupDirtyRenderTargets(tables); SetupDirtyColorMasks(tables); diff --git a/src/video_core/renderer_opengl/gl_state_tracker.h b/src/video_core/renderer_opengl/gl_state_tracker.h index 0f823288e..9d127548f 100644 --- a/src/video_core/renderer_opengl/gl_state_tracker.h +++ b/src/video_core/renderer_opengl/gl_state_tracker.h @@ -13,8 +13,8 @@ #include "video_core/dirty_flags.h" #include "video_core/engines/maxwell_3d.h" -namespace Core { -class System; +namespace Tegra { +class GPU; } namespace OpenGL { @@ -90,9 +90,7 @@ static_assert(Last <= std::numeric_limits<u8>::max()); class StateTracker { public: - explicit StateTracker(Core::System& system); - - void Initialize(); + explicit StateTracker(Tegra::GPU& gpu); void BindIndexBuffer(GLuint new_index_buffer) { if (index_buffer == new_index_buffer) { @@ -103,7 +101,6 @@ public: } void NotifyScreenDrawVertexArray() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::VertexFormats] = true; flags[OpenGL::Dirty::VertexFormat0 + 0] = true; flags[OpenGL::Dirty::VertexFormat0 + 1] = true; @@ -117,98 +114,81 @@ public: } void NotifyPolygonModes() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::PolygonModes] = true; flags[OpenGL::Dirty::PolygonModeFront] = true; flags[OpenGL::Dirty::PolygonModeBack] = true; } void NotifyViewport0() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::Viewports] = true; flags[OpenGL::Dirty::Viewport0] = true; } void NotifyScissor0() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::Scissors] = true; flags[OpenGL::Dirty::Scissor0] = true; } void NotifyColorMask0() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::ColorMasks] = true; flags[OpenGL::Dirty::ColorMask0] = true; } void NotifyBlend0() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::BlendStates] = true; flags[OpenGL::Dirty::BlendState0] = true; } void NotifyFramebuffer() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[VideoCommon::Dirty::RenderTargets] = true; } void NotifyFrontFace() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::FrontFace] = true; } void NotifyCullTest() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::CullTest] = true; } void NotifyDepthMask() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::DepthMask] = true; } void NotifyDepthTest() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::DepthTest] = true; } void NotifyStencilTest() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::StencilTest] = true; } void NotifyPolygonOffset() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::PolygonOffset] = true; } void NotifyRasterizeEnable() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::RasterizeEnable] = true; } void NotifyFramebufferSRGB() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::FramebufferSRGB] = true; } void NotifyLogicOp() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::LogicOp] = true; } void NotifyClipControl() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::ClipControl] = true; } void NotifyAlphaTest() { - auto& flags = system.GPU().Maxwell3D().dirty.flags; flags[OpenGL::Dirty::AlphaTest] = true; } private: - Core::System& system; + Tegra::Engines::Maxwell3D::DirtyState::Flags& flags; GLuint index_buffer = 0; }; diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp index 6ec328c53..887995cf4 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp +++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp @@ -2,11 +2,13 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <deque> +#include <tuple> #include <vector> + #include "common/alignment.h" #include "common/assert.h" #include "common/microprofile.h" +#include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_stream_buffer.h" MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning", @@ -14,8 +16,7 @@ MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning", namespace OpenGL { -OGLStreamBuffer::OGLStreamBuffer(GLsizeiptr size, bool vertex_data_usage, bool prefer_coherent, - bool use_persistent) +OGLStreamBuffer::OGLStreamBuffer(const Device& device, GLsizeiptr size, bool vertex_data_usage) : buffer_size(size) { gl_buffer.Create(); @@ -29,34 +30,22 @@ OGLStreamBuffer::OGLStreamBuffer(GLsizeiptr size, bool vertex_data_usage, bool p allocate_size *= 2; } - if (use_persistent) { - persistent = true; - coherent = prefer_coherent; - const GLbitfield flags = - GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | (coherent ? GL_MAP_COHERENT_BIT : 0); - glNamedBufferStorage(gl_buffer.handle, allocate_size, nullptr, flags); - mapped_ptr = static_cast<u8*>(glMapNamedBufferRange( - gl_buffer.handle, 0, buffer_size, flags | (coherent ? 0 : GL_MAP_FLUSH_EXPLICIT_BIT))); - } else { - glNamedBufferData(gl_buffer.handle, allocate_size, nullptr, GL_STREAM_DRAW); + static constexpr GLbitfield flags = GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT; + glNamedBufferStorage(gl_buffer.handle, allocate_size, nullptr, flags); + mapped_ptr = static_cast<u8*>( + glMapNamedBufferRange(gl_buffer.handle, 0, buffer_size, flags | GL_MAP_FLUSH_EXPLICIT_BIT)); + + if (device.UseAssemblyShaders() || device.HasVertexBufferUnifiedMemory()) { + glMakeNamedBufferResidentNV(gl_buffer.handle, GL_READ_ONLY); + glGetNamedBufferParameterui64vNV(gl_buffer.handle, GL_BUFFER_GPU_ADDRESS_NV, &gpu_address); } } OGLStreamBuffer::~OGLStreamBuffer() { - if (persistent) { - glUnmapNamedBuffer(gl_buffer.handle); - } + glUnmapNamedBuffer(gl_buffer.handle); gl_buffer.Release(); } -GLuint OGLStreamBuffer::GetHandle() const { - return gl_buffer.handle; -} - -GLsizeiptr OGLStreamBuffer::GetSize() const { - return buffer_size; -} - std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr alignment) { ASSERT(size <= buffer_size); ASSERT(alignment <= buffer_size); @@ -68,36 +57,21 @@ std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr a bool invalidate = false; if (buffer_pos + size > buffer_size) { + MICROPROFILE_SCOPE(OpenGL_StreamBuffer); + glInvalidateBufferData(gl_buffer.handle); + buffer_pos = 0; invalidate = true; - - if (persistent) { - glUnmapNamedBuffer(gl_buffer.handle); - } } - if (invalidate || !persistent) { - MICROPROFILE_SCOPE(OpenGL_StreamBuffer); - GLbitfield flags = GL_MAP_WRITE_BIT | (persistent ? GL_MAP_PERSISTENT_BIT : 0) | - (coherent ? GL_MAP_COHERENT_BIT : GL_MAP_FLUSH_EXPLICIT_BIT) | - (invalidate ? GL_MAP_INVALIDATE_BUFFER_BIT : GL_MAP_UNSYNCHRONIZED_BIT); - mapped_ptr = static_cast<u8*>( - glMapNamedBufferRange(gl_buffer.handle, buffer_pos, buffer_size - buffer_pos, flags)); - mapped_offset = buffer_pos; - } - - return std::make_tuple(mapped_ptr + buffer_pos - mapped_offset, buffer_pos, invalidate); + return std::make_tuple(mapped_ptr + buffer_pos, buffer_pos, invalidate); } void OGLStreamBuffer::Unmap(GLsizeiptr size) { ASSERT(size <= mapped_size); - if (!coherent && size > 0) { - glFlushMappedNamedBufferRange(gl_buffer.handle, buffer_pos - mapped_offset, size); - } - - if (!persistent) { - glUnmapNamedBuffer(gl_buffer.handle); + if (size > 0) { + glFlushMappedNamedBufferRange(gl_buffer.handle, buffer_pos, size); } buffer_pos += size; diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.h b/src/video_core/renderer_opengl/gl_stream_buffer.h index f8383cbd4..307a67113 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.h +++ b/src/video_core/renderer_opengl/gl_stream_buffer.h @@ -11,15 +11,13 @@ namespace OpenGL { +class Device; + class OGLStreamBuffer : private NonCopyable { public: - explicit OGLStreamBuffer(GLsizeiptr size, bool vertex_data_usage, bool prefer_coherent = false, - bool use_persistent = true); + explicit OGLStreamBuffer(const Device& device, GLsizeiptr size, bool vertex_data_usage); ~OGLStreamBuffer(); - GLuint GetHandle() const; - GLsizeiptr GetSize() const; - /* * Allocates a linear chunk of memory in the GPU buffer with at least "size" bytes * and the optional alignment requirement. @@ -32,15 +30,24 @@ public: void Unmap(GLsizeiptr size); + GLuint Handle() const { + return gl_buffer.handle; + } + + u64 Address() const { + return gpu_address; + } + + GLsizeiptr Size() const noexcept { + return buffer_size; + } + private: OGLBuffer gl_buffer; - bool coherent = false; - bool persistent = false; - + GLuint64EXT gpu_address = 0; GLintptr buffer_pos = 0; GLsizeiptr buffer_size = 0; - GLintptr mapped_offset = 0; GLsizeiptr mapped_size = 0; u8* mapped_ptr = nullptr; }; diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index 2729d1265..a863ef218 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -35,96 +35,109 @@ MICROPROFILE_DEFINE(OpenGL_Texture_Buffer_Copy, "OpenGL", "Texture Buffer Copy", namespace { struct FormatTuple { - GLint internal_format; + GLenum internal_format; GLenum format = GL_NONE; GLenum type = GL_NONE; }; constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{ - {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV}, // ABGR8U - {GL_RGBA8_SNORM, GL_RGBA, GL_BYTE}, // ABGR8S - {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE}, // ABGR8UI - {GL_RGB565, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV}, // B5G6R5U - {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV}, // A2B10G10R10U - {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV}, // A1B5G5R5U - {GL_R8, GL_RED, GL_UNSIGNED_BYTE}, // R8U - {GL_R8UI, GL_RED_INTEGER, GL_UNSIGNED_BYTE}, // R8UI - {GL_RGBA16F, GL_RGBA, GL_HALF_FLOAT}, // RGBA16F - {GL_RGBA16, GL_RGBA, GL_UNSIGNED_SHORT}, // RGBA16U - {GL_RGBA16_SNORM, GL_RGBA, GL_SHORT}, // RGBA16S - {GL_RGBA16UI, GL_RGBA_INTEGER, GL_UNSIGNED_SHORT}, // RGBA16UI - {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV}, // R11FG11FB10F - {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT}, // RGBA32UI - {GL_COMPRESSED_RGBA_S3TC_DXT1_EXT}, // DXT1 - {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT}, // DXT23 - {GL_COMPRESSED_RGBA_S3TC_DXT5_EXT}, // DXT45 - {GL_COMPRESSED_RED_RGTC1}, // DXN1 - {GL_COMPRESSED_RG_RGTC2}, // DXN2UNORM - {GL_COMPRESSED_SIGNED_RG_RGTC2}, // DXN2SNORM - {GL_COMPRESSED_RGBA_BPTC_UNORM}, // BC7U - {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT}, // BC6H_UF16 - {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT}, // BC6H_SF16 - {GL_COMPRESSED_RGBA_ASTC_4x4_KHR}, // ASTC_2D_4X4 - {GL_RGBA8, GL_BGRA, GL_UNSIGNED_BYTE}, // BGRA8 - {GL_RGBA32F, GL_RGBA, GL_FLOAT}, // RGBA32F - {GL_RG32F, GL_RG, GL_FLOAT}, // RG32F - {GL_R32F, GL_RED, GL_FLOAT}, // R32F - {GL_R16F, GL_RED, GL_HALF_FLOAT}, // R16F - {GL_R16, GL_RED, GL_UNSIGNED_SHORT}, // R16U - {GL_R16_SNORM, GL_RED, GL_SHORT}, // R16S - {GL_R16UI, GL_RED_INTEGER, GL_UNSIGNED_SHORT}, // R16UI - {GL_R16I, GL_RED_INTEGER, GL_SHORT}, // R16I - {GL_RG16, GL_RG, GL_UNSIGNED_SHORT}, // RG16 - {GL_RG16F, GL_RG, GL_HALF_FLOAT}, // RG16F - {GL_RG16UI, GL_RG_INTEGER, GL_UNSIGNED_SHORT}, // RG16UI - {GL_RG16I, GL_RG_INTEGER, GL_SHORT}, // RG16I - {GL_RG16_SNORM, GL_RG, GL_SHORT}, // RG16S - {GL_RGB32F, GL_RGB, GL_FLOAT}, // RGB32F - {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV}, // RGBA8_SRGB - {GL_RG8, GL_RG, GL_UNSIGNED_BYTE}, // RG8U - {GL_RG8_SNORM, GL_RG, GL_BYTE}, // RG8S - {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT}, // RG32UI - {GL_RGB16F, GL_RGBA, GL_HALF_FLOAT}, // RGBX16F - {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT}, // R32UI - {GL_R32I, GL_RED_INTEGER, GL_INT}, // R32I - {GL_COMPRESSED_RGBA_ASTC_8x8_KHR}, // ASTC_2D_8X8 - {GL_COMPRESSED_RGBA_ASTC_8x5_KHR}, // ASTC_2D_8X5 - {GL_COMPRESSED_RGBA_ASTC_5x4_KHR}, // ASTC_2D_5X4 - {GL_SRGB8_ALPHA8, GL_BGRA, GL_UNSIGNED_BYTE}, // BGRA8 + {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV}, // A8B8G8R8_UNORM + {GL_RGBA8_SNORM, GL_RGBA, GL_BYTE}, // A8B8G8R8_SNORM + {GL_RGBA8I, GL_RGBA_INTEGER, GL_BYTE}, // A8B8G8R8_SINT + {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE}, // A8B8G8R8_UINT + {GL_RGB565, GL_RGB, GL_UNSIGNED_SHORT_5_6_5}, // R5G6B5_UNORM + {GL_RGB565, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV}, // B5G6R5_UNORM + {GL_RGB5_A1, GL_BGRA, GL_UNSIGNED_SHORT_1_5_5_5_REV}, // A1R5G5B5_UNORM + {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV}, // A2B10G10R10_UNORM + {GL_RGB10_A2UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT_2_10_10_10_REV}, // A2B10G10R10_UINT + {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV}, // A1B5G5R5_UNORM + {GL_R8, GL_RED, GL_UNSIGNED_BYTE}, // R8_UNORM + {GL_R8_SNORM, GL_RED, GL_BYTE}, // R8_SNORM + {GL_R8I, GL_RED_INTEGER, GL_BYTE}, // R8_SINT + {GL_R8UI, GL_RED_INTEGER, GL_UNSIGNED_BYTE}, // R8_UINT + {GL_RGBA16F, GL_RGBA, GL_HALF_FLOAT}, // R16G16B16A16_FLOAT + {GL_RGBA16, GL_RGBA, GL_UNSIGNED_SHORT}, // R16G16B16A16_UNORM + {GL_RGBA16_SNORM, GL_RGBA, GL_SHORT}, // R16G16B16A16_SNORM + {GL_RGBA16I, GL_RGBA_INTEGER, GL_SHORT}, // R16G16B16A16_SINT + {GL_RGBA16UI, GL_RGBA_INTEGER, GL_UNSIGNED_SHORT}, // R16G16B16A16_UINT + {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV}, // B10G11R11_FLOAT + {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT}, // R32G32B32A32_UINT + {GL_COMPRESSED_RGBA_S3TC_DXT1_EXT}, // BC1_RGBA_UNORM + {GL_COMPRESSED_RGBA_S3TC_DXT3_EXT}, // BC2_UNORM + {GL_COMPRESSED_RGBA_S3TC_DXT5_EXT}, // BC3_UNORM + {GL_COMPRESSED_RED_RGTC1}, // BC4_UNORM + {GL_COMPRESSED_SIGNED_RED_RGTC1}, // BC4_SNORM + {GL_COMPRESSED_RG_RGTC2}, // BC5_UNORM + {GL_COMPRESSED_SIGNED_RG_RGTC2}, // BC5_SNORM + {GL_COMPRESSED_RGBA_BPTC_UNORM}, // BC7_UNORM + {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT}, // BC6H_UFLOAT + {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT}, // BC6H_SFLOAT + {GL_COMPRESSED_RGBA_ASTC_4x4_KHR}, // ASTC_2D_4X4_UNORM + {GL_RGBA8, GL_BGRA, GL_UNSIGNED_BYTE}, // B8G8R8A8_UNORM + {GL_RGBA32F, GL_RGBA, GL_FLOAT}, // R32G32B32A32_FLOAT + {GL_RGBA32I, GL_RGBA_INTEGER, GL_INT}, // R32G32B32A32_SINT + {GL_RG32F, GL_RG, GL_FLOAT}, // R32G32_FLOAT + {GL_RG32I, GL_RG_INTEGER, GL_INT}, // R32G32_SINT + {GL_R32F, GL_RED, GL_FLOAT}, // R32_FLOAT + {GL_R16F, GL_RED, GL_HALF_FLOAT}, // R16_FLOAT + {GL_R16, GL_RED, GL_UNSIGNED_SHORT}, // R16_UNORM + {GL_R16_SNORM, GL_RED, GL_SHORT}, // R16_SNORM + {GL_R16UI, GL_RED_INTEGER, GL_UNSIGNED_SHORT}, // R16_UINT + {GL_R16I, GL_RED_INTEGER, GL_SHORT}, // R16_SINT + {GL_RG16, GL_RG, GL_UNSIGNED_SHORT}, // R16G16_UNORM + {GL_RG16F, GL_RG, GL_HALF_FLOAT}, // R16G16_FLOAT + {GL_RG16UI, GL_RG_INTEGER, GL_UNSIGNED_SHORT}, // R16G16_UINT + {GL_RG16I, GL_RG_INTEGER, GL_SHORT}, // R16G16_SINT + {GL_RG16_SNORM, GL_RG, GL_SHORT}, // R16G16_SNORM + {GL_RGB32F, GL_RGB, GL_FLOAT}, // R32G32B32_FLOAT + {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV}, // A8B8G8R8_SRGB + {GL_RG8, GL_RG, GL_UNSIGNED_BYTE}, // R8G8_UNORM + {GL_RG8_SNORM, GL_RG, GL_BYTE}, // R8G8_SNORM + {GL_RG8I, GL_RG_INTEGER, GL_BYTE}, // R8G8_SINT + {GL_RG8UI, GL_RG_INTEGER, GL_UNSIGNED_BYTE}, // R8G8_UINT + {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT}, // R32G32_UINT + {GL_RGB16F, GL_RGBA, GL_HALF_FLOAT}, // R16G16B16X16_FLOAT + {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT}, // R32_UINT + {GL_R32I, GL_RED_INTEGER, GL_INT}, // R32_SINT + {GL_COMPRESSED_RGBA_ASTC_8x8_KHR}, // ASTC_2D_8X8_UNORM + {GL_COMPRESSED_RGBA_ASTC_8x5_KHR}, // ASTC_2D_8X5_UNORM + {GL_COMPRESSED_RGBA_ASTC_5x4_KHR}, // ASTC_2D_5X4_UNORM + {GL_SRGB8_ALPHA8, GL_BGRA, GL_UNSIGNED_BYTE}, // B8G8R8A8_UNORM // Compressed sRGB formats - {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT}, // DXT1_SRGB - {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT}, // DXT23_SRGB - {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT}, // DXT45_SRGB - {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM}, // BC7U_SRGB - {GL_RGBA4, GL_RGBA, GL_UNSIGNED_SHORT_4_4_4_4_REV}, // R4G4B4A4U + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT}, // BC1_RGBA_SRGB + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT}, // BC2_SRGB + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT}, // BC3_SRGB + {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM}, // BC7_SRGB + {GL_RGBA4, GL_RGBA, GL_UNSIGNED_SHORT_4_4_4_4_REV}, // A4B4G4R4_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_4x4_KHR}, // ASTC_2D_4X4_SRGB {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x8_KHR}, // ASTC_2D_8X8_SRGB {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x5_KHR}, // ASTC_2D_8X5_SRGB {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x4_KHR}, // ASTC_2D_5X4_SRGB - {GL_COMPRESSED_RGBA_ASTC_5x5_KHR}, // ASTC_2D_5X5 + {GL_COMPRESSED_RGBA_ASTC_5x5_KHR}, // ASTC_2D_5X5_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x5_KHR}, // ASTC_2D_5X5_SRGB - {GL_COMPRESSED_RGBA_ASTC_10x8_KHR}, // ASTC_2D_10X8 + {GL_COMPRESSED_RGBA_ASTC_10x8_KHR}, // ASTC_2D_10X8_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x8_KHR}, // ASTC_2D_10X8_SRGB - {GL_COMPRESSED_RGBA_ASTC_6x6_KHR}, // ASTC_2D_6X6 + {GL_COMPRESSED_RGBA_ASTC_6x6_KHR}, // ASTC_2D_6X6_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR}, // ASTC_2D_6X6_SRGB - {GL_COMPRESSED_RGBA_ASTC_10x10_KHR}, // ASTC_2D_10X10 + {GL_COMPRESSED_RGBA_ASTC_10x10_KHR}, // ASTC_2D_10X10_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR}, // ASTC_2D_10X10_SRGB - {GL_COMPRESSED_RGBA_ASTC_12x12_KHR}, // ASTC_2D_12X12 + {GL_COMPRESSED_RGBA_ASTC_12x12_KHR}, // ASTC_2D_12X12_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR}, // ASTC_2D_12X12_SRGB - {GL_COMPRESSED_RGBA_ASTC_8x6_KHR}, // ASTC_2D_8X6 + {GL_COMPRESSED_RGBA_ASTC_8x6_KHR}, // ASTC_2D_8X6_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x6_KHR}, // ASTC_2D_8X6_SRGB - {GL_COMPRESSED_RGBA_ASTC_6x5_KHR}, // ASTC_2D_6X5 + {GL_COMPRESSED_RGBA_ASTC_6x5_KHR}, // ASTC_2D_6X5_UNORM {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x5_KHR}, // ASTC_2D_6X5_SRGB - {GL_RGB9_E5, GL_RGB, GL_UNSIGNED_INT_5_9_9_9_REV}, // E5B9G9R9F + {GL_RGB9_E5, GL_RGB, GL_UNSIGNED_INT_5_9_9_9_REV}, // E5B9G9R9_FLOAT // Depth formats - {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT}, // Z32F - {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT}, // Z16 + {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT}, // D32_FLOAT + {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT}, // D16_UNORM // DepthStencil formats - {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // Z24S8 - {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // S8Z24 - {GL_DEPTH32F_STENCIL8, GL_DEPTH_STENCIL, GL_FLOAT_32_UNSIGNED_INT_24_8_REV}, // Z32FS8 + {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // D24_UNORM_S8_UINT + {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // S8_UINT_D24_UNORM + {GL_DEPTH32F_STENCIL8, GL_DEPTH_STENCIL, + GL_FLOAT_32_UNSIGNED_INT_24_8_REV}, // D32_FLOAT_S8_UINT }}; const FormatTuple& GetFormatTuple(PixelFormat pixel_format) { @@ -177,10 +190,10 @@ GLint GetSwizzleSource(SwizzleSource source) { GLenum GetComponent(PixelFormat format, bool is_first) { switch (format) { - case PixelFormat::Z24S8: - case PixelFormat::Z32FS8: + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::D32_FLOAT_S8_UINT: return is_first ? GL_DEPTH_COMPONENT : GL_STENCIL_INDEX; - case PixelFormat::S8Z24: + case PixelFormat::S8_UINT_D24_UNORM: return is_first ? GL_STENCIL_INDEX : GL_DEPTH_COMPONENT; default: UNREACHABLE(); @@ -237,6 +250,12 @@ OGLTexture CreateTexture(const SurfaceParams& params, GLenum target, GLenum inte return texture; } +constexpr u32 EncodeSwizzle(SwizzleSource x_source, SwizzleSource y_source, SwizzleSource z_source, + SwizzleSource w_source) { + return (static_cast<u32>(x_source) << 24) | (static_cast<u32>(y_source) << 16) | + (static_cast<u32>(z_source) << 8) | static_cast<u32>(w_source); +} + } // Anonymous namespace CachedSurface::CachedSurface(const GPUVAddr gpu_addr, const SurfaceParams& params, @@ -256,9 +275,14 @@ CachedSurface::CachedSurface(const GPUVAddr gpu_addr, const SurfaceParams& param target = GetTextureTarget(params.target); texture = CreateTexture(params, target, internal_format, texture_buffer); DecorateSurfaceName(); - main_view = CreateViewInner( - ViewParams(params.target, 0, params.is_layered ? params.depth : 1, 0, params.num_levels), - true); + + u32 num_layers = 1; + if (params.is_layered || params.target == SurfaceTarget::Texture3D) { + num_layers = params.depth; + } + + main_view = + CreateViewInner(ViewParams(params.target, 0, num_layers, 0, params.num_levels), true); } CachedSurface::~CachedSurface() = default; @@ -379,8 +403,8 @@ void CachedSurface::DecorateSurfaceName() { LabelGLObject(GL_TEXTURE, texture.handle, GetGpuAddr(), params.TargetName()); } -void CachedSurfaceView::DecorateViewName(GPUVAddr gpu_addr, std::string prefix) { - LabelGLObject(GL_TEXTURE, texture_view.handle, gpu_addr, prefix); +void CachedSurfaceView::DecorateViewName(GPUVAddr gpu_addr, const std::string& prefix) { + LabelGLObject(GL_TEXTURE, main_view.handle, gpu_addr, prefix); } View CachedSurface::CreateView(const ViewParams& view_key) { @@ -396,32 +420,33 @@ View CachedSurface::CreateViewInner(const ViewParams& view_key, const bool is_pr } CachedSurfaceView::CachedSurfaceView(CachedSurface& surface, const ViewParams& params, - const bool is_proxy) - : VideoCommon::ViewBase(params), surface{surface}, is_proxy{is_proxy} { - target = GetTextureTarget(params.target); - format = GetFormatTuple(surface.GetSurfaceParams().pixel_format).internal_format; + bool is_proxy) + : VideoCommon::ViewBase(params), surface{surface}, format{surface.internal_format}, + target{GetTextureTarget(params.target)}, is_proxy{is_proxy} { if (!is_proxy) { - texture_view = CreateTextureView(); + main_view = CreateTextureView(); } - swizzle = EncodeSwizzle(SwizzleSource::R, SwizzleSource::G, SwizzleSource::B, SwizzleSource::A); } CachedSurfaceView::~CachedSurfaceView() = default; -void CachedSurfaceView::Attach(GLenum attachment, GLenum target) const { +void CachedSurfaceView::Attach(GLenum attachment, GLenum fb_target) const { ASSERT(params.num_levels == 1); + if (params.target == SurfaceTarget::Texture3D) { + if (params.num_layers > 1) { + ASSERT(params.base_layer == 0); + glFramebufferTexture(fb_target, attachment, surface.texture.handle, params.base_level); + } else { + glFramebufferTexture3D(fb_target, attachment, target, surface.texture.handle, + params.base_level, params.base_layer); + } + return; + } + if (params.num_layers > 1) { - // Layered framebuffer attachments UNIMPLEMENTED_IF(params.base_layer != 0); - - switch (params.target) { - case SurfaceTarget::Texture2DArray: - glFramebufferTexture(target, attachment, GetTexture(), 0); - break; - default: - UNIMPLEMENTED(); - } + glFramebufferTexture(fb_target, attachment, GetTexture(), 0); return; } @@ -429,16 +454,16 @@ void CachedSurfaceView::Attach(GLenum attachment, GLenum target) const { const GLuint texture = surface.GetTexture(); switch (surface.GetSurfaceParams().target) { case SurfaceTarget::Texture1D: - glFramebufferTexture1D(target, attachment, view_target, texture, params.base_level); + glFramebufferTexture1D(fb_target, attachment, view_target, texture, params.base_level); break; case SurfaceTarget::Texture2D: - glFramebufferTexture2D(target, attachment, view_target, texture, params.base_level); + glFramebufferTexture2D(fb_target, attachment, view_target, texture, params.base_level); break; case SurfaceTarget::Texture1DArray: case SurfaceTarget::Texture2DArray: case SurfaceTarget::TextureCubemap: case SurfaceTarget::TextureCubeArray: - glFramebufferTextureLayer(target, attachment, texture, params.base_level, + glFramebufferTextureLayer(fb_target, attachment, texture, params.base_level, params.base_layer); break; default: @@ -446,44 +471,73 @@ void CachedSurfaceView::Attach(GLenum attachment, GLenum target) const { } } -void CachedSurfaceView::ApplySwizzle(SwizzleSource x_source, SwizzleSource y_source, +GLuint CachedSurfaceView::GetTexture(SwizzleSource x_source, SwizzleSource y_source, SwizzleSource z_source, SwizzleSource w_source) { - u32 new_swizzle = EncodeSwizzle(x_source, y_source, z_source, w_source); - if (new_swizzle == swizzle) - return; - swizzle = new_swizzle; - const std::array gl_swizzle = {GetSwizzleSource(x_source), GetSwizzleSource(y_source), - GetSwizzleSource(z_source), GetSwizzleSource(w_source)}; - const GLuint handle = GetTexture(); - const PixelFormat format = surface.GetSurfaceParams().pixel_format; - switch (format) { - case PixelFormat::Z24S8: - case PixelFormat::Z32FS8: - case PixelFormat::S8Z24: - glTextureParameteri(handle, GL_DEPTH_STENCIL_TEXTURE_MODE, + if (GetSurfaceParams().IsBuffer()) { + return GetTexture(); + } + const u32 new_swizzle = EncodeSwizzle(x_source, y_source, z_source, w_source); + if (current_swizzle == new_swizzle) { + return current_view; + } + current_swizzle = new_swizzle; + + const auto [entry, is_cache_miss] = view_cache.try_emplace(new_swizzle); + OGLTextureView& view = entry->second; + if (!is_cache_miss) { + current_view = view.handle; + return view.handle; + } + view = CreateTextureView(); + current_view = view.handle; + + std::array swizzle{x_source, y_source, z_source, w_source}; + + switch (const PixelFormat format = GetSurfaceParams().pixel_format) { + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::D32_FLOAT_S8_UINT: + case PixelFormat::S8_UINT_D24_UNORM: + UNIMPLEMENTED_IF(x_source != SwizzleSource::R && x_source != SwizzleSource::G); + glTextureParameteri(view.handle, GL_DEPTH_STENCIL_TEXTURE_MODE, GetComponent(format, x_source == SwizzleSource::R)); - break; - default: - glTextureParameteriv(handle, GL_TEXTURE_SWIZZLE_RGBA, gl_swizzle.data()); + + // Make sure we sample the first component + std::transform(swizzle.begin(), swizzle.end(), swizzle.begin(), [](SwizzleSource value) { + return value == SwizzleSource::G ? SwizzleSource::R : value; + }); + [[fallthrough]]; + default: { + const std::array gl_swizzle = {GetSwizzleSource(swizzle[0]), GetSwizzleSource(swizzle[1]), + GetSwizzleSource(swizzle[2]), GetSwizzleSource(swizzle[3])}; + glTextureParameteriv(view.handle, GL_TEXTURE_SWIZZLE_RGBA, gl_swizzle.data()); break; } + } + return view.handle; } OGLTextureView CachedSurfaceView::CreateTextureView() const { OGLTextureView texture_view; texture_view.Create(); - glTextureView(texture_view.handle, target, surface.texture.handle, format, params.base_level, - params.num_levels, params.base_layer, params.num_layers); + if (target == GL_TEXTURE_3D) { + glTextureView(texture_view.handle, target, surface.texture.handle, format, + params.base_level, params.num_levels, 0, 1); + } else { + glTextureView(texture_view.handle, target, surface.texture.handle, format, + params.base_level, params.num_levels, params.base_layer, params.num_layers); + } ApplyTextureDefaults(surface.GetSurfaceParams(), texture_view.handle); return texture_view; } -TextureCacheOpenGL::TextureCacheOpenGL(Core::System& system, - VideoCore::RasterizerInterface& rasterizer, - const Device& device, StateTracker& state_tracker) - : TextureCacheBase{system, rasterizer, device.HasASTC()}, state_tracker{state_tracker} { +TextureCacheOpenGL::TextureCacheOpenGL(VideoCore::RasterizerInterface& rasterizer, + Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::MemoryManager& gpu_memory, const Device& device, + StateTracker& state_tracker_) + : TextureCacheBase{rasterizer, maxwell3d, gpu_memory, device.HasASTC()}, state_tracker{ + state_tracker_} { src_framebuffer.Create(); dst_framebuffer.Create(); } @@ -517,8 +571,8 @@ void TextureCacheOpenGL::ImageBlit(View& src_view, View& dst_view, const Tegra::Engines::Fermi2D::Config& copy_config) { const auto& src_params{src_view->GetSurfaceParams()}; const auto& dst_params{dst_view->GetSurfaceParams()}; - UNIMPLEMENTED_IF(src_params.target == SurfaceTarget::Texture3D); - UNIMPLEMENTED_IF(dst_params.target == SurfaceTarget::Texture3D); + UNIMPLEMENTED_IF(src_params.depth != 1); + UNIMPLEMENTED_IF(dst_params.depth != 1); state_tracker.NotifyScissor0(); state_tracker.NotifyFramebuffer(); diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index 02d9981a1..7787134fc 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -80,15 +80,17 @@ public: explicit CachedSurfaceView(CachedSurface& surface, const ViewParams& params, bool is_proxy); ~CachedSurfaceView(); - /// Attaches this texture view to the current bound GL_DRAW_FRAMEBUFFER - void Attach(GLenum attachment, GLenum target) const; + /// @brief Attaches this texture view to the currently bound fb_target framebuffer + /// @param attachment Attachment to bind textures to + /// @param fb_target Framebuffer target to attach to (e.g. DRAW_FRAMEBUFFER) + void Attach(GLenum attachment, GLenum fb_target) const; - void ApplySwizzle(Tegra::Texture::SwizzleSource x_source, + GLuint GetTexture(Tegra::Texture::SwizzleSource x_source, Tegra::Texture::SwizzleSource y_source, Tegra::Texture::SwizzleSource z_source, Tegra::Texture::SwizzleSource w_source); - void DecorateViewName(GPUVAddr gpu_addr, std::string prefix); + void DecorateViewName(GPUVAddr gpu_addr, const std::string& prefix); void MarkAsModified(u64 tick) { surface.MarkAsModified(true, tick); @@ -98,7 +100,7 @@ public: if (is_proxy) { return surface.GetTexture(); } - return texture_view.handle; + return main_view.handle; } GLenum GetFormat() const { @@ -110,29 +112,27 @@ public: } private: - u32 EncodeSwizzle(Tegra::Texture::SwizzleSource x_source, - Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, - Tegra::Texture::SwizzleSource w_source) const { - return (static_cast<u32>(x_source) << 24) | (static_cast<u32>(y_source) << 16) | - (static_cast<u32>(z_source) << 8) | static_cast<u32>(w_source); - } - OGLTextureView CreateTextureView() const; CachedSurface& surface; - GLenum target{}; - GLenum format{}; + const GLenum format; + const GLenum target; + const bool is_proxy; + + std::unordered_map<u32, OGLTextureView> view_cache; + OGLTextureView main_view; - OGLTextureView texture_view; - u32 swizzle{}; - bool is_proxy{}; + // Use an invalid default so it always fails the comparison test + u32 current_swizzle = 0xffffffff; + GLuint current_view = 0; }; class TextureCacheOpenGL final : public TextureCacheBase { public: - explicit TextureCacheOpenGL(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - const Device& device, StateTracker& state_tracker); + explicit TextureCacheOpenGL(VideoCore::RasterizerInterface& rasterizer, + Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::MemoryManager& gpu_memory, const Device& device, + StateTracker& state_tracker); ~TextureCacheOpenGL(); protected: diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 89f0e04ef..a8be2aa37 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -24,10 +24,11 @@ namespace MaxwellToGL { using Maxwell = Tegra::Engines::Maxwell3D::Regs; -inline GLenum VertexType(Maxwell::VertexAttribute attrib) { +inline GLenum VertexFormat(Maxwell::VertexAttribute attrib) { switch (attrib.type) { - case Maxwell::VertexAttribute::Type::UnsignedInt: case Maxwell::VertexAttribute::Type::UnsignedNorm: + case Maxwell::VertexAttribute::Type::UnsignedScaled: + case Maxwell::VertexAttribute::Type::UnsignedInt: switch (attrib.size) { case Maxwell::VertexAttribute::Size::Size_8: case Maxwell::VertexAttribute::Size::Size_8_8: @@ -47,11 +48,12 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) { case Maxwell::VertexAttribute::Size::Size_10_10_10_2: return GL_UNSIGNED_INT_2_10_10_10_REV; default: - LOG_ERROR(Render_OpenGL, "Unimplemented vertex size={}", attrib.SizeString()); - return {}; + break; } - case Maxwell::VertexAttribute::Type::SignedInt: + break; case Maxwell::VertexAttribute::Type::SignedNorm: + case Maxwell::VertexAttribute::Type::SignedScaled: + case Maxwell::VertexAttribute::Type::SignedInt: switch (attrib.size) { case Maxwell::VertexAttribute::Size::Size_8: case Maxwell::VertexAttribute::Size::Size_8_8: @@ -71,9 +73,9 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) { case Maxwell::VertexAttribute::Size::Size_10_10_10_2: return GL_INT_2_10_10_10_REV; default: - LOG_ERROR(Render_OpenGL, "Unimplemented vertex size={}", attrib.SizeString()); - return {}; + break; } + break; case Maxwell::VertexAttribute::Type::Float: switch (attrib.size) { case Maxwell::VertexAttribute::Size::Size_16: @@ -87,45 +89,13 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) { case Maxwell::VertexAttribute::Size::Size_32_32_32_32: return GL_FLOAT; default: - LOG_ERROR(Render_OpenGL, "Unimplemented vertex size={}", attrib.SizeString()); - return {}; - } - case Maxwell::VertexAttribute::Type::UnsignedScaled: - switch (attrib.size) { - case Maxwell::VertexAttribute::Size::Size_8: - case Maxwell::VertexAttribute::Size::Size_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: - return GL_UNSIGNED_BYTE; - case Maxwell::VertexAttribute::Size::Size_16: - case Maxwell::VertexAttribute::Size::Size_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return GL_UNSIGNED_SHORT; - default: - LOG_ERROR(Render_OpenGL, "Unimplemented vertex size={}", attrib.SizeString()); - return {}; - } - case Maxwell::VertexAttribute::Type::SignedScaled: - switch (attrib.size) { - case Maxwell::VertexAttribute::Size::Size_8: - case Maxwell::VertexAttribute::Size::Size_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: - return GL_BYTE; - case Maxwell::VertexAttribute::Size::Size_16: - case Maxwell::VertexAttribute::Size::Size_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return GL_SHORT; - default: - LOG_ERROR(Render_OpenGL, "Unimplemented vertex size={}", attrib.SizeString()); - return {}; + break; } - default: - LOG_ERROR(Render_OpenGL, "Unimplemented vertex type={}", attrib.TypeString()); - return {}; + break; } + UNIMPLEMENTED_MSG("Unimplemented vertex format of type={} and size={}", attrib.TypeString(), + attrib.SizeString()); + return {}; } inline GLenum IndexFormat(Maxwell::IndexFormat index_format) { @@ -137,8 +107,7 @@ inline GLenum IndexFormat(Maxwell::IndexFormat index_format) { case Maxwell::IndexFormat::UnsignedInt: return GL_UNSIGNED_INT; } - LOG_CRITICAL(Render_OpenGL, "Unimplemented index_format={}", static_cast<u32>(index_format)); - UNREACHABLE(); + UNREACHABLE_MSG("Invalid index_format={}", static_cast<u32>(index_format)); return {}; } @@ -180,31 +149,32 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) { } inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode, - Tegra::Texture::TextureMipmapFilter mip_filter_mode) { + Tegra::Texture::TextureMipmapFilter mipmap_filter_mode) { switch (filter_mode) { - case Tegra::Texture::TextureFilter::Linear: { - switch (mip_filter_mode) { + case Tegra::Texture::TextureFilter::Nearest: + switch (mipmap_filter_mode) { case Tegra::Texture::TextureMipmapFilter::None: - return GL_LINEAR; + return GL_NEAREST; case Tegra::Texture::TextureMipmapFilter::Nearest: - return GL_LINEAR_MIPMAP_NEAREST; + return GL_NEAREST_MIPMAP_NEAREST; case Tegra::Texture::TextureMipmapFilter::Linear: - return GL_LINEAR_MIPMAP_LINEAR; + return GL_NEAREST_MIPMAP_LINEAR; } - } - case Tegra::Texture::TextureFilter::Nearest: { - switch (mip_filter_mode) { + break; + case Tegra::Texture::TextureFilter::Linear: + switch (mipmap_filter_mode) { case Tegra::Texture::TextureMipmapFilter::None: - return GL_NEAREST; + return GL_LINEAR; case Tegra::Texture::TextureMipmapFilter::Nearest: - return GL_NEAREST_MIPMAP_NEAREST; + return GL_LINEAR_MIPMAP_NEAREST; case Tegra::Texture::TextureMipmapFilter::Linear: - return GL_NEAREST_MIPMAP_LINEAR; + return GL_LINEAR_MIPMAP_LINEAR; } + break; } - } - LOG_ERROR(Render_OpenGL, "Unimplemented texture filter mode={}", static_cast<u32>(filter_mode)); - return GL_LINEAR; + UNREACHABLE_MSG("Invalid texture filter mode={} and mipmap filter mode={}", + static_cast<u32>(filter_mode), static_cast<u32>(mipmap_filter_mode)); + return GL_NEAREST; } inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) { @@ -227,10 +197,15 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) { } else { return GL_MIRROR_CLAMP_TO_EDGE; } - default: - LOG_ERROR(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode)); - return GL_REPEAT; + case Tegra::Texture::WrapMode::MirrorOnceClampOGL: + if (GL_EXT_texture_mirror_clamp) { + return GL_MIRROR_CLAMP_EXT; + } else { + return GL_MIRROR_CLAMP_TO_EDGE; + } } + UNIMPLEMENTED_MSG("Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode)); + return GL_REPEAT; } inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) { @@ -252,8 +227,7 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) { case Tegra::Texture::DepthCompareFunc::Always: return GL_ALWAYS; } - LOG_ERROR(Render_OpenGL, "Unimplemented texture depth compare function ={}", - static_cast<u32>(func)); + UNIMPLEMENTED_MSG("Unimplemented texture depth compare function={}", static_cast<u32>(func)); return GL_GREATER; } @@ -275,7 +249,7 @@ inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { case Maxwell::Blend::Equation::MaxGL: return GL_MAX; } - LOG_ERROR(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation)); + UNIMPLEMENTED_MSG("Unimplemented blend equation={}", static_cast<u32>(equation)); return GL_FUNC_ADD; } @@ -339,7 +313,7 @@ inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: return GL_ONE_MINUS_CONSTANT_ALPHA; } - LOG_ERROR(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor)); + UNIMPLEMENTED_MSG("Unimplemented blend factor={}", static_cast<u32>(factor)); return GL_ZERO; } @@ -359,7 +333,7 @@ inline GLenum SwizzleSource(Tegra::Texture::SwizzleSource source) { case Tegra::Texture::SwizzleSource::OneFloat: return GL_ONE; } - LOG_ERROR(Render_OpenGL, "Unimplemented swizzle source={}", static_cast<u32>(source)); + UNIMPLEMENTED_MSG("Unimplemented swizzle source={}", static_cast<u32>(source)); return GL_ZERO; } @@ -390,7 +364,7 @@ inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) { case Maxwell::ComparisonOp::AlwaysOld: return GL_ALWAYS; } - LOG_ERROR(Render_OpenGL, "Unimplemented comparison op={}", static_cast<u32>(comparison)); + UNIMPLEMENTED_MSG("Unimplemented comparison op={}", static_cast<u32>(comparison)); return GL_ALWAYS; } @@ -421,7 +395,7 @@ inline GLenum StencilOp(Maxwell::StencilOp stencil) { case Maxwell::StencilOp::DecrWrapOGL: return GL_DECR_WRAP; } - LOG_ERROR(Render_OpenGL, "Unimplemented stencil op={}", static_cast<u32>(stencil)); + UNIMPLEMENTED_MSG("Unimplemented stencil op={}", static_cast<u32>(stencil)); return GL_KEEP; } @@ -432,7 +406,7 @@ inline GLenum FrontFace(Maxwell::FrontFace front_face) { case Maxwell::FrontFace::CounterClockWise: return GL_CCW; } - LOG_ERROR(Render_OpenGL, "Unimplemented front face cull={}", static_cast<u32>(front_face)); + UNIMPLEMENTED_MSG("Unimplemented front face cull={}", static_cast<u32>(front_face)); return GL_CCW; } @@ -445,7 +419,7 @@ inline GLenum CullFace(Maxwell::CullFace cull_face) { case Maxwell::CullFace::FrontAndBack: return GL_FRONT_AND_BACK; } - LOG_ERROR(Render_OpenGL, "Unimplemented cull face={}", static_cast<u32>(cull_face)); + UNIMPLEMENTED_MSG("Unimplemented cull face={}", static_cast<u32>(cull_face)); return GL_BACK; } @@ -484,7 +458,7 @@ inline GLenum LogicOp(Maxwell::LogicOperation operation) { case Maxwell::LogicOperation::Set: return GL_SET; } - LOG_ERROR(Render_OpenGL, "Unimplemented logic operation={}", static_cast<u32>(operation)); + UNIMPLEMENTED_MSG("Unimplemented logic operation={}", static_cast<u32>(operation)); return GL_COPY; } @@ -501,5 +475,10 @@ inline GLenum PolygonMode(Maxwell::PolygonMode polygon_mode) { return GL_FILL; } +inline GLenum ViewportSwizzle(Maxwell::ViewportSwizzle swizzle) { + // Enumeration order matches register order. We can convert it arithmetically. + return GL_VIEWPORT_SWIZZLE_POSITIVE_X_NV + static_cast<GLenum>(swizzle); +} + } // namespace MaxwellToGL } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index b2a179746..2ccca1993 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -21,6 +21,8 @@ #include "core/perf_stats.h" #include "core/settings.h" #include "core/telemetry_session.h" +#include "video_core/host_shaders/opengl_present_frag.h" +#include "video_core/host_shaders/opengl_present_vert.h" #include "video_core/morton.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_shader_manager.h" @@ -30,60 +32,6 @@ namespace OpenGL { namespace { -constexpr std::size_t SWAP_CHAIN_SIZE = 3; - -struct Frame { - u32 width{}; /// Width of the frame (to detect resize) - u32 height{}; /// Height of the frame - bool color_reloaded{}; /// Texture attachment was recreated (ie: resized) - OpenGL::OGLRenderbuffer color{}; /// Buffer shared between the render/present FBO - OpenGL::OGLFramebuffer render{}; /// FBO created on the render thread - OpenGL::OGLFramebuffer present{}; /// FBO created on the present thread - GLsync render_fence{}; /// Fence created on the render thread - GLsync present_fence{}; /// Fence created on the presentation thread - bool is_srgb{}; /// Framebuffer is sRGB or RGB -}; - -constexpr char VERTEX_SHADER[] = R"( -#version 430 core - -out gl_PerVertex { - vec4 gl_Position; -}; - -layout (location = 0) in vec2 vert_position; -layout (location = 1) in vec2 vert_tex_coord; -layout (location = 0) out vec2 frag_tex_coord; - -// This is a truncated 3x3 matrix for 2D transformations: -// The upper-left 2x2 submatrix performs scaling/rotation/mirroring. -// The third column performs translation. -// The third row could be used for projection, which we don't need in 2D. It hence is assumed to -// implicitly be [0, 0, 1] -layout (location = 0) uniform mat3x2 modelview_matrix; - -void main() { - // Multiply input position by the rotscale part of the matrix and then manually translate by - // the last column. This is equivalent to using a full 3x3 matrix and expanding the vector - // to `vec3(vert_position.xy, 1.0)` - gl_Position = vec4(mat2(modelview_matrix) * vert_position + modelview_matrix[2], 0.0, 1.0); - frag_tex_coord = vert_tex_coord; -} -)"; - -constexpr char FRAGMENT_SHADER[] = R"( -#version 430 core - -layout (location = 0) in vec2 frag_tex_coord; -layout (location = 0) out vec4 color; - -layout (binding = 0) uniform sampler2D color_texture; - -void main() { - color = vec4(texture(color_texture, frag_tex_coord).rgb, 1.0f); -} -)"; - constexpr GLint PositionLocation = 0; constexpr GLint TexCoordLocation = 1; constexpr GLint ModelViewMatrixLocation = 0; @@ -96,24 +44,6 @@ struct ScreenRectVertex { std::array<GLfloat, 2> tex_coord; }; -/// Returns true if any debug tool is attached -bool HasDebugTool() { - const bool nsight = std::getenv("NVTX_INJECTION64_PATH") || std::getenv("NSIGHT_LAUNCHED"); - if (nsight) { - return true; - } - - GLint num_extensions; - glGetIntegerv(GL_NUM_EXTENSIONS, &num_extensions); - for (GLuint index = 0; index < static_cast<GLuint>(num_extensions); ++index) { - const auto name = reinterpret_cast<const char*>(glGetStringi(GL_EXTENSIONS, index)); - if (!std::strcmp(name, "GL_EXT_debug_tool")) { - return true; - } - } - return false; -} - /** * Defines a 1:1 pixel ortographic projection matrix with (0,0) on the top-left * corner and (width, height) on the lower-bottom. @@ -197,132 +127,15 @@ void APIENTRY DebugHandler(GLenum source, GLenum type, GLuint id, GLenum severit } // Anonymous namespace -/** - * For smooth Vsync rendering, we want to always present the latest frame that the core generates, - * but also make sure that rendering happens at the pace that the frontend dictates. This is a - * helper class that the renderer uses to sync frames between the render thread and the presentation - * thread - */ -class FrameMailbox { -public: - std::mutex swap_chain_lock; - std::condition_variable present_cv; - std::array<Frame, SWAP_CHAIN_SIZE> swap_chain{}; - std::queue<Frame*> free_queue; - std::deque<Frame*> present_queue; - Frame* previous_frame{}; - - FrameMailbox() { - for (auto& frame : swap_chain) { - free_queue.push(&frame); - } - } - - ~FrameMailbox() { - // lock the mutex and clear out the present and free_queues and notify any people who are - // blocked to prevent deadlock on shutdown - std::scoped_lock lock{swap_chain_lock}; - std::queue<Frame*>().swap(free_queue); - present_queue.clear(); - present_cv.notify_all(); - } - - void ReloadPresentFrame(Frame* frame, u32 height, u32 width) { - frame->present.Release(); - frame->present.Create(); - GLint previous_draw_fbo{}; - glGetIntegerv(GL_DRAW_FRAMEBUFFER_BINDING, &previous_draw_fbo); - glBindFramebuffer(GL_FRAMEBUFFER, frame->present.handle); - glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER, - frame->color.handle); - if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE) { - LOG_CRITICAL(Render_OpenGL, "Failed to recreate present FBO!"); - } - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, previous_draw_fbo); - frame->color_reloaded = false; - } - - void ReloadRenderFrame(Frame* frame, u32 width, u32 height) { - // Recreate the color texture attachment - frame->color.Release(); - frame->color.Create(); - const GLenum internal_format = frame->is_srgb ? GL_SRGB8 : GL_RGB8; - glNamedRenderbufferStorage(frame->color.handle, internal_format, width, height); - - // Recreate the FBO for the render target - frame->render.Release(); - frame->render.Create(); - glBindFramebuffer(GL_FRAMEBUFFER, frame->render.handle); - glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER, - frame->color.handle); - if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE) { - LOG_CRITICAL(Render_OpenGL, "Failed to recreate render FBO!"); - } - - frame->width = width; - frame->height = height; - frame->color_reloaded = true; - } - - Frame* GetRenderFrame() { - std::unique_lock lock{swap_chain_lock}; - - // If theres no free frames, we will reuse the oldest render frame - if (free_queue.empty()) { - auto frame = present_queue.back(); - present_queue.pop_back(); - return frame; - } - - Frame* frame = free_queue.front(); - free_queue.pop(); - return frame; - } - - void ReleaseRenderFrame(Frame* frame) { - std::unique_lock lock{swap_chain_lock}; - present_queue.push_front(frame); - present_cv.notify_one(); - } - - Frame* TryGetPresentFrame(int timeout_ms) { - std::unique_lock lock{swap_chain_lock}; - // wait for new entries in the present_queue - present_cv.wait_for(lock, std::chrono::milliseconds(timeout_ms), - [&] { return !present_queue.empty(); }); - if (present_queue.empty()) { - // timed out waiting for a frame to draw so return the previous frame - return previous_frame; - } - - // free the previous frame and add it back to the free queue - if (previous_frame) { - free_queue.push(previous_frame); - } - - // the newest entries are pushed to the front of the queue - Frame* frame = present_queue.front(); - present_queue.pop_front(); - // remove all old entries from the present queue and move them back to the free_queue - for (auto f : present_queue) { - free_queue.push(f); - } - present_queue.clear(); - previous_frame = frame; - return frame; - } -}; - -RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& emu_window, Core::System& system, - Core::Frontend::GraphicsContext& context) - : RendererBase{emu_window}, emu_window{emu_window}, system{system}, context{context}, - has_debug_tool{HasDebugTool()} {} +RendererOpenGL::RendererOpenGL(Core::TelemetrySession& telemetry_session_, + Core::Frontend::EmuWindow& emu_window_, + Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, + std::unique_ptr<Core::Frontend::GraphicsContext> context) + : RendererBase{emu_window_, std::move(context)}, telemetry_session{telemetry_session_}, + emu_window{emu_window_}, cpu_memory{cpu_memory_}, gpu{gpu_}, program_manager{device} {} RendererOpenGL::~RendererOpenGL() = default; -MICROPROFILE_DEFINE(OpenGL_RenderFrame, "OpenGL", "Render Frame", MP_RGB(128, 128, 64)); -MICROPROFILE_DEFINE(OpenGL_WaitPresent, "OpenGL", "Wait For Present", MP_RGB(128, 128, 128)); - void RendererOpenGL::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { if (!framebuffer) { return; @@ -331,79 +144,34 @@ void RendererOpenGL::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { PrepareRendertarget(framebuffer); RenderScreenshot(); - Frame* frame; - { - MICROPROFILE_SCOPE(OpenGL_WaitPresent); + glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0); + DrawScreen(emu_window.GetFramebufferLayout()); - frame = frame_mailbox->GetRenderFrame(); + ++m_current_frame; - // Clean up sync objects before drawing - - // INTEL driver workaround. We can't delete the previous render sync object until we are - // sure that the presentation is done - if (frame->present_fence) { - glClientWaitSync(frame->present_fence, 0, GL_TIMEOUT_IGNORED); - } - - // delete the draw fence if the frame wasn't presented - if (frame->render_fence) { - glDeleteSync(frame->render_fence); - frame->render_fence = 0; - } - - // wait for the presentation to be done - if (frame->present_fence) { - glWaitSync(frame->present_fence, 0, GL_TIMEOUT_IGNORED); - glDeleteSync(frame->present_fence); - frame->present_fence = 0; - } - } - - { - MICROPROFILE_SCOPE(OpenGL_RenderFrame); - const auto& layout = render_window.GetFramebufferLayout(); - - // Recreate the frame if the size of the window has changed - if (layout.width != frame->width || layout.height != frame->height || - screen_info.display_srgb != frame->is_srgb) { - LOG_DEBUG(Render_OpenGL, "Reloading render frame"); - frame->is_srgb = screen_info.display_srgb; - frame_mailbox->ReloadRenderFrame(frame, layout.width, layout.height); - } - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, frame->render.handle); - DrawScreen(layout); - // Create a fence for the frontend to wait on and swap this frame to OffTex - frame->render_fence = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0); - glFlush(); - frame_mailbox->ReleaseRenderFrame(frame); - m_current_frame++; - rasterizer->TickFrame(); - } + rasterizer->TickFrame(); render_window.PollEvents(); - if (has_debug_tool) { - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0); - Present(0); - context.SwapBuffers(); - } + context->SwapBuffers(); } void RendererOpenGL::PrepareRendertarget(const Tegra::FramebufferConfig* framebuffer) { - if (framebuffer) { - // If framebuffer is provided, reload it from memory to a texture - if (screen_info.texture.width != static_cast<GLsizei>(framebuffer->width) || - screen_info.texture.height != static_cast<GLsizei>(framebuffer->height) || - screen_info.texture.pixel_format != framebuffer->pixel_format || - gl_framebuffer_data.empty()) { - // Reallocate texture if the framebuffer size has changed. - // This is expected to not happen very often and hence should not be a - // performance problem. - ConfigureFramebufferTexture(screen_info.texture, *framebuffer); - } - - // Load the framebuffer from memory, draw it to the screen, and swap buffers - LoadFBToScreenInfo(*framebuffer); + if (!framebuffer) { + return; + } + // If framebuffer is provided, reload it from memory to a texture + if (screen_info.texture.width != static_cast<GLsizei>(framebuffer->width) || + screen_info.texture.height != static_cast<GLsizei>(framebuffer->height) || + screen_info.texture.pixel_format != framebuffer->pixel_format || + gl_framebuffer_data.empty()) { + // Reallocate texture if the framebuffer size has changed. + // This is expected to not happen very often and hence should not be a + // performance problem. + ConfigureFramebufferTexture(screen_info.texture, *framebuffer); } + + // Load the framebuffer from memory, draw it to the screen, and swap buffers + LoadFBToScreenInfo(*framebuffer); } void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer) { @@ -423,7 +191,7 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)}; const u32 bytes_per_pixel{VideoCore::Surface::GetBytesPerPixel(pixel_format)}; const u64 size_in_bytes{framebuffer.stride * framebuffer.height * bytes_per_pixel}; - u8* const host_ptr{system.Memory().GetPointer(framebuffer_addr)}; + u8* const host_ptr{cpu_memory.GetPointer(framebuffer_addr)}; rasterizer->FlushRegion(ToCacheAddr(host_ptr), size_in_bytes); // TODO(Rodrigo): Read this from HLE @@ -453,23 +221,22 @@ void RendererOpenGL::LoadColorToActiveGLTexture(u8 color_r, u8 color_g, u8 color } void RendererOpenGL::InitOpenGLObjects() { - frame_mailbox = std::make_unique<FrameMailbox>(); - - glClearColor(Settings::values.bg_red, Settings::values.bg_green, Settings::values.bg_blue, - 0.0f); + glClearColor(Settings::values.bg_red.GetValue(), Settings::values.bg_green.GetValue(), + Settings::values.bg_blue.GetValue(), 0.0f); // Create shader programs OGLShader vertex_shader; - vertex_shader.Create(VERTEX_SHADER, GL_VERTEX_SHADER); + vertex_shader.Create(HostShaders::OPENGL_PRESENT_VERT, GL_VERTEX_SHADER); OGLShader fragment_shader; - fragment_shader.Create(FRAGMENT_SHADER, GL_FRAGMENT_SHADER); + fragment_shader.Create(HostShaders::OPENGL_PRESENT_FRAG, GL_FRAGMENT_SHADER); vertex_program.Create(true, false, vertex_shader.handle); fragment_program.Create(true, false, fragment_shader.handle); - // Create program pipeline - program_manager.Create(); + pipeline.Create(); + glUseProgramStages(pipeline.handle, GL_VERTEX_SHADER_BIT, vertex_program.handle); + glUseProgramStages(pipeline.handle, GL_FRAGMENT_SHADER_BIT, fragment_program.handle); // Generate VBO handle for drawing vertex_buffer.Create(); @@ -487,6 +254,15 @@ void RendererOpenGL::InitOpenGLObjects() { // Clear screen to black LoadColorToActiveGLTexture(0, 0, 0, 0, screen_info.texture); + + // Enable unified vertex attributes and query vertex buffer address when the driver supports it + if (device.HasVertexBufferUnifiedMemory()) { + glEnableClientState(GL_VERTEX_ATTRIB_ARRAY_UNIFIED_NV); + + glMakeNamedBufferResidentNV(vertex_buffer.handle, GL_READ_ONLY); + glGetNamedBufferParameterui64vNV(vertex_buffer.handle, GL_BUFFER_GPU_ADDRESS_NV, + &vertex_buffer_address); + } } void RendererOpenGL::AddTelemetryFields() { @@ -498,18 +274,18 @@ void RendererOpenGL::AddTelemetryFields() { LOG_INFO(Render_OpenGL, "GL_VENDOR: {}", gpu_vendor); LOG_INFO(Render_OpenGL, "GL_RENDERER: {}", gpu_model); - auto& telemetry_session = system.TelemetrySession(); - telemetry_session.AddField(Telemetry::FieldType::UserSystem, "GPU_Vendor", gpu_vendor); - telemetry_session.AddField(Telemetry::FieldType::UserSystem, "GPU_Model", gpu_model); - telemetry_session.AddField(Telemetry::FieldType::UserSystem, "GPU_OpenGL_Version", gl_version); + constexpr auto user_system = Common::Telemetry::FieldType::UserSystem; + telemetry_session.AddField(user_system, "GPU_Vendor", gpu_vendor); + telemetry_session.AddField(user_system, "GPU_Model", gpu_model); + telemetry_session.AddField(user_system, "GPU_OpenGL_Version", gl_version); } void RendererOpenGL::CreateRasterizer() { if (rasterizer) { return; } - rasterizer = std::make_unique<RasterizerOpenGL>(system, emu_window, screen_info, - program_manager, state_tracker); + rasterizer = std::make_unique<RasterizerOpenGL>(emu_window, gpu, cpu_memory, device, + screen_info, program_manager, state_tracker); } void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture, @@ -525,12 +301,12 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture, GLint internal_format; switch (framebuffer.pixel_format) { - case Tegra::FramebufferConfig::PixelFormat::ABGR8: + case Tegra::FramebufferConfig::PixelFormat::A8B8G8R8_UNORM: internal_format = GL_RGBA8; texture.gl_format = GL_RGBA; texture.gl_type = GL_UNSIGNED_INT_8_8_8_8_REV; break; - case Tegra::FramebufferConfig::PixelFormat::RGB565: + case Tegra::FramebufferConfig::PixelFormat::RGB565_UNORM: internal_format = GL_RGB565; texture.gl_format = GL_RGB; texture.gl_type = GL_UNSIGNED_SHORT_5_6_5; @@ -551,8 +327,8 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture, void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) { if (renderer_settings.set_background_color) { // Update background color before drawing - glClearColor(Settings::values.bg_red, Settings::values.bg_green, Settings::values.bg_blue, - 0.0f); + glClearColor(Settings::values.bg_red.GetValue(), Settings::values.bg_green.GetValue(), + Settings::values.bg_blue.GetValue(), 0.0f); } // Set projection matrix @@ -620,10 +396,7 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) { state_tracker.NotifyClipControl(); state_tracker.NotifyAlphaTest(); - program_manager.UseVertexShader(vertex_program.handle); - program_manager.UseGeometryShader(0); - program_manager.UseFragmentShader(fragment_program.handle); - program_manager.BindGraphicsPipeline(); + program_manager.BindHostPipeline(pipeline.handle); glEnable(GL_CULL_FACE); if (screen_info.display_srgb) { @@ -658,58 +431,21 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) { offsetof(ScreenRectVertex, tex_coord)); glVertexAttribBinding(PositionLocation, 0); glVertexAttribBinding(TexCoordLocation, 0); - glBindVertexBuffer(0, vertex_buffer.handle, 0, sizeof(ScreenRectVertex)); + if (device.HasVertexBufferUnifiedMemory()) { + glBindVertexBuffer(0, 0, 0, sizeof(ScreenRectVertex)); + glBufferAddressRangeNV(GL_VERTEX_ATTRIB_ARRAY_ADDRESS_NV, 0, vertex_buffer_address, + sizeof(vertices)); + } else { + glBindVertexBuffer(0, vertex_buffer.handle, 0, sizeof(ScreenRectVertex)); + } glBindTextureUnit(0, screen_info.display_texture); glBindSampler(0, 0); glClear(GL_COLOR_BUFFER_BIT); glDrawArrays(GL_TRIANGLE_STRIP, 0, 4); -} -bool RendererOpenGL::TryPresent(int timeout_ms) { - if (has_debug_tool) { - LOG_DEBUG(Render_OpenGL, - "Skipping presentation because we are presenting on the main context"); - return false; - } - return Present(timeout_ms); -} - -bool RendererOpenGL::Present(int timeout_ms) { - const auto& layout = render_window.GetFramebufferLayout(); - auto frame = frame_mailbox->TryGetPresentFrame(timeout_ms); - if (!frame) { - LOG_DEBUG(Render_OpenGL, "TryGetPresentFrame returned no frame to present"); - return false; - } - - // Clearing before a full overwrite of a fbo can signal to drivers that they can avoid a - // readback since we won't be doing any blending - glClear(GL_COLOR_BUFFER_BIT); - - // Recreate the presentation FBO if the color attachment was changed - if (frame->color_reloaded) { - LOG_DEBUG(Render_OpenGL, "Reloading present frame"); - frame_mailbox->ReloadPresentFrame(frame, layout.width, layout.height); - } - glWaitSync(frame->render_fence, 0, GL_TIMEOUT_IGNORED); - // INTEL workaround. - // Normally we could just delete the draw fence here, but due to driver bugs, we can just delete - // it on the emulation thread without too much penalty - // glDeleteSync(frame.render_sync); - // frame.render_sync = 0; - - glBindFramebuffer(GL_READ_FRAMEBUFFER, frame->present.handle); - glBlitFramebuffer(0, 0, frame->width, frame->height, 0, 0, layout.width, layout.height, - GL_COLOR_BUFFER_BIT, GL_LINEAR); - - // Insert fence for the main thread to block on - frame->present_fence = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0); - glFlush(); - - glBindFramebuffer(GL_READ_FRAMEBUFFER, 0); - return true; + program_manager.RestoreGuestPipeline(); } void RendererOpenGL::RenderScreenshot() { @@ -726,7 +462,7 @@ void RendererOpenGL::RenderScreenshot() { screenshot_framebuffer.Create(); glBindFramebuffer(GL_FRAMEBUFFER, screenshot_framebuffer.handle); - Layout::FramebufferLayout layout{renderer_settings.screenshot_framebuffer_layout}; + const Layout::FramebufferLayout layout{renderer_settings.screenshot_framebuffer_layout}; GLuint renderbuffer; glGenRenderbuffers(1, &renderbuffer); @@ -751,8 +487,9 @@ void RendererOpenGL::RenderScreenshot() { } bool RendererOpenGL::Init() { - if (GLAD_GL_KHR_debug) { + if (Settings::values.renderer_debug && GLAD_GL_KHR_debug) { glEnable(GL_DEBUG_OUTPUT); + glEnable(GL_DEBUG_OUTPUT_SYNCHRONOUS); glDebugMessageCallback(DebugHandler, nullptr); } diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h index 50b647661..9ef181f95 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.h +++ b/src/video_core/renderer_opengl/renderer_opengl.h @@ -9,22 +9,32 @@ #include "common/common_types.h" #include "common/math_util.h" #include "video_core/renderer_base.h" +#include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/gl_state_tracker.h" namespace Core { class System; -} +class TelemetrySession; +} // namespace Core namespace Core::Frontend { class EmuWindow; } +namespace Core::Memory { +class Memory; +} + namespace Layout { struct FramebufferLayout; } +namespace Tegra { +class GPU; +} + namespace OpenGL { /// Structure used for storing information about the textures for the Switch screen @@ -45,24 +55,17 @@ struct ScreenInfo { TextureInfo texture; }; -struct PresentationTexture { - u32 width = 0; - u32 height = 0; - OGLTexture texture; -}; - -class FrameMailbox; - class RendererOpenGL final : public VideoCore::RendererBase { public: - explicit RendererOpenGL(Core::Frontend::EmuWindow& emu_window, Core::System& system, - Core::Frontend::GraphicsContext& context); + explicit RendererOpenGL(Core::TelemetrySession& telemetry_session, + Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory, + Tegra::GPU& gpu, + std::unique_ptr<Core::Frontend::GraphicsContext> context); ~RendererOpenGL() override; bool Init() override; void ShutDown() override; void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override; - bool TryPresent(int timeout_ms) override; private: /// Initializes the OpenGL state and creates persistent objects. @@ -90,37 +93,36 @@ private: void PrepareRendertarget(const Tegra::FramebufferConfig* framebuffer); - bool Present(int timeout_ms); - + Core::TelemetrySession& telemetry_session; Core::Frontend::EmuWindow& emu_window; - Core::System& system; - Core::Frontend::GraphicsContext& context; + Core::Memory::Memory& cpu_memory; + Tegra::GPU& gpu; - StateTracker state_tracker{system}; + const Device device; + StateTracker state_tracker{gpu}; // OpenGL object IDs OGLBuffer vertex_buffer; OGLProgram vertex_program; OGLProgram fragment_program; + OGLPipeline pipeline; OGLFramebuffer screenshot_framebuffer; + // GPU address of the vertex buffer + GLuint64EXT vertex_buffer_address = 0; + /// Display information for Switch screen ScreenInfo screen_info; /// Global dummy shader pipeline - GLShader::ProgramManager program_manager; + ProgramManager program_manager; /// OpenGL framebuffer data std::vector<u8> gl_framebuffer_data; /// Used for transforming the framebuffer orientation - Tegra::FramebufferConfig::TransformFlags framebuffer_transform_flags; + Tegra::FramebufferConfig::TransformFlags framebuffer_transform_flags{}; Common::Rectangle<int> framebuffer_crop_rect; - - /// Frame presentation mailbox - std::unique_ptr<FrameMailbox> frame_mailbox; - - bool has_debug_tool = false; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/utils.cpp b/src/video_core/renderer_opengl/utils.cpp index b751086fa..6d7bb16b2 100644 --- a/src/video_core/renderer_opengl/utils.cpp +++ b/src/video_core/renderer_opengl/utils.cpp @@ -14,68 +14,6 @@ namespace OpenGL { -struct VertexArrayPushBuffer::Entry { - GLuint binding_index{}; - const GLuint* buffer{}; - GLintptr offset{}; - GLsizei stride{}; -}; - -VertexArrayPushBuffer::VertexArrayPushBuffer(StateTracker& state_tracker) - : state_tracker{state_tracker} {} - -VertexArrayPushBuffer::~VertexArrayPushBuffer() = default; - -void VertexArrayPushBuffer::Setup() { - index_buffer = nullptr; - vertex_buffers.clear(); -} - -void VertexArrayPushBuffer::SetIndexBuffer(const GLuint* buffer) { - index_buffer = buffer; -} - -void VertexArrayPushBuffer::SetVertexBuffer(GLuint binding_index, const GLuint* buffer, - GLintptr offset, GLsizei stride) { - vertex_buffers.push_back(Entry{binding_index, buffer, offset, stride}); -} - -void VertexArrayPushBuffer::Bind() { - if (index_buffer) { - state_tracker.BindIndexBuffer(*index_buffer); - } - - for (const auto& entry : vertex_buffers) { - glBindVertexBuffer(entry.binding_index, *entry.buffer, entry.offset, entry.stride); - } -} - -struct BindBuffersRangePushBuffer::Entry { - GLuint binding; - const GLuint* buffer; - GLintptr offset; - GLsizeiptr size; -}; - -BindBuffersRangePushBuffer::BindBuffersRangePushBuffer(GLenum target) : target{target} {} - -BindBuffersRangePushBuffer::~BindBuffersRangePushBuffer() = default; - -void BindBuffersRangePushBuffer::Setup() { - entries.clear(); -} - -void BindBuffersRangePushBuffer::Push(GLuint binding, const GLuint* buffer, GLintptr offset, - GLsizeiptr size) { - entries.push_back(Entry{binding, buffer, offset, size}); -} - -void BindBuffersRangePushBuffer::Bind() { - for (const Entry& entry : entries) { - glBindBufferRange(target, entry.binding, *entry.buffer, entry.offset, entry.size); - } -} - void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string_view extra_info) { if (!GLAD_GL_KHR_debug) { // We don't need to throw an error as this is just for debugging diff --git a/src/video_core/renderer_opengl/utils.h b/src/video_core/renderer_opengl/utils.h index 47ee3177b..9c09ee12c 100644 --- a/src/video_core/renderer_opengl/utils.h +++ b/src/video_core/renderer_opengl/utils.h @@ -11,49 +11,6 @@ namespace OpenGL { -class StateTracker; - -class VertexArrayPushBuffer final { -public: - explicit VertexArrayPushBuffer(StateTracker& state_tracker); - ~VertexArrayPushBuffer(); - - void Setup(); - - void SetIndexBuffer(const GLuint* buffer); - - void SetVertexBuffer(GLuint binding_index, const GLuint* buffer, GLintptr offset, - GLsizei stride); - - void Bind(); - -private: - struct Entry; - - StateTracker& state_tracker; - - const GLuint* index_buffer{}; - std::vector<Entry> vertex_buffers; -}; - -class BindBuffersRangePushBuffer final { -public: - explicit BindBuffersRangePushBuffer(GLenum target); - ~BindBuffersRangePushBuffer(); - - void Setup(); - - void Push(GLuint binding, const GLuint* buffer, GLintptr offset, GLsizeiptr size); - - void Bind(); - -private: - struct Entry; - - GLenum target; - std::vector<Entry> entries; -}; - void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string_view extra_info = {}); } // namespace OpenGL diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp index 2bb376555..da5c550ea 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp @@ -2,10 +2,13 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> +#include <cstring> #include <tuple> #include <boost/functional/hash.hpp> +#include "common/cityhash.h" #include "common/common_types.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" @@ -13,289 +16,375 @@ namespace Vulkan { namespace { -constexpr FixedPipelineState::DepthStencil GetDepthStencilState(const Maxwell& regs) { - const FixedPipelineState::StencilFace front_stencil( - regs.stencil_front_op_fail, regs.stencil_front_op_zfail, regs.stencil_front_op_zpass, - regs.stencil_front_func_func); - const FixedPipelineState::StencilFace back_stencil = - regs.stencil_two_side_enable - ? FixedPipelineState::StencilFace(regs.stencil_back_op_fail, regs.stencil_back_op_zfail, - regs.stencil_back_op_zpass, - regs.stencil_back_func_func) - : front_stencil; - return FixedPipelineState::DepthStencil( - regs.depth_test_enable == 1, regs.depth_write_enabled == 1, regs.depth_bounds_enable == 1, - regs.stencil_enable == 1, regs.depth_test_func, front_stencil, back_stencil); -} +constexpr std::size_t POINT = 0; +constexpr std::size_t LINE = 1; +constexpr std::size_t POLYGON = 2; +constexpr std::array POLYGON_OFFSET_ENABLE_LUT = { + POINT, // Points + LINE, // Lines + LINE, // LineLoop + LINE, // LineStrip + POLYGON, // Triangles + POLYGON, // TriangleStrip + POLYGON, // TriangleFan + POLYGON, // Quads + POLYGON, // QuadStrip + POLYGON, // Polygon + LINE, // LinesAdjacency + LINE, // LineStripAdjacency + POLYGON, // TrianglesAdjacency + POLYGON, // TriangleStripAdjacency + POLYGON, // Patches +}; -constexpr FixedPipelineState::InputAssembly GetInputAssemblyState(const Maxwell& regs) { - return FixedPipelineState::InputAssembly( - regs.draw.topology, regs.primitive_restart.enabled, - regs.draw.topology == Maxwell::PrimitiveTopology::Points ? regs.point_size : 0.0f); -} +} // Anonymous namespace -constexpr FixedPipelineState::BlendingAttachment GetBlendingAttachmentState( - const Maxwell& regs, std::size_t render_target) { - const auto& mask = regs.color_mask[regs.color_mask_common ? 0 : render_target]; - const std::array components = {mask.R != 0, mask.G != 0, mask.B != 0, mask.A != 0}; - - const FixedPipelineState::BlendingAttachment default_blending( - false, Maxwell::Blend::Equation::Add, Maxwell::Blend::Factor::One, - Maxwell::Blend::Factor::Zero, Maxwell::Blend::Equation::Add, Maxwell::Blend::Factor::One, - Maxwell::Blend::Factor::Zero, components); - if (render_target >= regs.rt_control.count) { - return default_blending; +void FixedPipelineState::Fill(const Maxwell& regs, bool has_extended_dynamic_state) { + const std::array enabled_lut = {regs.polygon_offset_point_enable, + regs.polygon_offset_line_enable, + regs.polygon_offset_fill_enable}; + const u32 topology_index = static_cast<u32>(regs.draw.topology.Value()); + + raw = 0; + primitive_restart_enable.Assign(regs.primitive_restart.enabled != 0 ? 1 : 0); + depth_bias_enable.Assign(enabled_lut[POLYGON_OFFSET_ENABLE_LUT[topology_index]] != 0 ? 1 : 0); + depth_clamp_disabled.Assign(regs.view_volume_clip_control.depth_clamp_disabled.Value()); + ndc_minus_one_to_one.Assign(regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1 : 0); + polygon_mode.Assign(PackPolygonMode(regs.polygon_mode_front)); + patch_control_points_minus_one.Assign(regs.patch_vertices - 1); + tessellation_primitive.Assign(static_cast<u32>(regs.tess_mode.prim.Value())); + tessellation_spacing.Assign(static_cast<u32>(regs.tess_mode.spacing.Value())); + tessellation_clockwise.Assign(regs.tess_mode.cw.Value()); + logic_op_enable.Assign(regs.logic_op.enable != 0 ? 1 : 0); + logic_op.Assign(PackLogicOp(regs.logic_op.operation)); + rasterize_enable.Assign(regs.rasterize_enable != 0 ? 1 : 0); + topology.Assign(regs.draw.topology); + + std::memcpy(&point_size, ®s.point_size, sizeof(point_size)); // TODO: C++20 std::bit_cast + + for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { + binding_divisors[index] = + regs.instanced_arrays.IsInstancingEnabled(index) ? regs.vertex_array[index].divisor : 0; } - if (!regs.independent_blend_enable) { - const auto& src = regs.blend; - if (!src.enable[render_target]) { - return default_blending; - } - return FixedPipelineState::BlendingAttachment( - true, src.equation_rgb, src.factor_source_rgb, src.factor_dest_rgb, src.equation_a, - src.factor_source_a, src.factor_dest_a, components); + for (std::size_t index = 0; index < Maxwell::NumVertexAttributes; ++index) { + const auto& input = regs.vertex_attrib_format[index]; + auto& attribute = attributes[index]; + attribute.raw = 0; + attribute.enabled.Assign(input.IsConstant() ? 0 : 1); + attribute.buffer.Assign(input.buffer); + attribute.offset.Assign(input.offset); + attribute.type.Assign(static_cast<u32>(input.type.Value())); + attribute.size.Assign(static_cast<u32>(input.size.Value())); } - if (!regs.blend.enable[render_target]) { - return default_blending; + for (std::size_t index = 0; index < std::size(attachments); ++index) { + attachments[index].Fill(regs, index); } - const auto& src = regs.independent_blend[render_target]; - return FixedPipelineState::BlendingAttachment( - true, src.equation_rgb, src.factor_source_rgb, src.factor_dest_rgb, src.equation_a, - src.factor_source_a, src.factor_dest_a, components); -} -constexpr FixedPipelineState::ColorBlending GetColorBlendingState(const Maxwell& regs) { - return FixedPipelineState::ColorBlending( - {regs.blend_color.r, regs.blend_color.g, regs.blend_color.b, regs.blend_color.a}, - regs.rt_control.count, - {GetBlendingAttachmentState(regs, 0), GetBlendingAttachmentState(regs, 1), - GetBlendingAttachmentState(regs, 2), GetBlendingAttachmentState(regs, 3), - GetBlendingAttachmentState(regs, 4), GetBlendingAttachmentState(regs, 5), - GetBlendingAttachmentState(regs, 6), GetBlendingAttachmentState(regs, 7)}); -} + const auto& transform = regs.viewport_transform; + std::transform(transform.begin(), transform.end(), viewport_swizzles.begin(), + [](const auto& viewport) { return static_cast<u16>(viewport.swizzle.raw); }); -constexpr FixedPipelineState::Tessellation GetTessellationState(const Maxwell& regs) { - return FixedPipelineState::Tessellation(regs.patch_vertices, regs.tess_mode.prim, - regs.tess_mode.spacing, regs.tess_mode.cw != 0); + if (!has_extended_dynamic_state) { + no_extended_dynamic_state.Assign(1); + dynamic_state.Fill(regs); + } } -constexpr std::size_t Point = 0; -constexpr std::size_t Line = 1; -constexpr std::size_t Polygon = 2; -constexpr std::array PolygonOffsetEnableLUT = { - Point, // Points - Line, // Lines - Line, // LineLoop - Line, // LineStrip - Polygon, // Triangles - Polygon, // TriangleStrip - Polygon, // TriangleFan - Polygon, // Quads - Polygon, // QuadStrip - Polygon, // Polygon - Line, // LinesAdjacency - Line, // LineStripAdjacency - Polygon, // TrianglesAdjacency - Polygon, // TriangleStripAdjacency - Polygon, // Patches -}; +void FixedPipelineState::BlendingAttachment::Fill(const Maxwell& regs, std::size_t index) { + const auto& mask = regs.color_mask[regs.color_mask_common ? 0 : index]; -constexpr FixedPipelineState::Rasterizer GetRasterizerState(const Maxwell& regs) { - const std::array enabled_lut = {regs.polygon_offset_point_enable, - regs.polygon_offset_line_enable, - regs.polygon_offset_fill_enable}; - const auto topology = static_cast<std::size_t>(regs.draw.topology.Value()); - const bool depth_bias_enabled = enabled_lut[PolygonOffsetEnableLUT[topology]]; - - const auto& clip = regs.view_volume_clip_control; - const bool depth_clamp_enabled = clip.depth_clamp_near == 1 || clip.depth_clamp_far == 1; - - Maxwell::FrontFace front_face = regs.front_face; - if (regs.screen_y_control.triangle_rast_flip != 0 && - regs.viewport_transform[0].scale_y > 0.0f) { - if (front_face == Maxwell::FrontFace::CounterClockWise) - front_face = Maxwell::FrontFace::ClockWise; - else if (front_face == Maxwell::FrontFace::ClockWise) - front_face = Maxwell::FrontFace::CounterClockWise; - } + raw = 0; + mask_r.Assign(mask.R); + mask_g.Assign(mask.G); + mask_b.Assign(mask.B); + mask_a.Assign(mask.A); - const bool gl_ndc = regs.depth_mode == Maxwell::DepthMode::MinusOneToOne; - return FixedPipelineState::Rasterizer(regs.cull_test_enabled, depth_bias_enabled, - depth_clamp_enabled, gl_ndc, regs.cull_face, front_face); -} + // TODO: C++20 Use templated lambda to deduplicate code -} // Anonymous namespace - -std::size_t FixedPipelineState::VertexBinding::Hash() const noexcept { - return (index << stride) ^ divisor; -} + if (!regs.independent_blend_enable) { + const auto& src = regs.blend; + if (!src.enable[index]) { + return; + } + equation_rgb.Assign(PackBlendEquation(src.equation_rgb)); + equation_a.Assign(PackBlendEquation(src.equation_a)); + factor_source_rgb.Assign(PackBlendFactor(src.factor_source_rgb)); + factor_dest_rgb.Assign(PackBlendFactor(src.factor_dest_rgb)); + factor_source_a.Assign(PackBlendFactor(src.factor_source_a)); + factor_dest_a.Assign(PackBlendFactor(src.factor_dest_a)); + enable.Assign(1); + return; + } -bool FixedPipelineState::VertexBinding::operator==(const VertexBinding& rhs) const noexcept { - return std::tie(index, stride, divisor) == std::tie(rhs.index, rhs.stride, rhs.divisor); + if (!regs.blend.enable[index]) { + return; + } + const auto& src = regs.independent_blend[index]; + equation_rgb.Assign(PackBlendEquation(src.equation_rgb)); + equation_a.Assign(PackBlendEquation(src.equation_a)); + factor_source_rgb.Assign(PackBlendFactor(src.factor_source_rgb)); + factor_dest_rgb.Assign(PackBlendFactor(src.factor_dest_rgb)); + factor_source_a.Assign(PackBlendFactor(src.factor_source_a)); + factor_dest_a.Assign(PackBlendFactor(src.factor_dest_a)); + enable.Assign(1); } -std::size_t FixedPipelineState::VertexAttribute::Hash() const noexcept { - return static_cast<std::size_t>(index) ^ (static_cast<std::size_t>(buffer) << 13) ^ - (static_cast<std::size_t>(type) << 22) ^ (static_cast<std::size_t>(size) << 31) ^ - (static_cast<std::size_t>(offset) << 36); -} +void FixedPipelineState::DynamicState::Fill(const Maxwell& regs) { + u32 packed_front_face = PackFrontFace(regs.front_face); + if (regs.screen_y_control.triangle_rast_flip != 0) { + // Flip front face + packed_front_face = 1 - packed_front_face; + } -bool FixedPipelineState::VertexAttribute::operator==(const VertexAttribute& rhs) const noexcept { - return std::tie(index, buffer, type, size, offset) == - std::tie(rhs.index, rhs.buffer, rhs.type, rhs.size, rhs.offset); + raw1 = 0; + raw2 = 0; + front.action_stencil_fail.Assign(PackStencilOp(regs.stencil_front_op_fail)); + front.action_depth_fail.Assign(PackStencilOp(regs.stencil_front_op_zfail)); + front.action_depth_pass.Assign(PackStencilOp(regs.stencil_front_op_zpass)); + front.test_func.Assign(PackComparisonOp(regs.stencil_front_func_func)); + if (regs.stencil_two_side_enable) { + back.action_stencil_fail.Assign(PackStencilOp(regs.stencil_back_op_fail)); + back.action_depth_fail.Assign(PackStencilOp(regs.stencil_back_op_zfail)); + back.action_depth_pass.Assign(PackStencilOp(regs.stencil_back_op_zpass)); + back.test_func.Assign(PackComparisonOp(regs.stencil_back_func_func)); + } else { + back.action_stencil_fail.Assign(front.action_stencil_fail); + back.action_depth_fail.Assign(front.action_depth_fail); + back.action_depth_pass.Assign(front.action_depth_pass); + back.test_func.Assign(front.test_func); + } + stencil_enable.Assign(regs.stencil_enable); + depth_write_enable.Assign(regs.depth_write_enabled); + depth_bounds_enable.Assign(regs.depth_bounds_enable); + depth_test_enable.Assign(regs.depth_test_enable); + front_face.Assign(packed_front_face); + depth_test_func.Assign(PackComparisonOp(regs.depth_test_func)); + cull_face.Assign(PackCullFace(regs.cull_face)); + cull_enable.Assign(regs.cull_test_enabled != 0 ? 1 : 0); + + for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { + const auto& input = regs.vertex_array[index]; + VertexBinding& binding = vertex_bindings[index]; + binding.raw = 0; + binding.enabled.Assign(input.IsEnabled() ? 1 : 0); + binding.stride.Assign(static_cast<u16>(input.stride.Value())); + } } -std::size_t FixedPipelineState::StencilFace::Hash() const noexcept { - return static_cast<std::size_t>(action_stencil_fail) ^ - (static_cast<std::size_t>(action_depth_fail) << 4) ^ - (static_cast<std::size_t>(action_depth_fail) << 20) ^ - (static_cast<std::size_t>(action_depth_pass) << 36); +std::size_t FixedPipelineState::Hash() const noexcept { + const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), Size()); + return static_cast<std::size_t>(hash); } -bool FixedPipelineState::StencilFace::operator==(const StencilFace& rhs) const noexcept { - return std::tie(action_stencil_fail, action_depth_fail, action_depth_pass, test_func) == - std::tie(rhs.action_stencil_fail, rhs.action_depth_fail, rhs.action_depth_pass, - rhs.test_func); +bool FixedPipelineState::operator==(const FixedPipelineState& rhs) const noexcept { + return std::memcmp(this, &rhs, Size()) == 0; } -std::size_t FixedPipelineState::BlendingAttachment::Hash() const noexcept { - return static_cast<std::size_t>(enable) ^ (static_cast<std::size_t>(rgb_equation) << 5) ^ - (static_cast<std::size_t>(src_rgb_func) << 10) ^ - (static_cast<std::size_t>(dst_rgb_func) << 15) ^ - (static_cast<std::size_t>(a_equation) << 20) ^ - (static_cast<std::size_t>(src_a_func) << 25) ^ - (static_cast<std::size_t>(dst_a_func) << 30) ^ - (static_cast<std::size_t>(components[0]) << 35) ^ - (static_cast<std::size_t>(components[1]) << 36) ^ - (static_cast<std::size_t>(components[2]) << 37) ^ - (static_cast<std::size_t>(components[3]) << 38); +u32 FixedPipelineState::PackComparisonOp(Maxwell::ComparisonOp op) noexcept { + // OpenGL enums go from 0x200 to 0x207 and the others from 1 to 8 + // If we substract 0x200 to OpenGL enums and 1 to the others we get a 0-7 range. + // Perfect for a hash. + const u32 value = static_cast<u32>(op); + return value - (value >= 0x200 ? 0x200 : 1); } -bool FixedPipelineState::BlendingAttachment::operator==(const BlendingAttachment& rhs) const - noexcept { - return std::tie(enable, rgb_equation, src_rgb_func, dst_rgb_func, a_equation, src_a_func, - dst_a_func, components) == - std::tie(rhs.enable, rhs.rgb_equation, rhs.src_rgb_func, rhs.dst_rgb_func, - rhs.a_equation, rhs.src_a_func, rhs.dst_a_func, rhs.components); +Maxwell::ComparisonOp FixedPipelineState::UnpackComparisonOp(u32 packed) noexcept { + // Read PackComparisonOp for the logic behind this. + return static_cast<Maxwell::ComparisonOp>(packed + 1); } -std::size_t FixedPipelineState::VertexInput::Hash() const noexcept { - std::size_t hash = num_bindings ^ (num_attributes << 32); - for (std::size_t i = 0; i < num_bindings; ++i) { - boost::hash_combine(hash, bindings[i].Hash()); +u32 FixedPipelineState::PackStencilOp(Maxwell::StencilOp op) noexcept { + switch (op) { + case Maxwell::StencilOp::Keep: + case Maxwell::StencilOp::KeepOGL: + return 0; + case Maxwell::StencilOp::Zero: + case Maxwell::StencilOp::ZeroOGL: + return 1; + case Maxwell::StencilOp::Replace: + case Maxwell::StencilOp::ReplaceOGL: + return 2; + case Maxwell::StencilOp::Incr: + case Maxwell::StencilOp::IncrOGL: + return 3; + case Maxwell::StencilOp::Decr: + case Maxwell::StencilOp::DecrOGL: + return 4; + case Maxwell::StencilOp::Invert: + case Maxwell::StencilOp::InvertOGL: + return 5; + case Maxwell::StencilOp::IncrWrap: + case Maxwell::StencilOp::IncrWrapOGL: + return 6; + case Maxwell::StencilOp::DecrWrap: + case Maxwell::StencilOp::DecrWrapOGL: + return 7; } - for (std::size_t i = 0; i < num_attributes; ++i) { - boost::hash_combine(hash, attributes[i].Hash()); - } - return hash; + return 0; } -bool FixedPipelineState::VertexInput::operator==(const VertexInput& rhs) const noexcept { - return std::equal(bindings.begin(), bindings.begin() + num_bindings, rhs.bindings.begin(), - rhs.bindings.begin() + rhs.num_bindings) && - std::equal(attributes.begin(), attributes.begin() + num_attributes, - rhs.attributes.begin(), rhs.attributes.begin() + rhs.num_attributes); +Maxwell::StencilOp FixedPipelineState::UnpackStencilOp(u32 packed) noexcept { + static constexpr std::array LUT = {Maxwell::StencilOp::Keep, Maxwell::StencilOp::Zero, + Maxwell::StencilOp::Replace, Maxwell::StencilOp::Incr, + Maxwell::StencilOp::Decr, Maxwell::StencilOp::Invert, + Maxwell::StencilOp::IncrWrap, Maxwell::StencilOp::DecrWrap}; + return LUT[packed]; } -std::size_t FixedPipelineState::InputAssembly::Hash() const noexcept { - std::size_t point_size_int = 0; - std::memcpy(&point_size_int, &point_size, sizeof(point_size)); - return (static_cast<std::size_t>(topology) << 24) ^ (point_size_int << 32) ^ - static_cast<std::size_t>(primitive_restart_enable); +u32 FixedPipelineState::PackCullFace(Maxwell::CullFace cull) noexcept { + // FrontAndBack is 0x408, by substracting 0x406 in it we get 2. + // Individual cull faces are in 0x404 and 0x405, substracting 0x404 we get 0 and 1. + const u32 value = static_cast<u32>(cull); + return value - (value == 0x408 ? 0x406 : 0x404); } -bool FixedPipelineState::InputAssembly::operator==(const InputAssembly& rhs) const noexcept { - return std::tie(topology, primitive_restart_enable, point_size) == - std::tie(rhs.topology, rhs.primitive_restart_enable, rhs.point_size); +Maxwell::CullFace FixedPipelineState::UnpackCullFace(u32 packed) noexcept { + static constexpr std::array LUT = {Maxwell::CullFace::Front, Maxwell::CullFace::Back, + Maxwell::CullFace::FrontAndBack}; + return LUT[packed]; } -std::size_t FixedPipelineState::Tessellation::Hash() const noexcept { - return static_cast<std::size_t>(patch_control_points) ^ - (static_cast<std::size_t>(primitive) << 6) ^ (static_cast<std::size_t>(spacing) << 8) ^ - (static_cast<std::size_t>(clockwise) << 10); +u32 FixedPipelineState::PackFrontFace(Maxwell::FrontFace face) noexcept { + return static_cast<u32>(face) - 0x900; } -bool FixedPipelineState::Tessellation::operator==(const Tessellation& rhs) const noexcept { - return std::tie(patch_control_points, primitive, spacing, clockwise) == - std::tie(rhs.patch_control_points, rhs.primitive, rhs.spacing, rhs.clockwise); +Maxwell::FrontFace FixedPipelineState::UnpackFrontFace(u32 packed) noexcept { + return static_cast<Maxwell::FrontFace>(packed + 0x900); } -std::size_t FixedPipelineState::Rasterizer::Hash() const noexcept { - return static_cast<std::size_t>(cull_enable) ^ - (static_cast<std::size_t>(depth_bias_enable) << 1) ^ - (static_cast<std::size_t>(depth_clamp_enable) << 2) ^ - (static_cast<std::size_t>(ndc_minus_one_to_one) << 3) ^ - (static_cast<std::size_t>(cull_face) << 24) ^ - (static_cast<std::size_t>(front_face) << 48); +u32 FixedPipelineState::PackPolygonMode(Maxwell::PolygonMode mode) noexcept { + return static_cast<u32>(mode) - 0x1B00; } -bool FixedPipelineState::Rasterizer::operator==(const Rasterizer& rhs) const noexcept { - return std::tie(cull_enable, depth_bias_enable, depth_clamp_enable, ndc_minus_one_to_one, - cull_face, front_face) == - std::tie(rhs.cull_enable, rhs.depth_bias_enable, rhs.depth_clamp_enable, - rhs.ndc_minus_one_to_one, rhs.cull_face, rhs.front_face); +Maxwell::PolygonMode FixedPipelineState::UnpackPolygonMode(u32 packed) noexcept { + return static_cast<Maxwell::PolygonMode>(packed + 0x1B00); } -std::size_t FixedPipelineState::DepthStencil::Hash() const noexcept { - std::size_t hash = static_cast<std::size_t>(depth_test_enable) ^ - (static_cast<std::size_t>(depth_write_enable) << 1) ^ - (static_cast<std::size_t>(depth_bounds_enable) << 2) ^ - (static_cast<std::size_t>(stencil_enable) << 3) ^ - (static_cast<std::size_t>(depth_test_function) << 4); - boost::hash_combine(hash, front_stencil.Hash()); - boost::hash_combine(hash, back_stencil.Hash()); - return hash; +u32 FixedPipelineState::PackLogicOp(Maxwell::LogicOperation op) noexcept { + return static_cast<u32>(op) - 0x1500; } -bool FixedPipelineState::DepthStencil::operator==(const DepthStencil& rhs) const noexcept { - return std::tie(depth_test_enable, depth_write_enable, depth_bounds_enable, depth_test_function, - stencil_enable, front_stencil, back_stencil) == - std::tie(rhs.depth_test_enable, rhs.depth_write_enable, rhs.depth_bounds_enable, - rhs.depth_test_function, rhs.stencil_enable, rhs.front_stencil, - rhs.back_stencil); +Maxwell::LogicOperation FixedPipelineState::UnpackLogicOp(u32 packed) noexcept { + return static_cast<Maxwell::LogicOperation>(packed + 0x1500); } -std::size_t FixedPipelineState::ColorBlending::Hash() const noexcept { - std::size_t hash = attachments_count << 13; - for (std::size_t rt = 0; rt < static_cast<std::size_t>(attachments_count); ++rt) { - boost::hash_combine(hash, attachments[rt].Hash()); +u32 FixedPipelineState::PackBlendEquation(Maxwell::Blend::Equation equation) noexcept { + switch (equation) { + case Maxwell::Blend::Equation::Add: + case Maxwell::Blend::Equation::AddGL: + return 0; + case Maxwell::Blend::Equation::Subtract: + case Maxwell::Blend::Equation::SubtractGL: + return 1; + case Maxwell::Blend::Equation::ReverseSubtract: + case Maxwell::Blend::Equation::ReverseSubtractGL: + return 2; + case Maxwell::Blend::Equation::Min: + case Maxwell::Blend::Equation::MinGL: + return 3; + case Maxwell::Blend::Equation::Max: + case Maxwell::Blend::Equation::MaxGL: + return 4; } - return hash; + return 0; } -bool FixedPipelineState::ColorBlending::operator==(const ColorBlending& rhs) const noexcept { - return std::equal(attachments.begin(), attachments.begin() + attachments_count, - rhs.attachments.begin(), rhs.attachments.begin() + rhs.attachments_count); +Maxwell::Blend::Equation FixedPipelineState::UnpackBlendEquation(u32 packed) noexcept { + static constexpr std::array LUT = { + Maxwell::Blend::Equation::Add, Maxwell::Blend::Equation::Subtract, + Maxwell::Blend::Equation::ReverseSubtract, Maxwell::Blend::Equation::Min, + Maxwell::Blend::Equation::Max}; + return LUT[packed]; } -std::size_t FixedPipelineState::Hash() const noexcept { - std::size_t hash = 0; - boost::hash_combine(hash, vertex_input.Hash()); - boost::hash_combine(hash, input_assembly.Hash()); - boost::hash_combine(hash, tessellation.Hash()); - boost::hash_combine(hash, rasterizer.Hash()); - boost::hash_combine(hash, depth_stencil.Hash()); - boost::hash_combine(hash, color_blending.Hash()); - return hash; -} - -bool FixedPipelineState::operator==(const FixedPipelineState& rhs) const noexcept { - return std::tie(vertex_input, input_assembly, tessellation, rasterizer, depth_stencil, - color_blending) == std::tie(rhs.vertex_input, rhs.input_assembly, - rhs.tessellation, rhs.rasterizer, rhs.depth_stencil, - rhs.color_blending); +u32 FixedPipelineState::PackBlendFactor(Maxwell::Blend::Factor factor) noexcept { + switch (factor) { + case Maxwell::Blend::Factor::Zero: + case Maxwell::Blend::Factor::ZeroGL: + return 0; + case Maxwell::Blend::Factor::One: + case Maxwell::Blend::Factor::OneGL: + return 1; + case Maxwell::Blend::Factor::SourceColor: + case Maxwell::Blend::Factor::SourceColorGL: + return 2; + case Maxwell::Blend::Factor::OneMinusSourceColor: + case Maxwell::Blend::Factor::OneMinusSourceColorGL: + return 3; + case Maxwell::Blend::Factor::SourceAlpha: + case Maxwell::Blend::Factor::SourceAlphaGL: + return 4; + case Maxwell::Blend::Factor::OneMinusSourceAlpha: + case Maxwell::Blend::Factor::OneMinusSourceAlphaGL: + return 5; + case Maxwell::Blend::Factor::DestAlpha: + case Maxwell::Blend::Factor::DestAlphaGL: + return 6; + case Maxwell::Blend::Factor::OneMinusDestAlpha: + case Maxwell::Blend::Factor::OneMinusDestAlphaGL: + return 7; + case Maxwell::Blend::Factor::DestColor: + case Maxwell::Blend::Factor::DestColorGL: + return 8; + case Maxwell::Blend::Factor::OneMinusDestColor: + case Maxwell::Blend::Factor::OneMinusDestColorGL: + return 9; + case Maxwell::Blend::Factor::SourceAlphaSaturate: + case Maxwell::Blend::Factor::SourceAlphaSaturateGL: + return 10; + case Maxwell::Blend::Factor::Source1Color: + case Maxwell::Blend::Factor::Source1ColorGL: + return 11; + case Maxwell::Blend::Factor::OneMinusSource1Color: + case Maxwell::Blend::Factor::OneMinusSource1ColorGL: + return 12; + case Maxwell::Blend::Factor::Source1Alpha: + case Maxwell::Blend::Factor::Source1AlphaGL: + return 13; + case Maxwell::Blend::Factor::OneMinusSource1Alpha: + case Maxwell::Blend::Factor::OneMinusSource1AlphaGL: + return 14; + case Maxwell::Blend::Factor::ConstantColor: + case Maxwell::Blend::Factor::ConstantColorGL: + return 15; + case Maxwell::Blend::Factor::OneMinusConstantColor: + case Maxwell::Blend::Factor::OneMinusConstantColorGL: + return 16; + case Maxwell::Blend::Factor::ConstantAlpha: + case Maxwell::Blend::Factor::ConstantAlphaGL: + return 17; + case Maxwell::Blend::Factor::OneMinusConstantAlpha: + case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: + return 18; + } + return 0; } -FixedPipelineState GetFixedPipelineState(const Maxwell& regs) { - FixedPipelineState fixed_state; - fixed_state.input_assembly = GetInputAssemblyState(regs); - fixed_state.tessellation = GetTessellationState(regs); - fixed_state.rasterizer = GetRasterizerState(regs); - fixed_state.depth_stencil = GetDepthStencilState(regs); - fixed_state.color_blending = GetColorBlendingState(regs); - return fixed_state; +Maxwell::Blend::Factor FixedPipelineState::UnpackBlendFactor(u32 packed) noexcept { + static constexpr std::array LUT = { + Maxwell::Blend::Factor::Zero, + Maxwell::Blend::Factor::One, + Maxwell::Blend::Factor::SourceColor, + Maxwell::Blend::Factor::OneMinusSourceColor, + Maxwell::Blend::Factor::SourceAlpha, + Maxwell::Blend::Factor::OneMinusSourceAlpha, + Maxwell::Blend::Factor::DestAlpha, + Maxwell::Blend::Factor::OneMinusDestAlpha, + Maxwell::Blend::Factor::DestColor, + Maxwell::Blend::Factor::OneMinusDestColor, + Maxwell::Blend::Factor::SourceAlphaSaturate, + Maxwell::Blend::Factor::Source1Color, + Maxwell::Blend::Factor::OneMinusSource1Color, + Maxwell::Blend::Factor::Source1Alpha, + Maxwell::Blend::Factor::OneMinusSource1Alpha, + Maxwell::Blend::Factor::ConstantColor, + Maxwell::Blend::Factor::OneMinusConstantColor, + Maxwell::Blend::Factor::ConstantAlpha, + Maxwell::Blend::Factor::OneMinusConstantAlpha, + }; + return LUT[packed]; } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.h b/src/video_core/renderer_vulkan/fixed_pipeline_state.h index 4c8ba7f90..2c18eeaae 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.h +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.h @@ -7,6 +7,7 @@ #include <array> #include <type_traits> +#include "common/bit_field.h" #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" @@ -16,230 +17,184 @@ namespace Vulkan { using Maxwell = Tegra::Engines::Maxwell3D::Regs; -// TODO(Rodrigo): Optimize this structure. - struct FixedPipelineState { - using PixelFormat = VideoCore::Surface::PixelFormat; - - struct VertexBinding { - constexpr VertexBinding(u32 index, u32 stride, u32 divisor) - : index{index}, stride{stride}, divisor{divisor} {} - VertexBinding() = default; - - u32 index; - u32 stride; - u32 divisor; + static u32 PackComparisonOp(Maxwell::ComparisonOp op) noexcept; + static Maxwell::ComparisonOp UnpackComparisonOp(u32 packed) noexcept; - std::size_t Hash() const noexcept; + static u32 PackStencilOp(Maxwell::StencilOp op) noexcept; + static Maxwell::StencilOp UnpackStencilOp(u32 packed) noexcept; - bool operator==(const VertexBinding& rhs) const noexcept; + static u32 PackCullFace(Maxwell::CullFace cull) noexcept; + static Maxwell::CullFace UnpackCullFace(u32 packed) noexcept; - bool operator!=(const VertexBinding& rhs) const noexcept { - return !operator==(rhs); - } - }; + static u32 PackFrontFace(Maxwell::FrontFace face) noexcept; + static Maxwell::FrontFace UnpackFrontFace(u32 packed) noexcept; - struct VertexAttribute { - constexpr VertexAttribute(u32 index, u32 buffer, Maxwell::VertexAttribute::Type type, - Maxwell::VertexAttribute::Size size, u32 offset) - : index{index}, buffer{buffer}, type{type}, size{size}, offset{offset} {} - VertexAttribute() = default; + static u32 PackPolygonMode(Maxwell::PolygonMode mode) noexcept; + static Maxwell::PolygonMode UnpackPolygonMode(u32 packed) noexcept; - u32 index; - u32 buffer; - Maxwell::VertexAttribute::Type type; - Maxwell::VertexAttribute::Size size; - u32 offset; + static u32 PackLogicOp(Maxwell::LogicOperation op) noexcept; + static Maxwell::LogicOperation UnpackLogicOp(u32 packed) noexcept; - std::size_t Hash() const noexcept; + static u32 PackBlendEquation(Maxwell::Blend::Equation equation) noexcept; + static Maxwell::Blend::Equation UnpackBlendEquation(u32 packed) noexcept; - bool operator==(const VertexAttribute& rhs) const noexcept; + static u32 PackBlendFactor(Maxwell::Blend::Factor factor) noexcept; + static Maxwell::Blend::Factor UnpackBlendFactor(u32 packed) noexcept; - bool operator!=(const VertexAttribute& rhs) const noexcept { - return !operator==(rhs); + struct BlendingAttachment { + union { + u32 raw; + BitField<0, 1, u32> mask_r; + BitField<1, 1, u32> mask_g; + BitField<2, 1, u32> mask_b; + BitField<3, 1, u32> mask_a; + BitField<4, 3, u32> equation_rgb; + BitField<7, 3, u32> equation_a; + BitField<10, 5, u32> factor_source_rgb; + BitField<15, 5, u32> factor_dest_rgb; + BitField<20, 5, u32> factor_source_a; + BitField<25, 5, u32> factor_dest_a; + BitField<30, 1, u32> enable; + }; + + void Fill(const Maxwell& regs, std::size_t index); + + constexpr std::array<bool, 4> Mask() const noexcept { + return {mask_r != 0, mask_g != 0, mask_b != 0, mask_a != 0}; } - }; - - struct StencilFace { - constexpr StencilFace(Maxwell::StencilOp action_stencil_fail, - Maxwell::StencilOp action_depth_fail, - Maxwell::StencilOp action_depth_pass, Maxwell::ComparisonOp test_func) - : action_stencil_fail{action_stencil_fail}, action_depth_fail{action_depth_fail}, - action_depth_pass{action_depth_pass}, test_func{test_func} {} - StencilFace() = default; - - Maxwell::StencilOp action_stencil_fail; - Maxwell::StencilOp action_depth_fail; - Maxwell::StencilOp action_depth_pass; - Maxwell::ComparisonOp test_func; - std::size_t Hash() const noexcept; - - bool operator==(const StencilFace& rhs) const noexcept; - - bool operator!=(const StencilFace& rhs) const noexcept { - return !operator==(rhs); + Maxwell::Blend::Equation EquationRGB() const noexcept { + return UnpackBlendEquation(equation_rgb.Value()); } - }; - struct BlendingAttachment { - constexpr BlendingAttachment(bool enable, Maxwell::Blend::Equation rgb_equation, - Maxwell::Blend::Factor src_rgb_func, - Maxwell::Blend::Factor dst_rgb_func, - Maxwell::Blend::Equation a_equation, - Maxwell::Blend::Factor src_a_func, - Maxwell::Blend::Factor dst_a_func, - std::array<bool, 4> components) - : enable{enable}, rgb_equation{rgb_equation}, src_rgb_func{src_rgb_func}, - dst_rgb_func{dst_rgb_func}, a_equation{a_equation}, src_a_func{src_a_func}, - dst_a_func{dst_a_func}, components{components} {} - BlendingAttachment() = default; - - bool enable; - Maxwell::Blend::Equation rgb_equation; - Maxwell::Blend::Factor src_rgb_func; - Maxwell::Blend::Factor dst_rgb_func; - Maxwell::Blend::Equation a_equation; - Maxwell::Blend::Factor src_a_func; - Maxwell::Blend::Factor dst_a_func; - std::array<bool, 4> components; - - std::size_t Hash() const noexcept; - - bool operator==(const BlendingAttachment& rhs) const noexcept; - - bool operator!=(const BlendingAttachment& rhs) const noexcept { - return !operator==(rhs); + Maxwell::Blend::Equation EquationAlpha() const noexcept { + return UnpackBlendEquation(equation_a.Value()); } - }; - struct VertexInput { - std::size_t num_bindings = 0; - std::size_t num_attributes = 0; - std::array<VertexBinding, Maxwell::NumVertexArrays> bindings; - std::array<VertexAttribute, Maxwell::NumVertexAttributes> attributes; + Maxwell::Blend::Factor SourceRGBFactor() const noexcept { + return UnpackBlendFactor(factor_source_rgb.Value()); + } - std::size_t Hash() const noexcept; + Maxwell::Blend::Factor DestRGBFactor() const noexcept { + return UnpackBlendFactor(factor_dest_rgb.Value()); + } - bool operator==(const VertexInput& rhs) const noexcept; + Maxwell::Blend::Factor SourceAlphaFactor() const noexcept { + return UnpackBlendFactor(factor_source_a.Value()); + } - bool operator!=(const VertexInput& rhs) const noexcept { - return !operator==(rhs); + Maxwell::Blend::Factor DestAlphaFactor() const noexcept { + return UnpackBlendFactor(factor_dest_a.Value()); } }; - struct InputAssembly { - constexpr InputAssembly(Maxwell::PrimitiveTopology topology, bool primitive_restart_enable, - float point_size) - : topology{topology}, primitive_restart_enable{primitive_restart_enable}, - point_size{point_size} {} - InputAssembly() = default; - - Maxwell::PrimitiveTopology topology; - bool primitive_restart_enable; - float point_size; + union VertexAttribute { + u32 raw; + BitField<0, 1, u32> enabled; + BitField<1, 5, u32> buffer; + BitField<6, 14, u32> offset; + BitField<20, 3, u32> type; + BitField<23, 6, u32> size; - std::size_t Hash() const noexcept; - - bool operator==(const InputAssembly& rhs) const noexcept; + constexpr Maxwell::VertexAttribute::Type Type() const noexcept { + return static_cast<Maxwell::VertexAttribute::Type>(type.Value()); + } - bool operator!=(const InputAssembly& rhs) const noexcept { - return !operator==(rhs); + constexpr Maxwell::VertexAttribute::Size Size() const noexcept { + return static_cast<Maxwell::VertexAttribute::Size>(size.Value()); } }; - struct Tessellation { - constexpr Tessellation(u32 patch_control_points, Maxwell::TessellationPrimitive primitive, - Maxwell::TessellationSpacing spacing, bool clockwise) - : patch_control_points{patch_control_points}, primitive{primitive}, spacing{spacing}, - clockwise{clockwise} {} - Tessellation() = default; + template <std::size_t Position> + union StencilFace { + BitField<Position + 0, 3, u32> action_stencil_fail; + BitField<Position + 3, 3, u32> action_depth_fail; + BitField<Position + 6, 3, u32> action_depth_pass; + BitField<Position + 9, 3, u32> test_func; - u32 patch_control_points; - Maxwell::TessellationPrimitive primitive; - Maxwell::TessellationSpacing spacing; - bool clockwise; + Maxwell::StencilOp ActionStencilFail() const noexcept { + return UnpackStencilOp(action_stencil_fail); + } - std::size_t Hash() const noexcept; + Maxwell::StencilOp ActionDepthFail() const noexcept { + return UnpackStencilOp(action_depth_fail); + } - bool operator==(const Tessellation& rhs) const noexcept; + Maxwell::StencilOp ActionDepthPass() const noexcept { + return UnpackStencilOp(action_depth_pass); + } - bool operator!=(const Tessellation& rhs) const noexcept { - return !operator==(rhs); + Maxwell::ComparisonOp TestFunc() const noexcept { + return UnpackComparisonOp(test_func); } }; - struct Rasterizer { - constexpr Rasterizer(bool cull_enable, bool depth_bias_enable, bool depth_clamp_enable, - bool ndc_minus_one_to_one, Maxwell::CullFace cull_face, - Maxwell::FrontFace front_face) - : cull_enable{cull_enable}, depth_bias_enable{depth_bias_enable}, - depth_clamp_enable{depth_clamp_enable}, ndc_minus_one_to_one{ndc_minus_one_to_one}, - cull_face{cull_face}, front_face{front_face} {} - Rasterizer() = default; - - bool cull_enable; - bool depth_bias_enable; - bool depth_clamp_enable; - bool ndc_minus_one_to_one; - Maxwell::CullFace cull_face; - Maxwell::FrontFace front_face; - - std::size_t Hash() const noexcept; + union VertexBinding { + u16 raw; + BitField<0, 12, u16> stride; + BitField<12, 1, u16> enabled; + }; - bool operator==(const Rasterizer& rhs) const noexcept; + struct DynamicState { + union { + u32 raw1; + StencilFace<0> front; + StencilFace<12> back; + BitField<24, 1, u32> stencil_enable; + BitField<25, 1, u32> depth_write_enable; + BitField<26, 1, u32> depth_bounds_enable; + BitField<27, 1, u32> depth_test_enable; + BitField<28, 1, u32> front_face; + BitField<29, 3, u32> depth_test_func; + }; + union { + u32 raw2; + BitField<0, 2, u32> cull_face; + BitField<2, 1, u32> cull_enable; + }; + std::array<VertexBinding, Maxwell::NumVertexArrays> vertex_bindings; + + void Fill(const Maxwell& regs); + + Maxwell::ComparisonOp DepthTestFunc() const noexcept { + return UnpackComparisonOp(depth_test_func); + } - bool operator!=(const Rasterizer& rhs) const noexcept { - return !operator==(rhs); + Maxwell::CullFace CullFace() const noexcept { + return UnpackCullFace(cull_face.Value()); } - }; - struct DepthStencil { - constexpr DepthStencil(bool depth_test_enable, bool depth_write_enable, - bool depth_bounds_enable, bool stencil_enable, - Maxwell::ComparisonOp depth_test_function, StencilFace front_stencil, - StencilFace back_stencil) - : depth_test_enable{depth_test_enable}, depth_write_enable{depth_write_enable}, - depth_bounds_enable{depth_bounds_enable}, stencil_enable{stencil_enable}, - depth_test_function{depth_test_function}, front_stencil{front_stencil}, - back_stencil{back_stencil} {} - DepthStencil() = default; - - bool depth_test_enable; - bool depth_write_enable; - bool depth_bounds_enable; - bool stencil_enable; - Maxwell::ComparisonOp depth_test_function; - StencilFace front_stencil; - StencilFace back_stencil; - - std::size_t Hash() const noexcept; - - bool operator==(const DepthStencil& rhs) const noexcept; - - bool operator!=(const DepthStencil& rhs) const noexcept { - return !operator==(rhs); + Maxwell::FrontFace FrontFace() const noexcept { + return UnpackFrontFace(front_face.Value()); } }; - struct ColorBlending { - constexpr ColorBlending( - std::array<float, 4> blend_constants, std::size_t attachments_count, - std::array<BlendingAttachment, Maxwell::NumRenderTargets> attachments) - : attachments_count{attachments_count}, attachments{attachments} {} - ColorBlending() = default; - - std::size_t attachments_count; - std::array<BlendingAttachment, Maxwell::NumRenderTargets> attachments; - - std::size_t Hash() const noexcept; - - bool operator==(const ColorBlending& rhs) const noexcept; - - bool operator!=(const ColorBlending& rhs) const noexcept { - return !operator==(rhs); - } + union { + u32 raw; + BitField<0, 1, u32> no_extended_dynamic_state; + BitField<2, 1, u32> primitive_restart_enable; + BitField<3, 1, u32> depth_bias_enable; + BitField<4, 1, u32> depth_clamp_disabled; + BitField<5, 1, u32> ndc_minus_one_to_one; + BitField<6, 2, u32> polygon_mode; + BitField<8, 5, u32> patch_control_points_minus_one; + BitField<13, 2, u32> tessellation_primitive; + BitField<15, 2, u32> tessellation_spacing; + BitField<17, 1, u32> tessellation_clockwise; + BitField<18, 1, u32> logic_op_enable; + BitField<19, 4, u32> logic_op; + BitField<23, 1, u32> rasterize_enable; + BitField<24, 4, Maxwell::PrimitiveTopology> topology; }; + u32 point_size; + std::array<u32, Maxwell::NumVertexArrays> binding_divisors; + std::array<VertexAttribute, Maxwell::NumVertexAttributes> attributes; + std::array<BlendingAttachment, Maxwell::NumRenderTargets> attachments; + std::array<u16, Maxwell::NumViewports> viewport_swizzles; + DynamicState dynamic_state; + + void Fill(const Maxwell& regs, bool has_extended_dynamic_state); std::size_t Hash() const noexcept; @@ -249,26 +204,14 @@ struct FixedPipelineState { return !operator==(rhs); } - VertexInput vertex_input; - InputAssembly input_assembly; - Tessellation tessellation; - Rasterizer rasterizer; - DepthStencil depth_stencil; - ColorBlending color_blending; + std::size_t Size() const noexcept { + const std::size_t total_size = sizeof *this; + return total_size - (no_extended_dynamic_state != 0 ? 0 : sizeof(DynamicState)); + } }; -static_assert(std::is_trivially_copyable_v<FixedPipelineState::VertexBinding>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::VertexAttribute>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::StencilFace>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::BlendingAttachment>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::VertexInput>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::InputAssembly>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::Tessellation>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::Rasterizer>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::DepthStencil>); -static_assert(std::is_trivially_copyable_v<FixedPipelineState::ColorBlending>); +static_assert(std::has_unique_object_representations_v<FixedPipelineState>); static_assert(std::is_trivially_copyable_v<FixedPipelineState>); - -FixedPipelineState GetFixedPipelineState(const Maxwell& regs); +static_assert(std::is_trivially_constructible_v<FixedPipelineState>); } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index 8681b821f..d22de1d81 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -21,29 +21,29 @@ namespace Sampler { VkFilter Filter(Tegra::Texture::TextureFilter filter) { switch (filter) { - case Tegra::Texture::TextureFilter::Linear: - return VK_FILTER_LINEAR; case Tegra::Texture::TextureFilter::Nearest: return VK_FILTER_NEAREST; + case Tegra::Texture::TextureFilter::Linear: + return VK_FILTER_LINEAR; } - UNIMPLEMENTED_MSG("Unimplemented sampler filter={}", static_cast<u32>(filter)); + UNREACHABLE_MSG("Invalid sampler filter={}", static_cast<u32>(filter)); return {}; } VkSamplerMipmapMode MipmapMode(Tegra::Texture::TextureMipmapFilter mipmap_filter) { switch (mipmap_filter) { case Tegra::Texture::TextureMipmapFilter::None: - // TODO(Rodrigo): None seems to be mapped to OpenGL's mag and min filters without mipmapping - // (e.g. GL_NEAREST and GL_LINEAR). Vulkan doesn't have such a thing, find out if we have to - // use an image view with a single mipmap level to emulate this. - return VK_SAMPLER_MIPMAP_MODE_LINEAR; - ; - case Tegra::Texture::TextureMipmapFilter::Linear: - return VK_SAMPLER_MIPMAP_MODE_LINEAR; + // There are no Vulkan filter modes that directly correspond to OpenGL minification filters + // of GL_LINEAR or GL_NEAREST, but they can be emulated using + // VK_SAMPLER_MIPMAP_MODE_NEAREST, minLod = 0, and maxLod = 0.25, and using minFilter = + // VK_FILTER_LINEAR or minFilter = VK_FILTER_NEAREST, respectively. + return VK_SAMPLER_MIPMAP_MODE_NEAREST; case Tegra::Texture::TextureMipmapFilter::Nearest: return VK_SAMPLER_MIPMAP_MODE_NEAREST; + case Tegra::Texture::TextureMipmapFilter::Linear: + return VK_SAMPLER_MIPMAP_MODE_LINEAR; } - UNIMPLEMENTED_MSG("Unimplemented sampler mipmap mode={}", static_cast<u32>(mipmap_filter)); + UNREACHABLE_MSG("Invalid sampler mipmap mode={}", static_cast<u32>(mipmap_filter)); return {}; } @@ -118,89 +118,101 @@ struct FormatTuple { VkFormat format; ///< Vulkan format int usage = 0; ///< Describes image format usage } constexpr tex_format_tuples[] = { - {VK_FORMAT_A8B8G8R8_UNORM_PACK32, Attachable | Storage}, // ABGR8U - {VK_FORMAT_A8B8G8R8_SNORM_PACK32, Attachable | Storage}, // ABGR8S - {VK_FORMAT_A8B8G8R8_UINT_PACK32, Attachable | Storage}, // ABGR8UI - {VK_FORMAT_B5G6R5_UNORM_PACK16}, // B5G6R5U - {VK_FORMAT_A2B10G10R10_UNORM_PACK32, Attachable | Storage}, // A2B10G10R10U - {VK_FORMAT_A1R5G5B5_UNORM_PACK16, Attachable}, // A1B5G5R5U (flipped with swizzle) - {VK_FORMAT_R8_UNORM, Attachable | Storage}, // R8U - {VK_FORMAT_R8_UINT, Attachable | Storage}, // R8UI - {VK_FORMAT_R16G16B16A16_SFLOAT, Attachable | Storage}, // RGBA16F - {VK_FORMAT_R16G16B16A16_UNORM, Attachable | Storage}, // RGBA16U - {VK_FORMAT_R16G16B16A16_SNORM, Attachable | Storage}, // RGBA16S - {VK_FORMAT_R16G16B16A16_UINT, Attachable | Storage}, // RGBA16UI - {VK_FORMAT_B10G11R11_UFLOAT_PACK32, Attachable | Storage}, // R11FG11FB10F - {VK_FORMAT_R32G32B32A32_UINT, Attachable | Storage}, // RGBA32UI - {VK_FORMAT_BC1_RGBA_UNORM_BLOCK}, // DXT1 - {VK_FORMAT_BC2_UNORM_BLOCK}, // DXT23 - {VK_FORMAT_BC3_UNORM_BLOCK}, // DXT45 - {VK_FORMAT_BC4_UNORM_BLOCK}, // DXN1 - {VK_FORMAT_BC5_UNORM_BLOCK}, // DXN2UNORM - {VK_FORMAT_BC5_SNORM_BLOCK}, // DXN2SNORM - {VK_FORMAT_BC7_UNORM_BLOCK}, // BC7U - {VK_FORMAT_BC6H_UFLOAT_BLOCK}, // BC6H_UF16 - {VK_FORMAT_BC6H_SFLOAT_BLOCK}, // BC6H_SF16 - {VK_FORMAT_ASTC_4x4_UNORM_BLOCK}, // ASTC_2D_4X4 - {VK_FORMAT_B8G8R8A8_UNORM}, // BGRA8 - {VK_FORMAT_R32G32B32A32_SFLOAT, Attachable | Storage}, // RGBA32F - {VK_FORMAT_R32G32_SFLOAT, Attachable | Storage}, // RG32F - {VK_FORMAT_R32_SFLOAT, Attachable | Storage}, // R32F - {VK_FORMAT_R16_SFLOAT, Attachable | Storage}, // R16F - {VK_FORMAT_R16_UNORM, Attachable | Storage}, // R16U - {VK_FORMAT_UNDEFINED}, // R16S - {VK_FORMAT_UNDEFINED}, // R16UI - {VK_FORMAT_UNDEFINED}, // R16I - {VK_FORMAT_R16G16_UNORM, Attachable | Storage}, // RG16 - {VK_FORMAT_R16G16_SFLOAT, Attachable | Storage}, // RG16F - {VK_FORMAT_UNDEFINED}, // RG16UI - {VK_FORMAT_UNDEFINED}, // RG16I - {VK_FORMAT_R16G16_SNORM, Attachable | Storage}, // RG16S - {VK_FORMAT_UNDEFINED}, // RGB32F - {VK_FORMAT_R8G8B8A8_SRGB, Attachable}, // RGBA8_SRGB - {VK_FORMAT_R8G8_UNORM, Attachable | Storage}, // RG8U - {VK_FORMAT_R8G8_SNORM, Attachable | Storage}, // RG8S - {VK_FORMAT_R32G32_UINT, Attachable | Storage}, // RG32UI - {VK_FORMAT_UNDEFINED}, // RGBX16F - {VK_FORMAT_R32_UINT, Attachable | Storage}, // R32UI - {VK_FORMAT_R32_SINT, Attachable | Storage}, // R32I - {VK_FORMAT_ASTC_8x8_UNORM_BLOCK}, // ASTC_2D_8X8 - {VK_FORMAT_UNDEFINED}, // ASTC_2D_8X5 - {VK_FORMAT_UNDEFINED}, // ASTC_2D_5X4 - {VK_FORMAT_UNDEFINED}, // BGRA8_SRGB - {VK_FORMAT_BC1_RGBA_SRGB_BLOCK}, // DXT1_SRGB - {VK_FORMAT_BC2_SRGB_BLOCK}, // DXT23_SRGB - {VK_FORMAT_BC3_SRGB_BLOCK}, // DXT45_SRGB - {VK_FORMAT_BC7_SRGB_BLOCK}, // BC7U_SRGB - {VK_FORMAT_R4G4B4A4_UNORM_PACK16, Attachable}, // R4G4B4A4U - {VK_FORMAT_ASTC_4x4_SRGB_BLOCK}, // ASTC_2D_4X4_SRGB - {VK_FORMAT_ASTC_8x8_SRGB_BLOCK}, // ASTC_2D_8X8_SRGB - {VK_FORMAT_ASTC_8x5_SRGB_BLOCK}, // ASTC_2D_8X5_SRGB - {VK_FORMAT_ASTC_5x4_SRGB_BLOCK}, // ASTC_2D_5X4_SRGB - {VK_FORMAT_ASTC_5x5_UNORM_BLOCK}, // ASTC_2D_5X5 - {VK_FORMAT_ASTC_5x5_SRGB_BLOCK}, // ASTC_2D_5X5_SRGB - {VK_FORMAT_ASTC_10x8_UNORM_BLOCK}, // ASTC_2D_10X8 - {VK_FORMAT_ASTC_10x8_SRGB_BLOCK}, // ASTC_2D_10X8_SRGB - {VK_FORMAT_ASTC_6x6_UNORM_BLOCK}, // ASTC_2D_6X6 - {VK_FORMAT_ASTC_6x6_SRGB_BLOCK}, // ASTC_2D_6X6_SRGB - {VK_FORMAT_ASTC_10x10_UNORM_BLOCK}, // ASTC_2D_10X10 - {VK_FORMAT_ASTC_10x10_SRGB_BLOCK}, // ASTC_2D_10X10_SRGB - {VK_FORMAT_ASTC_12x12_UNORM_BLOCK}, // ASTC_2D_12X12 - {VK_FORMAT_ASTC_12x12_SRGB_BLOCK}, // ASTC_2D_12X12_SRGB - {VK_FORMAT_ASTC_8x6_UNORM_BLOCK}, // ASTC_2D_8X6 - {VK_FORMAT_ASTC_8x6_SRGB_BLOCK}, // ASTC_2D_8X6_SRGB - {VK_FORMAT_ASTC_6x5_UNORM_BLOCK}, // ASTC_2D_6X5 - {VK_FORMAT_ASTC_6x5_SRGB_BLOCK}, // ASTC_2D_6X5_SRGB - {VK_FORMAT_E5B9G9R9_UFLOAT_PACK32}, // E5B9G9R9F + {VK_FORMAT_A8B8G8R8_UNORM_PACK32, Attachable | Storage}, // A8B8G8R8_UNORM + {VK_FORMAT_A8B8G8R8_SNORM_PACK32, Attachable | Storage}, // A8B8G8R8_SNORM + {VK_FORMAT_A8B8G8R8_SINT_PACK32, Attachable | Storage}, // A8B8G8R8_SINT + {VK_FORMAT_A8B8G8R8_UINT_PACK32, Attachable | Storage}, // A8B8G8R8_UINT + {VK_FORMAT_R5G6B5_UNORM_PACK16, Attachable}, // R5G6B5_UNORM + {VK_FORMAT_B5G6R5_UNORM_PACK16, Attachable}, // B5G6R5_UNORM + {VK_FORMAT_A1R5G5B5_UNORM_PACK16, Attachable}, // A1R5G5B5_UNORM + {VK_FORMAT_A2B10G10R10_UNORM_PACK32, Attachable | Storage}, // A2B10G10R10_UNORM + {VK_FORMAT_A2B10G10R10_UINT_PACK32, Attachable | Storage}, // A2B10G10R10_UINT + {VK_FORMAT_A1R5G5B5_UNORM_PACK16, Attachable}, // A1B5G5R5_UNORM (flipped with swizzle) + {VK_FORMAT_R8_UNORM, Attachable | Storage}, // R8_UNORM + {VK_FORMAT_R8_SNORM, Attachable | Storage}, // R8_SNORM + {VK_FORMAT_R8_SINT, Attachable | Storage}, // R8_SINT + {VK_FORMAT_R8_UINT, Attachable | Storage}, // R8_UINT + {VK_FORMAT_R16G16B16A16_SFLOAT, Attachable | Storage}, // R16G16B16A16_FLOAT + {VK_FORMAT_R16G16B16A16_UNORM, Attachable | Storage}, // R16G16B16A16_UNORM + {VK_FORMAT_R16G16B16A16_SNORM, Attachable | Storage}, // R16G16B16A16_SNORM + {VK_FORMAT_R16G16B16A16_SINT, Attachable | Storage}, // R16G16B16A16_SINT + {VK_FORMAT_R16G16B16A16_UINT, Attachable | Storage}, // R16G16B16A16_UINT + {VK_FORMAT_B10G11R11_UFLOAT_PACK32, Attachable | Storage}, // B10G11R11_FLOAT + {VK_FORMAT_R32G32B32A32_UINT, Attachable | Storage}, // R32G32B32A32_UINT + {VK_FORMAT_BC1_RGBA_UNORM_BLOCK}, // BC1_RGBA_UNORM + {VK_FORMAT_BC2_UNORM_BLOCK}, // BC2_UNORM + {VK_FORMAT_BC3_UNORM_BLOCK}, // BC3_UNORM + {VK_FORMAT_BC4_UNORM_BLOCK}, // BC4_UNORM + {VK_FORMAT_BC4_SNORM_BLOCK}, // BC4_SNORM + {VK_FORMAT_BC5_UNORM_BLOCK}, // BC5_UNORM + {VK_FORMAT_BC5_SNORM_BLOCK}, // BC5_SNORM + {VK_FORMAT_BC7_UNORM_BLOCK}, // BC7_UNORM + {VK_FORMAT_BC6H_UFLOAT_BLOCK}, // BC6H_UFLOAT + {VK_FORMAT_BC6H_SFLOAT_BLOCK}, // BC6H_SFLOAT + {VK_FORMAT_ASTC_4x4_UNORM_BLOCK}, // ASTC_2D_4X4_UNORM + {VK_FORMAT_B8G8R8A8_UNORM, Attachable}, // B8G8R8A8_UNORM + {VK_FORMAT_R32G32B32A32_SFLOAT, Attachable | Storage}, // R32G32B32A32_FLOAT + {VK_FORMAT_R32G32B32A32_SINT, Attachable | Storage}, // R32G32B32A32_SINT + {VK_FORMAT_R32G32_SFLOAT, Attachable | Storage}, // R32G32_FLOAT + {VK_FORMAT_R32G32_SINT, Attachable | Storage}, // R32G32_SINT + {VK_FORMAT_R32_SFLOAT, Attachable | Storage}, // R32_FLOAT + {VK_FORMAT_R16_SFLOAT, Attachable | Storage}, // R16_FLOAT + {VK_FORMAT_R16_UNORM, Attachable | Storage}, // R16_UNORM + {VK_FORMAT_UNDEFINED}, // R16_SNORM + {VK_FORMAT_R16_UINT, Attachable | Storage}, // R16_UINT + {VK_FORMAT_UNDEFINED}, // R16_SINT + {VK_FORMAT_R16G16_UNORM, Attachable | Storage}, // R16G16_UNORM + {VK_FORMAT_R16G16_SFLOAT, Attachable | Storage}, // R16G16_FLOAT + {VK_FORMAT_UNDEFINED}, // R16G16_UINT + {VK_FORMAT_UNDEFINED}, // R16G16_SINT + {VK_FORMAT_R16G16_SNORM, Attachable | Storage}, // R16G16_SNORM + {VK_FORMAT_UNDEFINED}, // R32G32B32_FLOAT + {VK_FORMAT_R8G8B8A8_SRGB, Attachable}, // A8B8G8R8_SRGB + {VK_FORMAT_R8G8_UNORM, Attachable | Storage}, // R8G8_UNORM + {VK_FORMAT_R8G8_SNORM, Attachable | Storage}, // R8G8_SNORM + {VK_FORMAT_R8G8_SINT, Attachable | Storage}, // R8G8_SINT + {VK_FORMAT_R8G8_UINT, Attachable | Storage}, // R8G8_UINT + {VK_FORMAT_R32G32_UINT, Attachable | Storage}, // R32G32_UINT + {VK_FORMAT_UNDEFINED}, // R16G16B16X16_FLOAT + {VK_FORMAT_R32_UINT, Attachable | Storage}, // R32_UINT + {VK_FORMAT_R32_SINT, Attachable | Storage}, // R32_SINT + {VK_FORMAT_ASTC_8x8_UNORM_BLOCK}, // ASTC_2D_8X8_UNORM + {VK_FORMAT_UNDEFINED}, // ASTC_2D_8X5_UNORM + {VK_FORMAT_UNDEFINED}, // ASTC_2D_5X4_UNORM + {VK_FORMAT_B8G8R8A8_SRGB, Attachable}, // B8G8R8A8_SRGB + {VK_FORMAT_BC1_RGBA_SRGB_BLOCK}, // BC1_RGBA_SRGB + {VK_FORMAT_BC2_SRGB_BLOCK}, // BC2_SRGB + {VK_FORMAT_BC3_SRGB_BLOCK}, // BC3_SRGB + {VK_FORMAT_BC7_SRGB_BLOCK}, // BC7_SRGB + {VK_FORMAT_R4G4B4A4_UNORM_PACK16, Attachable}, // A4B4G4R4_UNORM + {VK_FORMAT_ASTC_4x4_SRGB_BLOCK}, // ASTC_2D_4X4_SRGB + {VK_FORMAT_ASTC_8x8_SRGB_BLOCK}, // ASTC_2D_8X8_SRGB + {VK_FORMAT_ASTC_8x5_SRGB_BLOCK}, // ASTC_2D_8X5_SRGB + {VK_FORMAT_ASTC_5x4_SRGB_BLOCK}, // ASTC_2D_5X4_SRGB + {VK_FORMAT_ASTC_5x5_UNORM_BLOCK}, // ASTC_2D_5X5_UNORM + {VK_FORMAT_ASTC_5x5_SRGB_BLOCK}, // ASTC_2D_5X5_SRGB + {VK_FORMAT_ASTC_10x8_UNORM_BLOCK}, // ASTC_2D_10X8_UNORM + {VK_FORMAT_ASTC_10x8_SRGB_BLOCK}, // ASTC_2D_10X8_SRGB + {VK_FORMAT_ASTC_6x6_UNORM_BLOCK}, // ASTC_2D_6X6_UNORM + {VK_FORMAT_ASTC_6x6_SRGB_BLOCK}, // ASTC_2D_6X6_SRGB + {VK_FORMAT_ASTC_10x10_UNORM_BLOCK}, // ASTC_2D_10X10_UNORM + {VK_FORMAT_ASTC_10x10_SRGB_BLOCK}, // ASTC_2D_10X10_SRGB + {VK_FORMAT_ASTC_12x12_UNORM_BLOCK}, // ASTC_2D_12X12_UNORM + {VK_FORMAT_ASTC_12x12_SRGB_BLOCK}, // ASTC_2D_12X12_SRGB + {VK_FORMAT_ASTC_8x6_UNORM_BLOCK}, // ASTC_2D_8X6_UNORM + {VK_FORMAT_ASTC_8x6_SRGB_BLOCK}, // ASTC_2D_8X6_SRGB + {VK_FORMAT_ASTC_6x5_UNORM_BLOCK}, // ASTC_2D_6X5_UNORM + {VK_FORMAT_ASTC_6x5_SRGB_BLOCK}, // ASTC_2D_6X5_SRGB + {VK_FORMAT_E5B9G9R9_UFLOAT_PACK32}, // E5B9G9R9_FLOAT // Depth formats - {VK_FORMAT_D32_SFLOAT, Attachable}, // Z32F - {VK_FORMAT_D16_UNORM, Attachable}, // Z16 + {VK_FORMAT_D32_SFLOAT, Attachable}, // D32_FLOAT + {VK_FORMAT_D16_UNORM, Attachable}, // D16_UNORM // DepthStencil formats - {VK_FORMAT_D24_UNORM_S8_UINT, Attachable}, // Z24S8 - {VK_FORMAT_D24_UNORM_S8_UINT, Attachable}, // S8Z24 (emulated) - {VK_FORMAT_D32_SFLOAT_S8_UINT, Attachable}, // Z32FS8 + {VK_FORMAT_D24_UNORM_S8_UINT, Attachable}, // D24_UNORM_S8_UINT + {VK_FORMAT_D24_UNORM_S8_UINT, Attachable}, // S8_UINT_D24_UNORM (emulated) + {VK_FORMAT_D32_SFLOAT_S8_UINT, Attachable}, // D32_FLOAT_S8_UINT }; static_assert(std::size(tex_format_tuples) == VideoCore::Surface::MaxPixelFormat); @@ -221,7 +233,7 @@ FormatInfo SurfaceFormat(const VKDevice& device, FormatType format_type, PixelFo return {VK_FORMAT_A8B8G8R8_UNORM_PACK32, true, true}; } - // Use ABGR8 on hardware that doesn't support ASTC natively + // Use A8B8G8R8_UNORM on hardware that doesn't support ASTC natively if (!device.IsOptimalAstcSupported() && VideoCore::Surface::IsPixelFormatASTC(pixel_format)) { tuple.format = VideoCore::Surface::IsPixelFormatSRGB(pixel_format) ? VK_FORMAT_A8B8G8R8_SRGB_PACK32 @@ -295,6 +307,30 @@ VkPrimitiveTopology PrimitiveTopology([[maybe_unused]] const VKDevice& device, VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttribute::Size size) { switch (type) { + case Maxwell::VertexAttribute::Type::UnsignedNorm: + switch (size) { + case Maxwell::VertexAttribute::Size::Size_8: + return VK_FORMAT_R8_UNORM; + case Maxwell::VertexAttribute::Size::Size_8_8: + return VK_FORMAT_R8G8_UNORM; + case Maxwell::VertexAttribute::Size::Size_8_8_8: + return VK_FORMAT_R8G8B8_UNORM; + case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + return VK_FORMAT_R8G8B8A8_UNORM; + case Maxwell::VertexAttribute::Size::Size_16: + return VK_FORMAT_R16_UNORM; + case Maxwell::VertexAttribute::Size::Size_16_16: + return VK_FORMAT_R16G16_UNORM; + case Maxwell::VertexAttribute::Size::Size_16_16_16: + return VK_FORMAT_R16G16B16_UNORM; + case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + return VK_FORMAT_R16G16B16A16_UNORM; + case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + return VK_FORMAT_A2B10G10R10_UNORM_PACK32; + default: + break; + } + break; case Maxwell::VertexAttribute::Type::SignedNorm: switch (size) { case Maxwell::VertexAttribute::Size::Size_8: @@ -319,44 +355,50 @@ VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttrib break; } break; - case Maxwell::VertexAttribute::Type::UnsignedNorm: + case Maxwell::VertexAttribute::Type::UnsignedScaled: switch (size) { case Maxwell::VertexAttribute::Size::Size_8: - return VK_FORMAT_R8_UNORM; + return VK_FORMAT_R8_USCALED; case Maxwell::VertexAttribute::Size::Size_8_8: - return VK_FORMAT_R8G8_UNORM; + return VK_FORMAT_R8G8_USCALED; case Maxwell::VertexAttribute::Size::Size_8_8_8: - return VK_FORMAT_R8G8B8_UNORM; + return VK_FORMAT_R8G8B8_USCALED; case Maxwell::VertexAttribute::Size::Size_8_8_8_8: - return VK_FORMAT_R8G8B8A8_UNORM; + return VK_FORMAT_R8G8B8A8_USCALED; case Maxwell::VertexAttribute::Size::Size_16: - return VK_FORMAT_R16_UNORM; + return VK_FORMAT_R16_USCALED; case Maxwell::VertexAttribute::Size::Size_16_16: - return VK_FORMAT_R16G16_UNORM; + return VK_FORMAT_R16G16_USCALED; case Maxwell::VertexAttribute::Size::Size_16_16_16: - return VK_FORMAT_R16G16B16_UNORM; + return VK_FORMAT_R16G16B16_USCALED; case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return VK_FORMAT_R16G16B16A16_UNORM; + return VK_FORMAT_R16G16B16A16_USCALED; case Maxwell::VertexAttribute::Size::Size_10_10_10_2: - return VK_FORMAT_A2B10G10R10_UNORM_PACK32; + return VK_FORMAT_A2B10G10R10_USCALED_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::SignedInt: + case Maxwell::VertexAttribute::Type::SignedScaled: switch (size) { - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return VK_FORMAT_R16G16B16A16_SINT; case Maxwell::VertexAttribute::Size::Size_8: - return VK_FORMAT_R8_SINT; + return VK_FORMAT_R8_SSCALED; case Maxwell::VertexAttribute::Size::Size_8_8: - return VK_FORMAT_R8G8_SINT; + return VK_FORMAT_R8G8_SSCALED; case Maxwell::VertexAttribute::Size::Size_8_8_8: - return VK_FORMAT_R8G8B8_SINT; + return VK_FORMAT_R8G8B8_SSCALED; case Maxwell::VertexAttribute::Size::Size_8_8_8_8: - return VK_FORMAT_R8G8B8A8_SINT; - case Maxwell::VertexAttribute::Size::Size_32: - return VK_FORMAT_R32_SINT; + return VK_FORMAT_R8G8B8A8_SSCALED; + case Maxwell::VertexAttribute::Size::Size_16: + return VK_FORMAT_R16_SSCALED; + case Maxwell::VertexAttribute::Size::Size_16_16: + return VK_FORMAT_R16G16_SSCALED; + case Maxwell::VertexAttribute::Size::Size_16_16_16: + return VK_FORMAT_R16G16B16_SSCALED; + case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + return VK_FORMAT_R16G16B16A16_SSCALED; + case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + return VK_FORMAT_A2B10G10R10_SSCALED_PACK32; default: break; } @@ -387,56 +429,54 @@ VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttrib return VK_FORMAT_R32G32B32_UINT; case Maxwell::VertexAttribute::Size::Size_32_32_32_32: return VK_FORMAT_R32G32B32A32_UINT; + case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + return VK_FORMAT_A2B10G10R10_UINT_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::UnsignedScaled: + case Maxwell::VertexAttribute::Type::SignedInt: switch (size) { case Maxwell::VertexAttribute::Size::Size_8: - return VK_FORMAT_R8_USCALED; + return VK_FORMAT_R8_SINT; case Maxwell::VertexAttribute::Size::Size_8_8: - return VK_FORMAT_R8G8_USCALED; + return VK_FORMAT_R8G8_SINT; case Maxwell::VertexAttribute::Size::Size_8_8_8: - return VK_FORMAT_R8G8B8_USCALED; + return VK_FORMAT_R8G8B8_SINT; case Maxwell::VertexAttribute::Size::Size_8_8_8_8: - return VK_FORMAT_R8G8B8A8_USCALED; + return VK_FORMAT_R8G8B8A8_SINT; case Maxwell::VertexAttribute::Size::Size_16: - return VK_FORMAT_R16_USCALED; + return VK_FORMAT_R16_SINT; case Maxwell::VertexAttribute::Size::Size_16_16: - return VK_FORMAT_R16G16_USCALED; + return VK_FORMAT_R16G16_SINT; case Maxwell::VertexAttribute::Size::Size_16_16_16: - return VK_FORMAT_R16G16B16_USCALED; + return VK_FORMAT_R16G16B16_SINT; case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return VK_FORMAT_R16G16B16A16_USCALED; + return VK_FORMAT_R16G16B16A16_SINT; + case Maxwell::VertexAttribute::Size::Size_32: + return VK_FORMAT_R32_SINT; + case Maxwell::VertexAttribute::Size::Size_32_32: + return VK_FORMAT_R32G32_SINT; + case Maxwell::VertexAttribute::Size::Size_32_32_32: + return VK_FORMAT_R32G32B32_SINT; + case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + return VK_FORMAT_R32G32B32A32_SINT; + case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + return VK_FORMAT_A2B10G10R10_SINT_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::SignedScaled: + case Maxwell::VertexAttribute::Type::Float: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: - return VK_FORMAT_R8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_8_8: - return VK_FORMAT_R8G8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_8_8_8: - return VK_FORMAT_R8G8B8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: - return VK_FORMAT_R8G8B8A8_SSCALED; case Maxwell::VertexAttribute::Size::Size_16: - return VK_FORMAT_R16_SSCALED; + return VK_FORMAT_R16_SFLOAT; case Maxwell::VertexAttribute::Size::Size_16_16: - return VK_FORMAT_R16G16_SSCALED; + return VK_FORMAT_R16G16_SFLOAT; case Maxwell::VertexAttribute::Size::Size_16_16_16: - return VK_FORMAT_R16G16B16_SSCALED; + return VK_FORMAT_R16G16B16_SFLOAT; case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return VK_FORMAT_R16G16B16A16_SSCALED; - default: - break; - } - break; - case Maxwell::VertexAttribute::Type::Float: - switch (size) { + return VK_FORMAT_R16G16B16A16_SFLOAT; case Maxwell::VertexAttribute::Size::Size_32: return VK_FORMAT_R32_SFLOAT; case Maxwell::VertexAttribute::Size::Size_32_32: @@ -445,14 +485,6 @@ VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttrib return VK_FORMAT_R32G32B32_SFLOAT; case Maxwell::VertexAttribute::Size::Size_32_32_32_32: return VK_FORMAT_R32G32B32A32_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16: - return VK_FORMAT_R16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16_16: - return VK_FORMAT_R16G16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16_16_16: - return VK_FORMAT_R16G16B16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: - return VK_FORMAT_R16G16B16A16_SFLOAT; default: break; } @@ -672,4 +704,27 @@ VkComponentSwizzle SwizzleSource(Tegra::Texture::SwizzleSource swizzle) { return {}; } +VkViewportCoordinateSwizzleNV ViewportSwizzle(Maxwell::ViewportSwizzle swizzle) { + switch (swizzle) { + case Maxwell::ViewportSwizzle::PositiveX: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_X_NV; + case Maxwell::ViewportSwizzle::NegativeX: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_X_NV; + case Maxwell::ViewportSwizzle::PositiveY: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_Y_NV; + case Maxwell::ViewportSwizzle::NegativeY: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_Y_NV; + case Maxwell::ViewportSwizzle::PositiveZ: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_Z_NV; + case Maxwell::ViewportSwizzle::NegativeZ: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_Z_NV; + case Maxwell::ViewportSwizzle::PositiveW: + return VK_VIEWPORT_COORDINATE_SWIZZLE_POSITIVE_W_NV; + case Maxwell::ViewportSwizzle::NegativeW: + return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_W_NV; + } + UNREACHABLE_MSG("Invalid swizzle={}", static_cast<int>(swizzle)); + return {}; +} + } // namespace Vulkan::MaxwellToVK diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.h b/src/video_core/renderer_vulkan/maxwell_to_vk.h index 81bce4c6c..7e213452f 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.h +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.h @@ -59,4 +59,6 @@ VkCullModeFlags CullFace(Maxwell::CullFace cull_face); VkComponentSwizzle SwizzleSource(Tegra::Texture::SwizzleSource swizzle); +VkViewportCoordinateSwizzleNV ViewportSwizzle(Maxwell::ViewportSwizzle swizzle); + } // namespace Vulkan::MaxwellToVK diff --git a/src/video_core/renderer_vulkan/nsight_aftermath_tracker.cpp b/src/video_core/renderer_vulkan/nsight_aftermath_tracker.cpp new file mode 100644 index 000000000..5b01020ec --- /dev/null +++ b/src/video_core/renderer_vulkan/nsight_aftermath_tracker.cpp @@ -0,0 +1,220 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#ifdef HAS_NSIGHT_AFTERMATH + +#include <mutex> +#include <string> +#include <string_view> +#include <utility> +#include <vector> + +#include <fmt/format.h> + +#define VK_NO_PROTOTYPES +#include <vulkan/vulkan.h> + +#include <GFSDK_Aftermath.h> +#include <GFSDK_Aftermath_Defines.h> +#include <GFSDK_Aftermath_GpuCrashDump.h> +#include <GFSDK_Aftermath_GpuCrashDumpDecoding.h> + +#include "common/common_paths.h" +#include "common/common_types.h" +#include "common/file_util.h" +#include "common/logging/log.h" +#include "common/scope_exit.h" + +#include "video_core/renderer_vulkan/nsight_aftermath_tracker.h" + +namespace Vulkan { + +static constexpr char AFTERMATH_LIB_NAME[] = "GFSDK_Aftermath_Lib.x64.dll"; + +NsightAftermathTracker::NsightAftermathTracker() = default; + +NsightAftermathTracker::~NsightAftermathTracker() { + if (initialized) { + (void)GFSDK_Aftermath_DisableGpuCrashDumps(); + } +} + +bool NsightAftermathTracker::Initialize() { + if (!dl.Open(AFTERMATH_LIB_NAME)) { + LOG_ERROR(Render_Vulkan, "Failed to load Nsight Aftermath DLL"); + return false; + } + + if (!dl.GetSymbol("GFSDK_Aftermath_DisableGpuCrashDumps", + &GFSDK_Aftermath_DisableGpuCrashDumps) || + !dl.GetSymbol("GFSDK_Aftermath_EnableGpuCrashDumps", + &GFSDK_Aftermath_EnableGpuCrashDumps) || + !dl.GetSymbol("GFSDK_Aftermath_GetShaderDebugInfoIdentifier", + &GFSDK_Aftermath_GetShaderDebugInfoIdentifier) || + !dl.GetSymbol("GFSDK_Aftermath_GetShaderHashSpirv", &GFSDK_Aftermath_GetShaderHashSpirv) || + !dl.GetSymbol("GFSDK_Aftermath_GpuCrashDump_CreateDecoder", + &GFSDK_Aftermath_GpuCrashDump_CreateDecoder) || + !dl.GetSymbol("GFSDK_Aftermath_GpuCrashDump_DestroyDecoder", + &GFSDK_Aftermath_GpuCrashDump_DestroyDecoder) || + !dl.GetSymbol("GFSDK_Aftermath_GpuCrashDump_GenerateJSON", + &GFSDK_Aftermath_GpuCrashDump_GenerateJSON) || + !dl.GetSymbol("GFSDK_Aftermath_GpuCrashDump_GetJSON", + &GFSDK_Aftermath_GpuCrashDump_GetJSON)) { + LOG_ERROR(Render_Vulkan, "Failed to load Nsight Aftermath function pointers"); + return false; + } + + dump_dir = Common::FS::GetUserPath(Common::FS::UserPath::LogDir) + "gpucrash"; + + (void)Common::FS::DeleteDirRecursively(dump_dir); + if (!Common::FS::CreateDir(dump_dir)) { + LOG_ERROR(Render_Vulkan, "Failed to create Nsight Aftermath dump directory"); + return false; + } + + if (!GFSDK_Aftermath_SUCCEED(GFSDK_Aftermath_EnableGpuCrashDumps( + GFSDK_Aftermath_Version_API, GFSDK_Aftermath_GpuCrashDumpWatchedApiFlags_Vulkan, + GFSDK_Aftermath_GpuCrashDumpFeatureFlags_Default, GpuCrashDumpCallback, + ShaderDebugInfoCallback, CrashDumpDescriptionCallback, this))) { + LOG_ERROR(Render_Vulkan, "GFSDK_Aftermath_EnableGpuCrashDumps failed"); + return false; + } + + LOG_INFO(Render_Vulkan, "Nsight Aftermath dump directory is \"{}\"", dump_dir); + + initialized = true; + return true; +} + +void NsightAftermathTracker::SaveShader(const std::vector<u32>& spirv) const { + if (!initialized) { + return; + } + + std::vector<u32> spirv_copy = spirv; + GFSDK_Aftermath_SpirvCode shader; + shader.pData = spirv_copy.data(); + shader.size = static_cast<u32>(spirv_copy.size() * 4); + + std::scoped_lock lock{mutex}; + + GFSDK_Aftermath_ShaderHash hash; + if (!GFSDK_Aftermath_SUCCEED( + GFSDK_Aftermath_GetShaderHashSpirv(GFSDK_Aftermath_Version_API, &shader, &hash))) { + LOG_ERROR(Render_Vulkan, "Failed to hash SPIR-V module"); + return; + } + + Common::FS::IOFile file(fmt::format("{}/source_{:016x}.spv", dump_dir, hash.hash), "wb"); + if (!file.IsOpen()) { + LOG_ERROR(Render_Vulkan, "Failed to dump SPIR-V module with hash={:016x}", hash.hash); + return; + } + if (file.WriteArray(spirv.data(), spirv.size()) != spirv.size()) { + LOG_ERROR(Render_Vulkan, "Failed to write SPIR-V module with hash={:016x}", hash.hash); + return; + } +} + +void NsightAftermathTracker::OnGpuCrashDumpCallback(const void* gpu_crash_dump, + u32 gpu_crash_dump_size) { + std::scoped_lock lock{mutex}; + + LOG_CRITICAL(Render_Vulkan, "called"); + + GFSDK_Aftermath_GpuCrashDump_Decoder decoder; + if (!GFSDK_Aftermath_SUCCEED(GFSDK_Aftermath_GpuCrashDump_CreateDecoder( + GFSDK_Aftermath_Version_API, gpu_crash_dump, gpu_crash_dump_size, &decoder))) { + LOG_ERROR(Render_Vulkan, "Failed to create decoder"); + return; + } + SCOPE_EXIT({ GFSDK_Aftermath_GpuCrashDump_DestroyDecoder(decoder); }); + + u32 json_size = 0; + if (!GFSDK_Aftermath_SUCCEED(GFSDK_Aftermath_GpuCrashDump_GenerateJSON( + decoder, GFSDK_Aftermath_GpuCrashDumpDecoderFlags_ALL_INFO, + GFSDK_Aftermath_GpuCrashDumpFormatterFlags_NONE, nullptr, nullptr, nullptr, nullptr, + this, &json_size))) { + LOG_ERROR(Render_Vulkan, "Failed to generate JSON"); + return; + } + std::vector<char> json(json_size); + if (!GFSDK_Aftermath_SUCCEED( + GFSDK_Aftermath_GpuCrashDump_GetJSON(decoder, json_size, json.data()))) { + LOG_ERROR(Render_Vulkan, "Failed to query JSON"); + return; + } + + const std::string base_name = [this] { + const int id = dump_id++; + if (id == 0) { + return fmt::format("{}/crash.nv-gpudmp", dump_dir); + } else { + return fmt::format("{}/crash_{}.nv-gpudmp", dump_dir, id); + } + }(); + + std::string_view dump_view(static_cast<const char*>(gpu_crash_dump), gpu_crash_dump_size); + if (Common::FS::WriteStringToFile(false, base_name, dump_view) != gpu_crash_dump_size) { + LOG_ERROR(Render_Vulkan, "Failed to write dump file"); + return; + } + const std::string_view json_view(json.data(), json.size()); + if (Common::FS::WriteStringToFile(true, base_name + ".json", json_view) != json.size()) { + LOG_ERROR(Render_Vulkan, "Failed to write JSON"); + return; + } +} + +void NsightAftermathTracker::OnShaderDebugInfoCallback(const void* shader_debug_info, + u32 shader_debug_info_size) { + std::scoped_lock lock{mutex}; + + GFSDK_Aftermath_ShaderDebugInfoIdentifier identifier; + if (!GFSDK_Aftermath_SUCCEED(GFSDK_Aftermath_GetShaderDebugInfoIdentifier( + GFSDK_Aftermath_Version_API, shader_debug_info, shader_debug_info_size, &identifier))) { + LOG_ERROR(Render_Vulkan, "GFSDK_Aftermath_GetShaderDebugInfoIdentifier failed"); + return; + } + + const std::string path = + fmt::format("{}/shader_{:016x}{:016x}.nvdbg", dump_dir, identifier.id[0], identifier.id[1]); + Common::FS::IOFile file(path, "wb"); + if (!file.IsOpen()) { + LOG_ERROR(Render_Vulkan, "Failed to create file {}", path); + return; + } + if (file.WriteBytes(static_cast<const u8*>(shader_debug_info), shader_debug_info_size) != + shader_debug_info_size) { + LOG_ERROR(Render_Vulkan, "Failed to write file {}", path); + return; + } +} + +void NsightAftermathTracker::OnCrashDumpDescriptionCallback( + PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription add_description) { + add_description(GFSDK_Aftermath_GpuCrashDumpDescriptionKey_ApplicationName, "yuzu"); +} + +void NsightAftermathTracker::GpuCrashDumpCallback(const void* gpu_crash_dump, + u32 gpu_crash_dump_size, void* user_data) { + static_cast<NsightAftermathTracker*>(user_data)->OnGpuCrashDumpCallback(gpu_crash_dump, + gpu_crash_dump_size); +} + +void NsightAftermathTracker::ShaderDebugInfoCallback(const void* shader_debug_info, + u32 shader_debug_info_size, void* user_data) { + static_cast<NsightAftermathTracker*>(user_data)->OnShaderDebugInfoCallback( + shader_debug_info, shader_debug_info_size); +} + +void NsightAftermathTracker::CrashDumpDescriptionCallback( + PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription add_description, void* user_data) { + static_cast<NsightAftermathTracker*>(user_data)->OnCrashDumpDescriptionCallback( + add_description); +} + +} // namespace Vulkan + +#endif // HAS_NSIGHT_AFTERMATH diff --git a/src/video_core/renderer_vulkan/nsight_aftermath_tracker.h b/src/video_core/renderer_vulkan/nsight_aftermath_tracker.h new file mode 100644 index 000000000..afe7ae99e --- /dev/null +++ b/src/video_core/renderer_vulkan/nsight_aftermath_tracker.h @@ -0,0 +1,87 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <mutex> +#include <string> +#include <vector> + +#define VK_NO_PROTOTYPES +#include <vulkan/vulkan.h> + +#ifdef HAS_NSIGHT_AFTERMATH +#include <GFSDK_Aftermath_Defines.h> +#include <GFSDK_Aftermath_GpuCrashDump.h> +#include <GFSDK_Aftermath_GpuCrashDumpDecoding.h> +#endif + +#include "common/common_types.h" +#include "common/dynamic_library.h" + +namespace Vulkan { + +class NsightAftermathTracker { +public: + NsightAftermathTracker(); + ~NsightAftermathTracker(); + + NsightAftermathTracker(const NsightAftermathTracker&) = delete; + NsightAftermathTracker& operator=(const NsightAftermathTracker&) = delete; + + // Delete move semantics because Aftermath initialization uses a pointer to this. + NsightAftermathTracker(NsightAftermathTracker&&) = delete; + NsightAftermathTracker& operator=(NsightAftermathTracker&&) = delete; + + bool Initialize(); + + void SaveShader(const std::vector<u32>& spirv) const; + +private: +#ifdef HAS_NSIGHT_AFTERMATH + static void GpuCrashDumpCallback(const void* gpu_crash_dump, u32 gpu_crash_dump_size, + void* user_data); + + static void ShaderDebugInfoCallback(const void* shader_debug_info, u32 shader_debug_info_size, + void* user_data); + + static void CrashDumpDescriptionCallback( + PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription add_description, void* user_data); + + void OnGpuCrashDumpCallback(const void* gpu_crash_dump, u32 gpu_crash_dump_size); + + void OnShaderDebugInfoCallback(const void* shader_debug_info, u32 shader_debug_info_size); + + void OnCrashDumpDescriptionCallback( + PFN_GFSDK_Aftermath_AddGpuCrashDumpDescription add_description); + + mutable std::mutex mutex; + + std::string dump_dir; + int dump_id = 0; + + bool initialized = false; + + Common::DynamicLibrary dl; + PFN_GFSDK_Aftermath_DisableGpuCrashDumps GFSDK_Aftermath_DisableGpuCrashDumps; + PFN_GFSDK_Aftermath_EnableGpuCrashDumps GFSDK_Aftermath_EnableGpuCrashDumps; + PFN_GFSDK_Aftermath_GetShaderDebugInfoIdentifier GFSDK_Aftermath_GetShaderDebugInfoIdentifier; + PFN_GFSDK_Aftermath_GetShaderHashSpirv GFSDK_Aftermath_GetShaderHashSpirv; + PFN_GFSDK_Aftermath_GpuCrashDump_CreateDecoder GFSDK_Aftermath_GpuCrashDump_CreateDecoder; + PFN_GFSDK_Aftermath_GpuCrashDump_DestroyDecoder GFSDK_Aftermath_GpuCrashDump_DestroyDecoder; + PFN_GFSDK_Aftermath_GpuCrashDump_GenerateJSON GFSDK_Aftermath_GpuCrashDump_GenerateJSON; + PFN_GFSDK_Aftermath_GpuCrashDump_GetJSON GFSDK_Aftermath_GpuCrashDump_GetJSON; +#endif +}; + +#ifndef HAS_NSIGHT_AFTERMATH +inline NsightAftermathTracker::NsightAftermathTracker() = default; +inline NsightAftermathTracker::~NsightAftermathTracker() = default; +inline bool NsightAftermathTracker::Initialize() { + return false; +} +inline void NsightAftermathTracker::SaveShader(const std::vector<u32>&) const {} +#endif + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp index dd590c38b..f2610868e 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp +++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp @@ -12,24 +12,22 @@ #include <fmt/format.h> -#include "common/assert.h" #include "common/dynamic_library.h" +#include "common/file_util.h" #include "common/logging/log.h" #include "common/telemetry.h" #include "core/core.h" #include "core/core_timing.h" #include "core/frontend/emu_window.h" -#include "core/memory.h" -#include "core/perf_stats.h" #include "core/settings.h" #include "core/telemetry_session.h" #include "video_core/gpu.h" #include "video_core/renderer_vulkan/renderer_vulkan.h" #include "video_core/renderer_vulkan/vk_blit_screen.h" #include "video_core/renderer_vulkan/vk_device.h" +#include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_state_tracker.h" #include "video_core/renderer_vulkan/vk_swapchain.h" @@ -42,7 +40,7 @@ #include <vulkan/vulkan_win32.h> #endif -#ifdef __linux__ +#if !defined(_WIN32) && !defined(__APPLE__) #include <X11/Xlib.h> #include <vulkan/vulkan_wayland.h> #include <vulkan/vulkan_xlib.h> @@ -58,7 +56,7 @@ VkBool32 DebugCallback(VkDebugUtilsMessageSeverityFlagBitsEXT severity, VkDebugUtilsMessageTypeFlagsEXT type, const VkDebugUtilsMessengerCallbackDataEXT* data, [[maybe_unused]] void* user_data) { - const char* message{data->pMessage}; + const char* const message{data->pMessage}; if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT) { LOG_CRITICAL(Render_Vulkan, "{}", message); @@ -79,7 +77,8 @@ Common::DynamicLibrary OpenVulkanLibrary() { char* libvulkan_env = getenv("LIBVULKAN_PATH"); if (!libvulkan_env || !library.Open(libvulkan_env)) { // Use the libvulkan.dylib from the application bundle. - std::string filename = File::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib"; + const std::string filename = + Common::FS::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib"; library.Open(filename.c_str()); } #else @@ -87,15 +86,15 @@ Common::DynamicLibrary OpenVulkanLibrary() { if (!library.Open(filename.c_str())) { // Android devices may not have libvulkan.so.1, only libvulkan.so. filename = Common::DynamicLibrary::GetVersionedFilename("vulkan"); - library.Open(filename.c_str()); + (void)library.Open(filename.c_str()); } #endif return library; } -vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatch& dld, - WindowSystemType window_type = WindowSystemType::Headless, - bool enable_layers = false) { +std::pair<vk::Instance, u32> CreateInstance( + Common::DynamicLibrary& library, vk::InstanceDispatch& dld, + WindowSystemType window_type = WindowSystemType::Headless, bool enable_layers = false) { if (!library.IsOpen()) { LOG_ERROR(Render_Vulkan, "Vulkan library not available"); return {}; @@ -119,7 +118,7 @@ vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatc extensions.push_back(VK_KHR_WIN32_SURFACE_EXTENSION_NAME); break; #endif -#ifdef __linux__ +#if !defined(_WIN32) && !defined(__APPLE__) case Core::Frontend::WindowSystemType::X11: extensions.push_back(VK_KHR_XLIB_SURFACE_EXTENSION_NAME); break; @@ -156,12 +155,35 @@ vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatc } } - static constexpr std::array layers_data{"VK_LAYER_LUNARG_standard_validation"}; - vk::Span<const char*> layers = layers_data; - if (!enable_layers) { - layers = {}; + std::vector<const char*> layers; + layers.reserve(1); + if (enable_layers) { + layers.push_back("VK_LAYER_KHRONOS_validation"); + } + + const std::optional layer_properties = vk::EnumerateInstanceLayerProperties(dld); + if (!layer_properties) { + LOG_ERROR(Render_Vulkan, "Failed to query layer properties, disabling layers"); + layers.clear(); + } + + for (auto layer_it = layers.begin(); layer_it != layers.end();) { + const char* const layer = *layer_it; + const auto it = std::find_if( + layer_properties->begin(), layer_properties->end(), + [layer](const VkLayerProperties& prop) { return !std::strcmp(layer, prop.layerName); }); + if (it == layer_properties->end()) { + LOG_ERROR(Render_Vulkan, "Layer {} not available, removing it", layer); + layer_it = layers.erase(layer_it); + } else { + ++layer_it; + } } - vk::Instance instance = vk::Instance::Create(layers, extensions, dld); + + // Limit the maximum version of Vulkan to avoid using untested version. + const u32 version = std::min(vk::AvailableVersion(dld), static_cast<u32>(VK_API_VERSION_1_1)); + + vk::Instance instance = vk::Instance::Create(version, layers, extensions, dld); if (!instance) { LOG_ERROR(Render_Vulkan, "Failed to create Vulkan instance"); return {}; @@ -169,7 +191,7 @@ vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatc if (!vk::Load(*instance, dld)) { LOG_ERROR(Render_Vulkan, "Failed to load Vulkan instance function pointers"); } - return instance; + return std::make_pair(std::move(instance), version); } std::string GetReadableVersion(u32 version) { @@ -218,8 +240,12 @@ std::string BuildCommaSeparatedExtensions(std::vector<std::string> available_ext } // Anonymous namespace -RendererVulkan::RendererVulkan(Core::Frontend::EmuWindow& window, Core::System& system) - : RendererBase(window), system{system} {} +RendererVulkan::RendererVulkan(Core::TelemetrySession& telemetry_session_, + Core::Frontend::EmuWindow& emu_window, + Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, + std::unique_ptr<Core::Frontend::GraphicsContext> context) + : RendererBase{emu_window, std::move(context)}, telemetry_session{telemetry_session_}, + cpu_memory{cpu_memory_}, gpu{gpu_} {} RendererVulkan::~RendererVulkan() { ShutDown(); @@ -246,11 +272,11 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { scheduler->WaitWorker(); swapchain->AcquireNextImage(); - const auto [fence, render_semaphore] = blit_screen->Draw(*framebuffer, use_accelerated); + const VkSemaphore render_semaphore = blit_screen->Draw(*framebuffer, use_accelerated); - scheduler->Flush(false, render_semaphore); + scheduler->Flush(render_semaphore); - if (swapchain->Present(render_semaphore, fence)) { + if (swapchain->Present(render_semaphore)) { blit_screen->Recreate(); } @@ -260,15 +286,10 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { render_window.PollEvents(); } -bool RendererVulkan::TryPresent(int /*timeout_ms*/) { - // TODO (bunnei): ImplementMe - return true; -} - bool RendererVulkan::Init() { library = OpenVulkanLibrary(); - instance = CreateInstance(library, dld, render_window.GetWindowInfo().type, - Settings::values.renderer_debug); + std::tie(instance, instance_version) = CreateInstance( + library, dld, render_window.GetWindowInfo().type, Settings::values.renderer_debug); if (!instance || !CreateDebugCallback() || !CreateSurface() || !PickDevices()) { return false; } @@ -277,23 +298,21 @@ bool RendererVulkan::Init() { memory_manager = std::make_unique<VKMemoryManager>(*device); - resource_manager = std::make_unique<VKResourceManager>(*device); + state_tracker = std::make_unique<StateTracker>(gpu); + + scheduler = std::make_unique<VKScheduler>(*device, *state_tracker); const auto& framebuffer = render_window.GetFramebufferLayout(); - swapchain = std::make_unique<VKSwapchain>(*surface, *device); + swapchain = std::make_unique<VKSwapchain>(*surface, *device, *scheduler); swapchain->Create(framebuffer.width, framebuffer.height, false); - state_tracker = std::make_unique<StateTracker>(system); - - scheduler = std::make_unique<VKScheduler>(*device, *resource_manager, *state_tracker); - - rasterizer = std::make_unique<RasterizerVulkan>(system, render_window, screen_info, *device, - *resource_manager, *memory_manager, - *state_tracker, *scheduler); + rasterizer = std::make_unique<RasterizerVulkan>(render_window, gpu, gpu.MemoryManager(), + cpu_memory, screen_info, *device, + *memory_manager, *state_tracker, *scheduler); - blit_screen = std::make_unique<VKBlitScreen>(system, render_window, *rasterizer, *device, - *resource_manager, *memory_manager, *swapchain, - *scheduler, screen_info); + blit_screen = + std::make_unique<VKBlitScreen>(cpu_memory, render_window, *rasterizer, *device, + *memory_manager, *swapchain, *scheduler, screen_info); return true; } @@ -311,7 +330,6 @@ void RendererVulkan::ShutDown() { scheduler.reset(); swapchain.reset(); memory_manager.reset(); - resource_manager.reset(); device.reset(); } @@ -345,7 +363,7 @@ bool RendererVulkan::CreateSurface() { } } #endif -#ifdef __linux__ +#if !defined(_WIN32) && !defined(__APPLE__) if (window_info.type == Core::Frontend::WindowSystemType::X11) { const VkXlibSurfaceCreateInfoKHR xlib_ci{ VK_STRUCTURE_TYPE_XLIB_SURFACE_CREATE_INFO_KHR, nullptr, 0, @@ -390,7 +408,7 @@ bool RendererVulkan::PickDevices() { return false; } - const s32 device_index = Settings::values.vulkan_device; + const s32 device_index = Settings::values.vulkan_device.GetValue(); if (device_index < 0 || device_index >= static_cast<s32>(devices->size())) { LOG_ERROR(Render_Vulkan, "Invalid device index {}!", device_index); return false; @@ -401,7 +419,8 @@ bool RendererVulkan::PickDevices() { return false; } - device = std::make_unique<VKDevice>(*instance, physical_device, *surface, dld); + device = + std::make_unique<VKDevice>(*instance, instance_version, physical_device, *surface, dld); return device->Create(); } @@ -411,7 +430,7 @@ void RendererVulkan::Report() const { const std::string driver_version = GetDriverVersion(*device); const std::string driver_name = fmt::format("{} {}", vendor_name, driver_version); - const std::string api_version = GetReadableVersion(device->GetApiVersion()); + const std::string api_version = GetReadableVersion(device->ApiVersion()); const std::string extensions = BuildCommaSeparatedExtensions(device->GetAvailableExtensions()); @@ -419,8 +438,7 @@ void RendererVulkan::Report() const { LOG_INFO(Render_Vulkan, "Device: {}", model_name); LOG_INFO(Render_Vulkan, "Vulkan: {}", api_version); - auto& telemetry_session = system.TelemetrySession(); - constexpr auto field = Telemetry::FieldType::UserSystem; + static constexpr auto field = Common::Telemetry::FieldType::UserSystem; telemetry_session.AddField(field, "GPU_Vendor", vendor_name); telemetry_session.AddField(field, "GPU_Model", model_name); telemetry_session.AddField(field, "GPU_Vulkan_Driver", driver_name); @@ -431,7 +449,7 @@ void RendererVulkan::Report() const { std::vector<std::string> RendererVulkan::EnumerateDevices() { vk::InstanceDispatch dld; Common::DynamicLibrary library = OpenVulkanLibrary(); - vk::Instance instance = CreateInstance(library, dld); + vk::Instance instance = CreateInstance(library, dld).first; if (!instance) { return {}; } diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.h b/src/video_core/renderer_vulkan/renderer_vulkan.h index 18270909b..1044ca124 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.h +++ b/src/video_core/renderer_vulkan/renderer_vulkan.h @@ -5,7 +5,6 @@ #pragma once #include <memory> -#include <optional> #include <string> #include <vector> @@ -15,7 +14,15 @@ #include "video_core/renderer_vulkan/wrapper.h" namespace Core { -class System; +class TelemetrySession; +} + +namespace Core::Memory { +class Memory; +} + +namespace Tegra { +class GPU; } namespace Vulkan { @@ -23,9 +30,7 @@ namespace Vulkan { class StateTracker; class VKBlitScreen; class VKDevice; -class VKFence; class VKMemoryManager; -class VKResourceManager; class VKSwapchain; class VKScheduler; class VKImage; @@ -39,13 +44,15 @@ struct VKScreenInfo { class RendererVulkan final : public VideoCore::RendererBase { public: - explicit RendererVulkan(Core::Frontend::EmuWindow& window, Core::System& system); + explicit RendererVulkan(Core::TelemetrySession& telemtry_session, + Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory, + Tegra::GPU& gpu, + std::unique_ptr<Core::Frontend::GraphicsContext> context); ~RendererVulkan() override; bool Init() override; void ShutDown() override; void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override; - bool TryPresent(int timeout_ms) override; static std::vector<std::string> EnumerateDevices(); @@ -58,23 +65,26 @@ private: void Report() const; - Core::System& system; + Core::TelemetrySession& telemetry_session; + Core::Memory::Memory& cpu_memory; + Tegra::GPU& gpu; Common::DynamicLibrary library; vk::InstanceDispatch dld; vk::Instance instance; + u32 instance_version{}; + vk::SurfaceKHR surface; VKScreenInfo screen_info; vk::DebugCallback debug_callback; std::unique_ptr<VKDevice> device; - std::unique_ptr<VKSwapchain> swapchain; std::unique_ptr<VKMemoryManager> memory_manager; - std::unique_ptr<VKResourceManager> resource_manager; std::unique_ptr<StateTracker> state_tracker; std::unique_ptr<VKScheduler> scheduler; + std::unique_ptr<VKSwapchain> swapchain; std::unique_ptr<VKBlitScreen> blit_screen; }; diff --git a/src/video_core/renderer_vulkan/shaders/quad_indexed.comp b/src/video_core/renderer_vulkan/shaders/quad_indexed.comp new file mode 100644 index 000000000..5a472ba9b --- /dev/null +++ b/src/video_core/renderer_vulkan/shaders/quad_indexed.comp @@ -0,0 +1,50 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +/* + * Build instructions: + * $ glslangValidator -V quad_indexed.comp -o output.spv + * $ spirv-opt -O --strip-debug output.spv -o optimized.spv + * $ xxd -i optimized.spv + * + * Then copy that bytecode to the C++ file + */ + +#version 460 core + +layout (local_size_x = 1024) in; + +layout (std430, set = 0, binding = 0) readonly buffer InputBuffer { + uint input_indexes[]; +}; + +layout (std430, set = 0, binding = 1) writeonly buffer OutputBuffer { + uint output_indexes[]; +}; + +layout (push_constant) uniform PushConstants { + uint base_vertex; + int index_shift; // 0: uint8, 1: uint16, 2: uint32 +}; + +void main() { + int primitive = int(gl_GlobalInvocationID.x); + if (primitive * 6 >= output_indexes.length()) { + return; + } + + int index_size = 8 << index_shift; + int flipped_shift = 2 - index_shift; + int mask = (1 << flipped_shift) - 1; + + const int quad_swizzle[6] = int[](0, 1, 2, 0, 2, 3); + for (uint vertex = 0; vertex < 6; ++vertex) { + int offset = primitive * 4 + quad_swizzle[vertex]; + int int_offset = offset >> flipped_shift; + int bit_offset = (offset & mask) * index_size; + uint packed_input = input_indexes[int_offset]; + uint index = bitfieldExtract(packed_input, bit_offset, index_size); + output_indexes[primitive * 6 + vertex] = index + base_vertex; + } +} diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.cpp b/src/video_core/renderer_vulkan/vk_blit_screen.cpp index fbd406f2b..b5b60309e 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.cpp +++ b/src/video_core/renderer_vulkan/vk_blit_screen.cpp @@ -12,11 +12,9 @@ #include "common/assert.h" #include "common/common_types.h" #include "common/math_util.h" - #include "core/core.h" #include "core/frontend/emu_window.h" #include "core/memory.h" - #include "video_core/gpu.h" #include "video_core/morton.h" #include "video_core/rasterizer_interface.h" @@ -24,8 +22,8 @@ #include "video_core/renderer_vulkan/vk_blit_screen.h" #include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_image.h" +#include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_shader_util.h" #include "video_core/renderer_vulkan/vk_swapchain.h" @@ -141,24 +139,28 @@ struct ScreenRectVertex { std::array<f32, 2> tex_coord; static VkVertexInputBindingDescription GetDescription() { - VkVertexInputBindingDescription description; - description.binding = 0; - description.stride = sizeof(ScreenRectVertex); - description.inputRate = VK_VERTEX_INPUT_RATE_VERTEX; - return description; + return { + .binding = 0, + .stride = sizeof(ScreenRectVertex), + .inputRate = VK_VERTEX_INPUT_RATE_VERTEX, + }; } static std::array<VkVertexInputAttributeDescription, 2> GetAttributes() { - std::array<VkVertexInputAttributeDescription, 2> attributes; - attributes[0].location = 0; - attributes[0].binding = 0; - attributes[0].format = VK_FORMAT_R32G32_SFLOAT; - attributes[0].offset = offsetof(ScreenRectVertex, position); - attributes[1].location = 1; - attributes[1].binding = 0; - attributes[1].format = VK_FORMAT_R32G32_SFLOAT; - attributes[1].offset = offsetof(ScreenRectVertex, tex_coord); - return attributes; + return {{ + { + .location = 0, + .binding = 0, + .format = VK_FORMAT_R32G32_SFLOAT, + .offset = offsetof(ScreenRectVertex, position), + }, + { + .location = 1, + .binding = 0, + .format = VK_FORMAT_R32G32_SFLOAT, + .offset = offsetof(ScreenRectVertex, tex_coord), + }, + }}; } }; @@ -183,9 +185,9 @@ std::size_t GetSizeInBytes(const Tegra::FramebufferConfig& framebuffer) { VkFormat GetFormat(const Tegra::FramebufferConfig& framebuffer) { switch (framebuffer.pixel_format) { - case Tegra::FramebufferConfig::PixelFormat::ABGR8: + case Tegra::FramebufferConfig::PixelFormat::A8B8G8R8_UNORM: return VK_FORMAT_A8B8G8R8_UNORM_PACK32; - case Tegra::FramebufferConfig::PixelFormat::RGB565: + case Tegra::FramebufferConfig::PixelFormat::RGB565_UNORM: return VK_FORMAT_R5G6B5_UNORM_PACK16; default: UNIMPLEMENTED_MSG("Unknown framebuffer pixel format: {}", @@ -206,17 +208,15 @@ struct VKBlitScreen::BufferData { // Unaligned image data goes here }; -VKBlitScreen::VKBlitScreen(Core::System& system, Core::Frontend::EmuWindow& render_window, - VideoCore::RasterizerInterface& rasterizer, const VKDevice& device, - VKResourceManager& resource_manager, VKMemoryManager& memory_manager, - VKSwapchain& swapchain, VKScheduler& scheduler, - const VKScreenInfo& screen_info) - : system{system}, render_window{render_window}, rasterizer{rasterizer}, device{device}, - resource_manager{resource_manager}, memory_manager{memory_manager}, swapchain{swapchain}, - scheduler{scheduler}, image_count{swapchain.GetImageCount()}, screen_info{screen_info} { - watches.resize(image_count); - std::generate(watches.begin(), watches.end(), - []() { return std::make_unique<VKFenceWatch>(); }); +VKBlitScreen::VKBlitScreen(Core::Memory::Memory& cpu_memory_, + Core::Frontend::EmuWindow& render_window_, + VideoCore::RasterizerInterface& rasterizer_, const VKDevice& device_, + VKMemoryManager& memory_manager_, VKSwapchain& swapchain_, + VKScheduler& scheduler_, const VKScreenInfo& screen_info_) + : cpu_memory{cpu_memory_}, render_window{render_window_}, rasterizer{rasterizer_}, + device{device_}, memory_manager{memory_manager_}, swapchain{swapchain_}, + scheduler{scheduler_}, image_count{swapchain.GetImageCount()}, screen_info{screen_info_} { + resource_ticks.resize(image_count); CreateStaticResources(); CreateDynamicResources(); @@ -228,15 +228,16 @@ void VKBlitScreen::Recreate() { CreateDynamicResources(); } -std::tuple<VKFence&, VkSemaphore> VKBlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer, - bool use_accelerated) { +VkSemaphore VKBlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer, bool use_accelerated) { RefreshResources(framebuffer); // Finish any pending renderpass scheduler.RequestOutsideRenderPassOperationContext(); const std::size_t image_index = swapchain.GetImageIndex(); - watches[image_index]->Watch(scheduler.GetFence()); + + scheduler.Wait(resource_ticks[image_index]); + resource_ticks[image_index] = scheduler.CurrentTick(); VKImage* blit_image = use_accelerated ? screen_info.image : raw_images[image_index].get(); @@ -255,7 +256,7 @@ std::tuple<VKFence&, VkSemaphore> VKBlitScreen::Draw(const Tegra::FramebufferCon const auto pixel_format = VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format); const VAddr framebuffer_addr = framebuffer.address + framebuffer.offset; - const auto host_ptr = system.Memory().GetPointer(framebuffer_addr); + const auto host_ptr = cpu_memory.GetPointer(framebuffer_addr); rasterizer.FlushRegion(ToCacheAddr(host_ptr), GetSizeInBytes(framebuffer)); // TODO(Rodrigo): Read this from HLE @@ -267,20 +268,25 @@ std::tuple<VKFence&, VkSemaphore> VKBlitScreen::Draw(const Tegra::FramebufferCon blit_image->Transition(0, 1, 0, 1, VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_WRITE_BIT, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL); - VkBufferImageCopy copy; - copy.bufferOffset = image_offset; - copy.bufferRowLength = 0; - copy.bufferImageHeight = 0; - copy.imageSubresource.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT; - copy.imageSubresource.mipLevel = 0; - copy.imageSubresource.baseArrayLayer = 0; - copy.imageSubresource.layerCount = 1; - copy.imageOffset.x = 0; - copy.imageOffset.y = 0; - copy.imageOffset.z = 0; - copy.imageExtent.width = framebuffer.width; - copy.imageExtent.height = framebuffer.height; - copy.imageExtent.depth = 1; + const VkBufferImageCopy copy{ + .bufferOffset = image_offset, + .bufferRowLength = 0, + .bufferImageHeight = 0, + .imageSubresource = + { + .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT, + .mipLevel = 0, + .baseArrayLayer = 0, + .layerCount = 1, + }, + .imageOffset = {.x = 0, .y = 0, .z = 0}, + .imageExtent = + { + .width = framebuffer.width, + .height = framebuffer.height, + .depth = 1, + }, + }; scheduler.Record( [buffer = *buffer, image = *blit_image->GetHandle(), copy](vk::CommandBuffer cmdbuf) { cmdbuf.CopyBufferToImage(buffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, copy); @@ -295,11 +301,9 @@ std::tuple<VKFence&, VkSemaphore> VKBlitScreen::Draw(const Tegra::FramebufferCon descriptor_set = descriptor_sets[image_index], buffer = *buffer, size = swapchain.GetSize(), pipeline = *pipeline, layout = *pipeline_layout](vk::CommandBuffer cmdbuf) { - VkClearValue clear_color; - clear_color.color.float32[0] = 0.0f; - clear_color.color.float32[1] = 0.0f; - clear_color.color.float32[2] = 0.0f; - clear_color.color.float32[3] = 0.0f; + const VkClearValue clear_color{ + .color = {.float32 = {0.0f, 0.0f, 0.0f, 0.0f}}, + }; VkRenderPassBeginInfo renderpass_bi; renderpass_bi.sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO; @@ -336,7 +340,7 @@ std::tuple<VKFence&, VkSemaphore> VKBlitScreen::Draw(const Tegra::FramebufferCon cmdbuf.EndRenderPass(); }); - return {scheduler.GetFence(), *semaphores[image_index]}; + return *semaphores[image_index]; } void VKBlitScreen::CreateStaticResources() { @@ -379,93 +383,109 @@ void VKBlitScreen::CreateSemaphores() { } void VKBlitScreen::CreateDescriptorPool() { - std::array<VkDescriptorPoolSize, 2> pool_sizes; - pool_sizes[0].type = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER; - pool_sizes[0].descriptorCount = static_cast<u32>(image_count); - pool_sizes[1].type = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER; - pool_sizes[1].descriptorCount = static_cast<u32>(image_count); - - VkDescriptorPoolCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT; - ci.maxSets = static_cast<u32>(image_count); - ci.poolSizeCount = static_cast<u32>(pool_sizes.size()); - ci.pPoolSizes = pool_sizes.data(); + const std::array<VkDescriptorPoolSize, 2> pool_sizes{{ + { + .type = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER, + .descriptorCount = static_cast<u32>(image_count), + }, + { + .type = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .descriptorCount = static_cast<u32>(image_count), + }, + }}; + + const VkDescriptorPoolCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO, + .pNext = nullptr, + .flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT, + .maxSets = static_cast<u32>(image_count), + .poolSizeCount = static_cast<u32>(pool_sizes.size()), + .pPoolSizes = pool_sizes.data(), + }; descriptor_pool = device.GetLogical().CreateDescriptorPool(ci); } void VKBlitScreen::CreateRenderPass() { - VkAttachmentDescription color_attachment; - color_attachment.flags = 0; - color_attachment.format = swapchain.GetImageFormat(); - color_attachment.samples = VK_SAMPLE_COUNT_1_BIT; - color_attachment.loadOp = VK_ATTACHMENT_LOAD_OP_CLEAR; - color_attachment.storeOp = VK_ATTACHMENT_STORE_OP_STORE; - color_attachment.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE; - color_attachment.stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE; - color_attachment.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED; - color_attachment.finalLayout = VK_IMAGE_LAYOUT_PRESENT_SRC_KHR; - - VkAttachmentReference color_attachment_ref; - color_attachment_ref.attachment = 0; - color_attachment_ref.layout = VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL; - - VkSubpassDescription subpass_description; - subpass_description.flags = 0; - subpass_description.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS; - subpass_description.inputAttachmentCount = 0; - subpass_description.pInputAttachments = nullptr; - subpass_description.colorAttachmentCount = 1; - subpass_description.pColorAttachments = &color_attachment_ref; - subpass_description.pResolveAttachments = nullptr; - subpass_description.pDepthStencilAttachment = nullptr; - subpass_description.preserveAttachmentCount = 0; - subpass_description.pPreserveAttachments = nullptr; - - VkSubpassDependency dependency; - dependency.srcSubpass = VK_SUBPASS_EXTERNAL; - dependency.dstSubpass = 0; - dependency.srcStageMask = VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT; - dependency.dstStageMask = VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT; - dependency.srcAccessMask = 0; - dependency.dstAccessMask = - VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT; - dependency.dependencyFlags = 0; - - VkRenderPassCreateInfo renderpass_ci; - renderpass_ci.sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO; - renderpass_ci.pNext = nullptr; - renderpass_ci.flags = 0; - renderpass_ci.attachmentCount = 1; - renderpass_ci.pAttachments = &color_attachment; - renderpass_ci.subpassCount = 1; - renderpass_ci.pSubpasses = &subpass_description; - renderpass_ci.dependencyCount = 1; - renderpass_ci.pDependencies = &dependency; + const VkAttachmentDescription color_attachment{ + .flags = 0, + .format = swapchain.GetImageFormat(), + .samples = VK_SAMPLE_COUNT_1_BIT, + .loadOp = VK_ATTACHMENT_LOAD_OP_CLEAR, + .storeOp = VK_ATTACHMENT_STORE_OP_STORE, + .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE, + .stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE, + .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, + .finalLayout = VK_IMAGE_LAYOUT_PRESENT_SRC_KHR, + }; + + const VkAttachmentReference color_attachment_ref{ + .attachment = 0, + .layout = VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL, + }; + + const VkSubpassDescription subpass_description{ + .flags = 0, + .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, + .inputAttachmentCount = 0, + .pInputAttachments = nullptr, + .colorAttachmentCount = 1, + .pColorAttachments = &color_attachment_ref, + .pResolveAttachments = nullptr, + .pDepthStencilAttachment = nullptr, + .preserveAttachmentCount = 0, + .pPreserveAttachments = nullptr, + }; + + const VkSubpassDependency dependency{ + .srcSubpass = VK_SUBPASS_EXTERNAL, + .dstSubpass = 0, + .srcStageMask = VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT, + .dstStageMask = VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT, + .srcAccessMask = 0, + .dstAccessMask = VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT, + .dependencyFlags = 0, + }; + + const VkRenderPassCreateInfo renderpass_ci{ + .sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .attachmentCount = 1, + .pAttachments = &color_attachment, + .subpassCount = 1, + .pSubpasses = &subpass_description, + .dependencyCount = 1, + .pDependencies = &dependency, + }; renderpass = device.GetLogical().CreateRenderPass(renderpass_ci); } void VKBlitScreen::CreateDescriptorSetLayout() { - std::array<VkDescriptorSetLayoutBinding, 2> layout_bindings; - layout_bindings[0].binding = 0; - layout_bindings[0].descriptorType = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER; - layout_bindings[0].descriptorCount = 1; - layout_bindings[0].stageFlags = VK_SHADER_STAGE_VERTEX_BIT; - layout_bindings[0].pImmutableSamplers = nullptr; - layout_bindings[1].binding = 1; - layout_bindings[1].descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER; - layout_bindings[1].descriptorCount = 1; - layout_bindings[1].stageFlags = VK_SHADER_STAGE_FRAGMENT_BIT; - layout_bindings[1].pImmutableSamplers = nullptr; - - VkDescriptorSetLayoutCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.bindingCount = static_cast<u32>(layout_bindings.size()); - ci.pBindings = layout_bindings.data(); + const std::array<VkDescriptorSetLayoutBinding, 2> layout_bindings{{ + { + .binding = 0, + .descriptorType = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_VERTEX_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_FRAGMENT_BIT, + .pImmutableSamplers = nullptr, + }, + }}; + + const VkDescriptorSetLayoutCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .bindingCount = static_cast<u32>(layout_bindings.size()), + .pBindings = layout_bindings.data(), + }; descriptor_set_layout = device.GetLogical().CreateDescriptorSetLayout(ci); } @@ -473,175 +493,192 @@ void VKBlitScreen::CreateDescriptorSetLayout() { void VKBlitScreen::CreateDescriptorSets() { const std::vector layouts(image_count, *descriptor_set_layout); - VkDescriptorSetAllocateInfo ai; - ai.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO; - ai.pNext = nullptr; - ai.descriptorPool = *descriptor_pool; - ai.descriptorSetCount = static_cast<u32>(image_count); - ai.pSetLayouts = layouts.data(); + const VkDescriptorSetAllocateInfo ai{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO, + .pNext = nullptr, + .descriptorPool = *descriptor_pool, + .descriptorSetCount = static_cast<u32>(image_count), + .pSetLayouts = layouts.data(), + }; + descriptor_sets = descriptor_pool.Allocate(ai); } void VKBlitScreen::CreatePipelineLayout() { - VkPipelineLayoutCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.setLayoutCount = 1; - ci.pSetLayouts = descriptor_set_layout.address(); - ci.pushConstantRangeCount = 0; - ci.pPushConstantRanges = nullptr; + const VkPipelineLayoutCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .setLayoutCount = 1, + .pSetLayouts = descriptor_set_layout.address(), + .pushConstantRangeCount = 0, + .pPushConstantRanges = nullptr, + }; pipeline_layout = device.GetLogical().CreatePipelineLayout(ci); } void VKBlitScreen::CreateGraphicsPipeline() { - std::array<VkPipelineShaderStageCreateInfo, 2> shader_stages; - shader_stages[0].sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO; - shader_stages[0].pNext = nullptr; - shader_stages[0].flags = 0; - shader_stages[0].stage = VK_SHADER_STAGE_VERTEX_BIT; - shader_stages[0].module = *vertex_shader; - shader_stages[0].pName = "main"; - shader_stages[0].pSpecializationInfo = nullptr; - shader_stages[1].sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO; - shader_stages[1].pNext = nullptr; - shader_stages[1].flags = 0; - shader_stages[1].stage = VK_SHADER_STAGE_FRAGMENT_BIT; - shader_stages[1].module = *fragment_shader; - shader_stages[1].pName = "main"; - shader_stages[1].pSpecializationInfo = nullptr; + const std::array<VkPipelineShaderStageCreateInfo, 2> shader_stages{{ + { + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = VK_SHADER_STAGE_VERTEX_BIT, + .module = *vertex_shader, + .pName = "main", + .pSpecializationInfo = nullptr, + }, + { + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = VK_SHADER_STAGE_FRAGMENT_BIT, + .module = *fragment_shader, + .pName = "main", + .pSpecializationInfo = nullptr, + }, + }}; const auto vertex_binding_description = ScreenRectVertex::GetDescription(); const auto vertex_attrs_description = ScreenRectVertex::GetAttributes(); - VkPipelineVertexInputStateCreateInfo vertex_input_ci; - vertex_input_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO; - vertex_input_ci.pNext = nullptr; - vertex_input_ci.flags = 0; - vertex_input_ci.vertexBindingDescriptionCount = 1; - vertex_input_ci.pVertexBindingDescriptions = &vertex_binding_description; - vertex_input_ci.vertexAttributeDescriptionCount = u32{vertex_attrs_description.size()}; - vertex_input_ci.pVertexAttributeDescriptions = vertex_attrs_description.data(); - - VkPipelineInputAssemblyStateCreateInfo input_assembly_ci; - input_assembly_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO; - input_assembly_ci.pNext = nullptr; - input_assembly_ci.flags = 0; - input_assembly_ci.topology = VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP; - input_assembly_ci.primitiveRestartEnable = VK_FALSE; - - VkPipelineViewportStateCreateInfo viewport_state_ci; - viewport_state_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO; - viewport_state_ci.pNext = nullptr; - viewport_state_ci.flags = 0; - viewport_state_ci.viewportCount = 1; - viewport_state_ci.pViewports = nullptr; - viewport_state_ci.scissorCount = 1; - viewport_state_ci.pScissors = nullptr; - - VkPipelineRasterizationStateCreateInfo rasterization_ci; - rasterization_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO; - rasterization_ci.pNext = nullptr; - rasterization_ci.flags = 0; - rasterization_ci.depthClampEnable = VK_FALSE; - rasterization_ci.rasterizerDiscardEnable = VK_FALSE; - rasterization_ci.polygonMode = VK_POLYGON_MODE_FILL; - rasterization_ci.cullMode = VK_CULL_MODE_NONE; - rasterization_ci.frontFace = VK_FRONT_FACE_CLOCKWISE; - rasterization_ci.depthBiasEnable = VK_FALSE; - rasterization_ci.depthBiasConstantFactor = 0.0f; - rasterization_ci.depthBiasClamp = 0.0f; - rasterization_ci.depthBiasSlopeFactor = 0.0f; - rasterization_ci.lineWidth = 1.0f; - - VkPipelineMultisampleStateCreateInfo multisampling_ci; - multisampling_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO; - multisampling_ci.pNext = nullptr; - multisampling_ci.flags = 0; - multisampling_ci.rasterizationSamples = VK_SAMPLE_COUNT_1_BIT; - multisampling_ci.sampleShadingEnable = VK_FALSE; - multisampling_ci.minSampleShading = 0.0f; - multisampling_ci.pSampleMask = nullptr; - multisampling_ci.alphaToCoverageEnable = VK_FALSE; - multisampling_ci.alphaToOneEnable = VK_FALSE; - - VkPipelineColorBlendAttachmentState color_blend_attachment; - color_blend_attachment.blendEnable = VK_FALSE; - color_blend_attachment.srcColorBlendFactor = VK_BLEND_FACTOR_ZERO; - color_blend_attachment.dstColorBlendFactor = VK_BLEND_FACTOR_ZERO; - color_blend_attachment.colorBlendOp = VK_BLEND_OP_ADD; - color_blend_attachment.srcAlphaBlendFactor = VK_BLEND_FACTOR_ZERO; - color_blend_attachment.dstAlphaBlendFactor = VK_BLEND_FACTOR_ZERO; - color_blend_attachment.alphaBlendOp = VK_BLEND_OP_ADD; - color_blend_attachment.colorWriteMask = VK_COLOR_COMPONENT_R_BIT | VK_COLOR_COMPONENT_G_BIT | - VK_COLOR_COMPONENT_B_BIT | VK_COLOR_COMPONENT_A_BIT; - - VkPipelineColorBlendStateCreateInfo color_blend_ci; - color_blend_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO; - color_blend_ci.flags = 0; - color_blend_ci.pNext = nullptr; - color_blend_ci.logicOpEnable = VK_FALSE; - color_blend_ci.logicOp = VK_LOGIC_OP_COPY; - color_blend_ci.attachmentCount = 1; - color_blend_ci.pAttachments = &color_blend_attachment; - color_blend_ci.blendConstants[0] = 0.0f; - color_blend_ci.blendConstants[1] = 0.0f; - color_blend_ci.blendConstants[2] = 0.0f; - color_blend_ci.blendConstants[3] = 0.0f; - - static constexpr std::array dynamic_states = {VK_DYNAMIC_STATE_VIEWPORT, - VK_DYNAMIC_STATE_SCISSOR}; - VkPipelineDynamicStateCreateInfo dynamic_state_ci; - dynamic_state_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO; - dynamic_state_ci.pNext = nullptr; - dynamic_state_ci.flags = 0; - dynamic_state_ci.dynamicStateCount = static_cast<u32>(dynamic_states.size()); - dynamic_state_ci.pDynamicStates = dynamic_states.data(); - - VkGraphicsPipelineCreateInfo pipeline_ci; - pipeline_ci.sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO; - pipeline_ci.pNext = nullptr; - pipeline_ci.flags = 0; - pipeline_ci.stageCount = static_cast<u32>(shader_stages.size()); - pipeline_ci.pStages = shader_stages.data(); - pipeline_ci.pVertexInputState = &vertex_input_ci; - pipeline_ci.pInputAssemblyState = &input_assembly_ci; - pipeline_ci.pTessellationState = nullptr; - pipeline_ci.pViewportState = &viewport_state_ci; - pipeline_ci.pRasterizationState = &rasterization_ci; - pipeline_ci.pMultisampleState = &multisampling_ci; - pipeline_ci.pDepthStencilState = nullptr; - pipeline_ci.pColorBlendState = &color_blend_ci; - pipeline_ci.pDynamicState = &dynamic_state_ci; - pipeline_ci.layout = *pipeline_layout; - pipeline_ci.renderPass = *renderpass; - pipeline_ci.subpass = 0; - pipeline_ci.basePipelineHandle = 0; - pipeline_ci.basePipelineIndex = 0; + const VkPipelineVertexInputStateCreateInfo vertex_input_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .vertexBindingDescriptionCount = 1, + .pVertexBindingDescriptions = &vertex_binding_description, + .vertexAttributeDescriptionCount = u32{vertex_attrs_description.size()}, + .pVertexAttributeDescriptions = vertex_attrs_description.data(), + }; + + const VkPipelineInputAssemblyStateCreateInfo input_assembly_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .topology = VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP, + .primitiveRestartEnable = VK_FALSE, + }; + + const VkPipelineViewportStateCreateInfo viewport_state_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .viewportCount = 1, + .pViewports = nullptr, + .scissorCount = 1, + .pScissors = nullptr, + }; + + const VkPipelineRasterizationStateCreateInfo rasterization_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .depthClampEnable = VK_FALSE, + .rasterizerDiscardEnable = VK_FALSE, + .polygonMode = VK_POLYGON_MODE_FILL, + .cullMode = VK_CULL_MODE_NONE, + .frontFace = VK_FRONT_FACE_CLOCKWISE, + .depthBiasEnable = VK_FALSE, + .depthBiasConstantFactor = 0.0f, + .depthBiasClamp = 0.0f, + .depthBiasSlopeFactor = 0.0f, + .lineWidth = 1.0f, + }; + + const VkPipelineMultisampleStateCreateInfo multisampling_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .rasterizationSamples = VK_SAMPLE_COUNT_1_BIT, + .sampleShadingEnable = VK_FALSE, + .minSampleShading = 0.0f, + .pSampleMask = nullptr, + .alphaToCoverageEnable = VK_FALSE, + .alphaToOneEnable = VK_FALSE, + }; + + const VkPipelineColorBlendAttachmentState color_blend_attachment{ + .blendEnable = VK_FALSE, + .srcColorBlendFactor = VK_BLEND_FACTOR_ZERO, + .dstColorBlendFactor = VK_BLEND_FACTOR_ZERO, + .colorBlendOp = VK_BLEND_OP_ADD, + .srcAlphaBlendFactor = VK_BLEND_FACTOR_ZERO, + .dstAlphaBlendFactor = VK_BLEND_FACTOR_ZERO, + .alphaBlendOp = VK_BLEND_OP_ADD, + .colorWriteMask = VK_COLOR_COMPONENT_R_BIT | VK_COLOR_COMPONENT_G_BIT | + VK_COLOR_COMPONENT_B_BIT | VK_COLOR_COMPONENT_A_BIT, + }; + + const VkPipelineColorBlendStateCreateInfo color_blend_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .logicOpEnable = VK_FALSE, + .logicOp = VK_LOGIC_OP_COPY, + .attachmentCount = 1, + .pAttachments = &color_blend_attachment, + .blendConstants = {0.0f, 0.0f, 0.0f, 0.0f}, + }; + + static constexpr std::array dynamic_states{ + VK_DYNAMIC_STATE_VIEWPORT, + VK_DYNAMIC_STATE_SCISSOR, + }; + const VkPipelineDynamicStateCreateInfo dynamic_state_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .dynamicStateCount = static_cast<u32>(dynamic_states.size()), + .pDynamicStates = dynamic_states.data(), + }; + + const VkGraphicsPipelineCreateInfo pipeline_ci{ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(shader_stages.size()), + .pStages = shader_stages.data(), + .pVertexInputState = &vertex_input_ci, + .pInputAssemblyState = &input_assembly_ci, + .pTessellationState = nullptr, + .pViewportState = &viewport_state_ci, + .pRasterizationState = &rasterization_ci, + .pMultisampleState = &multisampling_ci, + .pDepthStencilState = nullptr, + .pColorBlendState = &color_blend_ci, + .pDynamicState = &dynamic_state_ci, + .layout = *pipeline_layout, + .renderPass = *renderpass, + .subpass = 0, + .basePipelineHandle = 0, + .basePipelineIndex = 0, + }; pipeline = device.GetLogical().CreateGraphicsPipeline(pipeline_ci); } void VKBlitScreen::CreateSampler() { - VkSamplerCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.magFilter = VK_FILTER_LINEAR; - ci.minFilter = VK_FILTER_NEAREST; - ci.mipmapMode = VK_SAMPLER_MIPMAP_MODE_LINEAR; - ci.addressModeU = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER; - ci.addressModeV = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER; - ci.addressModeW = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER; - ci.mipLodBias = 0.0f; - ci.anisotropyEnable = VK_FALSE; - ci.maxAnisotropy = 0.0f; - ci.compareEnable = VK_FALSE; - ci.compareOp = VK_COMPARE_OP_NEVER; - ci.minLod = 0.0f; - ci.maxLod = 0.0f; - ci.borderColor = VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK; - ci.unnormalizedCoordinates = VK_FALSE; + const VkSamplerCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .magFilter = VK_FILTER_LINEAR, + .minFilter = VK_FILTER_NEAREST, + .mipmapMode = VK_SAMPLER_MIPMAP_MODE_LINEAR, + .addressModeU = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER, + .addressModeV = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER, + .addressModeW = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER, + .mipLodBias = 0.0f, + .anisotropyEnable = VK_FALSE, + .maxAnisotropy = 0.0f, + .compareEnable = VK_FALSE, + .compareOp = VK_COMPARE_OP_NEVER, + .minLod = 0.0f, + .maxLod = 0.0f, + .borderColor = VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK, + .unnormalizedCoordinates = VK_FALSE, + }; sampler = device.GetLogical().CreateSampler(ci); } @@ -650,15 +687,17 @@ void VKBlitScreen::CreateFramebuffers() { const VkExtent2D size{swapchain.GetSize()}; framebuffers.resize(image_count); - VkFramebufferCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.renderPass = *renderpass; - ci.attachmentCount = 1; - ci.width = size.width; - ci.height = size.height; - ci.layers = 1; + VkFramebufferCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .renderPass = *renderpass, + .attachmentCount = 1, + .pAttachments = nullptr, + .width = size.width, + .height = size.height, + .layers = 1, + }; for (std::size_t i = 0; i < image_count; ++i) { const VkImageView image_view{swapchain.GetImageViewIndex(i)}; @@ -669,7 +708,7 @@ void VKBlitScreen::CreateFramebuffers() { void VKBlitScreen::ReleaseRawImages() { for (std::size_t i = 0; i < raw_images.size(); ++i) { - watches[i]->Wait(); + scheduler.Wait(resource_ticks.at(i)); } raw_images.clear(); raw_buffer_commits.clear(); @@ -678,16 +717,17 @@ void VKBlitScreen::ReleaseRawImages() { } void VKBlitScreen::CreateStagingBuffer(const Tegra::FramebufferConfig& framebuffer) { - VkBufferCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.size = CalculateBufferSize(framebuffer); - ci.usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | - VK_BUFFER_USAGE_VERTEX_BUFFER_BIT | VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT; - ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - ci.queueFamilyIndexCount = 0; - ci.pQueueFamilyIndices = nullptr; + const VkBufferCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = CalculateBufferSize(framebuffer), + .usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | + VK_BUFFER_USAGE_VERTEX_BUFFER_BIT | VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }; buffer = device.GetLogical().CreateBuffer(ci); buffer_commit = memory_manager.Commit(buffer, true); @@ -697,24 +737,28 @@ void VKBlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) raw_images.resize(image_count); raw_buffer_commits.resize(image_count); - VkImageCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.imageType = VK_IMAGE_TYPE_2D; - ci.format = GetFormat(framebuffer); - ci.extent.width = framebuffer.width; - ci.extent.height = framebuffer.height; - ci.extent.depth = 1; - ci.mipLevels = 1; - ci.arrayLayers = 1; - ci.samples = VK_SAMPLE_COUNT_1_BIT; - ci.tiling = VK_IMAGE_TILING_LINEAR; - ci.usage = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_SAMPLED_BIT; - ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - ci.queueFamilyIndexCount = 0; - ci.pQueueFamilyIndices = nullptr; - ci.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED; + const VkImageCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .imageType = VK_IMAGE_TYPE_2D, + .format = GetFormat(framebuffer), + .extent = + { + .width = framebuffer.width, + .height = framebuffer.height, + .depth = 1, + }, + .mipLevels = 1, + .arrayLayers = 1, + .samples = VK_SAMPLE_COUNT_1_BIT, + .tiling = VK_IMAGE_TILING_LINEAR, + .usage = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_SAMPLED_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, + }; for (std::size_t i = 0; i < image_count; ++i) { raw_images[i] = std::make_unique<VKImage>(device, scheduler, ci, VK_IMAGE_ASPECT_COLOR_BIT); @@ -723,39 +767,43 @@ void VKBlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) } void VKBlitScreen::UpdateDescriptorSet(std::size_t image_index, VkImageView image_view) const { - VkDescriptorBufferInfo buffer_info; - buffer_info.buffer = *buffer; - buffer_info.offset = offsetof(BufferData, uniform); - buffer_info.range = sizeof(BufferData::uniform); - - VkWriteDescriptorSet ubo_write; - ubo_write.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET; - ubo_write.pNext = nullptr; - ubo_write.dstSet = descriptor_sets[image_index]; - ubo_write.dstBinding = 0; - ubo_write.dstArrayElement = 0; - ubo_write.descriptorCount = 1; - ubo_write.descriptorType = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER; - ubo_write.pImageInfo = nullptr; - ubo_write.pBufferInfo = &buffer_info; - ubo_write.pTexelBufferView = nullptr; - - VkDescriptorImageInfo image_info; - image_info.sampler = *sampler; - image_info.imageView = image_view; - image_info.imageLayout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL; - - VkWriteDescriptorSet sampler_write; - sampler_write.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET; - sampler_write.pNext = nullptr; - sampler_write.dstSet = descriptor_sets[image_index]; - sampler_write.dstBinding = 1; - sampler_write.dstArrayElement = 0; - sampler_write.descriptorCount = 1; - sampler_write.descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER; - sampler_write.pImageInfo = &image_info; - sampler_write.pBufferInfo = nullptr; - sampler_write.pTexelBufferView = nullptr; + const VkDescriptorBufferInfo buffer_info{ + .buffer = *buffer, + .offset = offsetof(BufferData, uniform), + .range = sizeof(BufferData::uniform), + }; + + const VkWriteDescriptorSet ubo_write{ + .sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, + .pNext = nullptr, + .dstSet = descriptor_sets[image_index], + .dstBinding = 0, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER, + .pImageInfo = nullptr, + .pBufferInfo = &buffer_info, + .pTexelBufferView = nullptr, + }; + + const VkDescriptorImageInfo image_info{ + .sampler = *sampler, + .imageView = image_view, + .imageLayout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL, + }; + + const VkWriteDescriptorSet sampler_write{ + .sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, + .pNext = nullptr, + .dstSet = descriptor_sets[image_index], + .dstBinding = 1, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .pImageInfo = &image_info, + .pBufferInfo = nullptr, + .pTexelBufferView = nullptr, + }; device.GetLogical().UpdateDescriptorSets(std::array{ubo_write, sampler_write}, {}); } diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.h b/src/video_core/renderer_vulkan/vk_blit_screen.h index 5eb544aea..8f2839214 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.h +++ b/src/video_core/renderer_vulkan/vk_blit_screen.h @@ -4,18 +4,19 @@ #pragma once -#include <array> #include <memory> -#include <tuple> #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Core { class System; } +namespace Core::Memory { +class Memory; +} + namespace Core::Frontend { class EmuWindow; } @@ -31,26 +32,26 @@ class RasterizerInterface; namespace Vulkan { struct ScreenInfo; + class RasterizerVulkan; class VKDevice; -class VKFence; class VKImage; class VKScheduler; class VKSwapchain; class VKBlitScreen final { public: - explicit VKBlitScreen(Core::System& system, Core::Frontend::EmuWindow& render_window, + explicit VKBlitScreen(Core::Memory::Memory& cpu_memory, + Core::Frontend::EmuWindow& render_window, VideoCore::RasterizerInterface& rasterizer, const VKDevice& device, - VKResourceManager& resource_manager, VKMemoryManager& memory_manager, - VKSwapchain& swapchain, VKScheduler& scheduler, - const VKScreenInfo& screen_info); + VKMemoryManager& memory_manager, VKSwapchain& swapchain, + VKScheduler& scheduler, const VKScreenInfo& screen_info); ~VKBlitScreen(); void Recreate(); - std::tuple<VKFence&, VkSemaphore> Draw(const Tegra::FramebufferConfig& framebuffer, - bool use_accelerated); + [[nodiscard]] VkSemaphore Draw(const Tegra::FramebufferConfig& framebuffer, + bool use_accelerated); private: struct BufferData; @@ -82,11 +83,10 @@ private: u64 GetRawImageOffset(const Tegra::FramebufferConfig& framebuffer, std::size_t image_index) const; - Core::System& system; + Core::Memory::Memory& cpu_memory; Core::Frontend::EmuWindow& render_window; VideoCore::RasterizerInterface& rasterizer; const VKDevice& device; - VKResourceManager& resource_manager; VKMemoryManager& memory_manager; VKSwapchain& swapchain; VKScheduler& scheduler; @@ -107,7 +107,7 @@ private: vk::Buffer buffer; VKMemoryCommit buffer_commit; - std::vector<std::unique_ptr<VKFenceWatch>> watches; + std::vector<u64> resource_ticks; std::vector<vk::Semaphore> semaphores; std::vector<std::unique_ptr<VKImage>> raw_images; diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index 0d167afbd..d9d3da9ea 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp @@ -5,12 +5,9 @@ #include <algorithm> #include <cstring> #include <memory> -#include <optional> -#include <tuple> -#include "common/assert.h" -#include "common/bit_util.h" #include "core/core.h" +#include "video_core/buffer_cache/buffer_cache.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" @@ -40,112 +37,88 @@ std::unique_ptr<VKStreamBuffer> CreateStreamBuffer(const VKDevice& device, VKSch } // Anonymous namespace -CachedBufferBlock::CachedBufferBlock(const VKDevice& device, VKMemoryManager& memory_manager, - VAddr cpu_addr, std::size_t size) - : VideoCommon::BufferBlock{cpu_addr, size} { - VkBufferCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.size = static_cast<VkDeviceSize>(size); - ci.usage = BUFFER_USAGE | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT; - ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - ci.queueFamilyIndexCount = 0; - ci.pQueueFamilyIndices = nullptr; +Buffer::Buffer(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler_, + VKStagingBufferPool& staging_pool_, VAddr cpu_addr, std::size_t size) + : BufferBlock{cpu_addr, size}, scheduler{scheduler_}, staging_pool{staging_pool_} { + const VkBufferCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = static_cast<VkDeviceSize>(size), + .usage = BUFFER_USAGE | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }; buffer.handle = device.GetLogical().CreateBuffer(ci); buffer.commit = memory_manager.Commit(buffer.handle, false); } -CachedBufferBlock::~CachedBufferBlock() = default; +Buffer::~Buffer() = default; -VKBufferCache::VKBufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system, - const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool) - : VideoCommon::BufferCache<Buffer, VkBuffer, VKStreamBuffer>{rasterizer, system, - CreateStreamBuffer(device, - scheduler)}, - device{device}, memory_manager{memory_manager}, scheduler{scheduler}, staging_pool{ - staging_pool} {} - -VKBufferCache::~VKBufferCache() = default; - -Buffer VKBufferCache::CreateBlock(VAddr cpu_addr, std::size_t size) { - return std::make_shared<CachedBufferBlock>(device, memory_manager, cpu_addr, size); -} - -const VkBuffer* VKBufferCache::ToHandle(const Buffer& buffer) { - return buffer->GetHandle(); -} - -const VkBuffer* VKBufferCache::GetEmptyBuffer(std::size_t size) { - size = std::max(size, std::size_t(4)); - const auto& empty = staging_pool.GetUnusedBuffer(size, false); - scheduler.RequestOutsideRenderPassOperationContext(); - scheduler.Record([size, buffer = *empty.handle](vk::CommandBuffer cmdbuf) { - cmdbuf.FillBuffer(buffer, 0, size, 0); - }); - return empty.handle.address(); -} - -void VKBufferCache::UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - const u8* data) { +void Buffer::Upload(std::size_t offset, std::size_t size, const u8* data) { const auto& staging = staging_pool.GetUnusedBuffer(size, true); std::memcpy(staging.commit->Map(size), data, size); scheduler.RequestOutsideRenderPassOperationContext(); - scheduler.Record([staging = *staging.handle, buffer = *buffer->GetHandle(), offset, - size](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyBuffer(staging, buffer, VkBufferCopy{0, offset, size}); - - VkBufferMemoryBarrier barrier; - barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; - barrier.pNext = nullptr; - barrier.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT; - barrier.dstAccessMask = UPLOAD_ACCESS_BARRIERS; - barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.buffer = buffer; - barrier.offset = offset; - barrier.size = size; + + const VkBuffer handle = Handle(); + scheduler.Record([staging = *staging.handle, handle, offset, size](vk::CommandBuffer cmdbuf) { + cmdbuf.CopyBuffer(staging, handle, VkBufferCopy{0, offset, size}); + + const VkBufferMemoryBarrier barrier{ + .sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = UPLOAD_ACCESS_BARRIERS, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .buffer = handle, + .offset = offset, + .size = size, + }; cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, UPLOAD_PIPELINE_STAGE, 0, {}, barrier, {}); }); } -void VKBufferCache::DownloadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - u8* data) { +void Buffer::Download(std::size_t offset, std::size_t size, u8* data) { const auto& staging = staging_pool.GetUnusedBuffer(size, true); scheduler.RequestOutsideRenderPassOperationContext(); - scheduler.Record([staging = *staging.handle, buffer = *buffer->GetHandle(), offset, - size](vk::CommandBuffer cmdbuf) { - VkBufferMemoryBarrier barrier; - barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; - barrier.pNext = nullptr; - barrier.srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT; - barrier.dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT; - barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.buffer = buffer; - barrier.offset = offset; - barrier.size = size; + + const VkBuffer handle = Handle(); + scheduler.Record([staging = *staging.handle, handle, offset, size](vk::CommandBuffer cmdbuf) { + const VkBufferMemoryBarrier barrier{ + .sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .buffer = handle, + .offset = offset, + .size = size, + }; cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, 0, {}, barrier, {}); - cmdbuf.CopyBuffer(buffer, staging, VkBufferCopy{offset, 0, size}); + cmdbuf.CopyBuffer(handle, staging, VkBufferCopy{offset, 0, size}); }); scheduler.Finish(); std::memcpy(data, staging.commit->Map(size), size); } -void VKBufferCache::CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset, - std::size_t dst_offset, std::size_t size) { +void Buffer::CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, + std::size_t size) { scheduler.RequestOutsideRenderPassOperationContext(); - scheduler.Record([src_buffer = *src->GetHandle(), dst_buffer = *dst->GetHandle(), src_offset, - dst_offset, size](vk::CommandBuffer cmdbuf) { + + const VkBuffer dst_buffer = Handle(); + scheduler.Record([src_buffer = src.Handle(), dst_buffer, src_offset, dst_offset, + size](vk::CommandBuffer cmdbuf) { cmdbuf.CopyBuffer(src_buffer, dst_buffer, VkBufferCopy{src_offset, dst_offset, size}); std::array<VkBufferMemoryBarrier, 2> barriers; @@ -172,4 +145,31 @@ void VKBufferCache::CopyBlock(const Buffer& src, const Buffer& dst, std::size_t }); } +VKBufferCache::VKBufferCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, + const VKDevice& device_, VKMemoryManager& memory_manager_, + VKScheduler& scheduler_, VKStagingBufferPool& staging_pool_) + : VideoCommon::BufferCache<Buffer, VkBuffer, VKStreamBuffer>{rasterizer, gpu_memory, cpu_memory, + CreateStreamBuffer(device_, + scheduler_)}, + device{device_}, memory_manager{memory_manager_}, scheduler{scheduler_}, staging_pool{ + staging_pool_} {} + +VKBufferCache::~VKBufferCache() = default; + +std::shared_ptr<Buffer> VKBufferCache::CreateBlock(VAddr cpu_addr, std::size_t size) { + return std::make_shared<Buffer>(device, memory_manager, scheduler, staging_pool, cpu_addr, + size); +} + +VKBufferCache::BufferInfo VKBufferCache::GetEmptyBuffer(std::size_t size) { + size = std::max(size, std::size_t(4)); + const auto& empty = staging_pool.GetUnusedBuffer(size, false); + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([size, buffer = *empty.handle](vk::CommandBuffer cmdbuf) { + cmdbuf.FillBuffer(buffer, 0, size, 0); + }); + return {*empty.handle, 0, 0}; +} + } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h index d3c23da98..7fb5ceedf 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.h +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h @@ -5,68 +5,60 @@ #pragma once #include <memory> -#include <unordered_map> -#include <vector> #include "common/common_types.h" #include "video_core/buffer_cache/buffer_cache.h" -#include "video_core/rasterizer_cache.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" #include "video_core/renderer_vulkan/wrapper.h" -namespace Core { -class System; -} - namespace Vulkan { class VKDevice; class VKMemoryManager; class VKScheduler; -class CachedBufferBlock final : public VideoCommon::BufferBlock { +class Buffer final : public VideoCommon::BufferBlock { public: - explicit CachedBufferBlock(const VKDevice& device, VKMemoryManager& memory_manager, - VAddr cpu_addr, std::size_t size); - ~CachedBufferBlock(); + explicit Buffer(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler, + VKStagingBufferPool& staging_pool, VAddr cpu_addr, std::size_t size); + ~Buffer(); + + void Upload(std::size_t offset, std::size_t size, const u8* data); + + void Download(std::size_t offset, std::size_t size, u8* data); + + void CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, + std::size_t size); - const VkBuffer* GetHandle() const { - return buffer.handle.address(); + VkBuffer Handle() const { + return *buffer.handle; + } + + u64 Address() const { + return 0; } private: + VKScheduler& scheduler; + VKStagingBufferPool& staging_pool; + VKBuffer buffer; }; -using Buffer = std::shared_ptr<CachedBufferBlock>; - class VKBufferCache final : public VideoCommon::BufferCache<Buffer, VkBuffer, VKStreamBuffer> { public: - explicit VKBufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system, + explicit VKBufferCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler, VKStagingBufferPool& staging_pool); ~VKBufferCache(); - const VkBuffer* GetEmptyBuffer(std::size_t size) override; + BufferInfo GetEmptyBuffer(std::size_t size) override; protected: - void WriteBarrier() override {} - - Buffer CreateBlock(VAddr cpu_addr, std::size_t size) override; - - const VkBuffer* ToHandle(const Buffer& buffer) override; - - void UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - const u8* data) override; - - void DownloadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size, - u8* data) override; - - void CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset, - std::size_t dst_offset, std::size_t size) override; + std::shared_ptr<Buffer> CreateBlock(VAddr cpu_addr, std::size_t size) override; private: const VKDevice& device; diff --git a/src/video_core/renderer_vulkan/vk_command_pool.cpp b/src/video_core/renderer_vulkan/vk_command_pool.cpp new file mode 100644 index 000000000..6339f4fe0 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_command_pool.cpp @@ -0,0 +1,46 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstddef> + +#include "video_core/renderer_vulkan/vk_command_pool.h" +#include "video_core/renderer_vulkan/vk_device.h" +#include "video_core/renderer_vulkan/wrapper.h" + +namespace Vulkan { + +constexpr size_t COMMAND_BUFFER_POOL_SIZE = 0x1000; + +struct CommandPool::Pool { + vk::CommandPool handle; + vk::CommandBuffers cmdbufs; +}; + +CommandPool::CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device) + : ResourcePool(master_semaphore, COMMAND_BUFFER_POOL_SIZE), device{device} {} + +CommandPool::~CommandPool() = default; + +void CommandPool::Allocate(size_t begin, size_t end) { + // Command buffers are going to be commited, recorded, executed every single usage cycle. + // They are also going to be reseted when commited. + Pool& pool = pools.emplace_back(); + pool.handle = device.GetLogical().CreateCommandPool({ + .sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO, + .pNext = nullptr, + .flags = + VK_COMMAND_POOL_CREATE_TRANSIENT_BIT | VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT, + .queueFamilyIndex = device.GetGraphicsFamily(), + }); + pool.cmdbufs = pool.handle.Allocate(COMMAND_BUFFER_POOL_SIZE); +} + +VkCommandBuffer CommandPool::Commit() { + const size_t index = CommitResource(); + const auto pool_index = index / COMMAND_BUFFER_POOL_SIZE; + const auto sub_index = index % COMMAND_BUFFER_POOL_SIZE; + return pools[pool_index].cmdbufs[sub_index]; +} + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_command_pool.h b/src/video_core/renderer_vulkan/vk_command_pool.h new file mode 100644 index 000000000..b9cb3fb5d --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_command_pool.h @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstddef> +#include <vector> + +#include "video_core/renderer_vulkan/vk_resource_pool.h" +#include "video_core/renderer_vulkan/wrapper.h" + +namespace Vulkan { + +class MasterSemaphore; +class VKDevice; + +class CommandPool final : public ResourcePool { +public: + explicit CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device); + ~CommandPool() override; + + void Allocate(size_t begin, size_t end) override; + + VkCommandBuffer Commit(); + +private: + struct Pool; + + const VKDevice& device; + std::vector<Pool> pools; +}; + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.cpp b/src/video_core/renderer_vulkan/vk_compute_pass.cpp index 9d92305f4..9637c6059 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pass.cpp @@ -6,7 +6,7 @@ #include <memory> #include <optional> #include <utility> -#include <vector> + #include "common/alignment.h" #include "common/assert.h" #include "common/common_types.h" @@ -112,35 +112,36 @@ constexpr u8 quad_array[] = { 0xf9, 0x00, 0x02, 0x00, 0x21, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x23, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x4b, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x4e, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x4c, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x4b, 0x00, 0x00, 0x00, - 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00}; + 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00, +}; VkDescriptorSetLayoutBinding BuildQuadArrayPassDescriptorSetLayoutBinding() { - VkDescriptorSetLayoutBinding binding; - binding.binding = 0; - binding.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER; - binding.descriptorCount = 1; - binding.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT; - binding.pImmutableSamplers = nullptr; - return binding; + return { + .binding = 0, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }; } VkDescriptorUpdateTemplateEntryKHR BuildQuadArrayPassDescriptorUpdateTemplateEntry() { - VkDescriptorUpdateTemplateEntryKHR entry; - entry.dstBinding = 0; - entry.dstArrayElement = 0; - entry.descriptorCount = 1; - entry.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER; - entry.offset = 0; - entry.stride = sizeof(DescriptorUpdateEntry); - return entry; + return { + .dstBinding = 0, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = 0, + .stride = sizeof(DescriptorUpdateEntry), + }; } -VkPushConstantRange BuildQuadArrayPassPushConstantRange() { - VkPushConstantRange range; - range.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT; - range.offset = 0; - range.size = sizeof(u32); - return range; +VkPushConstantRange BuildComputePushConstantRange(std::size_t size) { + return { + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .offset = 0, + .size = static_cast<u32>(size), + }; } // Uint8 SPIR-V module. Generated from the "shaders/" directory. @@ -218,32 +219,161 @@ constexpr u8 uint8_pass[] = { 0x2a, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, - 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00}; - -std::array<VkDescriptorSetLayoutBinding, 2> BuildUint8PassDescriptorSetBindings() { - std::array<VkDescriptorSetLayoutBinding, 2> bindings; - bindings[0].binding = 0; - bindings[0].descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER; - bindings[0].descriptorCount = 1; - bindings[0].stageFlags = VK_SHADER_STAGE_COMPUTE_BIT; - bindings[0].pImmutableSamplers = nullptr; - bindings[1].binding = 1; - bindings[1].descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER; - bindings[1].descriptorCount = 1; - bindings[1].stageFlags = VK_SHADER_STAGE_COMPUTE_BIT; - bindings[1].pImmutableSamplers = nullptr; - return bindings; + 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00, +}; + +// Quad indexed SPIR-V module. Generated from the "shaders/" directory. +constexpr u8 QUAD_INDEXED_SPV[] = { + 0x03, 0x02, 0x23, 0x07, 0x00, 0x00, 0x01, 0x00, 0x07, 0x00, 0x08, 0x00, 0x7c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x06, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x47, 0x4c, 0x53, 0x4c, 0x2e, 0x73, 0x74, 0x64, 0x2e, 0x34, 0x35, 0x30, + 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x06, 0x00, 0x05, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, + 0x00, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x10, 0x00, 0x06, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x11, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x47, 0x00, 0x04, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x47, 0x00, 0x04, 0x00, 0x15, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, + 0x48, 0x00, 0x04, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, + 0x48, 0x00, 0x05, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x16, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0x47, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x47, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x48, 0x00, 0x05, 0x00, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x22, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x56, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x04, 0x00, 0x00, 0x00, 0x48, 0x00, 0x04, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x57, 0x00, 0x00, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x59, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x59, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x72, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x19, 0x00, 0x00, 0x00, 0x13, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, 0x21, 0x00, 0x03, 0x00, + 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x07, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x0a, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, + 0x0c, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x13, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x16, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, + 0x20, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, + 0x3b, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x02, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x21, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x04, 0x00, 0x22, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x23, 0x00, 0x00, 0x00, + 0x24, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x25, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x26, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x2b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x3b, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x3f, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x04, 0x00, 0x41, 0x00, 0x00, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x42, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x43, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x09, 0x00, 0x41, 0x00, 0x00, 0x00, + 0x44, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, + 0x42, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, + 0x46, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x41, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, + 0x56, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x57, 0x00, 0x00, 0x00, + 0x56, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x58, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x57, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x58, 0x00, 0x00, 0x00, 0x59, 0x00, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x5b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x69, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x70, 0x00, 0x00, 0x00, + 0x00, 0x04, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x71, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x06, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x72, 0x00, 0x00, 0x00, + 0x70, 0x00, 0x00, 0x00, 0x71, 0x00, 0x00, 0x00, 0x71, 0x00, 0x00, 0x00, 0x36, 0x00, 0x05, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, + 0xf8, 0x00, 0x02, 0x00, 0x05, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x46, 0x00, 0x00, 0x00, + 0x47, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x74, 0x00, 0x00, 0x00, + 0xf8, 0x00, 0x02, 0x00, 0x74, 0x00, 0x00, 0x00, 0xf6, 0x00, 0x04, 0x00, 0x73, 0x00, 0x00, 0x00, + 0x76, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x75, 0x00, 0x00, 0x00, + 0xf8, 0x00, 0x02, 0x00, 0x75, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x0e, 0x00, 0x00, 0x00, + 0x0f, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, + 0x44, 0x00, 0x05, 0x00, 0x09, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, + 0x19, 0x00, 0x00, 0x00, 0xaf, 0x00, 0x05, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, + 0x14, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0xf7, 0x00, 0x03, 0x00, 0x1e, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xfa, 0x00, 0x04, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, + 0x73, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, + 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, + 0xc4, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, + 0x28, 0x00, 0x00, 0x00, 0x82, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, + 0x2b, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0xc4, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x31, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x82, 0x00, 0x05, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0xf9, 0x00, 0x02, 0x00, 0x35, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x35, 0x00, 0x00, 0x00, + 0xf5, 0x00, 0x07, 0x00, 0x09, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, + 0x1e, 0x00, 0x00, 0x00, 0x6f, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0xb0, 0x00, 0x05, 0x00, + 0x1b, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, + 0xf6, 0x00, 0x04, 0x00, 0x37, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xfa, 0x00, 0x04, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, + 0xf8, 0x00, 0x02, 0x00, 0x36, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x40, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, + 0x47, 0x00, 0x00, 0x00, 0x44, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x07, 0x00, 0x00, 0x00, + 0x48, 0x00, 0x00, 0x00, 0x47, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, + 0x06, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, + 0xc3, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x4e, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, + 0x2e, 0x00, 0x00, 0x00, 0xc7, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, + 0x4a, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, + 0x54, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, + 0x5b, 0x00, 0x00, 0x00, 0x5c, 0x00, 0x00, 0x00, 0x59, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, + 0x4e, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x5d, 0x00, 0x00, 0x00, + 0x5c, 0x00, 0x00, 0x00, 0xcb, 0x00, 0x06, 0x00, 0x09, 0x00, 0x00, 0x00, 0x62, 0x00, 0x00, 0x00, + 0x5d, 0x00, 0x00, 0x00, 0x54, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x65, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x67, 0x00, 0x00, 0x00, 0x65, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, + 0x41, 0x00, 0x05, 0x00, 0x69, 0x00, 0x00, 0x00, 0x6a, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, + 0x42, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x6b, 0x00, 0x00, 0x00, + 0x6a, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, 0x09, 0x00, 0x00, 0x00, 0x6c, 0x00, 0x00, 0x00, + 0x62, 0x00, 0x00, 0x00, 0x6b, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, 0x5b, 0x00, 0x00, 0x00, + 0x6d, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x67, 0x00, 0x00, 0x00, + 0x3e, 0x00, 0x03, 0x00, 0x6d, 0x00, 0x00, 0x00, 0x6c, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, + 0x09, 0x00, 0x00, 0x00, 0x6f, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, + 0xf9, 0x00, 0x02, 0x00, 0x35, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x37, 0x00, 0x00, 0x00, + 0xf9, 0x00, 0x02, 0x00, 0x73, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x76, 0x00, 0x00, 0x00, + 0xf9, 0x00, 0x02, 0x00, 0x74, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x73, 0x00, 0x00, 0x00, + 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00, +}; + +std::array<VkDescriptorSetLayoutBinding, 2> BuildInputOutputDescriptorSetBindings() { + return {{ + { + .binding = 0, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + { + .binding = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }, + }}; } -VkDescriptorUpdateTemplateEntryKHR BuildUint8PassDescriptorUpdateTemplateEntry() { - VkDescriptorUpdateTemplateEntryKHR entry; - entry.dstBinding = 0; - entry.dstArrayElement = 0; - entry.descriptorCount = 2; - entry.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER; - entry.offset = 0; - entry.stride = sizeof(DescriptorUpdateEntry); - return entry; +VkDescriptorUpdateTemplateEntryKHR BuildInputOutputDescriptorUpdateTemplate() { + return { + .dstBinding = 0, + .dstArrayElement = 0, + .descriptorCount = 2, + .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, + .offset = 0, + .stride = sizeof(DescriptorUpdateEntry), + }; } } // Anonymous namespace @@ -253,37 +383,37 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto vk::Span<VkDescriptorUpdateTemplateEntryKHR> templates, vk::Span<VkPushConstantRange> push_constants, std::size_t code_size, const u8* code) { - VkDescriptorSetLayoutCreateInfo descriptor_layout_ci; - descriptor_layout_ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO; - descriptor_layout_ci.pNext = nullptr; - descriptor_layout_ci.flags = 0; - descriptor_layout_ci.bindingCount = bindings.size(); - descriptor_layout_ci.pBindings = bindings.data(); - descriptor_set_layout = device.GetLogical().CreateDescriptorSetLayout(descriptor_layout_ci); - - VkPipelineLayoutCreateInfo pipeline_layout_ci; - pipeline_layout_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO; - pipeline_layout_ci.pNext = nullptr; - pipeline_layout_ci.flags = 0; - pipeline_layout_ci.setLayoutCount = 1; - pipeline_layout_ci.pSetLayouts = descriptor_set_layout.address(); - pipeline_layout_ci.pushConstantRangeCount = push_constants.size(); - pipeline_layout_ci.pPushConstantRanges = push_constants.data(); - layout = device.GetLogical().CreatePipelineLayout(pipeline_layout_ci); + descriptor_set_layout = device.GetLogical().CreateDescriptorSetLayout({ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .bindingCount = bindings.size(), + .pBindings = bindings.data(), + }); + + layout = device.GetLogical().CreatePipelineLayout({ + .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .setLayoutCount = 1, + .pSetLayouts = descriptor_set_layout.address(), + .pushConstantRangeCount = push_constants.size(), + .pPushConstantRanges = push_constants.data(), + }); if (!templates.empty()) { - VkDescriptorUpdateTemplateCreateInfoKHR template_ci; - template_ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR; - template_ci.pNext = nullptr; - template_ci.flags = 0; - template_ci.descriptorUpdateEntryCount = templates.size(); - template_ci.pDescriptorUpdateEntries = templates.data(); - template_ci.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR; - template_ci.descriptorSetLayout = *descriptor_set_layout; - template_ci.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS; - template_ci.pipelineLayout = *layout; - template_ci.set = 0; - descriptor_template = device.GetLogical().CreateDescriptorUpdateTemplateKHR(template_ci); + descriptor_template = device.GetLogical().CreateDescriptorUpdateTemplateKHR({ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR, + .pNext = nullptr, + .flags = 0, + .descriptorUpdateEntryCount = templates.size(), + .pDescriptorUpdateEntries = templates.data(), + .templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR, + .descriptorSetLayout = *descriptor_set_layout, + .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, + .pipelineLayout = *layout, + .set = 0, + }); descriptor_allocator.emplace(descriptor_pool, *descriptor_set_layout); } @@ -291,42 +421,42 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto auto code_copy = std::make_unique<u32[]>(code_size / sizeof(u32) + 1); std::memcpy(code_copy.get(), code, code_size); - VkShaderModuleCreateInfo module_ci; - module_ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO; - module_ci.pNext = nullptr; - module_ci.flags = 0; - module_ci.codeSize = code_size; - module_ci.pCode = code_copy.get(); - module = device.GetLogical().CreateShaderModule(module_ci); - - VkComputePipelineCreateInfo pipeline_ci; - pipeline_ci.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO; - pipeline_ci.pNext = nullptr; - pipeline_ci.flags = 0; - pipeline_ci.layout = *layout; - pipeline_ci.basePipelineHandle = nullptr; - pipeline_ci.basePipelineIndex = 0; - - VkPipelineShaderStageCreateInfo& stage_ci = pipeline_ci.stage; - stage_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO; - stage_ci.pNext = nullptr; - stage_ci.flags = 0; - stage_ci.stage = VK_SHADER_STAGE_COMPUTE_BIT; - stage_ci.module = *module; - stage_ci.pName = "main"; - stage_ci.pSpecializationInfo = nullptr; - - pipeline = device.GetLogical().CreateComputePipeline(pipeline_ci); + module = device.GetLogical().CreateShaderModule({ + .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .codeSize = code_size, + .pCode = code_copy.get(), + }); + + pipeline = device.GetLogical().CreateComputePipeline({ + .sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = + { + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = VK_SHADER_STAGE_COMPUTE_BIT, + .module = *module, + .pName = "main", + .pSpecializationInfo = nullptr, + }, + .layout = *layout, + .basePipelineHandle = nullptr, + .basePipelineIndex = 0, + }); } VKComputePass::~VKComputePass() = default; -VkDescriptorSet VKComputePass::CommitDescriptorSet(VKUpdateDescriptorQueue& update_descriptor_queue, - VKFence& fence) { +VkDescriptorSet VKComputePass::CommitDescriptorSet( + VKUpdateDescriptorQueue& update_descriptor_queue) { if (!descriptor_template) { return nullptr; } - const auto set = descriptor_allocator->Commit(fence); + const VkDescriptorSet set = descriptor_allocator->Commit(); update_descriptor_queue.Send(*descriptor_template, set); return set; } @@ -337,20 +467,20 @@ QuadArrayPass::QuadArrayPass(const VKDevice& device, VKScheduler& scheduler, VKUpdateDescriptorQueue& update_descriptor_queue) : VKComputePass(device, descriptor_pool, BuildQuadArrayPassDescriptorSetLayoutBinding(), BuildQuadArrayPassDescriptorUpdateTemplateEntry(), - BuildQuadArrayPassPushConstantRange(), std::size(quad_array), quad_array), + BuildComputePushConstantRange(sizeof(u32)), std::size(quad_array), quad_array), scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, update_descriptor_queue{update_descriptor_queue} {} QuadArrayPass::~QuadArrayPass() = default; -std::pair<const VkBuffer*, VkDeviceSize> QuadArrayPass::Assemble(u32 num_vertices, u32 first) { - const u32 num_triangle_vertices = num_vertices * 6 / 4; +std::pair<VkBuffer, VkDeviceSize> QuadArrayPass::Assemble(u32 num_vertices, u32 first) { + const u32 num_triangle_vertices = (num_vertices / 4) * 6; const std::size_t staging_size = num_triangle_vertices * sizeof(u32); auto& buffer = staging_buffer_pool.GetUnusedBuffer(staging_size, false); update_descriptor_queue.Acquire(); - update_descriptor_queue.AddBuffer(buffer.handle.address(), 0, staging_size); - const auto set = CommitDescriptorSet(update_descriptor_queue, scheduler.GetFence()); + update_descriptor_queue.AddBuffer(*buffer.handle, 0, staging_size); + const VkDescriptorSet set = CommitDescriptorSet(update_descriptor_queue); scheduler.RequestOutsideRenderPassOperationContext(); @@ -377,29 +507,29 @@ std::pair<const VkBuffer*, VkDeviceSize> QuadArrayPass::Assemble(u32 num_vertice cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_VERTEX_INPUT_BIT, 0, {}, {barrier}, {}); }); - return {buffer.handle.address(), 0}; + return {*buffer.handle, 0}; } Uint8Pass::Uint8Pass(const VKDevice& device, VKScheduler& scheduler, VKDescriptorPool& descriptor_pool, VKStagingBufferPool& staging_buffer_pool, VKUpdateDescriptorQueue& update_descriptor_queue) - : VKComputePass(device, descriptor_pool, BuildUint8PassDescriptorSetBindings(), - BuildUint8PassDescriptorUpdateTemplateEntry(), {}, std::size(uint8_pass), + : VKComputePass(device, descriptor_pool, BuildInputOutputDescriptorSetBindings(), + BuildInputOutputDescriptorUpdateTemplate(), {}, std::size(uint8_pass), uint8_pass), scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, update_descriptor_queue{update_descriptor_queue} {} Uint8Pass::~Uint8Pass() = default; -std::pair<const VkBuffer*, u64> Uint8Pass::Assemble(u32 num_vertices, VkBuffer src_buffer, - u64 src_offset) { - const auto staging_size = static_cast<u32>(num_vertices * sizeof(u16)); +std::pair<VkBuffer, u64> Uint8Pass::Assemble(u32 num_vertices, VkBuffer src_buffer, + u64 src_offset) { + const u32 staging_size = static_cast<u32>(num_vertices * sizeof(u16)); auto& buffer = staging_buffer_pool.GetUnusedBuffer(staging_size, false); update_descriptor_queue.Acquire(); - update_descriptor_queue.AddBuffer(&src_buffer, src_offset, num_vertices); - update_descriptor_queue.AddBuffer(buffer.handle.address(), 0, staging_size); - const auto set = CommitDescriptorSet(update_descriptor_queue, scheduler.GetFence()); + update_descriptor_queue.AddBuffer(src_buffer, src_offset, num_vertices); + update_descriptor_queue.AddBuffer(*buffer.handle, 0, staging_size); + const VkDescriptorSet set = CommitDescriptorSet(update_descriptor_queue); scheduler.RequestOutsideRenderPassOperationContext(); scheduler.Record([layout = *layout, pipeline = *pipeline, buffer = *buffer.handle, set, @@ -422,7 +552,73 @@ std::pair<const VkBuffer*, u64> Uint8Pass::Assemble(u32 num_vertices, VkBuffer s cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_VERTEX_INPUT_BIT, 0, {}, barrier, {}); }); - return {buffer.handle.address(), 0}; + return {*buffer.handle, 0}; +} + +QuadIndexedPass::QuadIndexedPass(const VKDevice& device, VKScheduler& scheduler, + VKDescriptorPool& descriptor_pool, + VKStagingBufferPool& staging_buffer_pool, + VKUpdateDescriptorQueue& update_descriptor_queue) + : VKComputePass(device, descriptor_pool, BuildInputOutputDescriptorSetBindings(), + BuildInputOutputDescriptorUpdateTemplate(), + BuildComputePushConstantRange(sizeof(u32) * 2), std::size(QUAD_INDEXED_SPV), + QUAD_INDEXED_SPV), + scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, + update_descriptor_queue{update_descriptor_queue} {} + +QuadIndexedPass::~QuadIndexedPass() = default; + +std::pair<VkBuffer, u64> QuadIndexedPass::Assemble( + Tegra::Engines::Maxwell3D::Regs::IndexFormat index_format, u32 num_vertices, u32 base_vertex, + VkBuffer src_buffer, u64 src_offset) { + const u32 index_shift = [index_format] { + switch (index_format) { + case Tegra::Engines::Maxwell3D::Regs::IndexFormat::UnsignedByte: + return 0; + case Tegra::Engines::Maxwell3D::Regs::IndexFormat::UnsignedShort: + return 1; + case Tegra::Engines::Maxwell3D::Regs::IndexFormat::UnsignedInt: + return 2; + } + UNREACHABLE(); + return 2; + }(); + const u32 input_size = num_vertices << index_shift; + const u32 num_tri_vertices = (num_vertices / 4) * 6; + + const std::size_t staging_size = num_tri_vertices * sizeof(u32); + auto& buffer = staging_buffer_pool.GetUnusedBuffer(staging_size, false); + + update_descriptor_queue.Acquire(); + update_descriptor_queue.AddBuffer(src_buffer, src_offset, input_size); + update_descriptor_queue.AddBuffer(*buffer.handle, 0, staging_size); + const VkDescriptorSet set = CommitDescriptorSet(update_descriptor_queue); + + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([layout = *layout, pipeline = *pipeline, buffer = *buffer.handle, set, + num_tri_vertices, base_vertex, index_shift](vk::CommandBuffer cmdbuf) { + static constexpr u32 dispatch_size = 1024; + const std::array push_constants = {base_vertex, index_shift}; + cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_COMPUTE, pipeline); + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, layout, 0, set, {}); + cmdbuf.PushConstants(layout, VK_SHADER_STAGE_COMPUTE_BIT, 0, sizeof(push_constants), + &push_constants); + cmdbuf.Dispatch(Common::AlignUp(num_tri_vertices, dispatch_size) / dispatch_size, 1, 1); + + VkBufferMemoryBarrier barrier; + barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; + barrier.pNext = nullptr; + barrier.srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT; + barrier.dstAccessMask = VK_ACCESS_VERTEX_ATTRIBUTE_READ_BIT; + barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; + barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; + barrier.buffer = buffer; + barrier.offset = 0; + barrier.size = static_cast<VkDeviceSize>(num_tri_vertices * sizeof(u32)); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, + VK_PIPELINE_STAGE_VERTEX_INPUT_BIT, 0, {}, barrier, {}); + }); + return {*buffer.handle, 0}; } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.h b/src/video_core/renderer_vulkan/vk_compute_pass.h index c62516bff..acc94f27e 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.h +++ b/src/video_core/renderer_vulkan/vk_compute_pass.h @@ -6,15 +6,15 @@ #include <optional> #include <utility> -#include <vector> + #include "common/common_types.h" +#include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { class VKDevice; -class VKFence; class VKScheduler; class VKStagingBufferPool; class VKUpdateDescriptorQueue; @@ -29,8 +29,7 @@ public: ~VKComputePass(); protected: - VkDescriptorSet CommitDescriptorSet(VKUpdateDescriptorQueue& update_descriptor_queue, - VKFence& fence); + VkDescriptorSet CommitDescriptorSet(VKUpdateDescriptorQueue& update_descriptor_queue); vk::DescriptorUpdateTemplateKHR descriptor_template; vk::PipelineLayout layout; @@ -50,7 +49,7 @@ public: VKUpdateDescriptorQueue& update_descriptor_queue); ~QuadArrayPass(); - std::pair<const VkBuffer*, VkDeviceSize> Assemble(u32 num_vertices, u32 first); + std::pair<VkBuffer, VkDeviceSize> Assemble(u32 num_vertices, u32 first); private: VKScheduler& scheduler; @@ -65,7 +64,25 @@ public: VKUpdateDescriptorQueue& update_descriptor_queue); ~Uint8Pass(); - std::pair<const VkBuffer*, u64> Assemble(u32 num_vertices, VkBuffer src_buffer, u64 src_offset); + std::pair<VkBuffer, u64> Assemble(u32 num_vertices, VkBuffer src_buffer, u64 src_offset); + +private: + VKScheduler& scheduler; + VKStagingBufferPool& staging_buffer_pool; + VKUpdateDescriptorQueue& update_descriptor_queue; +}; + +class QuadIndexedPass final : public VKComputePass { +public: + explicit QuadIndexedPass(const VKDevice& device, VKScheduler& scheduler, + VKDescriptorPool& descriptor_pool, + VKStagingBufferPool& staging_buffer_pool, + VKUpdateDescriptorQueue& update_descriptor_queue); + ~QuadIndexedPass(); + + std::pair<VkBuffer, u64> Assemble(Tegra::Engines::Maxwell3D::Regs::IndexFormat index_format, + u32 num_vertices, u32 base_vertex, VkBuffer src_buffer, + u64 src_offset); private: VKScheduler& scheduler; diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp index 23beafa4f..9be72dc9b 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp @@ -2,14 +2,12 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <memory> #include <vector> #include "video_core/renderer_vulkan/vk_compute_pipeline.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" @@ -34,7 +32,7 @@ VkDescriptorSet VKComputePipeline::CommitDescriptorSet() { if (!descriptor_template) { return {}; } - const auto set = descriptor_allocator.Commit(scheduler.GetFence()); + const VkDescriptorSet set = descriptor_allocator.Commit(); update_descriptor_queue.Send(*descriptor_template, set); return set; } @@ -45,39 +43,41 @@ vk::DescriptorSetLayout VKComputePipeline::CreateDescriptorSetLayout() const { const auto add_bindings = [&](VkDescriptorType descriptor_type, std::size_t num_entries) { // TODO(Rodrigo): Maybe make individual bindings here? for (u32 bindpoint = 0; bindpoint < static_cast<u32>(num_entries); ++bindpoint) { - VkDescriptorSetLayoutBinding& entry = bindings.emplace_back(); - entry.binding = binding++; - entry.descriptorType = descriptor_type; - entry.descriptorCount = 1; - entry.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT; - entry.pImmutableSamplers = nullptr; + bindings.push_back({ + .binding = binding++, + .descriptorType = descriptor_type, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, + .pImmutableSamplers = nullptr, + }); } }; add_bindings(VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER, entries.const_buffers.size()); add_bindings(VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, entries.global_buffers.size()); - add_bindings(VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER, entries.texel_buffers.size()); + add_bindings(VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER, entries.uniform_texels.size()); add_bindings(VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, entries.samplers.size()); + add_bindings(VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER, entries.storage_texels.size()); add_bindings(VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, entries.images.size()); - VkDescriptorSetLayoutCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.bindingCount = static_cast<u32>(bindings.size()); - ci.pBindings = bindings.data(); - return device.GetLogical().CreateDescriptorSetLayout(ci); + return device.GetLogical().CreateDescriptorSetLayout({ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .bindingCount = static_cast<u32>(bindings.size()), + .pBindings = bindings.data(), + }); } vk::PipelineLayout VKComputePipeline::CreatePipelineLayout() const { - VkPipelineLayoutCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.setLayoutCount = 1; - ci.pSetLayouts = descriptor_set_layout.address(); - ci.pushConstantRangeCount = 0; - ci.pPushConstantRanges = nullptr; - return device.GetLogical().CreatePipelineLayout(ci); + return device.GetLogical().CreatePipelineLayout({ + .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .setLayoutCount = 1, + .pSetLayouts = descriptor_set_layout.address(), + .pushConstantRangeCount = 0, + .pPushConstantRanges = nullptr, + }); } vk::DescriptorUpdateTemplateKHR VKComputePipeline::CreateDescriptorUpdateTemplate() const { @@ -90,57 +90,63 @@ vk::DescriptorUpdateTemplateKHR VKComputePipeline::CreateDescriptorUpdateTemplat return {}; } - VkDescriptorUpdateTemplateCreateInfoKHR ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR; - ci.pNext = nullptr; - ci.flags = 0; - ci.descriptorUpdateEntryCount = static_cast<u32>(template_entries.size()); - ci.pDescriptorUpdateEntries = template_entries.data(); - ci.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR; - ci.descriptorSetLayout = *descriptor_set_layout; - ci.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS; - ci.pipelineLayout = *layout; - ci.set = DESCRIPTOR_SET; - return device.GetLogical().CreateDescriptorUpdateTemplateKHR(ci); + return device.GetLogical().CreateDescriptorUpdateTemplateKHR({ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR, + .pNext = nullptr, + .flags = 0, + .descriptorUpdateEntryCount = static_cast<u32>(template_entries.size()), + .pDescriptorUpdateEntries = template_entries.data(), + .templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR, + .descriptorSetLayout = *descriptor_set_layout, + .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, + .pipelineLayout = *layout, + .set = DESCRIPTOR_SET, + }); } vk::ShaderModule VKComputePipeline::CreateShaderModule(const std::vector<u32>& code) const { - VkShaderModuleCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.codeSize = code.size() * sizeof(u32); - ci.pCode = code.data(); - return device.GetLogical().CreateShaderModule(ci); + device.SaveShader(code); + + return device.GetLogical().CreateShaderModule({ + .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .codeSize = code.size() * sizeof(u32), + .pCode = code.data(), + }); } vk::Pipeline VKComputePipeline::CreatePipeline() const { - VkComputePipelineCreateInfo ci; - VkPipelineShaderStageCreateInfo& stage_ci = ci.stage; - stage_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO; - stage_ci.pNext = nullptr; - stage_ci.flags = 0; - stage_ci.stage = VK_SHADER_STAGE_COMPUTE_BIT; - stage_ci.module = *shader_module; - stage_ci.pName = "main"; - stage_ci.pSpecializationInfo = nullptr; - - VkPipelineShaderStageRequiredSubgroupSizeCreateInfoEXT subgroup_size_ci; - subgroup_size_ci.sType = - VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_REQUIRED_SUBGROUP_SIZE_CREATE_INFO_EXT; - subgroup_size_ci.pNext = nullptr; - subgroup_size_ci.requiredSubgroupSize = GuestWarpSize; + + VkComputePipelineCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = + { + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = VK_SHADER_STAGE_COMPUTE_BIT, + .module = *shader_module, + .pName = "main", + .pSpecializationInfo = nullptr, + }, + .layout = *layout, + .basePipelineHandle = nullptr, + .basePipelineIndex = 0, + }; + + const VkPipelineShaderStageRequiredSubgroupSizeCreateInfoEXT subgroup_size_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_REQUIRED_SUBGROUP_SIZE_CREATE_INFO_EXT, + .pNext = nullptr, + .requiredSubgroupSize = GuestWarpSize, + }; if (entries.uses_warps && device.IsGuestWarpSizeSupported(VK_SHADER_STAGE_COMPUTE_BIT)) { - stage_ci.pNext = &subgroup_size_ci; + ci.stage.pNext = &subgroup_size_ci; } - ci.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.layout = *layout; - ci.basePipelineHandle = nullptr; - ci.basePipelineIndex = 0; return device.GetLogical().CreateComputePipeline(ci); } diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.h b/src/video_core/renderer_vulkan/vk_compute_pipeline.h index 33b9af29e..6e2f22a4a 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.h @@ -4,8 +4,6 @@ #pragma once -#include <memory> - #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" diff --git a/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp b/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp index e9d528aa6..f38e089d5 100644 --- a/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp @@ -2,13 +2,13 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <memory> #include <vector> #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" +#include "video_core/renderer_vulkan/vk_resource_pool.h" +#include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { @@ -16,14 +16,15 @@ namespace Vulkan { // Prefer small grow rates to avoid saturating the descriptor pool with barely used pipelines. constexpr std::size_t SETS_GROW_RATE = 0x20; -DescriptorAllocator::DescriptorAllocator(VKDescriptorPool& descriptor_pool, - VkDescriptorSetLayout layout) - : VKFencedPool{SETS_GROW_RATE}, descriptor_pool{descriptor_pool}, layout{layout} {} +DescriptorAllocator::DescriptorAllocator(VKDescriptorPool& descriptor_pool_, + VkDescriptorSetLayout layout_) + : ResourcePool(descriptor_pool_.master_semaphore, SETS_GROW_RATE), + descriptor_pool{descriptor_pool_}, layout{layout_} {} DescriptorAllocator::~DescriptorAllocator() = default; -VkDescriptorSet DescriptorAllocator::Commit(VKFence& fence) { - const std::size_t index = CommitResource(fence); +VkDescriptorSet DescriptorAllocator::Commit() { + const std::size_t index = CommitResource(); return descriptors_allocations[index / SETS_GROW_RATE][index % SETS_GROW_RATE]; } @@ -31,8 +32,9 @@ void DescriptorAllocator::Allocate(std::size_t begin, std::size_t end) { descriptors_allocations.push_back(descriptor_pool.AllocateDescriptors(layout, end - begin)); } -VKDescriptorPool::VKDescriptorPool(const VKDevice& device) - : device{device}, active_pool{AllocateNewPool()} {} +VKDescriptorPool::VKDescriptorPool(const VKDevice& device_, VKScheduler& scheduler) + : device{device_}, master_semaphore{scheduler.GetMasterSemaphore()}, active_pool{ + AllocateNewPool()} {} VKDescriptorPool::~VKDescriptorPool() = default; @@ -43,27 +45,31 @@ vk::DescriptorPool* VKDescriptorPool::AllocateNewPool() { {VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, num_sets * 60}, {VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER, num_sets * 64}, {VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, num_sets * 64}, - {VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, num_sets * 40}}; - - VkDescriptorPoolCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT; - ci.maxSets = num_sets; - ci.poolSizeCount = static_cast<u32>(std::size(pool_sizes)); - ci.pPoolSizes = std::data(pool_sizes); + {VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER, num_sets * 64}, + {VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, num_sets * 40}, + }; + + const VkDescriptorPoolCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO, + .pNext = nullptr, + .flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT, + .maxSets = num_sets, + .poolSizeCount = static_cast<u32>(std::size(pool_sizes)), + .pPoolSizes = std::data(pool_sizes), + }; return &pools.emplace_back(device.GetLogical().CreateDescriptorPool(ci)); } vk::DescriptorSets VKDescriptorPool::AllocateDescriptors(VkDescriptorSetLayout layout, std::size_t count) { const std::vector layout_copies(count, layout); - VkDescriptorSetAllocateInfo ai; - ai.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO; - ai.pNext = nullptr; - ai.descriptorPool = **active_pool; - ai.descriptorSetCount = static_cast<u32>(count); - ai.pSetLayouts = layout_copies.data(); + VkDescriptorSetAllocateInfo ai{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO, + .pNext = nullptr, + .descriptorPool = **active_pool, + .descriptorSetCount = static_cast<u32>(count), + .pSetLayouts = layout_copies.data(), + }; vk::DescriptorSets sets = active_pool->Allocate(ai); if (!sets.IsOutOfPoolMemory()) { diff --git a/src/video_core/renderer_vulkan/vk_descriptor_pool.h b/src/video_core/renderer_vulkan/vk_descriptor_pool.h index ab40c70f0..544f32a20 100644 --- a/src/video_core/renderer_vulkan/vk_descriptor_pool.h +++ b/src/video_core/renderer_vulkan/vk_descriptor_pool.h @@ -4,25 +4,26 @@ #pragma once -#include <memory> #include <vector> -#include "common/common_types.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" +#include "video_core/renderer_vulkan/vk_resource_pool.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { +class VKDevice; class VKDescriptorPool; +class VKScheduler; -class DescriptorAllocator final : public VKFencedPool { +class DescriptorAllocator final : public ResourcePool { public: explicit DescriptorAllocator(VKDescriptorPool& descriptor_pool, VkDescriptorSetLayout layout); ~DescriptorAllocator() override; + DescriptorAllocator& operator=(const DescriptorAllocator&) = delete; DescriptorAllocator(const DescriptorAllocator&) = delete; - VkDescriptorSet Commit(VKFence& fence); + VkDescriptorSet Commit(); protected: void Allocate(std::size_t begin, std::size_t end) override; @@ -38,15 +39,19 @@ class VKDescriptorPool final { friend DescriptorAllocator; public: - explicit VKDescriptorPool(const VKDevice& device); + explicit VKDescriptorPool(const VKDevice& device, VKScheduler& scheduler); ~VKDescriptorPool(); + VKDescriptorPool(const VKDescriptorPool&) = delete; + VKDescriptorPool& operator=(const VKDescriptorPool&) = delete; + private: vk::DescriptorPool* AllocateNewPool(); vk::DescriptorSets AllocateDescriptors(VkDescriptorSetLayout layout, std::size_t count); const VKDevice& device; + MasterSemaphore& master_semaphore; std::vector<vk::DescriptorPool> pools; vk::DescriptorPool* active_pool; diff --git a/src/video_core/renderer_vulkan/vk_device.cpp b/src/video_core/renderer_vulkan/vk_device.cpp index 52d29e49d..f34ed6735 100644 --- a/src/video_core/renderer_vulkan/vk_device.cpp +++ b/src/video_core/renderer_vulkan/vk_device.cpp @@ -4,11 +4,11 @@ #include <bitset> #include <chrono> -#include <cstdlib> #include <optional> #include <string_view> #include <thread> #include <unordered_set> +#include <utility> #include <vector> #include "common/assert.h" @@ -22,19 +22,30 @@ namespace { namespace Alternatives { -constexpr std::array Depth24UnormS8_UINT = {VK_FORMAT_D32_SFLOAT_S8_UINT, - VK_FORMAT_D16_UNORM_S8_UINT, VkFormat{}}; -constexpr std::array Depth16UnormS8_UINT = {VK_FORMAT_D24_UNORM_S8_UINT, - VK_FORMAT_D32_SFLOAT_S8_UINT, VkFormat{}}; +constexpr std::array Depth24UnormS8_UINT{ + VK_FORMAT_D32_SFLOAT_S8_UINT, + VK_FORMAT_D16_UNORM_S8_UINT, + VkFormat{}, +}; + +constexpr std::array Depth16UnormS8_UINT{ + VK_FORMAT_D24_UNORM_S8_UINT, + VK_FORMAT_D32_SFLOAT_S8_UINT, + VkFormat{}, +}; } // namespace Alternatives -constexpr std::array REQUIRED_EXTENSIONS = { +constexpr std::array REQUIRED_EXTENSIONS{ VK_KHR_SWAPCHAIN_EXTENSION_NAME, + VK_KHR_MAINTENANCE1_EXTENSION_NAME, + VK_KHR_STORAGE_BUFFER_STORAGE_CLASS_EXTENSION_NAME, + VK_KHR_SHADER_DRAW_PARAMETERS_EXTENSION_NAME, VK_KHR_16BIT_STORAGE_EXTENSION_NAME, VK_KHR_8BIT_STORAGE_EXTENSION_NAME, VK_KHR_DRIVER_PROPERTIES_EXTENSION_NAME, VK_KHR_DESCRIPTOR_UPDATE_TEMPLATE_EXTENSION_NAME, + VK_KHR_TIMELINE_SEMAPHORE_EXTENSION_NAME, VK_EXT_VERTEX_ATTRIBUTE_DIVISOR_EXTENSION_NAME, VK_EXT_SHADER_SUBGROUP_BALLOT_EXTENSION_NAME, VK_EXT_SHADER_SUBGROUP_VOTE_EXTENSION_NAME, @@ -71,76 +82,105 @@ VkFormatFeatureFlags GetFormatFeatures(VkFormatProperties properties, FormatType } } +[[nodiscard]] bool IsRDNA(std::string_view device_name, VkDriverIdKHR driver_id) { + static constexpr std::array RDNA_DEVICES{ + "5700", + "5600", + "5500", + "5300", + }; + if (driver_id != VK_DRIVER_ID_AMD_PROPRIETARY_KHR) { + return false; + } + return std::any_of(RDNA_DEVICES.begin(), RDNA_DEVICES.end(), [device_name](const char* name) { + return device_name.find(name) != std::string_view::npos; + }); +} + std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties( vk::PhysicalDevice physical, const vk::InstanceDispatch& dld) { - static constexpr std::array formats{VK_FORMAT_A8B8G8R8_UNORM_PACK32, - VK_FORMAT_A8B8G8R8_UINT_PACK32, - VK_FORMAT_A8B8G8R8_SNORM_PACK32, - VK_FORMAT_A8B8G8R8_SRGB_PACK32, - VK_FORMAT_B5G6R5_UNORM_PACK16, - VK_FORMAT_A2B10G10R10_UNORM_PACK32, - VK_FORMAT_A1R5G5B5_UNORM_PACK16, - VK_FORMAT_R32G32B32A32_SFLOAT, - VK_FORMAT_R32G32B32A32_UINT, - VK_FORMAT_R32G32_SFLOAT, - VK_FORMAT_R32G32_UINT, - VK_FORMAT_R16G16B16A16_UINT, - VK_FORMAT_R16G16B16A16_SNORM, - VK_FORMAT_R16G16B16A16_UNORM, - VK_FORMAT_R16G16_UNORM, - VK_FORMAT_R16G16_SNORM, - VK_FORMAT_R16G16_SFLOAT, - VK_FORMAT_R16_UNORM, - VK_FORMAT_R8G8B8A8_SRGB, - VK_FORMAT_R8G8_UNORM, - VK_FORMAT_R8G8_SNORM, - VK_FORMAT_R8_UNORM, - VK_FORMAT_R8_UINT, - VK_FORMAT_B10G11R11_UFLOAT_PACK32, - VK_FORMAT_R32_SFLOAT, - VK_FORMAT_R32_UINT, - VK_FORMAT_R32_SINT, - VK_FORMAT_R16_SFLOAT, - VK_FORMAT_R16G16B16A16_SFLOAT, - VK_FORMAT_B8G8R8A8_UNORM, - VK_FORMAT_R4G4B4A4_UNORM_PACK16, - VK_FORMAT_D32_SFLOAT, - VK_FORMAT_D16_UNORM, - VK_FORMAT_D16_UNORM_S8_UINT, - VK_FORMAT_D24_UNORM_S8_UINT, - VK_FORMAT_D32_SFLOAT_S8_UINT, - VK_FORMAT_BC1_RGBA_UNORM_BLOCK, - VK_FORMAT_BC2_UNORM_BLOCK, - VK_FORMAT_BC3_UNORM_BLOCK, - VK_FORMAT_BC4_UNORM_BLOCK, - VK_FORMAT_BC5_UNORM_BLOCK, - VK_FORMAT_BC5_SNORM_BLOCK, - VK_FORMAT_BC7_UNORM_BLOCK, - VK_FORMAT_BC6H_UFLOAT_BLOCK, - VK_FORMAT_BC6H_SFLOAT_BLOCK, - VK_FORMAT_BC1_RGBA_SRGB_BLOCK, - VK_FORMAT_BC2_SRGB_BLOCK, - VK_FORMAT_BC3_SRGB_BLOCK, - VK_FORMAT_BC7_SRGB_BLOCK, - VK_FORMAT_ASTC_4x4_SRGB_BLOCK, - VK_FORMAT_ASTC_8x8_SRGB_BLOCK, - VK_FORMAT_ASTC_8x5_SRGB_BLOCK, - VK_FORMAT_ASTC_5x4_SRGB_BLOCK, - VK_FORMAT_ASTC_5x5_UNORM_BLOCK, - VK_FORMAT_ASTC_5x5_SRGB_BLOCK, - VK_FORMAT_ASTC_10x8_UNORM_BLOCK, - VK_FORMAT_ASTC_10x8_SRGB_BLOCK, - VK_FORMAT_ASTC_6x6_UNORM_BLOCK, - VK_FORMAT_ASTC_6x6_SRGB_BLOCK, - VK_FORMAT_ASTC_10x10_UNORM_BLOCK, - VK_FORMAT_ASTC_10x10_SRGB_BLOCK, - VK_FORMAT_ASTC_12x12_UNORM_BLOCK, - VK_FORMAT_ASTC_12x12_SRGB_BLOCK, - VK_FORMAT_ASTC_8x6_UNORM_BLOCK, - VK_FORMAT_ASTC_8x6_SRGB_BLOCK, - VK_FORMAT_ASTC_6x5_UNORM_BLOCK, - VK_FORMAT_ASTC_6x5_SRGB_BLOCK, - VK_FORMAT_E5B9G9R9_UFLOAT_PACK32}; + static constexpr std::array formats{ + VK_FORMAT_A8B8G8R8_UNORM_PACK32, + VK_FORMAT_A8B8G8R8_UINT_PACK32, + VK_FORMAT_A8B8G8R8_SNORM_PACK32, + VK_FORMAT_A8B8G8R8_SINT_PACK32, + VK_FORMAT_A8B8G8R8_SRGB_PACK32, + VK_FORMAT_B5G6R5_UNORM_PACK16, + VK_FORMAT_A2B10G10R10_UNORM_PACK32, + VK_FORMAT_A2B10G10R10_UINT_PACK32, + VK_FORMAT_A1R5G5B5_UNORM_PACK16, + VK_FORMAT_R32G32B32A32_SFLOAT, + VK_FORMAT_R32G32B32A32_SINT, + VK_FORMAT_R32G32B32A32_UINT, + VK_FORMAT_R32G32_SFLOAT, + VK_FORMAT_R32G32_SINT, + VK_FORMAT_R32G32_UINT, + VK_FORMAT_R16G16B16A16_SINT, + VK_FORMAT_R16G16B16A16_UINT, + VK_FORMAT_R16G16B16A16_SNORM, + VK_FORMAT_R16G16B16A16_UNORM, + VK_FORMAT_R16G16_UNORM, + VK_FORMAT_R16G16_SNORM, + VK_FORMAT_R16G16_SFLOAT, + VK_FORMAT_R16_UNORM, + VK_FORMAT_R16_UINT, + VK_FORMAT_R8G8B8A8_SRGB, + VK_FORMAT_R8G8_UNORM, + VK_FORMAT_R8G8_SNORM, + VK_FORMAT_R8G8_SINT, + VK_FORMAT_R8G8_UINT, + VK_FORMAT_R8_UNORM, + VK_FORMAT_R8_SNORM, + VK_FORMAT_R8_SINT, + VK_FORMAT_R8_UINT, + VK_FORMAT_B10G11R11_UFLOAT_PACK32, + VK_FORMAT_R32_SFLOAT, + VK_FORMAT_R32_UINT, + VK_FORMAT_R32_SINT, + VK_FORMAT_R16_SFLOAT, + VK_FORMAT_R16G16B16A16_SFLOAT, + VK_FORMAT_B8G8R8A8_UNORM, + VK_FORMAT_B8G8R8A8_SRGB, + VK_FORMAT_R4G4B4A4_UNORM_PACK16, + VK_FORMAT_D32_SFLOAT, + VK_FORMAT_D16_UNORM, + VK_FORMAT_D16_UNORM_S8_UINT, + VK_FORMAT_D24_UNORM_S8_UINT, + VK_FORMAT_D32_SFLOAT_S8_UINT, + VK_FORMAT_BC1_RGBA_UNORM_BLOCK, + VK_FORMAT_BC2_UNORM_BLOCK, + VK_FORMAT_BC3_UNORM_BLOCK, + VK_FORMAT_BC4_UNORM_BLOCK, + VK_FORMAT_BC4_SNORM_BLOCK, + VK_FORMAT_BC5_UNORM_BLOCK, + VK_FORMAT_BC5_SNORM_BLOCK, + VK_FORMAT_BC7_UNORM_BLOCK, + VK_FORMAT_BC6H_UFLOAT_BLOCK, + VK_FORMAT_BC6H_SFLOAT_BLOCK, + VK_FORMAT_BC1_RGBA_SRGB_BLOCK, + VK_FORMAT_BC2_SRGB_BLOCK, + VK_FORMAT_BC3_SRGB_BLOCK, + VK_FORMAT_BC7_SRGB_BLOCK, + VK_FORMAT_ASTC_4x4_SRGB_BLOCK, + VK_FORMAT_ASTC_8x8_SRGB_BLOCK, + VK_FORMAT_ASTC_8x5_SRGB_BLOCK, + VK_FORMAT_ASTC_5x4_SRGB_BLOCK, + VK_FORMAT_ASTC_5x5_UNORM_BLOCK, + VK_FORMAT_ASTC_5x5_SRGB_BLOCK, + VK_FORMAT_ASTC_10x8_UNORM_BLOCK, + VK_FORMAT_ASTC_10x8_SRGB_BLOCK, + VK_FORMAT_ASTC_6x6_UNORM_BLOCK, + VK_FORMAT_ASTC_6x6_SRGB_BLOCK, + VK_FORMAT_ASTC_10x10_UNORM_BLOCK, + VK_FORMAT_ASTC_10x10_SRGB_BLOCK, + VK_FORMAT_ASTC_12x12_UNORM_BLOCK, + VK_FORMAT_ASTC_12x12_SRGB_BLOCK, + VK_FORMAT_ASTC_8x6_UNORM_BLOCK, + VK_FORMAT_ASTC_8x6_SRGB_BLOCK, + VK_FORMAT_ASTC_6x5_UNORM_BLOCK, + VK_FORMAT_ASTC_6x5_SRGB_BLOCK, + VK_FORMAT_E5B9G9R9_UFLOAT_PACK32, + }; std::unordered_map<VkFormat, VkFormatProperties> format_properties; for (const auto format : formats) { format_properties.emplace(format, physical.GetFormatProperties(format)); @@ -150,10 +190,10 @@ std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties( } // Anonymous namespace -VKDevice::VKDevice(VkInstance instance, vk::PhysicalDevice physical, VkSurfaceKHR surface, - const vk::InstanceDispatch& dld) - : dld{dld}, physical{physical}, properties{physical.GetProperties()}, - format_properties{GetFormatProperties(physical, dld)} { +VKDevice::VKDevice(VkInstance instance_, u32 instance_version_, vk::PhysicalDevice physical_, + VkSurfaceKHR surface, const vk::InstanceDispatch& dld_) + : dld{dld_}, physical{physical_}, properties{physical.GetProperties()}, + instance_version{instance_version_}, format_properties{GetFormatProperties(physical, dld)} { SetupFamilies(surface); SetupFeatures(); } @@ -164,107 +204,127 @@ bool VKDevice::Create() { const auto queue_cis = GetDeviceQueueCreateInfos(); const std::vector extensions = LoadExtensions(); - VkPhysicalDeviceFeatures2 features2; - features2.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2; - features2.pNext = nullptr; + VkPhysicalDeviceFeatures2 features2{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2, + .pNext = nullptr, + }; + const void* first_next = &features2; void** next = &features2.pNext; - auto& features = features2.features; - features.robustBufferAccess = false; - features.fullDrawIndexUint32 = false; - features.imageCubeArray = false; - features.independentBlend = true; - features.geometryShader = true; - features.tessellationShader = true; - features.sampleRateShading = false; - features.dualSrcBlend = false; - features.logicOp = false; - features.multiDrawIndirect = false; - features.drawIndirectFirstInstance = false; - features.depthClamp = true; - features.depthBiasClamp = true; - features.fillModeNonSolid = false; - features.depthBounds = false; - features.wideLines = false; - features.largePoints = true; - features.alphaToOne = false; - features.multiViewport = true; - features.samplerAnisotropy = true; - features.textureCompressionETC2 = false; - features.textureCompressionASTC_LDR = is_optimal_astc_supported; - features.textureCompressionBC = false; - features.occlusionQueryPrecise = true; - features.pipelineStatisticsQuery = false; - features.vertexPipelineStoresAndAtomics = true; - features.fragmentStoresAndAtomics = true; - features.shaderTessellationAndGeometryPointSize = false; - features.shaderImageGatherExtended = true; - features.shaderStorageImageExtendedFormats = false; - features.shaderStorageImageMultisample = false; - features.shaderStorageImageReadWithoutFormat = is_formatless_image_load_supported; - features.shaderStorageImageWriteWithoutFormat = true; - features.shaderUniformBufferArrayDynamicIndexing = false; - features.shaderSampledImageArrayDynamicIndexing = false; - features.shaderStorageBufferArrayDynamicIndexing = false; - features.shaderStorageImageArrayDynamicIndexing = false; - features.shaderClipDistance = false; - features.shaderCullDistance = false; - features.shaderFloat64 = false; - features.shaderInt64 = false; - features.shaderInt16 = false; - features.shaderResourceResidency = false; - features.shaderResourceMinLod = false; - features.sparseBinding = false; - features.sparseResidencyBuffer = false; - features.sparseResidencyImage2D = false; - features.sparseResidencyImage3D = false; - features.sparseResidency2Samples = false; - features.sparseResidency4Samples = false; - features.sparseResidency8Samples = false; - features.sparseResidency16Samples = false; - features.sparseResidencyAliased = false; - features.variableMultisampleRate = false; - features.inheritedQueries = false; - - VkPhysicalDevice16BitStorageFeaturesKHR bit16_storage; - bit16_storage.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_16BIT_STORAGE_FEATURES_KHR; - bit16_storage.pNext = nullptr; - bit16_storage.storageBuffer16BitAccess = false; - bit16_storage.uniformAndStorageBuffer16BitAccess = true; - bit16_storage.storagePushConstant16 = false; - bit16_storage.storageInputOutput16 = false; + features2.features = { + .robustBufferAccess = false, + .fullDrawIndexUint32 = false, + .imageCubeArray = false, + .independentBlend = true, + .geometryShader = true, + .tessellationShader = true, + .sampleRateShading = false, + .dualSrcBlend = false, + .logicOp = false, + .multiDrawIndirect = false, + .drawIndirectFirstInstance = false, + .depthClamp = true, + .depthBiasClamp = true, + .fillModeNonSolid = false, + .depthBounds = false, + .wideLines = false, + .largePoints = true, + .alphaToOne = false, + .multiViewport = true, + .samplerAnisotropy = true, + .textureCompressionETC2 = false, + .textureCompressionASTC_LDR = is_optimal_astc_supported, + .textureCompressionBC = false, + .occlusionQueryPrecise = true, + .pipelineStatisticsQuery = false, + .vertexPipelineStoresAndAtomics = true, + .fragmentStoresAndAtomics = true, + .shaderTessellationAndGeometryPointSize = false, + .shaderImageGatherExtended = true, + .shaderStorageImageExtendedFormats = false, + .shaderStorageImageMultisample = false, + .shaderStorageImageReadWithoutFormat = is_formatless_image_load_supported, + .shaderStorageImageWriteWithoutFormat = true, + .shaderUniformBufferArrayDynamicIndexing = false, + .shaderSampledImageArrayDynamicIndexing = false, + .shaderStorageBufferArrayDynamicIndexing = false, + .shaderStorageImageArrayDynamicIndexing = false, + .shaderClipDistance = false, + .shaderCullDistance = false, + .shaderFloat64 = false, + .shaderInt64 = false, + .shaderInt16 = false, + .shaderResourceResidency = false, + .shaderResourceMinLod = false, + .sparseBinding = false, + .sparseResidencyBuffer = false, + .sparseResidencyImage2D = false, + .sparseResidencyImage3D = false, + .sparseResidency2Samples = false, + .sparseResidency4Samples = false, + .sparseResidency8Samples = false, + .sparseResidency16Samples = false, + .sparseResidencyAliased = false, + .variableMultisampleRate = false, + .inheritedQueries = false, + }; + + VkPhysicalDeviceTimelineSemaphoreFeaturesKHR timeline_semaphore{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TIMELINE_SEMAPHORE_FEATURES_KHR, + .pNext = nullptr, + .timelineSemaphore = true, + }; + SetNext(next, timeline_semaphore); + + VkPhysicalDevice16BitStorageFeaturesKHR bit16_storage{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_16BIT_STORAGE_FEATURES_KHR, + .pNext = nullptr, + .storageBuffer16BitAccess = false, + .uniformAndStorageBuffer16BitAccess = true, + .storagePushConstant16 = false, + .storageInputOutput16 = false, + }; SetNext(next, bit16_storage); - VkPhysicalDevice8BitStorageFeaturesKHR bit8_storage; - bit8_storage.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_8BIT_STORAGE_FEATURES_KHR; - bit8_storage.pNext = nullptr; - bit8_storage.storageBuffer8BitAccess = false; - bit8_storage.uniformAndStorageBuffer8BitAccess = true; - bit8_storage.storagePushConstant8 = false; + VkPhysicalDevice8BitStorageFeaturesKHR bit8_storage{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_8BIT_STORAGE_FEATURES_KHR, + .pNext = nullptr, + .storageBuffer8BitAccess = false, + .uniformAndStorageBuffer8BitAccess = true, + .storagePushConstant8 = false, + }; SetNext(next, bit8_storage); - VkPhysicalDeviceHostQueryResetFeaturesEXT host_query_reset; - host_query_reset.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_HOST_QUERY_RESET_FEATURES_EXT; - host_query_reset.hostQueryReset = true; + VkPhysicalDeviceHostQueryResetFeaturesEXT host_query_reset{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_HOST_QUERY_RESET_FEATURES_EXT, + .hostQueryReset = true, + }; SetNext(next, host_query_reset); VkPhysicalDeviceFloat16Int8FeaturesKHR float16_int8; if (is_float16_supported) { - float16_int8.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FLOAT16_INT8_FEATURES_KHR; - float16_int8.pNext = nullptr; - float16_int8.shaderFloat16 = true; - float16_int8.shaderInt8 = false; + float16_int8 = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FLOAT16_INT8_FEATURES_KHR, + .pNext = nullptr, + .shaderFloat16 = true, + .shaderInt8 = false, + }; SetNext(next, float16_int8); } else { LOG_INFO(Render_Vulkan, "Device doesn't support float16 natively"); } + if (!nv_viewport_swizzle) { + LOG_INFO(Render_Vulkan, "Device doesn't support viewport swizzles"); + } + VkPhysicalDeviceUniformBufferStandardLayoutFeaturesKHR std430_layout; if (khr_uniform_buffer_standard_layout) { - std430_layout.sType = - VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_UNIFORM_BUFFER_STANDARD_LAYOUT_FEATURES_KHR; - std430_layout.pNext = nullptr; - std430_layout.uniformBufferStandardLayout = true; + std430_layout = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_UNIFORM_BUFFER_STANDARD_LAYOUT_FEATURES_KHR, + .pNext = nullptr, + .uniformBufferStandardLayout = true, + }; SetNext(next, std430_layout); } else { LOG_INFO(Render_Vulkan, "Device doesn't support packed UBOs"); @@ -272,9 +332,11 @@ bool VKDevice::Create() { VkPhysicalDeviceIndexTypeUint8FeaturesEXT index_type_uint8; if (ext_index_type_uint8) { - index_type_uint8.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_INDEX_TYPE_UINT8_FEATURES_EXT; - index_type_uint8.pNext = nullptr; - index_type_uint8.indexTypeUint8 = true; + index_type_uint8 = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_INDEX_TYPE_UINT8_FEATURES_EXT, + .pNext = nullptr, + .indexTypeUint8 = true, + }; SetNext(next, index_type_uint8); } else { LOG_INFO(Render_Vulkan, "Device doesn't support uint8 indexes"); @@ -282,21 +344,61 @@ bool VKDevice::Create() { VkPhysicalDeviceTransformFeedbackFeaturesEXT transform_feedback; if (ext_transform_feedback) { - transform_feedback.sType = - VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_FEATURES_EXT; - transform_feedback.pNext = nullptr; - transform_feedback.transformFeedback = true; - transform_feedback.geometryStreams = true; + transform_feedback = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_FEATURES_EXT, + .pNext = nullptr, + .transformFeedback = true, + .geometryStreams = true, + }; SetNext(next, transform_feedback); } else { LOG_INFO(Render_Vulkan, "Device doesn't support transform feedbacks"); } + VkPhysicalDeviceCustomBorderColorFeaturesEXT custom_border; + if (ext_custom_border_color) { + custom_border = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_CUSTOM_BORDER_COLOR_FEATURES_EXT, + .pNext = nullptr, + .customBorderColors = VK_TRUE, + .customBorderColorWithoutFormat = VK_TRUE, + }; + SetNext(next, custom_border); + } else { + LOG_INFO(Render_Vulkan, "Device doesn't support custom border colors"); + } + + VkPhysicalDeviceExtendedDynamicStateFeaturesEXT dynamic_state; + if (ext_extended_dynamic_state) { + dynamic_state = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTENDED_DYNAMIC_STATE_FEATURES_EXT, + .pNext = nullptr, + .extendedDynamicState = VK_TRUE, + }; + SetNext(next, dynamic_state); + } else { + LOG_INFO(Render_Vulkan, "Device doesn't support extended dynamic state"); + } + if (!ext_depth_range_unrestricted) { LOG_INFO(Render_Vulkan, "Device doesn't support depth range unrestricted"); } - logical = vk::Device::Create(physical, queue_cis, extensions, features2, dld); + VkDeviceDiagnosticsConfigCreateInfoNV diagnostics_nv; + if (nv_device_diagnostics_config) { + nsight_aftermath_tracker.Initialize(); + + diagnostics_nv = { + .sType = VK_STRUCTURE_TYPE_DEVICE_DIAGNOSTICS_CONFIG_CREATE_INFO_NV, + .pNext = &features2, + .flags = VK_DEVICE_DIAGNOSTICS_CONFIG_ENABLE_SHADER_DEBUG_INFO_BIT_NV | + VK_DEVICE_DIAGNOSTICS_CONFIG_ENABLE_RESOURCE_TRACKING_BIT_NV | + VK_DEVICE_DIAGNOSTICS_CONFIG_ENABLE_AUTOMATIC_CHECKPOINTS_BIT_NV, + }; + first_next = &diagnostics_nv; + } + + logical = vk::Device::Create(physical, queue_cis, extensions, first_next, dld); if (!logical) { LOG_ERROR(Render_Vulkan, "Failed to create logical device"); return false; @@ -304,8 +406,19 @@ bool VKDevice::Create() { CollectTelemetryParameters(); + if (ext_extended_dynamic_state && IsRDNA(properties.deviceName, driver_id)) { + // AMD's proprietary driver supports VK_EXT_extended_dynamic_state but on RDNA devices it + // seems to cause stability issues + LOG_WARNING( + Render_Vulkan, + "Blacklisting AMD proprietary on RDNA devices from VK_EXT_extended_dynamic_state"); + ext_extended_dynamic_state = false; + } + graphics_queue = logical.GetQueue(graphics_family); present_queue = logical.GetQueue(present_family); + + use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue(); return true; } @@ -344,17 +457,12 @@ VkFormat VKDevice::GetSupportedFormat(VkFormat wanted_format, VkFormatFeatureFla void VKDevice::ReportLoss() const { LOG_CRITICAL(Render_Vulkan, "Device loss occured!"); - // Wait some time to let the log flush - std::this_thread::sleep_for(std::chrono::seconds{1}); - - if (!nv_device_diagnostic_checkpoints) { - return; - } + // Wait for the log to flush and for Nsight Aftermath to dump the results + std::this_thread::sleep_for(std::chrono::seconds{3}); +} - [[maybe_unused]] const std::vector data = graphics_queue.GetCheckpointDataNV(dld); - // Catch here in debug builds (or with optimizations disabled) the last graphics pipeline to be - // executed. It can be done on a debugger by evaluating the expression: - // *(VKGraphicsPipeline*)data[0] +void VKDevice::SaveShader(const std::vector<u32>& spirv) const { + nsight_aftermath_tracker.SaveShader(spirv); } bool VKDevice::IsOptimalAstcSupported(const VkPhysicalDeviceFeatures& features) const { @@ -492,43 +600,44 @@ bool VKDevice::IsSuitable(vk::PhysicalDevice physical, VkSurfaceKHR surface) { std::vector<const char*> VKDevice::LoadExtensions() { std::vector<const char*> extensions; - const auto Test = [&](const VkExtensionProperties& extension, - std::optional<std::reference_wrapper<bool>> status, const char* name, - bool push) { - if (extension.extensionName != std::string_view(name)) { - return; - } - if (push) { - extensions.push_back(name); - } - if (status) { - status->get() = true; - } - }; - extensions.reserve(7 + REQUIRED_EXTENSIONS.size()); extensions.insert(extensions.begin(), REQUIRED_EXTENSIONS.begin(), REQUIRED_EXTENSIONS.end()); bool has_khr_shader_float16_int8{}; bool has_ext_subgroup_size_control{}; bool has_ext_transform_feedback{}; - for (const auto& extension : physical.EnumerateDeviceExtensionProperties()) { - Test(extension, khr_uniform_buffer_standard_layout, + bool has_ext_custom_border_color{}; + bool has_ext_extended_dynamic_state{}; + for (const VkExtensionProperties& extension : physical.EnumerateDeviceExtensionProperties()) { + const auto test = [&](std::optional<std::reference_wrapper<bool>> status, const char* name, + bool push) { + if (extension.extensionName != std::string_view(name)) { + return; + } + if (push) { + extensions.push_back(name); + } + if (status) { + status->get() = true; + } + }; + test(nv_viewport_swizzle, VK_NV_VIEWPORT_SWIZZLE_EXTENSION_NAME, true); + test(khr_uniform_buffer_standard_layout, VK_KHR_UNIFORM_BUFFER_STANDARD_LAYOUT_EXTENSION_NAME, true); - Test(extension, has_khr_shader_float16_int8, VK_KHR_SHADER_FLOAT16_INT8_EXTENSION_NAME, - false); - Test(extension, ext_depth_range_unrestricted, - VK_EXT_DEPTH_RANGE_UNRESTRICTED_EXTENSION_NAME, true); - Test(extension, ext_index_type_uint8, VK_EXT_INDEX_TYPE_UINT8_EXTENSION_NAME, true); - Test(extension, ext_shader_viewport_index_layer, - VK_EXT_SHADER_VIEWPORT_INDEX_LAYER_EXTENSION_NAME, true); - Test(extension, has_ext_subgroup_size_control, VK_EXT_SUBGROUP_SIZE_CONTROL_EXTENSION_NAME, - false); - Test(extension, has_ext_transform_feedback, VK_EXT_TRANSFORM_FEEDBACK_EXTENSION_NAME, - false); + test(has_khr_shader_float16_int8, VK_KHR_SHADER_FLOAT16_INT8_EXTENSION_NAME, false); + test(ext_depth_range_unrestricted, VK_EXT_DEPTH_RANGE_UNRESTRICTED_EXTENSION_NAME, true); + test(ext_index_type_uint8, VK_EXT_INDEX_TYPE_UINT8_EXTENSION_NAME, true); + test(ext_shader_viewport_index_layer, VK_EXT_SHADER_VIEWPORT_INDEX_LAYER_EXTENSION_NAME, + true); + test(has_ext_transform_feedback, VK_EXT_TRANSFORM_FEEDBACK_EXTENSION_NAME, false); + test(has_ext_custom_border_color, VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME, false); + test(has_ext_extended_dynamic_state, VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME, false); + if (instance_version >= VK_API_VERSION_1_1) { + test(has_ext_subgroup_size_control, VK_EXT_SUBGROUP_SIZE_CONTROL_EXTENSION_NAME, false); + } if (Settings::values.renderer_debug) { - Test(extension, nv_device_diagnostic_checkpoints, - VK_NV_DEVICE_DIAGNOSTIC_CHECKPOINTS_EXTENSION_NAME, true); + test(nv_device_diagnostics_config, VK_NV_DEVICE_DIAGNOSTICS_CONFIG_EXTENSION_NAME, + true); } } @@ -598,6 +707,32 @@ std::vector<const char*> VKDevice::LoadExtensions() { } } + if (has_ext_custom_border_color) { + VkPhysicalDeviceCustomBorderColorFeaturesEXT border_features; + border_features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_CUSTOM_BORDER_COLOR_FEATURES_EXT; + border_features.pNext = nullptr; + features.pNext = &border_features; + physical.GetFeatures2KHR(features); + + if (border_features.customBorderColors && border_features.customBorderColorWithoutFormat) { + extensions.push_back(VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME); + ext_custom_border_color = true; + } + } + + if (has_ext_extended_dynamic_state) { + VkPhysicalDeviceExtendedDynamicStateFeaturesEXT dynamic_state; + dynamic_state.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTENDED_DYNAMIC_STATE_FEATURES_EXT; + dynamic_state.pNext = nullptr; + features.pNext = &dynamic_state; + physical.GetFeatures2KHR(features); + + if (dynamic_state.extendedDynamicState) { + extensions.push_back(VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME); + ext_extended_dynamic_state = true; + } + } + return extensions; } @@ -633,14 +768,21 @@ void VKDevice::SetupFeatures() { } void VKDevice::CollectTelemetryParameters() { - VkPhysicalDeviceDriverPropertiesKHR driver; - driver.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DRIVER_PROPERTIES_KHR; - driver.pNext = nullptr; + VkPhysicalDeviceDriverPropertiesKHR driver{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DRIVER_PROPERTIES_KHR, + .pNext = nullptr, + .driverID = {}, + .driverName = {}, + .driverInfo = {}, + .conformanceVersion = {}, + }; - VkPhysicalDeviceProperties2KHR properties; - properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR; - properties.pNext = &driver; - physical.GetProperties2KHR(properties); + VkPhysicalDeviceProperties2KHR device_properties{ + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR, + .pNext = &driver, + .properties = {}, + }; + physical.GetProperties2KHR(device_properties); driver_id = driver.driverID; vendor_name = driver.driverName; @@ -648,23 +790,26 @@ void VKDevice::CollectTelemetryParameters() { const std::vector extensions = physical.EnumerateDeviceExtensionProperties(); reported_extensions.reserve(std::size(extensions)); for (const auto& extension : extensions) { - reported_extensions.push_back(extension.extensionName); + reported_extensions.emplace_back(extension.extensionName); } } std::vector<VkDeviceQueueCreateInfo> VKDevice::GetDeviceQueueCreateInfos() const { static constexpr float QUEUE_PRIORITY = 1.0f; - std::unordered_set<u32> unique_queue_families = {graphics_family, present_family}; + std::unordered_set<u32> unique_queue_families{graphics_family, present_family}; std::vector<VkDeviceQueueCreateInfo> queue_cis; + queue_cis.reserve(unique_queue_families.size()); for (const u32 queue_family : unique_queue_families) { - VkDeviceQueueCreateInfo& ci = queue_cis.emplace_back(); - ci.sType = VK_STRUCTURE_TYPE_DEVICE_QUEUE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.queueFamilyIndex = queue_family; - ci.queueCount = 1; + auto& ci = queue_cis.emplace_back(VkDeviceQueueCreateInfo{ + .sType = VK_STRUCTURE_TYPE_DEVICE_QUEUE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .queueFamilyIndex = queue_family, + .queueCount = 1, + .pQueuePriorities = nullptr, + }); ci.pQueuePriorities = &QUEUE_PRIORITY; } diff --git a/src/video_core/renderer_vulkan/vk_device.h b/src/video_core/renderer_vulkan/vk_device.h index 60d64572a..4286673d9 100644 --- a/src/video_core/renderer_vulkan/vk_device.h +++ b/src/video_core/renderer_vulkan/vk_device.h @@ -10,6 +10,7 @@ #include <vector> #include "common/common_types.h" +#include "video_core/renderer_vulkan/nsight_aftermath_tracker.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { @@ -23,8 +24,8 @@ const u32 GuestWarpSize = 32; /// Handles data specific to a physical device. class VKDevice final { public: - explicit VKDevice(VkInstance instance, vk::PhysicalDevice physical, VkSurfaceKHR surface, - const vk::InstanceDispatch& dld); + explicit VKDevice(VkInstance instance, u32 instance_version, vk::PhysicalDevice physical, + VkSurfaceKHR surface, const vk::InstanceDispatch& dld); ~VKDevice(); /// Initializes the device. Returns true on success. @@ -43,6 +44,9 @@ public: /// Reports a device loss. void ReportLoss() const; + /// Reports a shader to Nsight Aftermath. + void SaveShader(const std::vector<u32>& spirv) const; + /// Returns the dispatch loader with direct function pointers of the device. const vk::DeviceDispatch& GetDispatchLoader() const { return dld; @@ -78,13 +82,13 @@ public: return present_family; } - /// Returns true if the device is integrated with the host CPU. - bool IsIntegrated() const { - return properties.deviceType == VK_PHYSICAL_DEVICE_TYPE_INTEGRATED_GPU; + /// Returns the current instance Vulkan API version in Vulkan-formatted version numbers. + u32 InstanceApiVersion() const { + return instance_version; } /// Returns the current Vulkan API version provided in Vulkan-formatted version numbers. - u32 GetApiVersion() const { + u32 ApiVersion() const { return properties.apiVersion; } @@ -123,6 +127,11 @@ public: return properties.limits.maxPushConstantsSize; } + /// Returns the maximum size for shared memory. + u32 GetMaxComputeSharedMemorySize() const { + return properties.limits.maxComputeSharedMemorySize; + } + /// Returns true if ASTC is natively supported. bool IsOptimalAstcSupported() const { return is_optimal_astc_supported; @@ -148,6 +157,11 @@ public: return is_formatless_image_load_supported; } + /// Returns true if the device supports VK_NV_viewport_swizzle. + bool IsNvViewportSwizzleSupported() const { + return nv_viewport_swizzle; + } + /// Returns true if the device supports VK_EXT_scalar_block_layout. bool IsKhrUniformBufferStandardLayoutSupported() const { return khr_uniform_buffer_standard_layout; @@ -173,9 +187,14 @@ public: return ext_transform_feedback; } - /// Returns true if the device supports VK_NV_device_diagnostic_checkpoints. - bool IsNvDeviceDiagnosticCheckpoints() const { - return nv_device_diagnostic_checkpoints; + /// Returns true if the device supports VK_EXT_custom_border_color. + bool IsExtCustomBorderColorSupported() const { + return ext_custom_border_color; + } + + /// Returns true if the device supports VK_EXT_extended_dynamic_state. + bool IsExtExtendedDynamicStateSupported() const { + return ext_extended_dynamic_state; } /// Returns the vendor name reported from Vulkan. @@ -188,6 +207,11 @@ public: return reported_extensions; } + /// Returns true if the setting for async shader compilation is enabled. + bool UseAsynchronousShaders() const { + return use_asynchronous_shaders; + } + /// Checks if the physical device is suitable. static bool IsSuitable(vk::PhysicalDevice physical, VkSurfaceKHR surface); @@ -220,6 +244,7 @@ private: vk::Device logical; ///< Logical device. vk::Queue graphics_queue; ///< Main graphics queue. vk::Queue present_queue; ///< Main present queue. + u32 instance_version{}; ///< Vulkan onstance version. u32 graphics_family{}; ///< Main graphics queue family index. u32 present_family{}; ///< Main present queue family index. VkDriverIdKHR driver_id{}; ///< Driver ID. @@ -228,12 +253,18 @@ private: bool is_float16_supported{}; ///< Support for float16 arithmetics. bool is_warp_potentially_bigger{}; ///< Host warp size can be bigger than guest. bool is_formatless_image_load_supported{}; ///< Support for shader image read without format. + bool nv_viewport_swizzle{}; ///< Support for VK_NV_viewport_swizzle. bool khr_uniform_buffer_standard_layout{}; ///< Support for std430 on UBOs. bool ext_index_type_uint8{}; ///< Support for VK_EXT_index_type_uint8. bool ext_depth_range_unrestricted{}; ///< Support for VK_EXT_depth_range_unrestricted. bool ext_shader_viewport_index_layer{}; ///< Support for VK_EXT_shader_viewport_index_layer. bool ext_transform_feedback{}; ///< Support for VK_EXT_transform_feedback. - bool nv_device_diagnostic_checkpoints{}; ///< Support for VK_NV_device_diagnostic_checkpoints. + bool ext_custom_border_color{}; ///< Support for VK_EXT_custom_border_color. + bool ext_extended_dynamic_state{}; ///< Support for VK_EXT_extended_dynamic_state. + bool nv_device_diagnostics_config{}; ///< Support for VK_NV_device_diagnostics_config. + + // Asynchronous Graphics Pipeline setting + bool use_asynchronous_shaders{}; ///< Setting to use asynchronous shaders/graphics pipeline // Telemetry parameters std::string vendor_name; ///< Device's driver name. @@ -241,6 +272,9 @@ private: /// Format properties dictionary. std::unordered_map<VkFormat, VkFormatProperties> format_properties; + + /// Nsight Aftermath GPU crash tracker + NsightAftermathTracker nsight_aftermath_tracker; }; } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.cpp b/src/video_core/renderer_vulkan/vk_fence_manager.cpp new file mode 100644 index 000000000..5babbdd0b --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_fence_manager.cpp @@ -0,0 +1,101 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <memory> +#include <thread> + +#include "video_core/renderer_vulkan/vk_buffer_cache.h" +#include "video_core/renderer_vulkan/vk_device.h" +#include "video_core/renderer_vulkan/vk_fence_manager.h" +#include "video_core/renderer_vulkan/vk_scheduler.h" +#include "video_core/renderer_vulkan/vk_texture_cache.h" +#include "video_core/renderer_vulkan/wrapper.h" + +namespace Vulkan { + +InnerFence::InnerFence(const VKDevice& device, VKScheduler& scheduler, u32 payload, bool is_stubbed) + : VideoCommon::FenceBase(payload, is_stubbed), device{device}, scheduler{scheduler} {} + +InnerFence::InnerFence(const VKDevice& device, VKScheduler& scheduler, GPUVAddr address, + u32 payload, bool is_stubbed) + : VideoCommon::FenceBase(address, payload, is_stubbed), device{device}, scheduler{scheduler} {} + +InnerFence::~InnerFence() = default; + +void InnerFence::Queue() { + if (is_stubbed) { + return; + } + ASSERT(!event); + + event = device.GetLogical().CreateEvent(); + ticks = scheduler.CurrentTick(); + + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([event = *event](vk::CommandBuffer cmdbuf) { + cmdbuf.SetEvent(event, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT); + }); +} + +bool InnerFence::IsSignaled() const { + if (is_stubbed) { + return true; + } + ASSERT(event); + return IsEventSignalled(); +} + +void InnerFence::Wait() { + if (is_stubbed) { + return; + } + ASSERT(event); + + if (ticks >= scheduler.CurrentTick()) { + scheduler.Flush(); + } + while (!IsEventSignalled()) { + std::this_thread::yield(); + } +} + +bool InnerFence::IsEventSignalled() const { + switch (const VkResult result = event.GetStatus()) { + case VK_EVENT_SET: + return true; + case VK_EVENT_RESET: + return false; + default: + throw vk::Exception(result); + } +} + +VKFenceManager::VKFenceManager(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, + Tegra::MemoryManager& memory_manager, VKTextureCache& texture_cache, + VKBufferCache& buffer_cache, VKQueryCache& query_cache, + const VKDevice& device_, VKScheduler& scheduler_) + : GenericFenceManager(rasterizer, gpu, texture_cache, buffer_cache, query_cache), + device{device_}, scheduler{scheduler_} {} + +Fence VKFenceManager::CreateFence(u32 value, bool is_stubbed) { + return std::make_shared<InnerFence>(device, scheduler, value, is_stubbed); +} + +Fence VKFenceManager::CreateFence(GPUVAddr addr, u32 value, bool is_stubbed) { + return std::make_shared<InnerFence>(device, scheduler, addr, value, is_stubbed); +} + +void VKFenceManager::QueueFence(Fence& fence) { + fence->Queue(); +} + +bool VKFenceManager::IsFenceSignaled(Fence& fence) const { + return fence->IsSignaled(); +} + +void VKFenceManager::WaitFence(Fence& fence) { + fence->Wait(); +} + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.h b/src/video_core/renderer_vulkan/vk_fence_manager.h new file mode 100644 index 000000000..1547d6d30 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_fence_manager.h @@ -0,0 +1,75 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> + +#include "video_core/fence_manager.h" +#include "video_core/renderer_vulkan/vk_buffer_cache.h" +#include "video_core/renderer_vulkan/wrapper.h" + +namespace Core { +class System; +} + +namespace VideoCore { +class RasterizerInterface; +} + +namespace Vulkan { + +class VKBufferCache; +class VKDevice; +class VKQueryCache; +class VKScheduler; +class VKTextureCache; + +class InnerFence : public VideoCommon::FenceBase { +public: + explicit InnerFence(const VKDevice& device, VKScheduler& scheduler, u32 payload, + bool is_stubbed); + explicit InnerFence(const VKDevice& device, VKScheduler& scheduler, GPUVAddr address, + u32 payload, bool is_stubbed); + ~InnerFence(); + + void Queue(); + + bool IsSignaled() const; + + void Wait(); + +private: + bool IsEventSignalled() const; + + const VKDevice& device; + VKScheduler& scheduler; + vk::Event event; + u64 ticks = 0; +}; +using Fence = std::shared_ptr<InnerFence>; + +using GenericFenceManager = + VideoCommon::FenceManager<Fence, VKTextureCache, VKBufferCache, VKQueryCache>; + +class VKFenceManager final : public GenericFenceManager { +public: + explicit VKFenceManager(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, + Tegra::MemoryManager& memory_manager, VKTextureCache& texture_cache, + VKBufferCache& buffer_cache, VKQueryCache& query_cache, + const VKDevice& device, VKScheduler& scheduler); + +protected: + Fence CreateFence(u32 value, bool is_stubbed) override; + Fence CreateFence(GPUVAddr addr, u32 value, bool is_stubbed) override; + void QueueFence(Fence& fence) override; + bool IsFenceSignaled(Fence& fence) const override; + void WaitFence(Fence& fence) override; + +private: + const VKDevice& device; + VKScheduler& scheduler; +}; + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index b540b838d..0e8f9c352 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -2,11 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> #include <array> #include <cstring> #include <vector> -#include "common/assert.h" #include "common/common_types.h" #include "common/microprofile.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" @@ -26,16 +26,17 @@ MICROPROFILE_DECLARE(Vulkan_PipelineCache); namespace { -VkStencilOpState GetStencilFaceState(const FixedPipelineState::StencilFace& face) { - VkStencilOpState state; - state.failOp = MaxwellToVK::StencilOp(face.action_stencil_fail); - state.passOp = MaxwellToVK::StencilOp(face.action_depth_pass); - state.depthFailOp = MaxwellToVK::StencilOp(face.action_depth_fail); - state.compareOp = MaxwellToVK::ComparisonOp(face.test_func); - state.compareMask = 0; - state.writeMask = 0; - state.reference = 0; - return state; +template <class StencilFace> +VkStencilOpState GetStencilFaceState(const StencilFace& face) { + return { + .failOp = MaxwellToVK::StencilOp(face.ActionStencilFail()), + .passOp = MaxwellToVK::StencilOp(face.ActionDepthPass()), + .depthFailOp = MaxwellToVK::StencilOp(face.ActionDepthFail()), + .compareOp = MaxwellToVK::ComparisonOp(face.TestFunc()), + .compareMask = 0, + .writeMask = 0, + .reference = 0, + }; } bool SupportsPrimitiveRestart(VkPrimitiveTopology topology) { @@ -50,6 +51,24 @@ bool SupportsPrimitiveRestart(VkPrimitiveTopology topology) { topology) == std::end(unsupported_topologies); } +VkViewportSwizzleNV UnpackViewportSwizzle(u16 swizzle) { + union Swizzle { + u32 raw; + BitField<0, 3, Maxwell::ViewportSwizzle> x; + BitField<4, 3, Maxwell::ViewportSwizzle> y; + BitField<8, 3, Maxwell::ViewportSwizzle> z; + BitField<12, 3, Maxwell::ViewportSwizzle> w; + }; + const Swizzle unpacked{swizzle}; + + return { + .x = MaxwellToVK::ViewportSwizzle(unpacked.x), + .y = MaxwellToVK::ViewportSwizzle(unpacked.y), + .z = MaxwellToVK::ViewportSwizzle(unpacked.z), + .w = MaxwellToVK::ViewportSwizzle(unpacked.w), + }; +} + } // Anonymous namespace VKGraphicsPipeline::VKGraphicsPipeline(const VKDevice& device, VKScheduler& scheduler, @@ -59,15 +78,14 @@ VKGraphicsPipeline::VKGraphicsPipeline(const VKDevice& device, VKScheduler& sche const GraphicsPipelineCacheKey& key, vk::Span<VkDescriptorSetLayoutBinding> bindings, const SPIRVProgram& program) - : device{device}, scheduler{scheduler}, fixed_state{key.fixed_state}, hash{key.Hash()}, + : device{device}, scheduler{scheduler}, cache_key{key}, hash{cache_key.Hash()}, descriptor_set_layout{CreateDescriptorSetLayout(bindings)}, descriptor_allocator{descriptor_pool, *descriptor_set_layout}, update_descriptor_queue{update_descriptor_queue}, layout{CreatePipelineLayout()}, descriptor_template{CreateDescriptorUpdateTemplate(program)}, modules{CreateShaderModules( program)}, - renderpass{renderpass_cache.GetRenderPass(key.renderpass_params)}, pipeline{CreatePipeline( - key.renderpass_params, - program)} {} + renderpass{renderpass_cache.GetRenderPass(cache_key.renderpass_params)}, + pipeline{CreatePipeline(cache_key.renderpass_params, program)} {} VKGraphicsPipeline::~VKGraphicsPipeline() = default; @@ -75,31 +93,33 @@ VkDescriptorSet VKGraphicsPipeline::CommitDescriptorSet() { if (!descriptor_template) { return {}; } - const auto set = descriptor_allocator.Commit(scheduler.GetFence()); + const VkDescriptorSet set = descriptor_allocator.Commit(); update_descriptor_queue.Send(*descriptor_template, set); return set; } vk::DescriptorSetLayout VKGraphicsPipeline::CreateDescriptorSetLayout( vk::Span<VkDescriptorSetLayoutBinding> bindings) const { - VkDescriptorSetLayoutCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.bindingCount = bindings.size(); - ci.pBindings = bindings.data(); + const VkDescriptorSetLayoutCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .bindingCount = bindings.size(), + .pBindings = bindings.data(), + }; return device.GetLogical().CreateDescriptorSetLayout(ci); } vk::PipelineLayout VKGraphicsPipeline::CreatePipelineLayout() const { - VkPipelineLayoutCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.setLayoutCount = 1; - ci.pSetLayouts = descriptor_set_layout.address(); - ci.pushConstantRangeCount = 0; - ci.pPushConstantRanges = nullptr; + const VkPipelineLayoutCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .setLayoutCount = 1, + .pSetLayouts = descriptor_set_layout.address(), + .pushConstantRangeCount = 0, + .pPushConstantRanges = nullptr, + }; return device.GetLogical().CreatePipelineLayout(ci); } @@ -118,26 +138,29 @@ vk::DescriptorUpdateTemplateKHR VKGraphicsPipeline::CreateDescriptorUpdateTempla return {}; } - VkDescriptorUpdateTemplateCreateInfoKHR ci; - ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR; - ci.pNext = nullptr; - ci.flags = 0; - ci.descriptorUpdateEntryCount = static_cast<u32>(template_entries.size()); - ci.pDescriptorUpdateEntries = template_entries.data(); - ci.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR; - ci.descriptorSetLayout = *descriptor_set_layout; - ci.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS; - ci.pipelineLayout = *layout; - ci.set = DESCRIPTOR_SET; + const VkDescriptorUpdateTemplateCreateInfoKHR ci{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR, + .pNext = nullptr, + .flags = 0, + .descriptorUpdateEntryCount = static_cast<u32>(template_entries.size()), + .pDescriptorUpdateEntries = template_entries.data(), + .templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR, + .descriptorSetLayout = *descriptor_set_layout, + .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, + .pipelineLayout = *layout, + .set = DESCRIPTOR_SET, + }; return device.GetLogical().CreateDescriptorUpdateTemplateKHR(ci); } std::vector<vk::ShaderModule> VKGraphicsPipeline::CreateShaderModules( const SPIRVProgram& program) const { - VkShaderModuleCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; + VkShaderModuleCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .codeSize = 0, + }; std::vector<vk::ShaderModule> modules; modules.reserve(Maxwell::MaxShaderStage); @@ -147,6 +170,8 @@ std::vector<vk::ShaderModule> VKGraphicsPipeline::CreateShaderModules( continue; } + device.SaveShader(stage->code); + ci.codeSize = stage->code.size() * sizeof(u32); ci.pCode = stage->code.data(); modules.push_back(device.GetLogical().CreateShaderModule(ci)); @@ -156,186 +181,251 @@ std::vector<vk::ShaderModule> VKGraphicsPipeline::CreateShaderModules( vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpass_params, const SPIRVProgram& program) const { - const auto& vi = fixed_state.vertex_input; - const auto& ia = fixed_state.input_assembly; - const auto& ds = fixed_state.depth_stencil; - const auto& cd = fixed_state.color_blending; - const auto& ts = fixed_state.tessellation; - const auto& rs = fixed_state.rasterizer; + const auto& state = cache_key.fixed_state; + const auto& viewport_swizzles = state.viewport_swizzles; + + FixedPipelineState::DynamicState dynamic; + if (device.IsExtExtendedDynamicStateSupported()) { + // Insert dummy values, as long as they are valid they don't matter as extended dynamic + // state is ignored + dynamic.raw1 = 0; + dynamic.raw2 = 0; + for (FixedPipelineState::VertexBinding& binding : dynamic.vertex_bindings) { + // Enable all vertex bindings + binding.raw = 0; + binding.enabled.Assign(1); + } + } else { + dynamic = state.dynamic_state; + } std::vector<VkVertexInputBindingDescription> vertex_bindings; std::vector<VkVertexInputBindingDivisorDescriptionEXT> vertex_binding_divisors; - for (std::size_t i = 0; i < vi.num_bindings; ++i) { - const auto& binding = vi.bindings[i]; - const bool instanced = binding.divisor != 0; + for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { + const auto& binding = dynamic.vertex_bindings[index]; + if (!binding.enabled) { + continue; + } + const bool instanced = state.binding_divisors[index] != 0; const auto rate = instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX; - auto& vertex_binding = vertex_bindings.emplace_back(); - vertex_binding.binding = binding.index; - vertex_binding.stride = binding.stride; - vertex_binding.inputRate = rate; + vertex_bindings.push_back({ + .binding = static_cast<u32>(index), + .stride = binding.stride, + .inputRate = rate, + }); if (instanced) { - auto& binding_divisor = vertex_binding_divisors.emplace_back(); - binding_divisor.binding = binding.index; - binding_divisor.divisor = binding.divisor; + vertex_binding_divisors.push_back({ + .binding = static_cast<u32>(index), + .divisor = state.binding_divisors[index], + }); } } std::vector<VkVertexInputAttributeDescription> vertex_attributes; const auto& input_attributes = program[0]->entries.attributes; - for (std::size_t i = 0; i < vi.num_attributes; ++i) { - const auto& attribute = vi.attributes[i]; - if (input_attributes.find(attribute.index) == input_attributes.end()) { + for (std::size_t index = 0; index < state.attributes.size(); ++index) { + const auto& attribute = state.attributes[index]; + if (!attribute.enabled) { + continue; + } + if (input_attributes.find(static_cast<u32>(index)) == input_attributes.end()) { // Skip attributes not used by the vertex shaders. continue; } - auto& vertex_attribute = vertex_attributes.emplace_back(); - vertex_attribute.location = attribute.index; - vertex_attribute.binding = attribute.buffer; - vertex_attribute.format = MaxwellToVK::VertexFormat(attribute.type, attribute.size); - vertex_attribute.offset = attribute.offset; + vertex_attributes.push_back({ + .location = static_cast<u32>(index), + .binding = attribute.buffer, + .format = MaxwellToVK::VertexFormat(attribute.Type(), attribute.Size()), + .offset = attribute.offset, + }); } - VkPipelineVertexInputStateCreateInfo vertex_input_ci; - vertex_input_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO; - vertex_input_ci.pNext = nullptr; - vertex_input_ci.flags = 0; - vertex_input_ci.vertexBindingDescriptionCount = static_cast<u32>(vertex_bindings.size()); - vertex_input_ci.pVertexBindingDescriptions = vertex_bindings.data(); - vertex_input_ci.vertexAttributeDescriptionCount = static_cast<u32>(vertex_attributes.size()); - vertex_input_ci.pVertexAttributeDescriptions = vertex_attributes.data(); - - VkPipelineVertexInputDivisorStateCreateInfoEXT input_divisor_ci; - input_divisor_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_DIVISOR_STATE_CREATE_INFO_EXT; - input_divisor_ci.pNext = nullptr; - input_divisor_ci.vertexBindingDivisorCount = static_cast<u32>(vertex_binding_divisors.size()); - input_divisor_ci.pVertexBindingDivisors = vertex_binding_divisors.data(); + VkPipelineVertexInputStateCreateInfo vertex_input_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .vertexBindingDescriptionCount = static_cast<u32>(vertex_bindings.size()), + .pVertexBindingDescriptions = vertex_bindings.data(), + .vertexAttributeDescriptionCount = static_cast<u32>(vertex_attributes.size()), + .pVertexAttributeDescriptions = vertex_attributes.data(), + }; + + const VkPipelineVertexInputDivisorStateCreateInfoEXT input_divisor_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_DIVISOR_STATE_CREATE_INFO_EXT, + .pNext = nullptr, + .vertexBindingDivisorCount = static_cast<u32>(vertex_binding_divisors.size()), + .pVertexBindingDivisors = vertex_binding_divisors.data(), + }; if (!vertex_binding_divisors.empty()) { vertex_input_ci.pNext = &input_divisor_ci; } - VkPipelineInputAssemblyStateCreateInfo input_assembly_ci; - input_assembly_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO; - input_assembly_ci.pNext = nullptr; - input_assembly_ci.flags = 0; - input_assembly_ci.topology = MaxwellToVK::PrimitiveTopology(device, ia.topology); - input_assembly_ci.primitiveRestartEnable = - ia.primitive_restart_enable && SupportsPrimitiveRestart(input_assembly_ci.topology); - - VkPipelineTessellationStateCreateInfo tessellation_ci; - tessellation_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_TESSELLATION_STATE_CREATE_INFO; - tessellation_ci.pNext = nullptr; - tessellation_ci.flags = 0; - tessellation_ci.patchControlPoints = ts.patch_control_points; - - VkPipelineViewportStateCreateInfo viewport_ci; - viewport_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO; - viewport_ci.pNext = nullptr; - viewport_ci.flags = 0; - viewport_ci.viewportCount = Maxwell::NumViewports; - viewport_ci.pViewports = nullptr; - viewport_ci.scissorCount = Maxwell::NumViewports; - viewport_ci.pScissors = nullptr; - - VkPipelineRasterizationStateCreateInfo rasterization_ci; - rasterization_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO; - rasterization_ci.pNext = nullptr; - rasterization_ci.flags = 0; - rasterization_ci.depthClampEnable = rs.depth_clamp_enable; - rasterization_ci.rasterizerDiscardEnable = VK_FALSE; - rasterization_ci.polygonMode = VK_POLYGON_MODE_FILL; - rasterization_ci.cullMode = - rs.cull_enable ? MaxwellToVK::CullFace(rs.cull_face) : VK_CULL_MODE_NONE; - rasterization_ci.frontFace = MaxwellToVK::FrontFace(rs.front_face); - rasterization_ci.depthBiasEnable = rs.depth_bias_enable; - rasterization_ci.depthBiasConstantFactor = 0.0f; - rasterization_ci.depthBiasClamp = 0.0f; - rasterization_ci.depthBiasSlopeFactor = 0.0f; - rasterization_ci.lineWidth = 1.0f; - - VkPipelineMultisampleStateCreateInfo multisample_ci; - multisample_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO; - multisample_ci.pNext = nullptr; - multisample_ci.flags = 0; - multisample_ci.rasterizationSamples = VK_SAMPLE_COUNT_1_BIT; - multisample_ci.sampleShadingEnable = VK_FALSE; - multisample_ci.minSampleShading = 0.0f; - multisample_ci.pSampleMask = nullptr; - multisample_ci.alphaToCoverageEnable = VK_FALSE; - multisample_ci.alphaToOneEnable = VK_FALSE; - - VkPipelineDepthStencilStateCreateInfo depth_stencil_ci; - depth_stencil_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO; - depth_stencil_ci.pNext = nullptr; - depth_stencil_ci.flags = 0; - depth_stencil_ci.depthTestEnable = ds.depth_test_enable; - depth_stencil_ci.depthWriteEnable = ds.depth_write_enable; - depth_stencil_ci.depthCompareOp = ds.depth_test_enable - ? MaxwellToVK::ComparisonOp(ds.depth_test_function) - : VK_COMPARE_OP_ALWAYS; - depth_stencil_ci.depthBoundsTestEnable = ds.depth_bounds_enable; - depth_stencil_ci.stencilTestEnable = ds.stencil_enable; - depth_stencil_ci.front = GetStencilFaceState(ds.front_stencil); - depth_stencil_ci.back = GetStencilFaceState(ds.back_stencil); - depth_stencil_ci.minDepthBounds = 0.0f; - depth_stencil_ci.maxDepthBounds = 0.0f; + const auto input_assembly_topology = MaxwellToVK::PrimitiveTopology(device, state.topology); + const VkPipelineInputAssemblyStateCreateInfo input_assembly_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .topology = MaxwellToVK::PrimitiveTopology(device, state.topology), + .primitiveRestartEnable = state.primitive_restart_enable != 0 && + SupportsPrimitiveRestart(input_assembly_topology), + }; + + const VkPipelineTessellationStateCreateInfo tessellation_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_TESSELLATION_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .patchControlPoints = state.patch_control_points_minus_one.Value() + 1, + }; + + VkPipelineViewportStateCreateInfo viewport_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .viewportCount = Maxwell::NumViewports, + .pViewports = nullptr, + .scissorCount = Maxwell::NumViewports, + .pScissors = nullptr, + }; + + std::array<VkViewportSwizzleNV, Maxwell::NumViewports> swizzles; + std::transform(viewport_swizzles.begin(), viewport_swizzles.end(), swizzles.begin(), + UnpackViewportSwizzle); + VkPipelineViewportSwizzleStateCreateInfoNV swizzle_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_SWIZZLE_STATE_CREATE_INFO_NV, + .pNext = nullptr, + .flags = 0, + .viewportCount = Maxwell::NumViewports, + .pViewportSwizzles = swizzles.data(), + }; + if (device.IsNvViewportSwizzleSupported()) { + viewport_ci.pNext = &swizzle_ci; + } + + const VkPipelineRasterizationStateCreateInfo rasterization_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .depthClampEnable = + static_cast<VkBool32>(state.depth_clamp_disabled == 0 ? VK_TRUE : VK_FALSE), + .rasterizerDiscardEnable = + static_cast<VkBool32>(state.rasterize_enable == 0 ? VK_TRUE : VK_FALSE), + .polygonMode = VK_POLYGON_MODE_FILL, + .cullMode = + dynamic.cull_enable ? MaxwellToVK::CullFace(dynamic.CullFace()) : VK_CULL_MODE_NONE, + .frontFace = MaxwellToVK::FrontFace(dynamic.FrontFace()), + .depthBiasEnable = state.depth_bias_enable, + .depthBiasConstantFactor = 0.0f, + .depthBiasClamp = 0.0f, + .depthBiasSlopeFactor = 0.0f, + .lineWidth = 1.0f, + }; + + const VkPipelineMultisampleStateCreateInfo multisample_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .rasterizationSamples = VK_SAMPLE_COUNT_1_BIT, + .sampleShadingEnable = VK_FALSE, + .minSampleShading = 0.0f, + .pSampleMask = nullptr, + .alphaToCoverageEnable = VK_FALSE, + .alphaToOneEnable = VK_FALSE, + }; + + const VkPipelineDepthStencilStateCreateInfo depth_stencil_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .depthTestEnable = dynamic.depth_test_enable, + .depthWriteEnable = dynamic.depth_write_enable, + .depthCompareOp = dynamic.depth_test_enable + ? MaxwellToVK::ComparisonOp(dynamic.DepthTestFunc()) + : VK_COMPARE_OP_ALWAYS, + .depthBoundsTestEnable = dynamic.depth_bounds_enable, + .stencilTestEnable = dynamic.stencil_enable, + .front = GetStencilFaceState(dynamic.front), + .back = GetStencilFaceState(dynamic.back), + .minDepthBounds = 0.0f, + .maxDepthBounds = 0.0f, + }; std::array<VkPipelineColorBlendAttachmentState, Maxwell::NumRenderTargets> cb_attachments; - const std::size_t num_attachments = - std::min(cd.attachments_count, renderpass_params.color_attachments.size()); - for (std::size_t i = 0; i < num_attachments; ++i) { - static constexpr std::array component_table = { - VK_COLOR_COMPONENT_R_BIT, VK_COLOR_COMPONENT_G_BIT, VK_COLOR_COMPONENT_B_BIT, - VK_COLOR_COMPONENT_A_BIT}; - const auto& blend = cd.attachments[i]; + const auto num_attachments = static_cast<std::size_t>(renderpass_params.num_color_attachments); + for (std::size_t index = 0; index < num_attachments; ++index) { + static constexpr std::array COMPONENT_TABLE{ + VK_COLOR_COMPONENT_R_BIT, + VK_COLOR_COMPONENT_G_BIT, + VK_COLOR_COMPONENT_B_BIT, + VK_COLOR_COMPONENT_A_BIT, + }; + const auto& blend = state.attachments[index]; VkColorComponentFlags color_components = 0; - for (std::size_t j = 0; j < component_table.size(); ++j) { - if (blend.components[j]) { - color_components |= component_table[j]; + for (std::size_t i = 0; i < COMPONENT_TABLE.size(); ++i) { + if (blend.Mask()[i]) { + color_components |= COMPONENT_TABLE[i]; } } - VkPipelineColorBlendAttachmentState& attachment = cb_attachments[i]; - attachment.blendEnable = blend.enable; - attachment.srcColorBlendFactor = MaxwellToVK::BlendFactor(blend.src_rgb_func); - attachment.dstColorBlendFactor = MaxwellToVK::BlendFactor(blend.dst_rgb_func); - attachment.colorBlendOp = MaxwellToVK::BlendEquation(blend.rgb_equation); - attachment.srcAlphaBlendFactor = MaxwellToVK::BlendFactor(blend.src_a_func); - attachment.dstAlphaBlendFactor = MaxwellToVK::BlendFactor(blend.dst_a_func); - attachment.alphaBlendOp = MaxwellToVK::BlendEquation(blend.a_equation); - attachment.colorWriteMask = color_components; + cb_attachments[index] = { + .blendEnable = blend.enable != 0, + .srcColorBlendFactor = MaxwellToVK::BlendFactor(blend.SourceRGBFactor()), + .dstColorBlendFactor = MaxwellToVK::BlendFactor(blend.DestRGBFactor()), + .colorBlendOp = MaxwellToVK::BlendEquation(blend.EquationRGB()), + .srcAlphaBlendFactor = MaxwellToVK::BlendFactor(blend.SourceAlphaFactor()), + .dstAlphaBlendFactor = MaxwellToVK::BlendFactor(blend.DestAlphaFactor()), + .alphaBlendOp = MaxwellToVK::BlendEquation(blend.EquationAlpha()), + .colorWriteMask = color_components, + }; } - VkPipelineColorBlendStateCreateInfo color_blend_ci; - color_blend_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO; - color_blend_ci.pNext = nullptr; - color_blend_ci.flags = 0; - color_blend_ci.logicOpEnable = VK_FALSE; - color_blend_ci.logicOp = VK_LOGIC_OP_COPY; - color_blend_ci.attachmentCount = static_cast<u32>(num_attachments); - color_blend_ci.pAttachments = cb_attachments.data(); - std::memset(color_blend_ci.blendConstants, 0, sizeof(color_blend_ci.blendConstants)); - - static constexpr std::array dynamic_states = { + const VkPipelineColorBlendStateCreateInfo color_blend_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .logicOpEnable = VK_FALSE, + .logicOp = VK_LOGIC_OP_COPY, + .attachmentCount = static_cast<u32>(num_attachments), + .pAttachments = cb_attachments.data(), + .blendConstants = {}, + }; + + std::vector dynamic_states{ VK_DYNAMIC_STATE_VIEWPORT, VK_DYNAMIC_STATE_SCISSOR, VK_DYNAMIC_STATE_DEPTH_BIAS, VK_DYNAMIC_STATE_BLEND_CONSTANTS, VK_DYNAMIC_STATE_DEPTH_BOUNDS, VK_DYNAMIC_STATE_STENCIL_COMPARE_MASK, - VK_DYNAMIC_STATE_STENCIL_WRITE_MASK, VK_DYNAMIC_STATE_STENCIL_REFERENCE}; + VK_DYNAMIC_STATE_STENCIL_WRITE_MASK, VK_DYNAMIC_STATE_STENCIL_REFERENCE, + }; + if (device.IsExtExtendedDynamicStateSupported()) { + static constexpr std::array extended{ + VK_DYNAMIC_STATE_CULL_MODE_EXT, + VK_DYNAMIC_STATE_FRONT_FACE_EXT, + VK_DYNAMIC_STATE_VERTEX_INPUT_BINDING_STRIDE_EXT, + VK_DYNAMIC_STATE_DEPTH_TEST_ENABLE_EXT, + VK_DYNAMIC_STATE_DEPTH_WRITE_ENABLE_EXT, + VK_DYNAMIC_STATE_DEPTH_COMPARE_OP_EXT, + VK_DYNAMIC_STATE_DEPTH_BOUNDS_TEST_ENABLE_EXT, + VK_DYNAMIC_STATE_STENCIL_TEST_ENABLE_EXT, + VK_DYNAMIC_STATE_STENCIL_OP_EXT, + }; + dynamic_states.insert(dynamic_states.end(), extended.begin(), extended.end()); + } - VkPipelineDynamicStateCreateInfo dynamic_state_ci; - dynamic_state_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO; - dynamic_state_ci.pNext = nullptr; - dynamic_state_ci.flags = 0; - dynamic_state_ci.dynamicStateCount = static_cast<u32>(dynamic_states.size()); - dynamic_state_ci.pDynamicStates = dynamic_states.data(); + const VkPipelineDynamicStateCreateInfo dynamic_state_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .dynamicStateCount = static_cast<u32>(dynamic_states.size()), + .pDynamicStates = dynamic_states.data(), + }; - VkPipelineShaderStageRequiredSubgroupSizeCreateInfoEXT subgroup_size_ci; - subgroup_size_ci.sType = - VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_REQUIRED_SUBGROUP_SIZE_CREATE_INFO_EXT; - subgroup_size_ci.pNext = nullptr; - subgroup_size_ci.requiredSubgroupSize = GuestWarpSize; + const VkPipelineShaderStageRequiredSubgroupSizeCreateInfoEXT subgroup_size_ci{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_REQUIRED_SUBGROUP_SIZE_CREATE_INFO_EXT, + .pNext = nullptr, + .requiredSubgroupSize = GuestWarpSize, + }; std::vector<VkPipelineShaderStageCreateInfo> shader_stages; std::size_t module_index = 0; @@ -343,6 +433,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa if (!program[stage]) { continue; } + VkPipelineShaderStageCreateInfo& stage_ci = shader_stages.emplace_back(); stage_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO; stage_ci.pNext = nullptr; @@ -357,26 +448,27 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa } } - VkGraphicsPipelineCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.stageCount = static_cast<u32>(shader_stages.size()); - ci.pStages = shader_stages.data(); - ci.pVertexInputState = &vertex_input_ci; - ci.pInputAssemblyState = &input_assembly_ci; - ci.pTessellationState = &tessellation_ci; - ci.pViewportState = &viewport_ci; - ci.pRasterizationState = &rasterization_ci; - ci.pMultisampleState = &multisample_ci; - ci.pDepthStencilState = &depth_stencil_ci; - ci.pColorBlendState = &color_blend_ci; - ci.pDynamicState = &dynamic_state_ci; - ci.layout = *layout; - ci.renderPass = renderpass; - ci.subpass = 0; - ci.basePipelineHandle = nullptr; - ci.basePipelineIndex = 0; + const VkGraphicsPipelineCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(shader_stages.size()), + .pStages = shader_stages.data(), + .pVertexInputState = &vertex_input_ci, + .pInputAssemblyState = &input_assembly_ci, + .pTessellationState = &tessellation_ci, + .pViewportState = &viewport_ci, + .pRasterizationState = &rasterization_ci, + .pMultisampleState = &multisample_ci, + .pDepthStencilState = &depth_stencil_ci, + .pColorBlendState = &color_blend_ci, + .pDynamicState = &dynamic_state_ci, + .layout = *layout, + .renderPass = renderpass, + .subpass = 0, + .basePipelineHandle = nullptr, + .basePipelineIndex = 0, + }; return device.GetLogical().CreateGraphicsPipeline(ci); } diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h index 7aba70960..58aa35efd 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h @@ -5,16 +5,13 @@ #pragma once #include <array> -#include <memory> #include <optional> -#include <unordered_map> #include <vector> #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -22,7 +19,27 @@ namespace Vulkan { using Maxwell = Tegra::Engines::Maxwell3D::Regs; -struct GraphicsPipelineCacheKey; +struct GraphicsPipelineCacheKey { + RenderPassParams renderpass_params; + u32 padding; + std::array<GPUVAddr, Maxwell::MaxShaderProgram> shaders; + FixedPipelineState fixed_state; + + std::size_t Hash() const noexcept; + + bool operator==(const GraphicsPipelineCacheKey& rhs) const noexcept; + + bool operator!=(const GraphicsPipelineCacheKey& rhs) const noexcept { + return !operator==(rhs); + } + + std::size_t Size() const noexcept { + return sizeof(renderpass_params) + sizeof(padding) + sizeof(shaders) + fixed_state.Size(); + } +}; +static_assert(std::has_unique_object_representations_v<GraphicsPipelineCacheKey>); +static_assert(std::is_trivially_copyable_v<GraphicsPipelineCacheKey>); +static_assert(std::is_trivially_constructible_v<GraphicsPipelineCacheKey>); class VKDescriptorPool; class VKDevice; @@ -57,6 +74,10 @@ public: return renderpass; } + GraphicsPipelineCacheKey GetCacheKey() const { + return cache_key; + } + private: vk::DescriptorSetLayout CreateDescriptorSetLayout( vk::Span<VkDescriptorSetLayoutBinding> bindings) const; @@ -73,7 +94,7 @@ private: const VKDevice& device; VKScheduler& scheduler; - const FixedPipelineState fixed_state; + const GraphicsPipelineCacheKey cache_key; const u64 hash; vk::DescriptorSetLayout descriptor_set_layout; diff --git a/src/video_core/renderer_vulkan/vk_image.cpp b/src/video_core/renderer_vulkan/vk_image.cpp index 9bceb3861..1c418ea17 100644 --- a/src/video_core/renderer_vulkan/vk_image.cpp +++ b/src/video_core/renderer_vulkan/vk_image.cpp @@ -102,21 +102,29 @@ bool VKImage::HasChanged(u32 base_layer, u32 num_layers, u32 base_level, u32 num void VKImage::CreatePresentView() { // Image type has to be 2D to be presented. - VkImageViewCreateInfo image_view_ci; - image_view_ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO; - image_view_ci.pNext = nullptr; - image_view_ci.flags = 0; - image_view_ci.image = *image; - image_view_ci.viewType = VK_IMAGE_VIEW_TYPE_2D; - image_view_ci.format = format; - image_view_ci.components = {VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY, - VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY}; - image_view_ci.subresourceRange.aspectMask = aspect_mask; - image_view_ci.subresourceRange.baseMipLevel = 0; - image_view_ci.subresourceRange.levelCount = 1; - image_view_ci.subresourceRange.baseArrayLayer = 0; - image_view_ci.subresourceRange.layerCount = 1; - present_view = device.GetLogical().CreateImageView(image_view_ci); + present_view = device.GetLogical().CreateImageView({ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = *image, + .viewType = VK_IMAGE_VIEW_TYPE_2D, + .format = format, + .components = + { + .r = VK_COMPONENT_SWIZZLE_IDENTITY, + .g = VK_COMPONENT_SWIZZLE_IDENTITY, + .b = VK_COMPONENT_SWIZZLE_IDENTITY, + .a = VK_COMPONENT_SWIZZLE_IDENTITY, + }, + .subresourceRange = + { + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = 1, + .baseArrayLayer = 0, + .layerCount = 1, + }, + }); } VKImage::SubrangeState& VKImage::GetSubrangeState(u32 layer, u32 level) noexcept { diff --git a/src/video_core/renderer_vulkan/vk_master_semaphore.cpp b/src/video_core/renderer_vulkan/vk_master_semaphore.cpp new file mode 100644 index 000000000..ae26e558d --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_master_semaphore.cpp @@ -0,0 +1,56 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <atomic> +#include <chrono> + +#include "core/settings.h" +#include "video_core/renderer_vulkan/vk_device.h" +#include "video_core/renderer_vulkan/vk_master_semaphore.h" +#include "video_core/renderer_vulkan/wrapper.h" + +namespace Vulkan { + +using namespace std::chrono_literals; + +MasterSemaphore::MasterSemaphore(const VKDevice& device) { + static constexpr VkSemaphoreTypeCreateInfoKHR semaphore_type_ci{ + .sType = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO_KHR, + .pNext = nullptr, + .semaphoreType = VK_SEMAPHORE_TYPE_TIMELINE_KHR, + .initialValue = 0, + }; + static constexpr VkSemaphoreCreateInfo semaphore_ci{ + .sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO, + .pNext = &semaphore_type_ci, + .flags = 0, + }; + semaphore = device.GetLogical().CreateSemaphore(semaphore_ci); + + if (!Settings::values.renderer_debug) { + return; + } + // Validation layers have a bug where they fail to track resource usage when using timeline + // semaphores and synchronizing with GetSemaphoreCounterValueKHR. To workaround this issue, have + // a separate thread waiting for each timeline semaphore value. + debug_thread = std::thread([this] { + u64 counter = 0; + while (!shutdown) { + if (semaphore.Wait(counter, 10'000'000)) { + ++counter; + } + } + }); +} + +MasterSemaphore::~MasterSemaphore() { + shutdown = true; + + // This thread might not be started + if (debug_thread.joinable()) { + debug_thread.join(); + } +} + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_master_semaphore.h b/src/video_core/renderer_vulkan/vk_master_semaphore.h new file mode 100644 index 000000000..0e93706d7 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_master_semaphore.h @@ -0,0 +1,70 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <atomic> +#include <thread> + +#include "common/common_types.h" +#include "video_core/renderer_vulkan/wrapper.h" + +namespace Vulkan { + +class VKDevice; + +class MasterSemaphore { +public: + explicit MasterSemaphore(const VKDevice& device); + ~MasterSemaphore(); + + /// Returns the current logical tick. + [[nodiscard]] u64 CurrentTick() const noexcept { + return current_tick; + } + + /// Returns the timeline semaphore handle. + [[nodiscard]] VkSemaphore Handle() const noexcept { + return *semaphore; + } + + /// Returns true when a tick has been hit by the GPU. + [[nodiscard]] bool IsFree(u64 tick) { + return gpu_tick >= tick; + } + + /// Advance to the logical tick. + void NextTick() noexcept { + ++current_tick; + } + + /// Refresh the known GPU tick + void Refresh() { + gpu_tick = semaphore.GetCounter(); + } + + /// Waits for a tick to be hit on the GPU + void Wait(u64 tick) { + // No need to wait if the GPU is ahead of the tick + if (IsFree(tick)) { + return; + } + // Update the GPU tick and try again + Refresh(); + if (IsFree(tick)) { + return; + } + // If none of the above is hit, fallback to a regular wait + semaphore.Wait(tick); + } + +private: + vk::Semaphore semaphore; ///< Timeline semaphore. + std::atomic<u64> gpu_tick{0}; ///< Current known GPU tick. + std::atomic<u64> current_tick{1}; ///< Current logical tick. + std::atomic<bool> shutdown{false}; ///< True when the object is being destroyed. + std::thread debug_thread; ///< Debug thread to workaround validation layer bugs. +}; + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.cpp b/src/video_core/renderer_vulkan/vk_memory_manager.cpp index 6a9e658bf..24c8960ac 100644 --- a/src/video_core/renderer_vulkan/vk_memory_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_memory_manager.cpp @@ -118,8 +118,7 @@ private: }; VKMemoryManager::VKMemoryManager(const VKDevice& device) - : device{device}, properties{device.GetPhysical().GetMemoryProperties()}, - is_memory_unified{GetMemoryUnified(properties)} {} + : device{device}, properties{device.GetPhysical().GetMemoryProperties()} {} VKMemoryManager::~VKMemoryManager() = default; @@ -179,13 +178,12 @@ bool VKMemoryManager::AllocMemory(VkMemoryPropertyFlags wanted_properties, u32 t }(); // Try to allocate found type. - VkMemoryAllocateInfo memory_ai; - memory_ai.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO; - memory_ai.pNext = nullptr; - memory_ai.allocationSize = size; - memory_ai.memoryTypeIndex = type; - - vk::DeviceMemory memory = device.GetLogical().TryAllocateMemory(memory_ai); + vk::DeviceMemory memory = device.GetLogical().TryAllocateMemory({ + .sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO, + .pNext = nullptr, + .allocationSize = size, + .memoryTypeIndex = type, + }); if (!memory) { LOG_CRITICAL(Render_Vulkan, "Device allocation failed!"); return false; @@ -209,16 +207,6 @@ VKMemoryCommit VKMemoryManager::TryAllocCommit(const VkMemoryRequirements& requi return {}; } -bool VKMemoryManager::GetMemoryUnified(const VkPhysicalDeviceMemoryProperties& properties) { - for (u32 heap_index = 0; heap_index < properties.memoryHeapCount; ++heap_index) { - if (!(properties.memoryHeaps[heap_index].flags & VK_MEMORY_HEAP_DEVICE_LOCAL_BIT)) { - // Memory is considered unified when heaps are device local only. - return false; - } - } - return true; -} - VKMemoryCommitImpl::VKMemoryCommitImpl(const VKDevice& device, VKMemoryAllocation* allocation, const vk::DeviceMemory& memory, u64 begin, u64 end) : device{device}, memory{memory}, interval{begin, end}, allocation{allocation} {} diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.h b/src/video_core/renderer_vulkan/vk_memory_manager.h index 35ee54d30..1af88e3d4 100644 --- a/src/video_core/renderer_vulkan/vk_memory_manager.h +++ b/src/video_core/renderer_vulkan/vk_memory_manager.h @@ -32,7 +32,7 @@ public: * memory. When passing false, it will try to allocate device local memory. * @returns A memory commit. */ - VKMemoryCommit Commit(const VkMemoryRequirements& reqs, bool host_visible); + VKMemoryCommit Commit(const VkMemoryRequirements& requirements, bool host_visible); /// Commits memory required by the buffer and binds it. VKMemoryCommit Commit(const vk::Buffer& buffer, bool host_visible); @@ -40,11 +40,6 @@ public: /// Commits memory required by the image and binds it. VKMemoryCommit Commit(const vk::Image& image, bool host_visible); - /// Returns true if the memory allocations are done always in host visible and coherent memory. - bool IsMemoryUnified() const { - return is_memory_unified; - } - private: /// Allocates a chunk of memory. bool AllocMemory(VkMemoryPropertyFlags wanted_properties, u32 type_mask, u64 size); @@ -53,12 +48,8 @@ private: VKMemoryCommit TryAllocCommit(const VkMemoryRequirements& requirements, VkMemoryPropertyFlags wanted_properties); - /// Returns true if the device uses an unified memory model. - static bool GetMemoryUnified(const VkPhysicalDeviceMemoryProperties& properties); - - const VKDevice& device; ///< Device handler. - const VkPhysicalDeviceMemoryProperties properties; ///< Physical device properties. - const bool is_memory_unified; ///< True if memory model is unified. + const VKDevice& device; ///< Device handler. + const VkPhysicalDeviceMemoryProperties properties; ///< Physical device properties. std::vector<std::unique_ptr<VKMemoryAllocation>> allocations; ///< Current allocations. }; diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index 90e3a8edd..dedc9c466 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -22,17 +22,24 @@ #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/shader/compiler_settings.h" +#include "video_core/shader/memory_util.h" +#include "video_core/shader_cache.h" +#include "video_core/shader_notify.h" namespace Vulkan { MICROPROFILE_DECLARE(Vulkan_PipelineCache); using Tegra::Engines::ShaderType; +using VideoCommon::Shader::GetShaderAddress; +using VideoCommon::Shader::GetShaderCode; +using VideoCommon::Shader::KERNEL_MAIN_OFFSET; +using VideoCommon::Shader::ProgramCode; +using VideoCommon::Shader::STAGE_MAIN_OFFSET; namespace { @@ -40,65 +47,12 @@ constexpr VkDescriptorType UNIFORM_BUFFER = VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER; constexpr VkDescriptorType STORAGE_BUFFER = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER; constexpr VkDescriptorType UNIFORM_TEXEL_BUFFER = VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER; constexpr VkDescriptorType COMBINED_IMAGE_SAMPLER = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER; +constexpr VkDescriptorType STORAGE_TEXEL_BUFFER = VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER; constexpr VkDescriptorType STORAGE_IMAGE = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE; constexpr VideoCommon::Shader::CompilerSettings compiler_settings{ VideoCommon::Shader::CompileDepth::FullDecompile}; -/// Gets the address for the specified shader stage program -GPUVAddr GetShaderAddress(Core::System& system, Maxwell::ShaderProgram program) { - const auto& gpu{system.GPU().Maxwell3D()}; - const auto& shader_config{gpu.regs.shader_config[static_cast<std::size_t>(program)]}; - return gpu.regs.code_address.CodeAddress() + shader_config.offset; -} - -/// Gets if the current instruction offset is a scheduler instruction -constexpr bool IsSchedInstruction(std::size_t offset, std::size_t main_offset) { - // Sched instructions appear once every 4 instructions. - constexpr std::size_t SchedPeriod = 4; - const std::size_t absolute_offset = offset - main_offset; - return (absolute_offset % SchedPeriod) == 0; -} - -/// Calculates the size of a program stream -std::size_t CalculateProgramSize(const ProgramCode& program, bool is_compute) { - const std::size_t start_offset = is_compute ? 0 : 10; - // This is the encoded version of BRA that jumps to itself. All Nvidia - // shaders end with one. - constexpr u64 self_jumping_branch = 0xE2400FFFFF07000FULL; - constexpr u64 mask = 0xFFFFFFFFFF7FFFFFULL; - std::size_t offset = start_offset; - while (offset < program.size()) { - const u64 instruction = program[offset]; - if (!IsSchedInstruction(offset, start_offset)) { - if ((instruction & mask) == self_jumping_branch) { - // End on Maxwell's "nop" instruction - break; - } - if (instruction == 0) { - break; - } - } - ++offset; - } - // The last instruction is included in the program size - return std::min(offset + 1, program.size()); -} - -/// Gets the shader program code from memory for the specified address -ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, const GPUVAddr gpu_addr, - const u8* host_ptr, bool is_compute) { - ProgramCode program_code(VideoCommon::Shader::MAX_PROGRAM_LENGTH); - ASSERT_OR_EXECUTE(host_ptr != nullptr, { - std::fill(program_code.begin(), program_code.end(), 0); - return program_code; - }); - memory_manager.ReadBlockUnsafe(gpu_addr, program_code.data(), - program_code.size() * sizeof(u64)); - program_code.resize(CalculateProgramSize(program_code, is_compute)); - return program_code; -} - constexpr std::size_t GetStageFromProgram(std::size_t program) { return program == 0 ? 0 : program - 1; } @@ -133,14 +87,15 @@ void AddBindings(std::vector<VkDescriptorSetLayoutBinding>& bindings, u32& bindi u32 count = 1; if constexpr (descriptor_type == VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER) { // Combined image samplers can be arrayed. - count = container[i].Size(); + count = container[i].size; } - VkDescriptorSetLayoutBinding& entry = bindings.emplace_back(); - entry.binding = binding++; - entry.descriptorType = descriptor_type; - entry.descriptorCount = count; - entry.stageFlags = stage_flags; - entry.pImmutableSamplers = nullptr; + bindings.push_back({ + .binding = binding++, + .descriptorType = descriptor_type, + .descriptorCount = count, + .stageFlags = stage_flags, + .pImmutableSamplers = nullptr, + }); } } @@ -153,96 +108,133 @@ u32 FillDescriptorLayout(const ShaderEntries& entries, u32 binding = base_binding; AddBindings<UNIFORM_BUFFER>(bindings, binding, flags, entries.const_buffers); AddBindings<STORAGE_BUFFER>(bindings, binding, flags, entries.global_buffers); - AddBindings<UNIFORM_TEXEL_BUFFER>(bindings, binding, flags, entries.texel_buffers); + AddBindings<UNIFORM_TEXEL_BUFFER>(bindings, binding, flags, entries.uniform_texels); AddBindings<COMBINED_IMAGE_SAMPLER>(bindings, binding, flags, entries.samplers); + AddBindings<STORAGE_TEXEL_BUFFER>(bindings, binding, flags, entries.storage_texels); AddBindings<STORAGE_IMAGE>(bindings, binding, flags, entries.images); return binding; } } // Anonymous namespace -CachedShader::CachedShader(Core::System& system, Tegra::Engines::ShaderType stage, - GPUVAddr gpu_addr, VAddr cpu_addr, ProgramCode program_code, - u32 main_offset) - : RasterizerCacheObject{cpu_addr}, gpu_addr{gpu_addr}, program_code{std::move(program_code)}, - registry{stage, GetEngine(system, stage)}, shader_ir{this->program_code, main_offset, - compiler_settings, registry}, - entries{GenerateShaderEntries(shader_ir)} {} - -CachedShader::~CachedShader() = default; - -Tegra::Engines::ConstBufferEngineInterface& CachedShader::GetEngine( - Core::System& system, Tegra::Engines::ShaderType stage) { - if (stage == Tegra::Engines::ShaderType::Compute) { - return system.GPU().KeplerCompute(); - } else { - return system.GPU().Maxwell3D(); - } +std::size_t GraphicsPipelineCacheKey::Hash() const noexcept { + const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), Size()); + return static_cast<std::size_t>(hash); +} + +bool GraphicsPipelineCacheKey::operator==(const GraphicsPipelineCacheKey& rhs) const noexcept { + return std::memcmp(&rhs, this, Size()) == 0; +} + +std::size_t ComputePipelineCacheKey::Hash() const noexcept { + const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this); + return static_cast<std::size_t>(hash); +} + +bool ComputePipelineCacheKey::operator==(const ComputePipelineCacheKey& rhs) const noexcept { + return std::memcmp(&rhs, this, sizeof *this) == 0; } -VKPipelineCache::VKPipelineCache(Core::System& system, RasterizerVulkan& rasterizer, - const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - VKRenderPassCache& renderpass_cache) - : RasterizerCache{rasterizer}, system{system}, device{device}, scheduler{scheduler}, - descriptor_pool{descriptor_pool}, update_descriptor_queue{update_descriptor_queue}, - renderpass_cache{renderpass_cache} {} +Shader::Shader(Tegra::Engines::ConstBufferEngineInterface& engine, Tegra::Engines::ShaderType stage, + GPUVAddr gpu_addr_, VAddr cpu_addr, VideoCommon::Shader::ProgramCode program_code_, + u32 main_offset) + : gpu_addr(gpu_addr_), program_code(std::move(program_code_)), registry(stage, engine), + shader_ir(program_code, main_offset, compiler_settings, registry), + entries(GenerateShaderEntries(shader_ir)) {} + +Shader::~Shader() = default; + +VKPipelineCache::VKPipelineCache(RasterizerVulkan& rasterizer, Tegra::GPU& gpu_, + Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::Engines::KeplerCompute& kepler_compute_, + Tegra::MemoryManager& gpu_memory_, const VKDevice& device_, + VKScheduler& scheduler_, VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + VKRenderPassCache& renderpass_cache_) + : VideoCommon::ShaderCache<Shader>{rasterizer}, gpu{gpu_}, maxwell3d{maxwell3d_}, + kepler_compute{kepler_compute_}, gpu_memory{gpu_memory_}, device{device_}, + scheduler{scheduler_}, descriptor_pool{descriptor_pool_}, + update_descriptor_queue{update_descriptor_queue_}, renderpass_cache{renderpass_cache_} {} VKPipelineCache::~VKPipelineCache() = default; -std::array<Shader, Maxwell::MaxShaderProgram> VKPipelineCache::GetShaders() { - const auto& gpu = system.GPU().Maxwell3D(); +std::array<Shader*, Maxwell::MaxShaderProgram> VKPipelineCache::GetShaders() { + std::array<Shader*, Maxwell::MaxShaderProgram> shaders{}; - std::array<Shader, Maxwell::MaxShaderProgram> shaders; for (std::size_t index = 0; index < Maxwell::MaxShaderProgram; ++index) { const auto program{static_cast<Maxwell::ShaderProgram>(index)}; // Skip stages that are not enabled - if (!gpu.regs.IsShaderConfigEnabled(index)) { + if (!maxwell3d.regs.IsShaderConfigEnabled(index)) { continue; } - auto& memory_manager{system.GPU().MemoryManager()}; - const GPUVAddr program_addr{GetShaderAddress(system, program)}; - const std::optional cpu_addr = memory_manager.GpuToCpuAddress(program_addr); + const GPUVAddr gpu_addr{GetShaderAddress(maxwell3d, program)}; + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); ASSERT(cpu_addr); - auto shader = cpu_addr ? TryGet(*cpu_addr) : nullptr; - if (!shader) { - const auto host_ptr{memory_manager.GetPointer(program_addr)}; - // No shader found - create a new one - constexpr u32 stage_offset = 10; - const auto stage = static_cast<Tegra::Engines::ShaderType>(index == 0 ? 0 : index - 1); - auto code = GetShaderCode(memory_manager, program_addr, host_ptr, false); + Shader* result = cpu_addr ? TryGet(*cpu_addr) : null_shader.get(); + if (!result) { + const u8* const host_ptr{gpu_memory.GetPointer(gpu_addr)}; - shader = std::make_shared<CachedShader>(system, stage, program_addr, *cpu_addr, - std::move(code), stage_offset); - Register(shader); + // No shader found - create a new one + static constexpr u32 stage_offset = STAGE_MAIN_OFFSET; + const auto stage = static_cast<ShaderType>(index == 0 ? 0 : index - 1); + ProgramCode code = GetShaderCode(gpu_memory, gpu_addr, host_ptr, false); + const std::size_t size_in_bytes = code.size() * sizeof(u64); + + auto shader = std::make_unique<Shader>(maxwell3d, stage, gpu_addr, *cpu_addr, + std::move(code), stage_offset); + result = shader.get(); + + if (cpu_addr) { + Register(std::move(shader), *cpu_addr, size_in_bytes); + } else { + null_shader = std::move(shader); + } } - shaders[index] = std::move(shader); + shaders[index] = result; } return last_shaders = shaders; } -VKGraphicsPipeline& VKPipelineCache::GetGraphicsPipeline(const GraphicsPipelineCacheKey& key) { +VKGraphicsPipeline* VKPipelineCache::GetGraphicsPipeline( + const GraphicsPipelineCacheKey& key, VideoCommon::Shader::AsyncShaders& async_shaders) { MICROPROFILE_SCOPE(Vulkan_PipelineCache); if (last_graphics_pipeline && last_graphics_key == key) { - return *last_graphics_pipeline; + return last_graphics_pipeline; } last_graphics_key = key; + if (device.UseAsynchronousShaders() && async_shaders.IsShaderAsync(gpu)) { + std::unique_lock lock{pipeline_cache}; + const auto [pair, is_cache_miss] = graphics_cache.try_emplace(key); + if (is_cache_miss) { + gpu.ShaderNotify().MarkSharderBuilding(); + LOG_INFO(Render_Vulkan, "Compile 0x{:016X}", key.Hash()); + const auto [program, bindings] = DecompileShaders(key.fixed_state); + async_shaders.QueueVulkanShader(this, device, scheduler, descriptor_pool, + update_descriptor_queue, renderpass_cache, bindings, + program, key); + } + last_graphics_pipeline = pair->second.get(); + return last_graphics_pipeline; + } + const auto [pair, is_cache_miss] = graphics_cache.try_emplace(key); auto& entry = pair->second; if (is_cache_miss) { + gpu.ShaderNotify().MarkSharderBuilding(); LOG_INFO(Render_Vulkan, "Compile 0x{:016X}", key.Hash()); - const auto [program, bindings] = DecompileShaders(key); + const auto [program, bindings] = DecompileShaders(key.fixed_state); entry = std::make_unique<VKGraphicsPipeline>(device, scheduler, descriptor_pool, update_descriptor_queue, renderpass_cache, key, bindings, program); + gpu.ShaderNotify().MarkShaderComplete(); } - return *(last_graphics_pipeline = entry.get()); + last_graphics_pipeline = entry.get(); + return last_graphics_pipeline; } VKComputePipeline& VKPipelineCache::GetComputePipeline(const ComputePipelineCacheKey& key) { @@ -255,29 +247,39 @@ VKComputePipeline& VKPipelineCache::GetComputePipeline(const ComputePipelineCach } LOG_INFO(Render_Vulkan, "Compile 0x{:016X}", key.Hash()); - auto& memory_manager = system.GPU().MemoryManager(); - const auto program_addr = key.shader; + const GPUVAddr gpu_addr = key.shader; - const auto cpu_addr = memory_manager.GpuToCpuAddress(program_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); ASSERT(cpu_addr); - auto shader = cpu_addr ? TryGet(*cpu_addr) : nullptr; + Shader* shader = cpu_addr ? TryGet(*cpu_addr) : null_kernel.get(); if (!shader) { // No shader found - create a new one - const auto host_ptr = memory_manager.GetPointer(program_addr); - - auto code = GetShaderCode(memory_manager, program_addr, host_ptr, true); - constexpr u32 kernel_main_offset = 0; - shader = std::make_shared<CachedShader>(system, Tegra::Engines::ShaderType::Compute, - program_addr, *cpu_addr, std::move(code), - kernel_main_offset); - Register(shader); - } + const auto host_ptr = gpu_memory.GetPointer(gpu_addr); - Specialization specialization; - specialization.workgroup_size = key.workgroup_size; - specialization.shared_memory_size = key.shared_memory_size; + ProgramCode code = GetShaderCode(gpu_memory, gpu_addr, host_ptr, true); + const std::size_t size_in_bytes = code.size() * sizeof(u64); + + auto shader_info = std::make_unique<Shader>(kepler_compute, ShaderType::Compute, gpu_addr, + *cpu_addr, std::move(code), KERNEL_MAIN_OFFSET); + shader = shader_info.get(); + if (cpu_addr) { + Register(std::move(shader_info), *cpu_addr, size_in_bytes); + } else { + null_kernel = std::move(shader_info); + } + } + + const Specialization specialization{ + .base_binding = 0, + .workgroup_size = key.workgroup_size, + .shared_memory_size = key.shared_memory_size, + .point_size = std::nullopt, + .enabled_attributes = {}, + .attribute_types = {}, + .ndc_minus_one_to_one = false, + }; const SPIRVShader spirv_shader{Decompile(device, shader->GetIR(), ShaderType::Compute, shader->GetRegistry(), specialization), shader->GetEntries()}; @@ -286,7 +288,13 @@ VKComputePipeline& VKPipelineCache::GetComputePipeline(const ComputePipelineCach return *entry; } -void VKPipelineCache::Unregister(const Shader& shader) { +void VKPipelineCache::EmplacePipeline(std::unique_ptr<VKGraphicsPipeline> pipeline) { + gpu.ShaderNotify().MarkShaderComplete(); + std::unique_lock lock{pipeline_cache}; + graphics_cache.at(pipeline->GetCacheKey()) = std::move(pipeline); +} + +void VKPipelineCache::OnShaderRemoval(Shader* shader) { bool finished = false; const auto Finish = [&] { // TODO(Rodrigo): Instead of finishing here, wait for the fences that use this pipeline and @@ -318,25 +326,23 @@ void VKPipelineCache::Unregister(const Shader& shader) { Finish(); it = compute_cache.erase(it); } - - RasterizerCache::Unregister(shader); } std::pair<SPIRVProgram, std::vector<VkDescriptorSetLayoutBinding>> -VKPipelineCache::DecompileShaders(const GraphicsPipelineCacheKey& key) { - const auto& fixed_state = key.fixed_state; - auto& memory_manager = system.GPU().MemoryManager(); - const auto& gpu = system.GPU().Maxwell3D(); - +VKPipelineCache::DecompileShaders(const FixedPipelineState& fixed_state) { Specialization specialization; - if (fixed_state.input_assembly.topology == Maxwell::PrimitiveTopology::Points) { - ASSERT(fixed_state.input_assembly.point_size != 0.0f); - specialization.point_size = fixed_state.input_assembly.point_size; + if (fixed_state.topology == Maxwell::PrimitiveTopology::Points) { + float point_size; + std::memcpy(&point_size, &fixed_state.point_size, sizeof(float)); + specialization.point_size = point_size; + ASSERT(point_size != 0.0f); } for (std::size_t i = 0; i < Maxwell::NumVertexAttributes; ++i) { - specialization.attribute_types[i] = fixed_state.vertex_input.attributes[i].type; + const auto& attribute = fixed_state.attributes[i]; + specialization.enabled_attributes[i] = attribute.enabled.Value() != 0; + specialization.attribute_types[i] = attribute.Type(); } - specialization.ndc_minus_one_to_one = fixed_state.rasterizer.ndc_minus_one_to_one; + specialization.ndc_minus_one_to_one = fixed_state.ndc_minus_one_to_one; SPIRVProgram program; std::vector<VkDescriptorSetLayoutBinding> bindings; @@ -345,18 +351,16 @@ VKPipelineCache::DecompileShaders(const GraphicsPipelineCacheKey& key) { const auto program_enum = static_cast<Maxwell::ShaderProgram>(index); // Skip stages that are not enabled - if (!gpu.regs.IsShaderConfigEnabled(index)) { + if (!maxwell3d.regs.IsShaderConfigEnabled(index)) { continue; } - const GPUVAddr gpu_addr = GetShaderAddress(system, program_enum); - const auto cpu_addr = memory_manager.GpuToCpuAddress(gpu_addr); - ASSERT(cpu_addr); - const auto shader = TryGet(*cpu_addr); - ASSERT(shader); + const GPUVAddr gpu_addr = GetShaderAddress(maxwell3d, program_enum); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); + Shader* const shader = cpu_addr ? TryGet(*cpu_addr) : null_shader.get(); const std::size_t stage = index == 0 ? 0 : index - 1; // Stage indices are 0 - 5 - const auto program_type = GetShaderType(program_enum); + const ShaderType program_type = GetShaderType(program_enum); const auto& entries = shader->GetEntries(); program[stage] = { Decompile(device, shader->GetIR(), program_type, shader->GetRegistry(), specialization), @@ -383,14 +387,15 @@ void AddEntry(std::vector<VkDescriptorUpdateTemplateEntry>& template_entries, u3 if constexpr (descriptor_type == COMBINED_IMAGE_SAMPLER) { for (u32 i = 0; i < count; ++i) { - const u32 num_samplers = container[i].Size(); - VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back(); - entry.dstBinding = binding; - entry.dstArrayElement = 0; - entry.descriptorCount = num_samplers; - entry.descriptorType = descriptor_type; - entry.offset = offset; - entry.stride = entry_size; + const u32 num_samplers = container[i].size; + template_entries.push_back({ + .dstBinding = binding, + .dstArrayElement = 0, + .descriptorCount = num_samplers, + .descriptorType = descriptor_type, + .offset = offset, + .stride = entry_size, + }); ++binding; offset += num_samplers * entry_size; @@ -398,26 +403,29 @@ void AddEntry(std::vector<VkDescriptorUpdateTemplateEntry>& template_entries, u3 return; } - if constexpr (descriptor_type == UNIFORM_TEXEL_BUFFER) { - // Nvidia has a bug where updating multiple uniform texels at once causes the driver to - // crash. + if constexpr (descriptor_type == UNIFORM_TEXEL_BUFFER || + descriptor_type == STORAGE_TEXEL_BUFFER) { + // Nvidia has a bug where updating multiple texels at once causes the driver to crash. + // Note: Fixed in driver Windows 443.24, Linux 440.66.15 for (u32 i = 0; i < count; ++i) { - VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back(); - entry.dstBinding = binding + i; - entry.dstArrayElement = 0; - entry.descriptorCount = 1; - entry.descriptorType = descriptor_type; - entry.offset = offset + i * entry_size; - entry.stride = entry_size; + template_entries.push_back({ + .dstBinding = binding + i, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = descriptor_type, + .offset = static_cast<std::size_t>(offset + i * entry_size), + .stride = entry_size, + }); } } else if (count > 0) { - VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back(); - entry.dstBinding = binding; - entry.dstArrayElement = 0; - entry.descriptorCount = count; - entry.descriptorType = descriptor_type; - entry.offset = offset; - entry.stride = entry_size; + template_entries.push_back({ + .dstBinding = binding, + .dstArrayElement = 0, + .descriptorCount = count, + .descriptorType = descriptor_type, + .offset = offset, + .stride = entry_size, + }); } offset += count * entry_size; binding += count; @@ -428,8 +436,9 @@ void FillDescriptorUpdateTemplateEntries( std::vector<VkDescriptorUpdateTemplateEntryKHR>& template_entries) { AddEntry<UNIFORM_BUFFER>(template_entries, offset, binding, entries.const_buffers); AddEntry<STORAGE_BUFFER>(template_entries, offset, binding, entries.global_buffers); - AddEntry<UNIFORM_TEXEL_BUFFER>(template_entries, offset, binding, entries.texel_buffers); + AddEntry<UNIFORM_TEXEL_BUFFER>(template_entries, offset, binding, entries.uniform_texels); AddEntry<COMBINED_IMAGE_SAMPLER>(template_entries, offset, binding, entries.samplers); + AddEntry<STORAGE_TEXEL_BUFFER>(template_entries, offset, binding, entries.storage_texels); AddEntry<STORAGE_IMAGE>(template_entries, offset, binding, entries.images); } diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.h b/src/video_core/renderer_vulkan/vk_pipeline_cache.h index 7ccdb7083..e558e6658 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.h +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.h @@ -7,7 +7,6 @@ #include <array> #include <cstddef> #include <memory> -#include <tuple> #include <type_traits> #include <unordered_map> #include <utility> @@ -18,16 +17,16 @@ #include "common/common_types.h" #include "video_core/engines/const_buffer_engine_interface.h" #include "video_core/engines/maxwell_3d.h" -#include "video_core/rasterizer_cache.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_graphics_pipeline.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/shader/async_shaders.h" +#include "video_core/shader/memory_util.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" -#include "video_core/surface.h" +#include "video_core/shader_cache.h" namespace Core { class System; @@ -39,54 +38,27 @@ class RasterizerVulkan; class VKComputePipeline; class VKDescriptorPool; class VKDevice; -class VKFence; class VKScheduler; class VKUpdateDescriptorQueue; -class CachedShader; -using Shader = std::shared_ptr<CachedShader>; using Maxwell = Tegra::Engines::Maxwell3D::Regs; -using ProgramCode = std::vector<u64>; +struct ComputePipelineCacheKey { + GPUVAddr shader; + u32 shared_memory_size; + std::array<u32, 3> workgroup_size; -struct GraphicsPipelineCacheKey { - FixedPipelineState fixed_state; - std::array<GPUVAddr, Maxwell::MaxShaderProgram> shaders; - RenderPassParams renderpass_params; + std::size_t Hash() const noexcept; - std::size_t Hash() const noexcept { - std::size_t hash = fixed_state.Hash(); - for (const auto& shader : shaders) { - boost::hash_combine(hash, shader); - } - boost::hash_combine(hash, renderpass_params.Hash()); - return hash; - } + bool operator==(const ComputePipelineCacheKey& rhs) const noexcept; - bool operator==(const GraphicsPipelineCacheKey& rhs) const noexcept { - return std::tie(fixed_state, shaders, renderpass_params) == - std::tie(rhs.fixed_state, rhs.shaders, rhs.renderpass_params); - } -}; - -struct ComputePipelineCacheKey { - GPUVAddr shader{}; - u32 shared_memory_size{}; - std::array<u32, 3> workgroup_size{}; - - std::size_t Hash() const noexcept { - return static_cast<std::size_t>(shader) ^ - ((static_cast<std::size_t>(shared_memory_size) >> 7) << 40) ^ - static_cast<std::size_t>(workgroup_size[0]) ^ - (static_cast<std::size_t>(workgroup_size[1]) << 16) ^ - (static_cast<std::size_t>(workgroup_size[2]) << 24); - } - - bool operator==(const ComputePipelineCacheKey& rhs) const noexcept { - return std::tie(shader, shared_memory_size, workgroup_size) == - std::tie(rhs.shader, rhs.shared_memory_size, rhs.workgroup_size); + bool operator!=(const ComputePipelineCacheKey& rhs) const noexcept { + return !operator==(rhs); } }; +static_assert(std::has_unique_object_representations_v<ComputePipelineCacheKey>); +static_assert(std::is_trivially_copyable_v<ComputePipelineCacheKey>); +static_assert(std::is_trivially_constructible_v<ComputePipelineCacheKey>); } // namespace Vulkan @@ -110,21 +82,22 @@ struct hash<Vulkan::ComputePipelineCacheKey> { namespace Vulkan { -class CachedShader final : public RasterizerCacheObject { +class Shader { public: - explicit CachedShader(Core::System& system, Tegra::Engines::ShaderType stage, GPUVAddr gpu_addr, - VAddr cpu_addr, ProgramCode program_code, u32 main_offset); - ~CachedShader(); + explicit Shader(Tegra::Engines::ConstBufferEngineInterface& engine, + Tegra::Engines::ShaderType stage, GPUVAddr gpu_addr, VAddr cpu_addr, + VideoCommon::Shader::ProgramCode program_code, u32 main_offset); + ~Shader(); GPUVAddr GetGpuAddr() const { return gpu_addr; } - std::size_t GetSizeInBytes() const override { - return program_code.size() * sizeof(u64); + VideoCommon::Shader::ShaderIR& GetIR() { + return shader_ir; } - VideoCommon::Shader::ShaderIR& GetIR() { + const VideoCommon::Shader::ShaderIR& GetIR() const { return shader_ir; } @@ -132,61 +105,65 @@ public: return registry; } - const VideoCommon::Shader::ShaderIR& GetIR() const { - return shader_ir; - } - const ShaderEntries& GetEntries() const { return entries; } private: - static Tegra::Engines::ConstBufferEngineInterface& GetEngine(Core::System& system, - Tegra::Engines::ShaderType stage); - GPUVAddr gpu_addr{}; - ProgramCode program_code; + VideoCommon::Shader::ProgramCode program_code; VideoCommon::Shader::Registry registry; VideoCommon::Shader::ShaderIR shader_ir; ShaderEntries entries; }; -class VKPipelineCache final : public RasterizerCache<Shader> { +class VKPipelineCache final : public VideoCommon::ShaderCache<Shader> { public: - explicit VKPipelineCache(Core::System& system, RasterizerVulkan& rasterizer, - const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, + explicit VKPipelineCache(RasterizerVulkan& rasterizer, Tegra::GPU& gpu, + Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::Engines::KeplerCompute& kepler_compute, + Tegra::MemoryManager& gpu_memory, const VKDevice& device, + VKScheduler& scheduler, VKDescriptorPool& descriptor_pool, VKUpdateDescriptorQueue& update_descriptor_queue, VKRenderPassCache& renderpass_cache); - ~VKPipelineCache(); + ~VKPipelineCache() override; - std::array<Shader, Maxwell::MaxShaderProgram> GetShaders(); + std::array<Shader*, Maxwell::MaxShaderProgram> GetShaders(); - VKGraphicsPipeline& GetGraphicsPipeline(const GraphicsPipelineCacheKey& key); + VKGraphicsPipeline* GetGraphicsPipeline(const GraphicsPipelineCacheKey& key, + VideoCommon::Shader::AsyncShaders& async_shaders); VKComputePipeline& GetComputePipeline(const ComputePipelineCacheKey& key); -protected: - void Unregister(const Shader& shader) override; + void EmplacePipeline(std::unique_ptr<VKGraphicsPipeline> pipeline); - void FlushObjectInner(const Shader& object) override {} +protected: + void OnShaderRemoval(Shader* shader) final; private: std::pair<SPIRVProgram, std::vector<VkDescriptorSetLayoutBinding>> DecompileShaders( - const GraphicsPipelineCacheKey& key); + const FixedPipelineState& fixed_state); + + Tegra::GPU& gpu; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::Engines::KeplerCompute& kepler_compute; + Tegra::MemoryManager& gpu_memory; - Core::System& system; const VKDevice& device; VKScheduler& scheduler; VKDescriptorPool& descriptor_pool; VKUpdateDescriptorQueue& update_descriptor_queue; VKRenderPassCache& renderpass_cache; - std::array<Shader, Maxwell::MaxShaderProgram> last_shaders; + std::unique_ptr<Shader> null_shader; + std::unique_ptr<Shader> null_kernel; + + std::array<Shader*, Maxwell::MaxShaderProgram> last_shaders{}; GraphicsPipelineCacheKey last_graphics_key; VKGraphicsPipeline* last_graphics_pipeline = nullptr; + std::mutex pipeline_cache; std::unordered_map<GraphicsPipelineCacheKey, std::unique_ptr<VKGraphicsPipeline>> graphics_cache; std::unordered_map<ComputePipelineCacheKey, std::unique_ptr<VKComputePipeline>> compute_cache; diff --git a/src/video_core/renderer_vulkan/vk_query_cache.cpp b/src/video_core/renderer_vulkan/vk_query_cache.cpp index 0966c7ff7..ee2d871e3 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_query_cache.cpp @@ -4,41 +4,38 @@ #include <algorithm> #include <cstddef> -#include <cstdint> #include <utility> #include <vector> #include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_query_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" +#include "video_core/renderer_vulkan/vk_resource_pool.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { +using VideoCore::QueryType; + namespace { constexpr std::array QUERY_TARGETS = {VK_QUERY_TYPE_OCCLUSION}; -constexpr VkQueryType GetTarget(VideoCore::QueryType type) { +constexpr VkQueryType GetTarget(QueryType type) { return QUERY_TARGETS[static_cast<std::size_t>(type)]; } } // Anonymous namespace -QueryPool::QueryPool() : VKFencedPool{GROW_STEP} {} +QueryPool::QueryPool(const VKDevice& device_, VKScheduler& scheduler, QueryType type_) + : ResourcePool{scheduler.GetMasterSemaphore(), GROW_STEP}, device{device_}, type{type_} {} QueryPool::~QueryPool() = default; -void QueryPool::Initialize(const VKDevice& device_, VideoCore::QueryType type_) { - device = &device_; - type = type_; -} - -std::pair<VkQueryPool, u32> QueryPool::Commit(VKFence& fence) { +std::pair<VkQueryPool, u32> QueryPool::Commit() { std::size_t index; do { - index = CommitResource(fence); + index = CommitResource(); } while (usage[index]); usage[index] = true; @@ -48,14 +45,14 @@ std::pair<VkQueryPool, u32> QueryPool::Commit(VKFence& fence) { void QueryPool::Allocate(std::size_t begin, std::size_t end) { usage.resize(end); - VkQueryPoolCreateInfo query_pool_ci; - query_pool_ci.sType = VK_STRUCTURE_TYPE_QUERY_POOL_CREATE_INFO; - query_pool_ci.pNext = nullptr; - query_pool_ci.flags = 0; - query_pool_ci.queryType = GetTarget(type); - query_pool_ci.queryCount = static_cast<u32>(end - begin); - query_pool_ci.pipelineStatistics = 0; - pools.push_back(device->GetLogical().CreateQueryPool(query_pool_ci)); + pools.push_back(device.GetLogical().CreateQueryPool({ + .sType = VK_STRUCTURE_TYPE_QUERY_POOL_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .queryType = GetTarget(type), + .queryCount = static_cast<u32>(end - begin), + .pipelineStatistics = 0, + })); } void QueryPool::Reserve(std::pair<VkQueryPool, u32> query) { @@ -69,30 +66,39 @@ void QueryPool::Reserve(std::pair<VkQueryPool, u32> query) { usage[pool_index * GROW_STEP + static_cast<std::ptrdiff_t>(query.second)] = false; } -VKQueryCache::VKQueryCache(Core::System& system, VideoCore::RasterizerInterface& rasterizer, +VKQueryCache::VKQueryCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, const VKDevice& device, VKScheduler& scheduler) - : VideoCommon::QueryCacheBase<VKQueryCache, CachedQuery, CounterStream, HostCounter, - QueryPool>{system, rasterizer}, - device{device}, scheduler{scheduler} { - for (std::size_t i = 0; i < static_cast<std::size_t>(VideoCore::NumQueryTypes); ++i) { - query_pools[i].Initialize(device, static_cast<VideoCore::QueryType>(i)); + : VideoCommon::QueryCacheBase<VKQueryCache, CachedQuery, CounterStream, + HostCounter>{rasterizer, maxwell3d, gpu_memory}, + device{device}, scheduler{scheduler}, query_pools{ + QueryPool{device, scheduler, + QueryType::SamplesPassed}, + } {} + +VKQueryCache::~VKQueryCache() { + // TODO(Rodrigo): This is a hack to destroy all HostCounter instances before the base class + // destructor is called. The query cache should be redesigned to have a proper ownership model + // instead of using shared pointers. + for (size_t query_type = 0; query_type < VideoCore::NumQueryTypes; ++query_type) { + auto& stream = Stream(static_cast<QueryType>(query_type)); + stream.Update(false); + stream.Reset(); } } -VKQueryCache::~VKQueryCache() = default; - -std::pair<VkQueryPool, u32> VKQueryCache::AllocateQuery(VideoCore::QueryType type) { - return query_pools[static_cast<std::size_t>(type)].Commit(scheduler.GetFence()); +std::pair<VkQueryPool, u32> VKQueryCache::AllocateQuery(QueryType type) { + return query_pools[static_cast<std::size_t>(type)].Commit(); } -void VKQueryCache::Reserve(VideoCore::QueryType type, std::pair<VkQueryPool, u32> query) { +void VKQueryCache::Reserve(QueryType type, std::pair<VkQueryPool, u32> query) { query_pools[static_cast<std::size_t>(type)].Reserve(query); } HostCounter::HostCounter(VKQueryCache& cache, std::shared_ptr<HostCounter> dependency, - VideoCore::QueryType type) + QueryType type) : VideoCommon::HostCounterBase<VKQueryCache, HostCounter>{std::move(dependency)}, cache{cache}, - type{type}, query{cache.AllocateQuery(type)}, ticks{cache.Scheduler().Ticks()} { + type{type}, query{cache.AllocateQuery(type)}, tick{cache.Scheduler().CurrentTick()} { const vk::Device* logical = &cache.Device().GetLogical(); cache.Scheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) { logical->ResetQueryPoolEXT(query.first, query.second, 1); @@ -110,11 +116,22 @@ void HostCounter::EndQuery() { } u64 HostCounter::BlockingQuery() const { - if (ticks >= cache.Scheduler().Ticks()) { + if (tick >= cache.Scheduler().CurrentTick()) { cache.Scheduler().Flush(); } - return cache.Device().GetLogical().GetQueryResult<u64>( - query.first, query.second, VK_QUERY_RESULT_64_BIT | VK_QUERY_RESULT_WAIT_BIT); + u64 data; + const VkResult result = cache.Device().GetLogical().GetQueryResults( + query.first, query.second, 1, sizeof(data), &data, sizeof(data), + VK_QUERY_RESULT_64_BIT | VK_QUERY_RESULT_WAIT_BIT); + switch (result) { + case VK_SUCCESS: + return data; + case VK_ERROR_DEVICE_LOST: + cache.Device().ReportLoss(); + [[fallthrough]]; + default: + throw vk::Exception(result); + } } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_query_cache.h b/src/video_core/renderer_vulkan/vk_query_cache.h index b63784f4b..2e57fb75d 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.h +++ b/src/video_core/renderer_vulkan/vk_query_cache.h @@ -5,14 +5,13 @@ #pragma once #include <cstddef> -#include <cstdint> #include <memory> #include <utility> #include <vector> #include "common/common_types.h" #include "video_core/query_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" +#include "video_core/renderer_vulkan/vk_resource_pool.h" #include "video_core/renderer_vulkan/wrapper.h" namespace VideoCore { @@ -29,14 +28,12 @@ class VKScheduler; using CounterStream = VideoCommon::CounterStreamBase<VKQueryCache, HostCounter>; -class QueryPool final : public VKFencedPool { +class QueryPool final : public ResourcePool { public: - explicit QueryPool(); + explicit QueryPool(const VKDevice& device, VKScheduler& scheduler, VideoCore::QueryType type); ~QueryPool() override; - void Initialize(const VKDevice& device, VideoCore::QueryType type); - - std::pair<VkQueryPool, u32> Commit(VKFence& fence); + std::pair<VkQueryPool, u32> Commit(); void Reserve(std::pair<VkQueryPool, u32> query); @@ -46,18 +43,18 @@ protected: private: static constexpr std::size_t GROW_STEP = 512; - const VKDevice* device = nullptr; - VideoCore::QueryType type = {}; + const VKDevice& device; + const VideoCore::QueryType type; std::vector<vk::QueryPool> pools; std::vector<bool> usage; }; class VKQueryCache final - : public VideoCommon::QueryCacheBase<VKQueryCache, CachedQuery, CounterStream, HostCounter, - QueryPool> { + : public VideoCommon::QueryCacheBase<VKQueryCache, CachedQuery, CounterStream, HostCounter> { public: - explicit VKQueryCache(Core::System& system, VideoCore::RasterizerInterface& rasterizer, + explicit VKQueryCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, const VKDevice& device, VKScheduler& scheduler); ~VKQueryCache(); @@ -76,6 +73,7 @@ public: private: const VKDevice& device; VKScheduler& scheduler; + std::array<QueryPool, VideoCore::NumQueryTypes> query_pools; }; class HostCounter final : public VideoCommon::HostCounterBase<VKQueryCache, HostCounter> { @@ -92,7 +90,7 @@ private: VKQueryCache& cache; const VideoCore::QueryType type; const std::pair<VkQueryPool, u32> query; - const u64 ticks; + const u64 tick; }; class CachedQuery : public VideoCommon::CachedQueryBase<HostCounter> { diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index 774ba1f26..e0fb8693f 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -9,14 +9,14 @@ #include <vector> #include <boost/container/static_vector.hpp> -#include <boost/functional/hash.hpp> #include "common/alignment.h" #include "common/assert.h" #include "common/logging/log.h" #include "common/microprofile.h" +#include "common/scope_exit.h" #include "core/core.h" -#include "core/memory.h" +#include "core/settings.h" #include "video_core/engines/kepler_compute.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" @@ -31,7 +31,6 @@ #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_sampler_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" @@ -39,6 +38,7 @@ #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" #include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/shader_cache.h" namespace Vulkan { @@ -64,20 +64,22 @@ VkViewport GetViewportState(const VKDevice& device, const Maxwell& regs, std::si const auto& src = regs.viewport_transform[index]; const float width = src.scale_x * 2.0f; const float height = src.scale_y * 2.0f; + const float reduce_z = regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1.0f : 0.0f; - VkViewport viewport; - viewport.x = src.translate_x - src.scale_x; - viewport.y = src.translate_y - src.scale_y; - viewport.width = width != 0.0f ? width : 1.0f; - viewport.height = height != 0.0f ? height : 1.0f; + VkViewport viewport{ + .x = src.translate_x - src.scale_x, + .y = src.translate_y - src.scale_y, + .width = width != 0.0f ? width : 1.0f, + .height = height != 0.0f ? height : 1.0f, + .minDepth = src.translate_z - src.scale_z * reduce_z, + .maxDepth = src.translate_z + src.scale_z, + }; - const float reduce_z = regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1.0f : 0.0f; - viewport.minDepth = src.translate_z - src.scale_z * reduce_z; - viewport.maxDepth = src.translate_z + src.scale_z; if (!device.IsExtDepthRangeUnrestrictedSupported()) { viewport.minDepth = std::clamp(viewport.minDepth, 0.0f, 1.0f); viewport.maxDepth = std::clamp(viewport.maxDepth, 0.0f, 1.0f); } + return viewport; } @@ -99,7 +101,7 @@ VkRect2D GetScissorState(const Maxwell& regs, std::size_t index) { } std::array<GPUVAddr, Maxwell::MaxShaderProgram> GetShaderAddresses( - const std::array<Shader, Maxwell::MaxShaderProgram>& shaders) { + const std::array<Shader*, Maxwell::MaxShaderProgram>& shaders) { std::array<GPUVAddr, Maxwell::MaxShaderProgram> addresses; for (std::size_t i = 0; i < std::size(addresses); ++i) { addresses[i] = shaders[i] ? shaders[i]->GetGpuAddr() : 0; @@ -118,14 +120,24 @@ template <typename Engine, typename Entry> Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry, std::size_t stage, std::size_t index = 0) { const auto stage_type = static_cast<Tegra::Engines::ShaderType>(stage); - if (entry.IsBindless()) { - const Tegra::Texture::TextureHandle tex_handle = - engine.AccessConstBuffer32(stage_type, entry.GetBuffer(), entry.GetOffset()); + if constexpr (std::is_same_v<Entry, SamplerEntry>) { + if (entry.is_separated) { + const u32 buffer_1 = entry.buffer; + const u32 buffer_2 = entry.secondary_buffer; + const u32 offset_1 = entry.offset; + const u32 offset_2 = entry.secondary_offset; + const u32 handle_1 = engine.AccessConstBuffer32(stage_type, buffer_1, offset_1); + const u32 handle_2 = engine.AccessConstBuffer32(stage_type, buffer_2, offset_2); + return engine.GetTextureInfo(handle_1 | handle_2); + } + } + if (entry.is_bindless) { + const auto tex_handle = engine.AccessConstBuffer32(stage_type, entry.buffer, entry.offset); return engine.GetTextureInfo(tex_handle); } const auto& gpu_profile = engine.AccessGuestDriverProfile(); const u32 entry_offset = static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); - const u32 offset = entry.GetOffset() + entry_offset; + const u32 offset = entry.offset + entry_offset; if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) { return engine.GetStageTexture(stage_type, offset); } else { @@ -133,92 +145,144 @@ Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry } } +/// @brief Determine if an attachment to be updated has to preserve contents +/// @param is_clear True when a clear is being executed +/// @param regs 3D registers +/// @return True when the contents have to be preserved +bool HasToPreserveColorContents(bool is_clear, const Maxwell& regs) { + if (!is_clear) { + return true; + } + // First we have to make sure all clear masks are enabled. + if (!regs.clear_buffers.R || !regs.clear_buffers.G || !regs.clear_buffers.B || + !regs.clear_buffers.A) { + return true; + } + // If scissors are disabled, the whole screen is cleared + if (!regs.clear_flags.scissor) { + return false; + } + // Then we have to confirm scissor testing clears the whole image + const std::size_t index = regs.clear_buffers.RT; + const auto& scissor = regs.scissor_test[0]; + return scissor.min_x > 0 || scissor.min_y > 0 || scissor.max_x < regs.rt[index].width || + scissor.max_y < regs.rt[index].height; +} + +/// @brief Determine if an attachment to be updated has to preserve contents +/// @param is_clear True when a clear is being executed +/// @param regs 3D registers +/// @return True when the contents have to be preserved +bool HasToPreserveDepthContents(bool is_clear, const Maxwell& regs) { + // If we are not clearing, the contents have to be preserved + if (!is_clear) { + return true; + } + // For depth stencil clears we only have to confirm scissor test covers the whole image + if (!regs.clear_flags.scissor) { + return false; + } + // Make sure the clear cover the whole image + const auto& scissor = regs.scissor_test[0]; + return scissor.min_x > 0 || scissor.min_y > 0 || scissor.max_x < regs.zeta_width || + scissor.max_y < regs.zeta_height; +} + +template <std::size_t N> +std::array<VkDeviceSize, N> ExpandStrides(const std::array<u16, N>& strides) { + std::array<VkDeviceSize, N> expanded; + std::copy(strides.begin(), strides.end(), expanded.begin()); + return expanded; +} + } // Anonymous namespace class BufferBindings final { public: - void AddVertexBinding(const VkBuffer* buffer, VkDeviceSize offset) { - vertex.buffer_ptrs[vertex.num_buffers] = buffer; + void AddVertexBinding(VkBuffer buffer, VkDeviceSize offset, VkDeviceSize size, u32 stride) { + vertex.buffers[vertex.num_buffers] = buffer; vertex.offsets[vertex.num_buffers] = offset; + vertex.sizes[vertex.num_buffers] = size; + vertex.strides[vertex.num_buffers] = static_cast<u16>(stride); ++vertex.num_buffers; } - void SetIndexBinding(const VkBuffer* buffer, VkDeviceSize offset, VkIndexType type) { + void SetIndexBinding(VkBuffer buffer, VkDeviceSize offset, VkIndexType type) { index.buffer = buffer; index.offset = offset; index.type = type; } - void Bind(VKScheduler& scheduler) const { + void Bind(const VKDevice& device, VKScheduler& scheduler) const { // Use this large switch case to avoid dispatching more memory in the record lambda than // what we need. It looks horrible, but it's the best we can do on standard C++. switch (vertex.num_buffers) { case 0: - return BindStatic<0>(scheduler); + return BindStatic<0>(device, scheduler); case 1: - return BindStatic<1>(scheduler); + return BindStatic<1>(device, scheduler); case 2: - return BindStatic<2>(scheduler); + return BindStatic<2>(device, scheduler); case 3: - return BindStatic<3>(scheduler); + return BindStatic<3>(device, scheduler); case 4: - return BindStatic<4>(scheduler); + return BindStatic<4>(device, scheduler); case 5: - return BindStatic<5>(scheduler); + return BindStatic<5>(device, scheduler); case 6: - return BindStatic<6>(scheduler); + return BindStatic<6>(device, scheduler); case 7: - return BindStatic<7>(scheduler); + return BindStatic<7>(device, scheduler); case 8: - return BindStatic<8>(scheduler); + return BindStatic<8>(device, scheduler); case 9: - return BindStatic<9>(scheduler); + return BindStatic<9>(device, scheduler); case 10: - return BindStatic<10>(scheduler); + return BindStatic<10>(device, scheduler); case 11: - return BindStatic<11>(scheduler); + return BindStatic<11>(device, scheduler); case 12: - return BindStatic<12>(scheduler); + return BindStatic<12>(device, scheduler); case 13: - return BindStatic<13>(scheduler); + return BindStatic<13>(device, scheduler); case 14: - return BindStatic<14>(scheduler); + return BindStatic<14>(device, scheduler); case 15: - return BindStatic<15>(scheduler); + return BindStatic<15>(device, scheduler); case 16: - return BindStatic<16>(scheduler); + return BindStatic<16>(device, scheduler); case 17: - return BindStatic<17>(scheduler); + return BindStatic<17>(device, scheduler); case 18: - return BindStatic<18>(scheduler); + return BindStatic<18>(device, scheduler); case 19: - return BindStatic<19>(scheduler); + return BindStatic<19>(device, scheduler); case 20: - return BindStatic<20>(scheduler); + return BindStatic<20>(device, scheduler); case 21: - return BindStatic<21>(scheduler); + return BindStatic<21>(device, scheduler); case 22: - return BindStatic<22>(scheduler); + return BindStatic<22>(device, scheduler); case 23: - return BindStatic<23>(scheduler); + return BindStatic<23>(device, scheduler); case 24: - return BindStatic<24>(scheduler); + return BindStatic<24>(device, scheduler); case 25: - return BindStatic<25>(scheduler); + return BindStatic<25>(device, scheduler); case 26: - return BindStatic<26>(scheduler); + return BindStatic<26>(device, scheduler); case 27: - return BindStatic<27>(scheduler); + return BindStatic<27>(device, scheduler); case 28: - return BindStatic<28>(scheduler); + return BindStatic<28>(device, scheduler); case 29: - return BindStatic<29>(scheduler); + return BindStatic<29>(device, scheduler); case 30: - return BindStatic<30>(scheduler); + return BindStatic<30>(device, scheduler); case 31: - return BindStatic<31>(scheduler); + return BindStatic<31>(device, scheduler); case 32: - return BindStatic<32>(scheduler); + return BindStatic<32>(device, scheduler); } UNREACHABLE(); } @@ -227,26 +291,36 @@ private: // Some of these fields are intentionally left uninitialized to avoid initializing them twice. struct { std::size_t num_buffers = 0; - std::array<const VkBuffer*, Maxwell::NumVertexArrays> buffer_ptrs; + std::array<VkBuffer, Maxwell::NumVertexArrays> buffers; std::array<VkDeviceSize, Maxwell::NumVertexArrays> offsets; + std::array<VkDeviceSize, Maxwell::NumVertexArrays> sizes; + std::array<u16, Maxwell::NumVertexArrays> strides; } vertex; struct { - const VkBuffer* buffer = nullptr; + VkBuffer buffer = nullptr; VkDeviceSize offset; VkIndexType type; } index; template <std::size_t N> - void BindStatic(VKScheduler& scheduler) const { - if (index.buffer != nullptr) { - BindStatic<N, true>(scheduler); + void BindStatic(const VKDevice& device, VKScheduler& scheduler) const { + if (device.IsExtExtendedDynamicStateSupported()) { + if (index.buffer) { + BindStatic<N, true, true>(scheduler); + } else { + BindStatic<N, false, true>(scheduler); + } } else { - BindStatic<N, false>(scheduler); + if (index.buffer) { + BindStatic<N, true, false>(scheduler); + } else { + BindStatic<N, false, false>(scheduler); + } } } - template <std::size_t N, bool is_indexed> + template <std::size_t N, bool is_indexed, bool has_extended_dynamic_state> void BindStatic(VKScheduler& scheduler) const { static_assert(N <= Maxwell::NumVertexArrays); if constexpr (N == 0) { @@ -254,18 +328,39 @@ private: } std::array<VkBuffer, N> buffers; - std::transform(vertex.buffer_ptrs.begin(), vertex.buffer_ptrs.begin() + N, buffers.begin(), - [](const auto ptr) { return *ptr; }); - std::array<VkDeviceSize, N> offsets; + std::copy(vertex.buffers.begin(), vertex.buffers.begin() + N, buffers.begin()); std::copy(vertex.offsets.begin(), vertex.offsets.begin() + N, offsets.begin()); + if constexpr (has_extended_dynamic_state) { + // With extended dynamic states we can specify the length and stride of a vertex buffer + std::array<VkDeviceSize, N> sizes; + std::array<u16, N> strides; + std::copy(vertex.sizes.begin(), vertex.sizes.begin() + N, sizes.begin()); + std::copy(vertex.strides.begin(), vertex.strides.begin() + N, strides.begin()); + + if constexpr (is_indexed) { + scheduler.Record( + [buffers, offsets, sizes, strides, index = index](vk::CommandBuffer cmdbuf) { + cmdbuf.BindIndexBuffer(index.buffer, index.offset, index.type); + cmdbuf.BindVertexBuffers2EXT(0, static_cast<u32>(N), buffers.data(), + offsets.data(), sizes.data(), + ExpandStrides(strides).data()); + }); + } else { + scheduler.Record([buffers, offsets, sizes, strides](vk::CommandBuffer cmdbuf) { + cmdbuf.BindVertexBuffers2EXT(0, static_cast<u32>(N), buffers.data(), + offsets.data(), sizes.data(), + ExpandStrides(strides).data()); + }); + } + return; + } + if constexpr (is_indexed) { // Indexed draw - scheduler.Record([buffers, offsets, index_buffer = *index.buffer, - index_offset = index.offset, - index_type = index.type](vk::CommandBuffer cmdbuf) { - cmdbuf.BindIndexBuffer(index_buffer, index_offset, index_type); + scheduler.Record([buffers, offsets, index = index](vk::CommandBuffer cmdbuf) { + cmdbuf.BindIndexBuffer(index.buffer, index.offset, index.type); cmdbuf.BindVertexBuffers(0, static_cast<u32>(N), buffers.data(), offsets.data()); }); } else { @@ -285,25 +380,32 @@ void RasterizerVulkan::DrawParameters::Draw(vk::CommandBuffer cmdbuf) const { } } -RasterizerVulkan::RasterizerVulkan(Core::System& system, Core::Frontend::EmuWindow& renderer, - VKScreenInfo& screen_info, const VKDevice& device, - VKResourceManager& resource_manager, - VKMemoryManager& memory_manager, StateTracker& state_tracker, - VKScheduler& scheduler) - : RasterizerAccelerated{system.Memory()}, system{system}, render_window{renderer}, - screen_info{screen_info}, device{device}, resource_manager{resource_manager}, - memory_manager{memory_manager}, state_tracker{state_tracker}, scheduler{scheduler}, - staging_pool(device, memory_manager, scheduler), descriptor_pool(device), - update_descriptor_queue(device, scheduler), renderpass_cache(device), +RasterizerVulkan::RasterizerVulkan(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu_, + Tegra::MemoryManager& gpu_memory_, + Core::Memory::Memory& cpu_memory, VKScreenInfo& screen_info_, + const VKDevice& device_, VKMemoryManager& memory_manager_, + StateTracker& state_tracker_, VKScheduler& scheduler_) + : RasterizerAccelerated(cpu_memory), gpu(gpu_), gpu_memory(gpu_memory_), + maxwell3d(gpu.Maxwell3D()), kepler_compute(gpu.KeplerCompute()), screen_info(screen_info_), + device(device_), memory_manager(memory_manager_), state_tracker(state_tracker_), + scheduler(scheduler_), staging_pool(device, memory_manager, scheduler), + descriptor_pool(device, scheduler_), update_descriptor_queue(device, scheduler), + renderpass_cache(device), quad_array_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue), + quad_indexed_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue), uint8_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue), - texture_cache(system, *this, device, resource_manager, memory_manager, scheduler, - staging_pool), - pipeline_cache(system, *this, device, scheduler, descriptor_pool, update_descriptor_queue, - renderpass_cache), - buffer_cache(*this, system, device, memory_manager, scheduler, staging_pool), - sampler_cache(device), query_cache(system, *this, device, scheduler) { + texture_cache(*this, maxwell3d, gpu_memory, device, memory_manager, scheduler, staging_pool), + pipeline_cache(*this, gpu, maxwell3d, kepler_compute, gpu_memory, device, scheduler, + descriptor_pool, update_descriptor_queue, renderpass_cache), + buffer_cache(*this, gpu_memory, cpu_memory, device, memory_manager, scheduler, staging_pool), + sampler_cache(device), query_cache(*this, maxwell3d, gpu_memory, device, scheduler), + fence_manager(*this, gpu, gpu_memory, texture_cache, buffer_cache, query_cache, device, + scheduler), + wfi_event(device.GetLogical().CreateEvent()), async_shaders(emu_window) { scheduler.SetQueryCache(query_cache); + if (device.UseAsynchronousShaders()) { + async_shaders.AllocateWorkers(); + } } RasterizerVulkan::~RasterizerVulkan() = default; @@ -311,12 +413,13 @@ RasterizerVulkan::~RasterizerVulkan() = default; void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) { MICROPROFILE_SCOPE(Vulkan_Drawing); + SCOPE_EXIT({ gpu.TickWork(); }); FlushWork(); query_cache.UpdateCounters(); - const auto& gpu = system.GPU().Maxwell3D(); - GraphicsPipelineCacheKey key{GetFixedPipelineState(gpu.regs)}; + GraphicsPipelineCacheKey key; + key.fixed_state.Fill(maxwell3d.regs, device.IsExtExtendedDynamicStateSupported()); buffer_cache.Map(CalculateGraphicsStreamBufferSize(is_indexed)); @@ -334,31 +437,32 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) { buffer_cache.Unmap(); - const auto texceptions = UpdateAttachments(); + const Texceptions texceptions = UpdateAttachments(false); SetupImageTransitions(texceptions, color_attachments, zeta_attachment); key.renderpass_params = GetRenderPassParams(texceptions); + key.padding = 0; + + auto* pipeline = pipeline_cache.GetGraphicsPipeline(key, async_shaders); + if (pipeline == nullptr || pipeline->GetHandle() == VK_NULL_HANDLE) { + // Async graphics pipeline was not ready. + return; + } - auto& pipeline = pipeline_cache.GetGraphicsPipeline(key); - scheduler.BindGraphicsPipeline(pipeline.GetHandle()); + scheduler.BindGraphicsPipeline(pipeline->GetHandle()); - const auto renderpass = pipeline.GetRenderPass(); + const auto renderpass = pipeline->GetRenderPass(); const auto [framebuffer, render_area] = ConfigureFramebuffers(renderpass); scheduler.RequestRenderpass(renderpass, framebuffer, render_area); UpdateDynamicStates(); - buffer_bindings.Bind(scheduler); - - if (device.IsNvDeviceDiagnosticCheckpoints()) { - scheduler.Record( - [&pipeline](vk::CommandBuffer cmdbuf) { cmdbuf.SetCheckpointNV(&pipeline); }); - } + buffer_bindings.Bind(device, scheduler); BeginTransformFeedback(); - const auto pipeline_layout = pipeline.GetLayout(); - const auto descriptor_set = pipeline.CommitDescriptorSet(); + const auto pipeline_layout = pipeline->GetLayout(); + const auto descriptor_set = pipeline->CommitDescriptorSet(); scheduler.Record([pipeline_layout, descriptor_set, draw_params](vk::CommandBuffer cmdbuf) { if (descriptor_set) { cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline_layout, @@ -373,8 +477,7 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) { void RasterizerVulkan::Clear() { MICROPROFILE_SCOPE(Vulkan_Clearing); - const auto& gpu = system.GPU().Maxwell3D(); - if (!system.GPU().Maxwell3D().ShouldExecute()) { + if (!maxwell3d.ShouldExecute()) { return; } @@ -383,7 +486,7 @@ void RasterizerVulkan::Clear() { query_cache.UpdateCounters(); - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; const bool use_color = regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B || regs.clear_buffers.A; const bool use_depth = regs.clear_buffers.Z; @@ -392,7 +495,7 @@ void RasterizerVulkan::Clear() { return; } - [[maybe_unused]] const auto texceptions = UpdateAttachments(); + [[maybe_unused]] const auto texceptions = UpdateAttachments(true); DEBUG_ASSERT(texceptions.none()); SetupImageTransitions(0, color_attachments, zeta_attachment); @@ -413,10 +516,11 @@ void RasterizerVulkan::Clear() { const u32 color_attachment = regs.clear_buffers.RT; scheduler.Record([color_attachment, clear_value, clear_rect](vk::CommandBuffer cmdbuf) { - VkClearAttachment attachment; - attachment.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT; - attachment.colorAttachment = color_attachment; - attachment.clearValue = clear_value; + const VkClearAttachment attachment{ + .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT, + .colorAttachment = color_attachment, + .clearValue = clear_value, + }; cmdbuf.ClearAttachments(attachment, clear_rect); }); } @@ -434,10 +538,6 @@ void RasterizerVulkan::Clear() { scheduler.Record([clear_depth = regs.clear_depth, clear_stencil = regs.clear_stencil, clear_rect, aspect_flags](vk::CommandBuffer cmdbuf) { - VkClearValue clear_value; - clear_value.depthStencil.depth = clear_depth; - clear_value.depthStencil.stencil = clear_stencil; - VkClearAttachment attachment; attachment.aspectMask = aspect_flags; attachment.colorAttachment = 0; @@ -455,12 +555,17 @@ void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) { query_cache.UpdateCounters(); - const auto& launch_desc = system.GPU().KeplerCompute().launch_description; - const ComputePipelineCacheKey key{ - code_addr, - launch_desc.shared_alloc, - {launch_desc.block_dim_x, launch_desc.block_dim_y, launch_desc.block_dim_z}}; - auto& pipeline = pipeline_cache.GetComputePipeline(key); + const auto& launch_desc = kepler_compute.launch_description; + auto& pipeline = pipeline_cache.GetComputePipeline({ + .shader = code_addr, + .shared_memory_size = launch_desc.shared_alloc, + .workgroup_size = + { + launch_desc.block_dim_x, + launch_desc.block_dim_y, + launch_desc.block_dim_z, + }, + }); // Compute dispatches can't be executed inside a renderpass scheduler.RequestOutsideRenderPassOperationContext(); @@ -470,8 +575,9 @@ void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) { const auto& entries = pipeline.GetEntries(); SetupComputeConstBuffers(entries); SetupComputeGlobalBuffers(entries); - SetupComputeTexelBuffers(entries); + SetupComputeUniformTexels(entries); SetupComputeTextures(entries); + SetupComputeStorageTexels(entries); SetupComputeImages(entries); buffer_cache.Unmap(); @@ -481,11 +587,6 @@ void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) { TransitionImages(image_views, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT); - if (device.IsNvDeviceDiagnosticCheckpoints()) { - scheduler.Record( - [&pipeline](vk::CommandBuffer cmdbuf) { cmdbuf.SetCheckpointNV(nullptr); }); - } - scheduler.Record([grid_x = launch_desc.grid_dim_x, grid_y = launch_desc.grid_dim_y, grid_z = launch_desc.grid_dim_z, pipeline_handle = pipeline.GetHandle(), layout = pipeline.GetLayout(), @@ -517,6 +618,13 @@ void RasterizerVulkan::FlushRegion(VAddr addr, u64 size) { query_cache.FlushRegion(addr, size); } +bool RasterizerVulkan::MustFlushRegion(VAddr addr, u64 size) { + if (!Settings::IsGPULevelHigh()) { + return buffer_cache.MustFlushRegion(addr, size); + } + return texture_cache.MustFlushRegion(addr, size) || buffer_cache.MustFlushRegion(addr, size); +} + void RasterizerVulkan::InvalidateRegion(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; @@ -527,11 +635,71 @@ void RasterizerVulkan::InvalidateRegion(VAddr addr, u64 size) { query_cache.InvalidateRegion(addr, size); } +void RasterizerVulkan::OnCPUWrite(VAddr addr, u64 size) { + if (addr == 0 || size == 0) { + return; + } + texture_cache.OnCPUWrite(addr, size); + pipeline_cache.OnCPUWrite(addr, size); + buffer_cache.OnCPUWrite(addr, size); +} + +void RasterizerVulkan::SyncGuestHost() { + texture_cache.SyncGuestHost(); + buffer_cache.SyncGuestHost(); + pipeline_cache.SyncGuestHost(); +} + +void RasterizerVulkan::SignalSemaphore(GPUVAddr addr, u32 value) { + if (!gpu.IsAsync()) { + gpu_memory.Write<u32>(addr, value); + return; + } + fence_manager.SignalSemaphore(addr, value); +} + +void RasterizerVulkan::SignalSyncPoint(u32 value) { + if (!gpu.IsAsync()) { + gpu.IncrementSyncPoint(value); + return; + } + fence_manager.SignalSyncPoint(value); +} + +void RasterizerVulkan::ReleaseFences() { + if (!gpu.IsAsync()) { + return; + } + fence_manager.WaitPendingFences(); +} + void RasterizerVulkan::FlushAndInvalidateRegion(VAddr addr, u64 size) { - FlushRegion(addr, size); + if (Settings::IsGPULevelExtreme()) { + FlushRegion(addr, size); + } InvalidateRegion(addr, size); } +void RasterizerVulkan::WaitForIdle() { + // Everything but wait pixel operations. This intentionally includes FRAGMENT_SHADER_BIT because + // fragment shaders can still write storage buffers. + VkPipelineStageFlags flags = + VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT | VK_PIPELINE_STAGE_VERTEX_INPUT_BIT | + VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT | + VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT | + VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT | VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | + VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | VK_PIPELINE_STAGE_TRANSFER_BIT; + if (device.IsExtTransformFeedbackSupported()) { + flags |= VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT; + } + + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([event = *wfi_event, flags](vk::CommandBuffer cmdbuf) { + cmdbuf.SetEvent(event, flags); + cmdbuf.WaitEvents(event, flags, VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT, {}, {}, {}); + }); +} + void RasterizerVulkan::FlushCommands() { if (draw_counter > 0) { draw_counter = 0; @@ -576,10 +744,6 @@ bool RasterizerVulkan::AccelerateDisplay(const Tegra::FramebufferConfig& config, return true; } -void RasterizerVulkan::SetupDirtyFlags() { - state_tracker.Initialize(); -} - void RasterizerVulkan::FlushWork() { static constexpr u32 DRAWS_TO_DISPATCH = 4096; @@ -601,9 +765,11 @@ void RasterizerVulkan::FlushWork() { draw_counter = 0; } -RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments() { +RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments(bool is_clear) { MICROPROFILE_SCOPE(Vulkan_RenderTargets); - auto& dirty = system.GPU().Maxwell3D().dirty.flags; + + const auto& regs = maxwell3d.regs; + auto& dirty = maxwell3d.dirty.flags; const bool update_rendertargets = dirty[VideoCommon::Dirty::RenderTargets]; dirty[VideoCommon::Dirty::RenderTargets] = false; @@ -612,7 +778,8 @@ RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments() { Texceptions texceptions; for (std::size_t rt = 0; rt < Maxwell::NumRenderTargets; ++rt) { if (update_rendertargets) { - color_attachments[rt] = texture_cache.GetColorBufferSurface(rt); + const bool preserve_contents = HasToPreserveColorContents(is_clear, regs); + color_attachments[rt] = texture_cache.GetColorBufferSurface(rt, preserve_contents); } if (color_attachments[rt] && WalkAttachmentOverlaps(*color_attachments[rt])) { texceptions[rt] = true; @@ -620,7 +787,8 @@ RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments() { } if (update_rendertargets) { - zeta_attachment = texture_cache.GetDepthBufferSurface(); + const bool preserve_contents = HasToPreserveDepthContents(is_clear, regs); + zeta_attachment = texture_cache.GetDepthBufferSurface(preserve_contents); } if (zeta_attachment && WalkAttachmentOverlaps(*zeta_attachment)) { texceptions[ZETA_TEXCEPTION_INDEX] = true; @@ -645,21 +813,28 @@ bool RasterizerVulkan::WalkAttachmentOverlaps(const CachedSurfaceView& attachmen std::tuple<VkFramebuffer, VkExtent2D> RasterizerVulkan::ConfigureFramebuffers( VkRenderPass renderpass) { - FramebufferCacheKey key{renderpass, std::numeric_limits<u32>::max(), - std::numeric_limits<u32>::max(), std::numeric_limits<u32>::max()}; + FramebufferCacheKey key{ + .renderpass = renderpass, + .width = std::numeric_limits<u32>::max(), + .height = std::numeric_limits<u32>::max(), + .layers = std::numeric_limits<u32>::max(), + .views = {}, + }; - const auto try_push = [&](const View& view) { + const auto try_push = [&key](const View& view) { if (!view) { return false; } - key.views.push_back(view->GetHandle()); + key.views.push_back(view->GetAttachment()); key.width = std::min(key.width, view->GetWidth()); key.height = std::min(key.height, view->GetHeight()); key.layers = std::min(key.layers, view->GetNumLayers()); return true; }; - for (std::size_t index = 0; index < std::size(color_attachments); ++index) { + const auto& regs = maxwell3d.regs; + const std::size_t num_attachments = static_cast<std::size_t>(regs.rt_control.count); + for (std::size_t index = 0; index < num_attachments; ++index) { if (try_push(color_attachments[index])) { texture_cache.MarkColorBufferInUse(index); } @@ -671,17 +846,17 @@ std::tuple<VkFramebuffer, VkExtent2D> RasterizerVulkan::ConfigureFramebuffers( const auto [fbentry, is_cache_miss] = framebuffer_cache.try_emplace(key); auto& framebuffer = fbentry->second; if (is_cache_miss) { - VkFramebufferCreateInfo framebuffer_ci; - framebuffer_ci.sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO; - framebuffer_ci.pNext = nullptr; - framebuffer_ci.flags = 0; - framebuffer_ci.renderPass = key.renderpass; - framebuffer_ci.attachmentCount = static_cast<u32>(key.views.size()); - framebuffer_ci.pAttachments = key.views.data(); - framebuffer_ci.width = key.width; - framebuffer_ci.height = key.height; - framebuffer_ci.layers = key.layers; - framebuffer = device.GetLogical().CreateFramebuffer(framebuffer_ci); + framebuffer = device.GetLogical().CreateFramebuffer({ + .sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .renderPass = key.renderpass, + .attachmentCount = static_cast<u32>(key.views.size()), + .pAttachments = key.views.data(), + .width = key.width, + .height = key.height, + .layers = key.layers, + }); } return {*framebuffer, VkExtent2D{key.width, key.height}}; @@ -693,13 +868,12 @@ RasterizerVulkan::DrawParameters RasterizerVulkan::SetupGeometry(FixedPipelineSt bool is_instanced) { MICROPROFILE_SCOPE(Vulkan_Geometry); - const auto& gpu = system.GPU().Maxwell3D(); - const auto& regs = gpu.regs; + const auto& regs = maxwell3d.regs; - SetupVertexArrays(fixed_state.vertex_input, buffer_bindings); + SetupVertexArrays(buffer_bindings); const u32 base_instance = regs.vb_base_instance; - const u32 num_instances = is_instanced ? gpu.mme_draw.instance_count : 1; + const u32 num_instances = is_instanced ? maxwell3d.mme_draw.instance_count : 1; const u32 base_vertex = is_indexed ? regs.vb_element_base : regs.vertex_buffer.first; const u32 num_vertices = is_indexed ? regs.index_array.count : regs.vertex_buffer.count; @@ -710,20 +884,21 @@ RasterizerVulkan::DrawParameters RasterizerVulkan::SetupGeometry(FixedPipelineSt } void RasterizerVulkan::SetupShaderDescriptors( - const std::array<Shader, Maxwell::MaxShaderProgram>& shaders) { + const std::array<Shader*, Maxwell::MaxShaderProgram>& shaders) { texture_cache.GuardSamplers(true); for (std::size_t stage = 0; stage < Maxwell::MaxShaderStage; ++stage) { // Skip VertexA stage - const auto& shader = shaders[stage + 1]; + Shader* const shader = shaders[stage + 1]; if (!shader) { continue; } const auto& entries = shader->GetEntries(); SetupGraphicsConstBuffers(entries, stage); SetupGraphicsGlobalBuffers(entries, stage); - SetupGraphicsTexelBuffers(entries, stage); + SetupGraphicsUniformTexels(entries, stage); SetupGraphicsTextures(entries, stage); + SetupGraphicsStorageTexels(entries, stage); SetupGraphicsImages(entries, stage); } texture_cache.GuardSamplers(false); @@ -759,20 +934,34 @@ void RasterizerVulkan::SetupImageTransitions( } void RasterizerVulkan::UpdateDynamicStates() { - auto& regs = system.GPU().Maxwell3D().regs; + auto& regs = maxwell3d.regs; UpdateViewportsState(regs); UpdateScissorsState(regs); UpdateDepthBias(regs); UpdateBlendConstants(regs); UpdateDepthBounds(regs); UpdateStencilFaces(regs); + if (device.IsExtExtendedDynamicStateSupported()) { + UpdateCullMode(regs); + UpdateDepthBoundsTestEnable(regs); + UpdateDepthTestEnable(regs); + UpdateDepthWriteEnable(regs); + UpdateDepthCompareOp(regs); + UpdateFrontFace(regs); + UpdateStencilOp(regs); + UpdateStencilTestEnable(regs); + } } void RasterizerVulkan::BeginTransformFeedback() { - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; if (regs.tfb_enabled == 0) { return; } + if (!device.IsExtTransformFeedbackSupported()) { + LOG_ERROR(Render_Vulkan, "Transform feedbacks used but not supported"); + return; + } UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationControl) || regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationEval) || @@ -787,90 +976,92 @@ void RasterizerVulkan::BeginTransformFeedback() { UNIMPLEMENTED_IF(binding.buffer_offset != 0); const GPUVAddr gpu_addr = binding.Address(); - const std::size_t size = binding.buffer_size; - const auto [buffer, offset] = buffer_cache.UploadMemory(gpu_addr, size, 4, true); + const VkDeviceSize size = static_cast<VkDeviceSize>(binding.buffer_size); + const auto info = buffer_cache.UploadMemory(gpu_addr, size, 4, true); - scheduler.Record([buffer = *buffer, offset = offset, size](vk::CommandBuffer cmdbuf) { + scheduler.Record([buffer = info.handle, offset = info.offset, size](vk::CommandBuffer cmdbuf) { cmdbuf.BindTransformFeedbackBuffersEXT(0, 1, &buffer, &offset, &size); cmdbuf.BeginTransformFeedbackEXT(0, 0, nullptr, nullptr); }); } void RasterizerVulkan::EndTransformFeedback() { - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; if (regs.tfb_enabled == 0) { return; } + if (!device.IsExtTransformFeedbackSupported()) { + return; + } scheduler.Record( [](vk::CommandBuffer cmdbuf) { cmdbuf.EndTransformFeedbackEXT(0, 0, nullptr, nullptr); }); } -void RasterizerVulkan::SetupVertexArrays(FixedPipelineState::VertexInput& vertex_input, - BufferBindings& buffer_bindings) { - const auto& regs = system.GPU().Maxwell3D().regs; - - for (u32 index = 0; index < static_cast<u32>(Maxwell::NumVertexAttributes); ++index) { - const auto& attrib = regs.vertex_attrib_format[index]; - if (!attrib.IsValid()) { - continue; - } +void RasterizerVulkan::SetupVertexArrays(BufferBindings& buffer_bindings) { + const auto& regs = maxwell3d.regs; - const auto& buffer = regs.vertex_array[attrib.buffer]; - ASSERT(buffer.IsEnabled()); - - vertex_input.attributes[vertex_input.num_attributes++] = - FixedPipelineState::VertexAttribute(index, attrib.buffer, attrib.type, attrib.size, - attrib.offset); - } - - for (u32 index = 0; index < static_cast<u32>(Maxwell::NumVertexArrays); ++index) { + for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { const auto& vertex_array = regs.vertex_array[index]; if (!vertex_array.IsEnabled()) { continue; } - const GPUVAddr start{vertex_array.StartAddress()}; const GPUVAddr end{regs.vertex_array_limit[index].LimitAddress()}; - ASSERT(end > start); - const std::size_t size{end - start + 1}; - const auto [buffer, offset] = buffer_cache.UploadMemory(start, size); - - vertex_input.bindings[vertex_input.num_bindings++] = FixedPipelineState::VertexBinding( - index, vertex_array.stride, - regs.instanced_arrays.IsInstancingEnabled(index) ? vertex_array.divisor : 0); - buffer_bindings.AddVertexBinding(buffer, offset); + ASSERT(end >= start); + const std::size_t size = end - start; + if (size == 0) { + buffer_bindings.AddVertexBinding(DefaultBuffer(), 0, DEFAULT_BUFFER_SIZE, 0); + continue; + } + const auto info = buffer_cache.UploadMemory(start, size); + buffer_bindings.AddVertexBinding(info.handle, info.offset, size, vertex_array.stride); } } void RasterizerVulkan::SetupIndexBuffer(BufferBindings& buffer_bindings, DrawParameters& params, bool is_indexed) { - const auto& regs = system.GPU().Maxwell3D().regs; + if (params.num_vertices == 0) { + return; + } + const auto& regs = maxwell3d.regs; switch (regs.draw.topology) { - case Maxwell::PrimitiveTopology::Quads: - if (params.is_indexed) { - UNIMPLEMENTED(); - } else { + case Maxwell::PrimitiveTopology::Quads: { + if (!params.is_indexed) { const auto [buffer, offset] = quad_array_pass.Assemble(params.num_vertices, params.base_vertex); buffer_bindings.SetIndexBinding(buffer, offset, VK_INDEX_TYPE_UINT32); params.base_vertex = 0; params.num_vertices = params.num_vertices * 6 / 4; params.is_indexed = true; + break; } + const GPUVAddr gpu_addr = regs.index_array.IndexStart(); + const auto info = buffer_cache.UploadMemory(gpu_addr, CalculateIndexBufferSize()); + VkBuffer buffer = info.handle; + u64 offset = info.offset; + std::tie(buffer, offset) = quad_indexed_pass.Assemble( + regs.index_array.format, params.num_vertices, params.base_vertex, buffer, offset); + + buffer_bindings.SetIndexBinding(buffer, offset, VK_INDEX_TYPE_UINT32); + params.num_vertices = (params.num_vertices / 4) * 6; + params.base_vertex = 0; break; + } default: { if (!is_indexed) { break; } const GPUVAddr gpu_addr = regs.index_array.IndexStart(); - auto [buffer, offset] = buffer_cache.UploadMemory(gpu_addr, CalculateIndexBufferSize()); + const auto info = buffer_cache.UploadMemory(gpu_addr, CalculateIndexBufferSize()); + VkBuffer buffer = info.handle; + u64 offset = info.offset; auto format = regs.index_array.format; const bool is_uint8 = format == Maxwell::IndexFormat::UnsignedByte; if (is_uint8 && !device.IsExtIndexTypeUint8Supported()) { - std::tie(buffer, offset) = uint8_pass.Assemble(params.num_vertices, *buffer, offset); + std::tie(buffer, offset) = uint8_pass.Assemble(params.num_vertices, buffer, offset); format = Maxwell::IndexFormat::UnsignedShort; } @@ -882,8 +1073,7 @@ void RasterizerVulkan::SetupIndexBuffer(BufferBindings& buffer_bindings, DrawPar void RasterizerVulkan::SetupGraphicsConstBuffers(const ShaderEntries& entries, std::size_t stage) { MICROPROFILE_SCOPE(Vulkan_ConstBuffers); - const auto& gpu = system.GPU().Maxwell3D(); - const auto& shader_stage = gpu.state.shader_stages[stage]; + const auto& shader_stage = maxwell3d.state.shader_stages[stage]; for (const auto& entry : entries.const_buffers) { SetupConstBuffer(entry, shader_stage.const_buffers[entry.GetIndex()]); } @@ -891,8 +1081,7 @@ void RasterizerVulkan::SetupGraphicsConstBuffers(const ShaderEntries& entries, s void RasterizerVulkan::SetupGraphicsGlobalBuffers(const ShaderEntries& entries, std::size_t stage) { MICROPROFILE_SCOPE(Vulkan_GlobalBuffers); - auto& gpu{system.GPU()}; - const auto cbufs{gpu.Maxwell3D().state.shader_stages[stage]}; + const auto& cbufs{maxwell3d.state.shader_stages[stage]}; for (const auto& entry : entries.global_buffers) { const auto addr = cbufs.const_buffers[entry.GetCbufIndex()].address + entry.GetCbufOffset(); @@ -900,38 +1089,43 @@ void RasterizerVulkan::SetupGraphicsGlobalBuffers(const ShaderEntries& entries, } } -void RasterizerVulkan::SetupGraphicsTexelBuffers(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsUniformTexels(const ShaderEntries& entries, std::size_t stage) { MICROPROFILE_SCOPE(Vulkan_Textures); - const auto& gpu = system.GPU().Maxwell3D(); - for (const auto& entry : entries.texel_buffers) { - const auto image = GetTextureInfo(gpu, entry, stage).tic; - SetupTexelBuffer(image, entry); + for (const auto& entry : entries.uniform_texels) { + const auto image = GetTextureInfo(maxwell3d, entry, stage).tic; + SetupUniformTexels(image, entry); } } void RasterizerVulkan::SetupGraphicsTextures(const ShaderEntries& entries, std::size_t stage) { MICROPROFILE_SCOPE(Vulkan_Textures); - const auto& gpu = system.GPU().Maxwell3D(); for (const auto& entry : entries.samplers) { - for (std::size_t i = 0; i < entry.Size(); ++i) { - const auto texture = GetTextureInfo(gpu, entry, stage, i); + for (std::size_t i = 0; i < entry.size; ++i) { + const auto texture = GetTextureInfo(maxwell3d, entry, stage, i); SetupTexture(texture, entry); } } } +void RasterizerVulkan::SetupGraphicsStorageTexels(const ShaderEntries& entries, std::size_t stage) { + MICROPROFILE_SCOPE(Vulkan_Textures); + for (const auto& entry : entries.storage_texels) { + const auto image = GetTextureInfo(maxwell3d, entry, stage).tic; + SetupStorageTexel(image, entry); + } +} + void RasterizerVulkan::SetupGraphicsImages(const ShaderEntries& entries, std::size_t stage) { MICROPROFILE_SCOPE(Vulkan_Images); - const auto& gpu = system.GPU().Maxwell3D(); for (const auto& entry : entries.images) { - const auto tic = GetTextureInfo(gpu, entry, stage).tic; + const auto tic = GetTextureInfo(maxwell3d, entry, stage).tic; SetupImage(tic, entry); } } void RasterizerVulkan::SetupComputeConstBuffers(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_ConstBuffers); - const auto& launch_desc = system.GPU().KeplerCompute().launch_description; + const auto& launch_desc = kepler_compute.launch_description; for (const auto& entry : entries.const_buffers) { const auto& config = launch_desc.const_buffer_config[entry.GetIndex()]; const std::bitset<8> mask = launch_desc.const_buffer_enable_mask.Value(); @@ -945,38 +1139,43 @@ void RasterizerVulkan::SetupComputeConstBuffers(const ShaderEntries& entries) { void RasterizerVulkan::SetupComputeGlobalBuffers(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_GlobalBuffers); - const auto cbufs{system.GPU().KeplerCompute().launch_description.const_buffer_config}; + const auto& cbufs{kepler_compute.launch_description.const_buffer_config}; for (const auto& entry : entries.global_buffers) { const auto addr{cbufs[entry.GetCbufIndex()].Address() + entry.GetCbufOffset()}; SetupGlobalBuffer(entry, addr); } } -void RasterizerVulkan::SetupComputeTexelBuffers(const ShaderEntries& entries) { +void RasterizerVulkan::SetupComputeUniformTexels(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Textures); - const auto& gpu = system.GPU().KeplerCompute(); - for (const auto& entry : entries.texel_buffers) { - const auto image = GetTextureInfo(gpu, entry, ComputeShaderIndex).tic; - SetupTexelBuffer(image, entry); + for (const auto& entry : entries.uniform_texels) { + const auto image = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex).tic; + SetupUniformTexels(image, entry); } } void RasterizerVulkan::SetupComputeTextures(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Textures); - const auto& gpu = system.GPU().KeplerCompute(); for (const auto& entry : entries.samplers) { - for (std::size_t i = 0; i < entry.Size(); ++i) { - const auto texture = GetTextureInfo(gpu, entry, ComputeShaderIndex, i); + for (std::size_t i = 0; i < entry.size; ++i) { + const auto texture = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex, i); SetupTexture(texture, entry); } } } +void RasterizerVulkan::SetupComputeStorageTexels(const ShaderEntries& entries) { + MICROPROFILE_SCOPE(Vulkan_Textures); + for (const auto& entry : entries.storage_texels) { + const auto image = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex).tic; + SetupStorageTexel(image, entry); + } +} + void RasterizerVulkan::SetupComputeImages(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Images); - const auto& gpu = system.GPU().KeplerCompute(); for (const auto& entry : entries.images) { - const auto tic = GetTextureInfo(gpu, entry, ComputeShaderIndex).tic; + const auto tic = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex).tic; SetupImage(tic, entry); } } @@ -985,8 +1184,7 @@ void RasterizerVulkan::SetupConstBuffer(const ConstBufferEntry& entry, const Tegra::Engines::ConstBufferInfo& buffer) { if (!buffer.enabled) { // Set values to zero to unbind buffers - update_descriptor_queue.AddBuffer(buffer_cache.GetEmptyBuffer(sizeof(float)), 0, - sizeof(float)); + update_descriptor_queue.AddBuffer(DefaultBuffer(), 0, DEFAULT_BUFFER_SIZE); return; } @@ -995,33 +1193,33 @@ void RasterizerVulkan::SetupConstBuffer(const ConstBufferEntry& entry, Common::AlignUp(CalculateConstBufferSize(entry, buffer), 4 * sizeof(float)); ASSERT(size <= MaxConstbufferSize); - const auto [buffer_handle, offset] = + const auto info = buffer_cache.UploadMemory(buffer.address, size, device.GetUniformBufferAlignment()); - - update_descriptor_queue.AddBuffer(buffer_handle, offset, size); + update_descriptor_queue.AddBuffer(info.handle, info.offset, size); } void RasterizerVulkan::SetupGlobalBuffer(const GlobalBufferEntry& entry, GPUVAddr address) { - auto& memory_manager{system.GPU().MemoryManager()}; - const auto actual_addr = memory_manager.Read<u64>(address); - const auto size = memory_manager.Read<u32>(address + 8); + const u64 actual_addr = gpu_memory.Read<u64>(address); + const u32 size = gpu_memory.Read<u32>(address + 8); if (size == 0) { - // Sometimes global memory pointers don't have a proper size. Upload a dummy entry because - // Vulkan doesn't like empty buffers. - constexpr std::size_t dummy_size = 4; - const auto buffer = buffer_cache.GetEmptyBuffer(dummy_size); - update_descriptor_queue.AddBuffer(buffer, 0, dummy_size); + // Sometimes global memory pointers don't have a proper size. Upload a dummy entry + // because Vulkan doesn't like empty buffers. + // Note: Do *not* use DefaultBuffer() here, storage buffers can be written breaking the + // default buffer. + static constexpr std::size_t dummy_size = 4; + const auto info = buffer_cache.GetEmptyBuffer(dummy_size); + update_descriptor_queue.AddBuffer(info.handle, info.offset, dummy_size); return; } - const auto [buffer, offset] = buffer_cache.UploadMemory( + const auto info = buffer_cache.UploadMemory( actual_addr, size, device.GetStorageBufferAlignment(), entry.IsWritten()); - update_descriptor_queue.AddBuffer(buffer, offset, size); + update_descriptor_queue.AddBuffer(info.handle, info.offset, size); } -void RasterizerVulkan::SetupTexelBuffer(const Tegra::Texture::TICEntry& tic, - const TexelBufferEntry& entry) { +void RasterizerVulkan::SetupUniformTexels(const Tegra::Texture::TICEntry& tic, + const UniformTexelEntry& entry) { const auto view = texture_cache.GetTextureSurface(tic, entry); ASSERT(view->IsBufferView()); @@ -1033,29 +1231,38 @@ void RasterizerVulkan::SetupTexture(const Tegra::Texture::FullTextureInfo& textu auto view = texture_cache.GetTextureSurface(texture.tic, entry); ASSERT(!view->IsBufferView()); - const auto image_view = view->GetHandle(texture.tic.x_source, texture.tic.y_source, - texture.tic.z_source, texture.tic.w_source); + const VkImageView image_view = view->GetImageView(texture.tic.x_source, texture.tic.y_source, + texture.tic.z_source, texture.tic.w_source); const auto sampler = sampler_cache.GetSampler(texture.tsc); update_descriptor_queue.AddSampledImage(sampler, image_view); - const auto image_layout = update_descriptor_queue.GetLastImageLayout(); + VkImageLayout* const image_layout = update_descriptor_queue.LastImageLayout(); *image_layout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL; sampled_views.push_back(ImageView{std::move(view), image_layout}); } +void RasterizerVulkan::SetupStorageTexel(const Tegra::Texture::TICEntry& tic, + const StorageTexelEntry& entry) { + const auto view = texture_cache.GetImageSurface(tic, entry); + ASSERT(view->IsBufferView()); + + update_descriptor_queue.AddTexelBuffer(view->GetBufferView()); +} + void RasterizerVulkan::SetupImage(const Tegra::Texture::TICEntry& tic, const ImageEntry& entry) { auto view = texture_cache.GetImageSurface(tic, entry); - if (entry.IsWritten()) { + if (entry.is_written) { view->MarkAsModified(texture_cache.Tick()); } UNIMPLEMENTED_IF(tic.IsBuffer()); - const auto image_view = view->GetHandle(tic.x_source, tic.y_source, tic.z_source, tic.w_source); + const VkImageView image_view = + view->GetImageView(tic.x_source, tic.y_source, tic.z_source, tic.w_source); update_descriptor_queue.AddImage(image_view); - const auto image_layout = update_descriptor_queue.GetLastImageLayout(); + VkImageLayout* const image_layout = update_descriptor_queue.LastImageLayout(); *image_layout = VK_IMAGE_LAYOUT_GENERAL; image_views.push_back(ImageView{std::move(view), image_layout}); } @@ -1150,6 +1357,107 @@ void RasterizerVulkan::UpdateStencilFaces(Tegra::Engines::Maxwell3D::Regs& regs) } } +void RasterizerVulkan::UpdateCullMode(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchCullMode()) { + return; + } + scheduler.Record( + [enabled = regs.cull_test_enabled, cull_face = regs.cull_face](vk::CommandBuffer cmdbuf) { + cmdbuf.SetCullModeEXT(enabled ? MaxwellToVK::CullFace(cull_face) : VK_CULL_MODE_NONE); + }); +} + +void RasterizerVulkan::UpdateDepthBoundsTestEnable(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchDepthBoundsTestEnable()) { + return; + } + scheduler.Record([enable = regs.depth_bounds_enable](vk::CommandBuffer cmdbuf) { + cmdbuf.SetDepthBoundsTestEnableEXT(enable); + }); +} + +void RasterizerVulkan::UpdateDepthTestEnable(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchDepthTestEnable()) { + return; + } + scheduler.Record([enable = regs.depth_test_enable](vk::CommandBuffer cmdbuf) { + cmdbuf.SetDepthTestEnableEXT(enable); + }); +} + +void RasterizerVulkan::UpdateDepthWriteEnable(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchDepthWriteEnable()) { + return; + } + scheduler.Record([enable = regs.depth_write_enabled](vk::CommandBuffer cmdbuf) { + cmdbuf.SetDepthWriteEnableEXT(enable); + }); +} + +void RasterizerVulkan::UpdateDepthCompareOp(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchDepthCompareOp()) { + return; + } + scheduler.Record([func = regs.depth_test_func](vk::CommandBuffer cmdbuf) { + cmdbuf.SetDepthCompareOpEXT(MaxwellToVK::ComparisonOp(func)); + }); +} + +void RasterizerVulkan::UpdateFrontFace(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchFrontFace()) { + return; + } + + VkFrontFace front_face = MaxwellToVK::FrontFace(regs.front_face); + if (regs.screen_y_control.triangle_rast_flip != 0) { + front_face = front_face == VK_FRONT_FACE_CLOCKWISE ? VK_FRONT_FACE_COUNTER_CLOCKWISE + : VK_FRONT_FACE_CLOCKWISE; + } + scheduler.Record( + [front_face](vk::CommandBuffer cmdbuf) { cmdbuf.SetFrontFaceEXT(front_face); }); +} + +void RasterizerVulkan::UpdateStencilOp(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchStencilOp()) { + return; + } + const Maxwell::StencilOp fail = regs.stencil_front_op_fail; + const Maxwell::StencilOp zfail = regs.stencil_front_op_zfail; + const Maxwell::StencilOp zpass = regs.stencil_front_op_zpass; + const Maxwell::ComparisonOp compare = regs.stencil_front_func_func; + if (regs.stencil_two_side_enable) { + scheduler.Record([fail, zfail, zpass, compare](vk::CommandBuffer cmdbuf) { + cmdbuf.SetStencilOpEXT(VK_STENCIL_FACE_FRONT_AND_BACK, MaxwellToVK::StencilOp(fail), + MaxwellToVK::StencilOp(zpass), MaxwellToVK::StencilOp(zfail), + MaxwellToVK::ComparisonOp(compare)); + }); + } else { + const Maxwell::StencilOp back_fail = regs.stencil_back_op_fail; + const Maxwell::StencilOp back_zfail = regs.stencil_back_op_zfail; + const Maxwell::StencilOp back_zpass = regs.stencil_back_op_zpass; + const Maxwell::ComparisonOp back_compare = regs.stencil_back_func_func; + scheduler.Record([fail, zfail, zpass, compare, back_fail, back_zfail, back_zpass, + back_compare](vk::CommandBuffer cmdbuf) { + cmdbuf.SetStencilOpEXT(VK_STENCIL_FACE_FRONT_BIT, MaxwellToVK::StencilOp(fail), + MaxwellToVK::StencilOp(zpass), MaxwellToVK::StencilOp(zfail), + MaxwellToVK::ComparisonOp(compare)); + cmdbuf.SetStencilOpEXT(VK_STENCIL_FACE_BACK_BIT, MaxwellToVK::StencilOp(back_fail), + MaxwellToVK::StencilOp(back_zpass), + MaxwellToVK::StencilOp(back_zfail), + MaxwellToVK::ComparisonOp(back_compare)); + }); + } +} + +void RasterizerVulkan::UpdateStencilTestEnable(Tegra::Engines::Maxwell3D::Regs& regs) { + if (!state_tracker.TouchStencilTestEnable()) { + return; + } + scheduler.Record([enable = regs.stencil_enable](vk::CommandBuffer cmdbuf) { + cmdbuf.SetStencilTestEnableEXT(enable); + }); +} + std::size_t RasterizerVulkan::CalculateGraphicsStreamBufferSize(bool is_indexed) const { std::size_t size = CalculateVertexArraysSize(); if (is_indexed) { @@ -1165,7 +1473,7 @@ std::size_t RasterizerVulkan::CalculateComputeStreamBufferSize() const { } std::size_t RasterizerVulkan::CalculateVertexArraysSize() const { - const auto& regs = system.GPU().Maxwell3D().regs; + const auto& regs = maxwell3d.regs; std::size_t size = 0; for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) { @@ -1174,15 +1482,14 @@ std::size_t RasterizerVulkan::CalculateVertexArraysSize() const { const GPUVAddr end{regs.vertex_array_limit[index].LimitAddress()}; DEBUG_ASSERT(end >= start); - size += (end - start + 1) * regs.vertex_array[index].enable; + size += (end - start) * regs.vertex_array[index].enable; } return size; } std::size_t RasterizerVulkan::CalculateIndexBufferSize() const { - const auto& regs = system.GPU().Maxwell3D().regs; - return static_cast<std::size_t>(regs.index_array.count) * - static_cast<std::size_t>(regs.index_array.FormatSizeInBytes()); + return static_cast<std::size_t>(maxwell3d.regs.index_array.count) * + static_cast<std::size_t>(maxwell3d.regs.index_array.FormatSizeInBytes()); } std::size_t RasterizerVulkan::CalculateConstBufferSize( @@ -1197,28 +1504,54 @@ std::size_t RasterizerVulkan::CalculateConstBufferSize( } RenderPassParams RasterizerVulkan::GetRenderPassParams(Texceptions texceptions) const { - using namespace VideoCore::Surface; + const auto& regs = maxwell3d.regs; + const std::size_t num_attachments = static_cast<std::size_t>(regs.rt_control.count); - const auto& regs = system.GPU().Maxwell3D().regs; - RenderPassParams renderpass_params; + RenderPassParams params; + params.color_formats = {}; + std::size_t color_texceptions = 0; - for (std::size_t rt = 0; rt < static_cast<std::size_t>(regs.rt_control.count); ++rt) { + std::size_t index = 0; + for (std::size_t rt = 0; rt < num_attachments; ++rt) { const auto& rendertarget = regs.rt[rt]; if (rendertarget.Address() == 0 || rendertarget.format == Tegra::RenderTargetFormat::NONE) { continue; } - renderpass_params.color_attachments.push_back(RenderPassParams::ColorAttachment{ - static_cast<u32>(rt), PixelFormatFromRenderTargetFormat(rendertarget.format), - texceptions[rt]}); + params.color_formats[index] = static_cast<u8>(rendertarget.format); + color_texceptions |= (texceptions[rt] ? 1ULL : 0ULL) << index; + ++index; } + params.num_color_attachments = static_cast<u8>(index); + params.texceptions = static_cast<u8>(color_texceptions); - renderpass_params.has_zeta = regs.zeta_enable; - if (renderpass_params.has_zeta) { - renderpass_params.zeta_pixel_format = PixelFormatFromDepthFormat(regs.zeta.format); - renderpass_params.zeta_texception = texceptions[ZETA_TEXCEPTION_INDEX]; - } + params.zeta_format = regs.zeta_enable ? static_cast<u8>(regs.zeta.format) : 0; + params.zeta_texception = texceptions[ZETA_TEXCEPTION_INDEX]; + return params; +} + +VkBuffer RasterizerVulkan::DefaultBuffer() { + if (default_buffer) { + return *default_buffer; + } + + default_buffer = device.GetLogical().CreateBuffer({ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = DEFAULT_BUFFER_SIZE, + .usage = VK_BUFFER_USAGE_TRANSFER_DST_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT | + VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }); + default_buffer_commit = memory_manager.Commit(default_buffer, false); - return renderpass_params; + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([buffer = *default_buffer](vk::CommandBuffer cmdbuf) { + cmdbuf.FillBuffer(buffer, 0, DEFAULT_BUFFER_SIZE, 0); + }); + return *default_buffer; } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h index 46037860a..237e51fa4 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.h +++ b/src/video_core/renderer_vulkan/vk_rasterizer.h @@ -14,24 +14,24 @@ #include <boost/functional/hash.hpp> #include "common/common_types.h" -#include "video_core/memory_manager.h" #include "video_core/rasterizer_accelerated.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_compute_pass.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" +#include "video_core/renderer_vulkan/vk_fence_manager.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_query_cache.h" #include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_sampler_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" #include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/shader/async_shaders.h" namespace Core { class System; @@ -105,10 +105,11 @@ struct ImageView { class RasterizerVulkan final : public VideoCore::RasterizerAccelerated { public: - explicit RasterizerVulkan(Core::System& system, Core::Frontend::EmuWindow& render_window, + explicit RasterizerVulkan(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu, + Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, VKScreenInfo& screen_info, const VKDevice& device, - VKResourceManager& resource_manager, VKMemoryManager& memory_manager, - StateTracker& state_tracker, VKScheduler& scheduler); + VKMemoryManager& memory_manager, StateTracker& state_tracker, + VKScheduler& scheduler); ~RasterizerVulkan() override; void Draw(bool is_indexed, bool is_instanced) override; @@ -118,8 +119,15 @@ public: void Query(GPUVAddr gpu_addr, VideoCore::QueryType type, std::optional<u64> timestamp) override; void FlushAll() override; void FlushRegion(VAddr addr, u64 size) override; + bool MustFlushRegion(VAddr addr, u64 size) override; void InvalidateRegion(VAddr addr, u64 size) override; + void OnCPUWrite(VAddr addr, u64 size) override; + void SyncGuestHost() override; + void SignalSemaphore(GPUVAddr addr, u32 value) override; + void SignalSyncPoint(u32 value) override; + void ReleaseFences() override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; + void WaitForIdle() override; void FlushCommands() override; void TickFrame() override; bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, @@ -127,7 +135,14 @@ public: const Tegra::Engines::Fermi2D::Config& copy_config) override; bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, u32 pixel_stride) override; - void SetupDirtyFlags() override; + + VideoCommon::Shader::AsyncShaders& GetAsyncShaders() { + return async_shaders; + } + + const VideoCommon::Shader::AsyncShaders& GetAsyncShaders() const { + return async_shaders; + } /// Maximum supported size that a constbuffer can have in bytes. static constexpr std::size_t MaxConstbufferSize = 0x10000; @@ -148,10 +163,14 @@ private: using Texceptions = std::bitset<Maxwell::NumRenderTargets + 1>; static constexpr std::size_t ZETA_TEXCEPTION_INDEX = 8; + static constexpr VkDeviceSize DEFAULT_BUFFER_SIZE = 4 * sizeof(float); void FlushWork(); - Texceptions UpdateAttachments(); + /// @brief Updates the currently bound attachments + /// @param is_clear True when the framebuffer is updated as a clear + /// @return Bitfield of attachments being used as sampled textures + Texceptions UpdateAttachments(bool is_clear); std::tuple<VkFramebuffer, VkExtent2D> ConfigureFramebuffers(VkRenderPass renderpass); @@ -160,7 +179,7 @@ private: bool is_indexed, bool is_instanced); /// Setup descriptors in the graphics pipeline. - void SetupShaderDescriptors(const std::array<Shader, Maxwell::MaxShaderProgram>& shaders); + void SetupShaderDescriptors(const std::array<Shader*, Maxwell::MaxShaderProgram>& shaders); void SetupImageTransitions(Texceptions texceptions, const std::array<View, Maxwell::NumRenderTargets>& color_attachments, @@ -174,8 +193,7 @@ private: bool WalkAttachmentOverlaps(const CachedSurfaceView& attachment); - void SetupVertexArrays(FixedPipelineState::VertexInput& vertex_input, - BufferBindings& buffer_bindings); + void SetupVertexArrays(BufferBindings& buffer_bindings); void SetupIndexBuffer(BufferBindings& buffer_bindings, DrawParameters& params, bool is_indexed); @@ -185,12 +203,15 @@ private: /// Setup global buffers in the graphics pipeline. void SetupGraphicsGlobalBuffers(const ShaderEntries& entries, std::size_t stage); - /// Setup texel buffers in the graphics pipeline. - void SetupGraphicsTexelBuffers(const ShaderEntries& entries, std::size_t stage); + /// Setup uniform texels in the graphics pipeline. + void SetupGraphicsUniformTexels(const ShaderEntries& entries, std::size_t stage); /// Setup textures in the graphics pipeline. void SetupGraphicsTextures(const ShaderEntries& entries, std::size_t stage); + /// Setup storage texels in the graphics pipeline. + void SetupGraphicsStorageTexels(const ShaderEntries& entries, std::size_t stage); + /// Setup images in the graphics pipeline. void SetupGraphicsImages(const ShaderEntries& entries, std::size_t stage); @@ -201,11 +222,14 @@ private: void SetupComputeGlobalBuffers(const ShaderEntries& entries); /// Setup texel buffers in the compute pipeline. - void SetupComputeTexelBuffers(const ShaderEntries& entries); + void SetupComputeUniformTexels(const ShaderEntries& entries); /// Setup textures in the compute pipeline. void SetupComputeTextures(const ShaderEntries& entries); + /// Setup storage texels in the compute pipeline. + void SetupComputeStorageTexels(const ShaderEntries& entries); + /// Setup images in the compute pipeline. void SetupComputeImages(const ShaderEntries& entries); @@ -214,10 +238,12 @@ private: void SetupGlobalBuffer(const GlobalBufferEntry& entry, GPUVAddr address); - void SetupTexelBuffer(const Tegra::Texture::TICEntry& image, const TexelBufferEntry& entry); + void SetupUniformTexels(const Tegra::Texture::TICEntry& image, const UniformTexelEntry& entry); void SetupTexture(const Tegra::Texture::FullTextureInfo& texture, const SamplerEntry& entry); + void SetupStorageTexel(const Tegra::Texture::TICEntry& tic, const StorageTexelEntry& entry); + void SetupImage(const Tegra::Texture::TICEntry& tic, const ImageEntry& entry); void UpdateViewportsState(Tegra::Engines::Maxwell3D::Regs& regs); @@ -227,6 +253,15 @@ private: void UpdateDepthBounds(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateStencilFaces(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateCullMode(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateDepthBoundsTestEnable(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateDepthTestEnable(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateDepthWriteEnable(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateDepthCompareOp(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateFrontFace(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateStencilOp(Tegra::Engines::Maxwell3D::Regs& regs); + void UpdateStencilTestEnable(Tegra::Engines::Maxwell3D::Regs& regs); + std::size_t CalculateGraphicsStreamBufferSize(bool is_indexed) const; std::size_t CalculateComputeStreamBufferSize() const; @@ -240,11 +275,15 @@ private: RenderPassParams GetRenderPassParams(Texceptions texceptions) const; - Core::System& system; - Core::Frontend::EmuWindow& render_window; + VkBuffer DefaultBuffer(); + + Tegra::GPU& gpu; + Tegra::MemoryManager& gpu_memory; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::Engines::KeplerCompute& kepler_compute; + VKScreenInfo& screen_info; const VKDevice& device; - VKResourceManager& resource_manager; VKMemoryManager& memory_manager; StateTracker& state_tracker; VKScheduler& scheduler; @@ -254,6 +293,7 @@ private: VKUpdateDescriptorQueue update_descriptor_queue; VKRenderPassCache renderpass_cache; QuadArrayPass quad_array_pass; + QuadIndexedPass quad_indexed_pass; Uint8Pass uint8_pass; VKTextureCache texture_cache; @@ -261,6 +301,12 @@ private: VKBufferCache buffer_cache; VKSamplerCache sampler_cache; VKQueryCache query_cache; + VKFenceManager fence_manager; + + vk::Buffer default_buffer; + VKMemoryCommit default_buffer_commit; + vk::Event wfi_event; + VideoCommon::Shader::AsyncShaders async_shaders; std::array<View, Maxwell::NumRenderTargets> color_attachments; View zeta_attachment; diff --git a/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp b/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp index 4e5286a69..80284cf92 100644 --- a/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp @@ -2,9 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <cstring> #include <memory> #include <vector> +#include "common/cityhash.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/vk_device.h" @@ -13,6 +15,15 @@ namespace Vulkan { +std::size_t RenderPassParams::Hash() const noexcept { + const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this); + return static_cast<std::size_t>(hash); +} + +bool RenderPassParams::operator==(const RenderPassParams& rhs) const noexcept { + return std::memcmp(&rhs, this, sizeof *this) == 0; +} + VKRenderPassCache::VKRenderPassCache(const VKDevice& device) : device{device} {} VKRenderPassCache::~VKRenderPassCache() = default; @@ -27,72 +38,86 @@ VkRenderPass VKRenderPassCache::GetRenderPass(const RenderPassParams& params) { } vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& params) const { + using namespace VideoCore::Surface; + const std::size_t num_attachments = static_cast<std::size_t>(params.num_color_attachments); + std::vector<VkAttachmentDescription> descriptors; + descriptors.reserve(num_attachments); + std::vector<VkAttachmentReference> color_references; + color_references.reserve(num_attachments); - for (std::size_t rt = 0; rt < params.color_attachments.size(); ++rt) { - const auto attachment = params.color_attachments[rt]; - const auto format = - MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, attachment.pixel_format); + for (std::size_t rt = 0; rt < num_attachments; ++rt) { + const auto guest_format = static_cast<Tegra::RenderTargetFormat>(params.color_formats[rt]); + const PixelFormat pixel_format = PixelFormatFromRenderTargetFormat(guest_format); + const auto format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format); ASSERT_MSG(format.attachable, "Trying to attach a non-attachable format with format={}", - static_cast<u32>(attachment.pixel_format)); - - // TODO(Rodrigo): Add eMayAlias when it's needed. - const auto color_layout = attachment.is_texception - ? VK_IMAGE_LAYOUT_GENERAL - : VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL; - VkAttachmentDescription& descriptor = descriptors.emplace_back(); - descriptor.flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT; - descriptor.format = format.format; - descriptor.samples = VK_SAMPLE_COUNT_1_BIT; - descriptor.loadOp = VK_ATTACHMENT_LOAD_OP_LOAD; - descriptor.storeOp = VK_ATTACHMENT_STORE_OP_STORE; - descriptor.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE; - descriptor.stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE; - descriptor.initialLayout = color_layout; - descriptor.finalLayout = color_layout; - - VkAttachmentReference& reference = color_references.emplace_back(); - reference.attachment = static_cast<u32>(rt); - reference.layout = color_layout; + static_cast<int>(pixel_format)); + + // TODO(Rodrigo): Add MAY_ALIAS_BIT when it's needed. + const VkImageLayout color_layout = ((params.texceptions >> rt) & 1) != 0 + ? VK_IMAGE_LAYOUT_GENERAL + : VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL; + descriptors.push_back({ + .flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT, + .format = format.format, + .samples = VK_SAMPLE_COUNT_1_BIT, + .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD, + .storeOp = VK_ATTACHMENT_STORE_OP_STORE, + .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE, + .stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE, + .initialLayout = color_layout, + .finalLayout = color_layout, + }); + + color_references.push_back({ + .attachment = static_cast<u32>(rt), + .layout = color_layout, + }); } VkAttachmentReference zeta_attachment_ref; - if (params.has_zeta) { - const auto format = - MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, params.zeta_pixel_format); + const bool has_zeta = params.zeta_format != 0; + if (has_zeta) { + const auto guest_format = static_cast<Tegra::DepthFormat>(params.zeta_format); + const PixelFormat pixel_format = PixelFormatFromDepthFormat(guest_format); + const auto format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format); ASSERT_MSG(format.attachable, "Trying to attach a non-attachable format with format={}", - static_cast<u32>(params.zeta_pixel_format)); - - const auto zeta_layout = params.zeta_texception - ? VK_IMAGE_LAYOUT_GENERAL - : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL; - VkAttachmentDescription& descriptor = descriptors.emplace_back(); - descriptor.flags = 0; - descriptor.format = format.format; - descriptor.samples = VK_SAMPLE_COUNT_1_BIT; - descriptor.loadOp = VK_ATTACHMENT_LOAD_OP_LOAD; - descriptor.storeOp = VK_ATTACHMENT_STORE_OP_STORE; - descriptor.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD; - descriptor.stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE; - descriptor.initialLayout = zeta_layout; - descriptor.finalLayout = zeta_layout; - - zeta_attachment_ref.attachment = static_cast<u32>(params.color_attachments.size()); - zeta_attachment_ref.layout = zeta_layout; + static_cast<int>(pixel_format)); + + const VkImageLayout zeta_layout = params.zeta_texception != 0 + ? VK_IMAGE_LAYOUT_GENERAL + : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL; + descriptors.push_back({ + .flags = 0, + .format = format.format, + .samples = VK_SAMPLE_COUNT_1_BIT, + .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD, + .storeOp = VK_ATTACHMENT_STORE_OP_STORE, + .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD, + .stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE, + .initialLayout = zeta_layout, + .finalLayout = zeta_layout, + }); + + zeta_attachment_ref = { + .attachment = static_cast<u32>(num_attachments), + .layout = zeta_layout, + }; } - VkSubpassDescription subpass_description; - subpass_description.flags = 0; - subpass_description.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS; - subpass_description.inputAttachmentCount = 0; - subpass_description.pInputAttachments = nullptr; - subpass_description.colorAttachmentCount = static_cast<u32>(color_references.size()); - subpass_description.pColorAttachments = color_references.data(); - subpass_description.pResolveAttachments = nullptr; - subpass_description.pDepthStencilAttachment = params.has_zeta ? &zeta_attachment_ref : nullptr; - subpass_description.preserveAttachmentCount = 0; - subpass_description.pPreserveAttachments = nullptr; + const VkSubpassDescription subpass_description{ + .flags = 0, + .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, + .inputAttachmentCount = 0, + .pInputAttachments = nullptr, + .colorAttachmentCount = static_cast<u32>(color_references.size()), + .pColorAttachments = color_references.data(), + .pResolveAttachments = nullptr, + .pDepthStencilAttachment = has_zeta ? &zeta_attachment_ref : nullptr, + .preserveAttachmentCount = 0, + .pPreserveAttachments = nullptr, + }; VkAccessFlags access = 0; VkPipelineStageFlags stage = 0; @@ -101,32 +126,33 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param stage |= VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT; } - if (params.has_zeta) { + if (has_zeta) { access |= VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT; stage |= VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT; } - VkSubpassDependency subpass_dependency; - subpass_dependency.srcSubpass = VK_SUBPASS_EXTERNAL; - subpass_dependency.dstSubpass = 0; - subpass_dependency.srcStageMask = stage; - subpass_dependency.dstStageMask = stage; - subpass_dependency.srcAccessMask = 0; - subpass_dependency.dstAccessMask = access; - subpass_dependency.dependencyFlags = 0; - - VkRenderPassCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.attachmentCount = static_cast<u32>(descriptors.size()); - ci.pAttachments = descriptors.data(); - ci.subpassCount = 1; - ci.pSubpasses = &subpass_description; - ci.dependencyCount = 1; - ci.pDependencies = &subpass_dependency; - return device.GetLogical().CreateRenderPass(ci); + const VkSubpassDependency subpass_dependency{ + .srcSubpass = VK_SUBPASS_EXTERNAL, + .dstSubpass = 0, + .srcStageMask = stage, + .dstStageMask = stage, + .srcAccessMask = 0, + .dstAccessMask = access, + .dependencyFlags = 0, + }; + + return device.GetLogical().CreateRenderPass({ + .sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .attachmentCount = static_cast<u32>(descriptors.size()), + .pAttachments = descriptors.data(), + .subpassCount = 1, + .pSubpasses = &subpass_description, + .dependencyCount = 1, + .pDependencies = &subpass_dependency, + }); } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_renderpass_cache.h b/src/video_core/renderer_vulkan/vk_renderpass_cache.h index 921b6efb5..8b0fec720 100644 --- a/src/video_core/renderer_vulkan/vk_renderpass_cache.h +++ b/src/video_core/renderer_vulkan/vk_renderpass_cache.h @@ -4,8 +4,7 @@ #pragma once -#include <memory> -#include <tuple> +#include <type_traits> #include <unordered_map> #include <boost/container/static_vector.hpp> @@ -19,51 +18,25 @@ namespace Vulkan { class VKDevice; -// TODO(Rodrigo): Optimize this structure for faster hashing - struct RenderPassParams { - struct ColorAttachment { - u32 index = 0; - VideoCore::Surface::PixelFormat pixel_format = VideoCore::Surface::PixelFormat::Invalid; - bool is_texception = false; - - std::size_t Hash() const noexcept { - return static_cast<std::size_t>(pixel_format) | - static_cast<std::size_t>(is_texception) << 6 | - static_cast<std::size_t>(index) << 7; - } - - bool operator==(const ColorAttachment& rhs) const noexcept { - return std::tie(index, pixel_format, is_texception) == - std::tie(rhs.index, rhs.pixel_format, rhs.is_texception); - } - }; - - boost::container::static_vector<ColorAttachment, - Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> - color_attachments{}; - // TODO(Rodrigo): Unify has_zeta into zeta_pixel_format and zeta_component_type. - VideoCore::Surface::PixelFormat zeta_pixel_format = VideoCore::Surface::PixelFormat::Invalid; - bool has_zeta = false; - bool zeta_texception = false; - - std::size_t Hash() const noexcept { - std::size_t hash = 0; - for (const auto& rt : color_attachments) { - boost::hash_combine(hash, rt.Hash()); - } - boost::hash_combine(hash, zeta_pixel_format); - boost::hash_combine(hash, has_zeta); - boost::hash_combine(hash, zeta_texception); - return hash; - } + std::array<u8, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> color_formats; + u8 num_color_attachments; + u8 texceptions; + + u8 zeta_format; + u8 zeta_texception; + + std::size_t Hash() const noexcept; + + bool operator==(const RenderPassParams& rhs) const noexcept; - bool operator==(const RenderPassParams& rhs) const { - return std::tie(color_attachments, zeta_pixel_format, has_zeta, zeta_texception) == - std::tie(rhs.color_attachments, rhs.zeta_pixel_format, rhs.has_zeta, - rhs.zeta_texception); + bool operator!=(const RenderPassParams& rhs) const noexcept { + return !operator==(rhs); } }; +static_assert(std::has_unique_object_representations_v<RenderPassParams>); +static_assert(std::is_trivially_copyable_v<RenderPassParams>); +static_assert(std::is_trivially_constructible_v<RenderPassParams>); } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_resource_manager.cpp b/src/video_core/renderer_vulkan/vk_resource_manager.cpp deleted file mode 100644 index dc06f545a..000000000 --- a/src/video_core/renderer_vulkan/vk_resource_manager.cpp +++ /dev/null @@ -1,312 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <optional> -#include "common/assert.h" -#include "common/logging/log.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" -#include "video_core/renderer_vulkan/wrapper.h" - -namespace Vulkan { - -namespace { - -// TODO(Rodrigo): Fine tune these numbers. -constexpr std::size_t COMMAND_BUFFER_POOL_SIZE = 0x1000; -constexpr std::size_t FENCES_GROW_STEP = 0x40; - -VkFenceCreateInfo BuildFenceCreateInfo() { - VkFenceCreateInfo fence_ci; - fence_ci.sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO; - fence_ci.pNext = nullptr; - fence_ci.flags = 0; - return fence_ci; -} - -} // Anonymous namespace - -class CommandBufferPool final : public VKFencedPool { -public: - CommandBufferPool(const VKDevice& device) - : VKFencedPool(COMMAND_BUFFER_POOL_SIZE), device{device} {} - - void Allocate(std::size_t begin, std::size_t end) override { - // Command buffers are going to be commited, recorded, executed every single usage cycle. - // They are also going to be reseted when commited. - VkCommandPoolCreateInfo command_pool_ci; - command_pool_ci.sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO; - command_pool_ci.pNext = nullptr; - command_pool_ci.flags = - VK_COMMAND_POOL_CREATE_TRANSIENT_BIT | VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT; - command_pool_ci.queueFamilyIndex = device.GetGraphicsFamily(); - - Pool& pool = pools.emplace_back(); - pool.handle = device.GetLogical().CreateCommandPool(command_pool_ci); - pool.cmdbufs = pool.handle.Allocate(COMMAND_BUFFER_POOL_SIZE); - } - - VkCommandBuffer Commit(VKFence& fence) { - const std::size_t index = CommitResource(fence); - const auto pool_index = index / COMMAND_BUFFER_POOL_SIZE; - const auto sub_index = index % COMMAND_BUFFER_POOL_SIZE; - return pools[pool_index].cmdbufs[sub_index]; - } - -private: - struct Pool { - vk::CommandPool handle; - vk::CommandBuffers cmdbufs; - }; - - const VKDevice& device; - std::vector<Pool> pools; -}; - -VKResource::VKResource() = default; - -VKResource::~VKResource() = default; - -VKFence::VKFence(const VKDevice& device) - : device{device}, handle{device.GetLogical().CreateFence(BuildFenceCreateInfo())} {} - -VKFence::~VKFence() = default; - -void VKFence::Wait() { - switch (const VkResult result = handle.Wait()) { - case VK_SUCCESS: - return; - case VK_ERROR_DEVICE_LOST: - device.ReportLoss(); - [[fallthrough]]; - default: - throw vk::Exception(result); - } -} - -void VKFence::Release() { - ASSERT(is_owned); - is_owned = false; -} - -void VKFence::Commit() { - is_owned = true; - is_used = true; -} - -bool VKFence::Tick(bool gpu_wait, bool owner_wait) { - if (!is_used) { - // If a fence is not used it's always free. - return true; - } - if (is_owned && !owner_wait) { - // The fence is still being owned (Release has not been called) and ownership wait has - // not been asked. - return false; - } - - if (gpu_wait) { - // Wait for the fence if it has been requested. - (void)handle.Wait(); - } else { - if (handle.GetStatus() != VK_SUCCESS) { - // Vulkan fence is not ready, not much it can do here - return false; - } - } - - // Broadcast resources their free state. - for (auto* resource : protected_resources) { - resource->OnFenceRemoval(this); - } - protected_resources.clear(); - - // Prepare fence for reusage. - handle.Reset(); - is_used = false; - return true; -} - -void VKFence::Protect(VKResource* resource) { - protected_resources.push_back(resource); -} - -void VKFence::Unprotect(VKResource* resource) { - const auto it = std::find(protected_resources.begin(), protected_resources.end(), resource); - ASSERT(it != protected_resources.end()); - - resource->OnFenceRemoval(this); - protected_resources.erase(it); -} - -void VKFence::RedirectProtection(VKResource* old_resource, VKResource* new_resource) noexcept { - std::replace(std::begin(protected_resources), std::end(protected_resources), old_resource, - new_resource); -} - -VKFenceWatch::VKFenceWatch() = default; - -VKFenceWatch::VKFenceWatch(VKFence& initial_fence) { - Watch(initial_fence); -} - -VKFenceWatch::VKFenceWatch(VKFenceWatch&& rhs) noexcept { - fence = std::exchange(rhs.fence, nullptr); - if (fence) { - fence->RedirectProtection(&rhs, this); - } -} - -VKFenceWatch& VKFenceWatch::operator=(VKFenceWatch&& rhs) noexcept { - fence = std::exchange(rhs.fence, nullptr); - if (fence) { - fence->RedirectProtection(&rhs, this); - } - return *this; -} - -VKFenceWatch::~VKFenceWatch() { - if (fence) { - fence->Unprotect(this); - } -} - -void VKFenceWatch::Wait() { - if (fence == nullptr) { - return; - } - fence->Wait(); - fence->Unprotect(this); -} - -void VKFenceWatch::Watch(VKFence& new_fence) { - Wait(); - fence = &new_fence; - fence->Protect(this); -} - -bool VKFenceWatch::TryWatch(VKFence& new_fence) { - if (fence) { - return false; - } - fence = &new_fence; - fence->Protect(this); - return true; -} - -void VKFenceWatch::OnFenceRemoval(VKFence* signaling_fence) { - ASSERT_MSG(signaling_fence == fence, "Removing the wrong fence"); - fence = nullptr; -} - -VKFencedPool::VKFencedPool(std::size_t grow_step) : grow_step{grow_step} {} - -VKFencedPool::~VKFencedPool() = default; - -std::size_t VKFencedPool::CommitResource(VKFence& fence) { - const auto Search = [&](std::size_t begin, std::size_t end) -> std::optional<std::size_t> { - for (std::size_t iterator = begin; iterator < end; ++iterator) { - if (watches[iterator]->TryWatch(fence)) { - // The resource is now being watched, a free resource was successfully found. - return iterator; - } - } - return {}; - }; - // Try to find a free resource from the hinted position to the end. - auto found = Search(free_iterator, watches.size()); - if (!found) { - // Search from beginning to the hinted position. - found = Search(0, free_iterator); - if (!found) { - // Both searches failed, the pool is full; handle it. - const std::size_t free_resource = ManageOverflow(); - - // Watch will wait for the resource to be free. - watches[free_resource]->Watch(fence); - found = free_resource; - } - } - // Free iterator is hinted to the resource after the one that's been commited. - free_iterator = (*found + 1) % watches.size(); - return *found; -} - -std::size_t VKFencedPool::ManageOverflow() { - const std::size_t old_capacity = watches.size(); - Grow(); - - // The last entry is guaranted to be free, since it's the first element of the freshly - // allocated resources. - return old_capacity; -} - -void VKFencedPool::Grow() { - const std::size_t old_capacity = watches.size(); - watches.resize(old_capacity + grow_step); - std::generate(watches.begin() + old_capacity, watches.end(), - []() { return std::make_unique<VKFenceWatch>(); }); - Allocate(old_capacity, old_capacity + grow_step); -} - -VKResourceManager::VKResourceManager(const VKDevice& device) : device{device} { - GrowFences(FENCES_GROW_STEP); - command_buffer_pool = std::make_unique<CommandBufferPool>(device); -} - -VKResourceManager::~VKResourceManager() = default; - -VKFence& VKResourceManager::CommitFence() { - const auto StepFences = [&](bool gpu_wait, bool owner_wait) -> VKFence* { - const auto Tick = [=](auto& fence) { return fence->Tick(gpu_wait, owner_wait); }; - const auto hinted = fences.begin() + fences_iterator; - - auto it = std::find_if(hinted, fences.end(), Tick); - if (it == fences.end()) { - it = std::find_if(fences.begin(), hinted, Tick); - if (it == hinted) { - return nullptr; - } - } - fences_iterator = std::distance(fences.begin(), it) + 1; - if (fences_iterator >= fences.size()) - fences_iterator = 0; - - auto& fence = *it; - fence->Commit(); - return fence.get(); - }; - - VKFence* found_fence = StepFences(false, false); - if (!found_fence) { - // Try again, this time waiting. - found_fence = StepFences(true, false); - - if (!found_fence) { - // Allocate new fences and try again. - LOG_INFO(Render_Vulkan, "Allocating new fences {} -> {}", fences.size(), - fences.size() + FENCES_GROW_STEP); - - GrowFences(FENCES_GROW_STEP); - found_fence = StepFences(true, false); - ASSERT(found_fence != nullptr); - } - } - return *found_fence; -} - -VkCommandBuffer VKResourceManager::CommitCommandBuffer(VKFence& fence) { - return command_buffer_pool->Commit(fence); -} - -void VKResourceManager::GrowFences(std::size_t new_fences_count) { - const std::size_t previous_size = fences.size(); - fences.resize(previous_size + new_fences_count); - - std::generate(fences.begin() + previous_size, fences.end(), - [this] { return std::make_unique<VKFence>(device); }); -} - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_resource_manager.h b/src/video_core/renderer_vulkan/vk_resource_manager.h deleted file mode 100644 index f683d2276..000000000 --- a/src/video_core/renderer_vulkan/vk_resource_manager.h +++ /dev/null @@ -1,196 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <cstddef> -#include <memory> -#include <vector> -#include "video_core/renderer_vulkan/wrapper.h" - -namespace Vulkan { - -class VKDevice; -class VKFence; -class VKResourceManager; - -class CommandBufferPool; - -/// Interface for a Vulkan resource -class VKResource { -public: - explicit VKResource(); - virtual ~VKResource(); - - /** - * Signals the object that an owning fence has been signaled. - * @param signaling_fence Fence that signals its usage end. - */ - virtual void OnFenceRemoval(VKFence* signaling_fence) = 0; -}; - -/** - * Fences take ownership of objects, protecting them from GPU-side or driver-side concurrent access. - * They must be commited from the resource manager. Their usage flow is: commit the fence from the - * resource manager, protect resources with it and use them, send the fence to an execution queue - * and Wait for it if needed and then call Release. Used resources will automatically be signaled - * when they are free to be reused. - * @brief Protects resources for concurrent usage and signals its release. - */ -class VKFence { - friend class VKResourceManager; - -public: - explicit VKFence(const VKDevice& device); - ~VKFence(); - - /** - * Waits for the fence to be signaled. - * @warning You must have ownership of the fence and it has to be previously sent to a queue to - * call this function. - */ - void Wait(); - - /** - * Releases ownership of the fence. Pass after it has been sent to an execution queue. - * Unmanaged usage of the fence after the call will result in undefined behavior because it may - * be being used for something else. - */ - void Release(); - - /// Protects a resource with this fence. - void Protect(VKResource* resource); - - /// Removes protection for a resource. - void Unprotect(VKResource* resource); - - /// Redirects one protected resource to a new address. - void RedirectProtection(VKResource* old_resource, VKResource* new_resource) noexcept; - - /// Retreives the fence. - operator VkFence() const { - return *handle; - } - -private: - /// Take ownership of the fence. - void Commit(); - - /** - * Updates the fence status. - * @warning Waiting for the owner might soft lock the execution. - * @param gpu_wait Wait for the fence to be signaled by the driver. - * @param owner_wait Wait for the owner to signal its freedom. - * @returns True if the fence is free. Waiting for gpu and owner will always return true. - */ - bool Tick(bool gpu_wait, bool owner_wait); - - const VKDevice& device; ///< Device handler - vk::Fence handle; ///< Vulkan fence - std::vector<VKResource*> protected_resources; ///< List of resources protected by this fence - bool is_owned = false; ///< The fence has been commited but not released yet. - bool is_used = false; ///< The fence has been commited but it has not been checked to be free. -}; - -/** - * A fence watch is used to keep track of the usage of a fence and protect a resource or set of - * resources without having to inherit VKResource from their handlers. - */ -class VKFenceWatch final : public VKResource { -public: - explicit VKFenceWatch(); - VKFenceWatch(VKFence& initial_fence); - VKFenceWatch(VKFenceWatch&&) noexcept; - VKFenceWatch(const VKFenceWatch&) = delete; - ~VKFenceWatch() override; - - VKFenceWatch& operator=(VKFenceWatch&&) noexcept; - - /// Waits for the fence to be released. - void Wait(); - - /** - * Waits for a previous fence and watches a new one. - * @param new_fence New fence to wait to. - */ - void Watch(VKFence& new_fence); - - /** - * Checks if it's currently being watched and starts watching it if it's available. - * @returns True if a watch has started, false if it's being watched. - */ - bool TryWatch(VKFence& new_fence); - - void OnFenceRemoval(VKFence* signaling_fence) override; - - /** - * Do not use it paired with Watch. Use TryWatch instead. - * Returns true when the watch is free. - */ - bool IsUsed() const { - return fence != nullptr; - } - -private: - VKFence* fence{}; ///< Fence watching this resource. nullptr when the watch is free. -}; - -/** - * Handles a pool of resources protected by fences. Manages resource overflow allocating more - * resources. - */ -class VKFencedPool { -public: - explicit VKFencedPool(std::size_t grow_step); - virtual ~VKFencedPool(); - -protected: - /** - * Commits a free resource and protects it with a fence. It may allocate new resources. - * @param fence Fence that protects the commited resource. - * @returns Index of the resource commited. - */ - std::size_t CommitResource(VKFence& fence); - - /// Called when a chunk of resources have to be allocated. - virtual void Allocate(std::size_t begin, std::size_t end) = 0; - -private: - /// Manages pool overflow allocating new resources. - std::size_t ManageOverflow(); - - /// Allocates a new page of resources. - void Grow(); - - std::size_t grow_step = 0; ///< Number of new resources created after an overflow - std::size_t free_iterator = 0; ///< Hint to where the next free resources is likely to be found - std::vector<std::unique_ptr<VKFenceWatch>> watches; ///< Set of watched resources -}; - -/** - * The resource manager handles all resources that can be protected with a fence avoiding - * driver-side or GPU-side concurrent usage. Usage is documented in VKFence. - */ -class VKResourceManager final { -public: - explicit VKResourceManager(const VKDevice& device); - ~VKResourceManager(); - - /// Commits a fence. It has to be sent to a queue and released. - VKFence& CommitFence(); - - /// Commits an unused command buffer and protects it with a fence. - VkCommandBuffer CommitCommandBuffer(VKFence& fence); - -private: - /// Allocates new fences. - void GrowFences(std::size_t new_fences_count); - - const VKDevice& device; ///< Device handler. - std::size_t fences_iterator = 0; ///< Index where a free fence is likely to be found. - std::vector<std::unique_ptr<VKFence>> fences; ///< Pool of fences. - std::unique_ptr<CommandBufferPool> command_buffer_pool; ///< Pool of command buffers. -}; - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_resource_pool.cpp b/src/video_core/renderer_vulkan/vk_resource_pool.cpp new file mode 100644 index 000000000..ee274ac59 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_resource_pool.cpp @@ -0,0 +1,63 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <optional> + +#include "video_core/renderer_vulkan/vk_master_semaphore.h" +#include "video_core/renderer_vulkan/vk_resource_pool.h" + +namespace Vulkan { + +ResourcePool::ResourcePool(MasterSemaphore& master_semaphore_, size_t grow_step_) + : master_semaphore{master_semaphore_}, grow_step{grow_step_} {} + +ResourcePool::~ResourcePool() = default; + +size_t ResourcePool::CommitResource() { + // Refresh semaphore to query updated results + master_semaphore.Refresh(); + + const auto search = [this](size_t begin, size_t end) -> std::optional<size_t> { + for (size_t iterator = begin; iterator < end; ++iterator) { + if (master_semaphore.IsFree(ticks[iterator])) { + ticks[iterator] = master_semaphore.CurrentTick(); + return iterator; + } + } + return {}; + }; + // Try to find a free resource from the hinted position to the end. + auto found = search(free_iterator, ticks.size()); + if (!found) { + // Search from beginning to the hinted position. + found = search(0, free_iterator); + if (!found) { + // Both searches failed, the pool is full; handle it. + const size_t free_resource = ManageOverflow(); + + ticks[free_resource] = master_semaphore.CurrentTick(); + found = free_resource; + } + } + // Free iterator is hinted to the resource after the one that's been commited. + free_iterator = (*found + 1) % ticks.size(); + return *found; +} + +size_t ResourcePool::ManageOverflow() { + const size_t old_capacity = ticks.size(); + Grow(); + + // The last entry is guaranted to be free, since it's the first element of the freshly + // allocated resources. + return old_capacity; +} + +void ResourcePool::Grow() { + const size_t old_capacity = ticks.size(); + ticks.resize(old_capacity + grow_step); + Allocate(old_capacity, old_capacity + grow_step); +} + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_resource_pool.h b/src/video_core/renderer_vulkan/vk_resource_pool.h new file mode 100644 index 000000000..a018c7ec2 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_resource_pool.h @@ -0,0 +1,43 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> + +#include "common/common_types.h" + +namespace Vulkan { + +class MasterSemaphore; + +/** + * Handles a pool of resources protected by fences. Manages resource overflow allocating more + * resources. + */ +class ResourcePool { +public: + explicit ResourcePool(MasterSemaphore& master_semaphore, size_t grow_step); + virtual ~ResourcePool(); + +protected: + size_t CommitResource(); + + /// Called when a chunk of resources have to be allocated. + virtual void Allocate(size_t begin, size_t end) = 0; + +private: + /// Manages pool overflow allocating new resources. + size_t ManageOverflow(); + + /// Allocates a new page of resources. + void Grow(); + + MasterSemaphore& master_semaphore; + size_t grow_step = 0; ///< Number of new resources created after an overflow + size_t free_iterator = 0; ///< Hint to where the next free resources is likely to be found + std::vector<u64> ticks; ///< Ticks for each resource +}; + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp index 07bbcf520..b068888f9 100644 --- a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp @@ -2,16 +2,15 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> -#include <optional> #include <unordered_map> -#include "common/assert.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/vk_sampler_cache.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/textures/texture.h" +using Tegra::Texture::TextureMipmapFilter; + namespace Vulkan { namespace { @@ -42,26 +41,39 @@ VKSamplerCache::VKSamplerCache(const VKDevice& device) : device{device} {} VKSamplerCache::~VKSamplerCache() = default; vk::Sampler VKSamplerCache::CreateSampler(const Tegra::Texture::TSCEntry& tsc) const { - VkSamplerCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter); - ci.minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter); - ci.mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter); - ci.addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter); - ci.addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter); - ci.addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter); - ci.mipLodBias = tsc.GetLodBias(); - ci.anisotropyEnable = tsc.GetMaxAnisotropy() > 1.0f ? VK_TRUE : VK_FALSE; - ci.maxAnisotropy = tsc.GetMaxAnisotropy(); - ci.compareEnable = tsc.depth_compare_enabled; - ci.compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func); - ci.minLod = tsc.GetMinLod(); - ci.maxLod = tsc.GetMaxLod(); - ci.borderColor = ConvertBorderColor(tsc.GetBorderColor()); - ci.unnormalizedCoordinates = VK_FALSE; - return device.GetLogical().CreateSampler(ci); + const bool arbitrary_borders = device.IsExtCustomBorderColorSupported(); + const std::array color = tsc.GetBorderColor(); + + VkSamplerCustomBorderColorCreateInfoEXT border{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CUSTOM_BORDER_COLOR_CREATE_INFO_EXT, + .pNext = nullptr, + .customBorderColor = {}, + .format = VK_FORMAT_UNDEFINED, + }; + std::memcpy(&border.customBorderColor, color.data(), sizeof(color)); + + return device.GetLogical().CreateSampler({ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, + .pNext = arbitrary_borders ? &border : nullptr, + .flags = 0, + .magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter), + .minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter), + .mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter), + .addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter), + .addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter), + .addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter), + .mipLodBias = tsc.GetLodBias(), + .anisotropyEnable = + static_cast<VkBool32>(tsc.GetMaxAnisotropy() > 1.0f ? VK_TRUE : VK_FALSE), + .maxAnisotropy = tsc.GetMaxAnisotropy(), + .compareEnable = tsc.depth_compare_enabled, + .compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func), + .minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.GetMinLod(), + .maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.GetMaxLod(), + .borderColor = + arbitrary_borders ? VK_BORDER_COLOR_INT_CUSTOM_EXT : ConvertBorderColor(color), + .unnormalizedCoordinates = VK_FALSE, + }); } VkSampler VKSamplerCache::ToSamplerType(const vk::Sampler& sampler) const { diff --git a/src/video_core/renderer_vulkan/vk_scheduler.cpp b/src/video_core/renderer_vulkan/vk_scheduler.cpp index 900f551b3..1a483dc71 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.cpp +++ b/src/video_core/renderer_vulkan/vk_scheduler.cpp @@ -8,11 +8,12 @@ #include <thread> #include <utility> -#include "common/assert.h" #include "common/microprofile.h" +#include "common/thread.h" +#include "video_core/renderer_vulkan/vk_command_pool.h" #include "video_core/renderer_vulkan/vk_device.h" +#include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/renderer_vulkan/vk_query_cache.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_state_tracker.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -35,10 +36,10 @@ void VKScheduler::CommandChunk::ExecuteAll(vk::CommandBuffer cmdbuf) { last = nullptr; } -VKScheduler::VKScheduler(const VKDevice& device, VKResourceManager& resource_manager, - StateTracker& state_tracker) - : device{device}, resource_manager{resource_manager}, state_tracker{state_tracker}, - next_fence{&resource_manager.CommitFence()} { +VKScheduler::VKScheduler(const VKDevice& device_, StateTracker& state_tracker_) + : device{device_}, state_tracker{state_tracker_}, + master_semaphore{std::make_unique<MasterSemaphore>(device)}, + command_pool{std::make_unique<CommandPool>(*master_semaphore, device)} { AcquireNewChunk(); AllocateNewContext(); worker_thread = std::thread(&VKScheduler::WorkerThread, this); @@ -50,20 +51,27 @@ VKScheduler::~VKScheduler() { worker_thread.join(); } -void VKScheduler::Flush(bool release_fence, VkSemaphore semaphore) { +u64 VKScheduler::CurrentTick() const noexcept { + return master_semaphore->CurrentTick(); +} + +bool VKScheduler::IsFree(u64 tick) const noexcept { + return master_semaphore->IsFree(tick); +} + +void VKScheduler::Wait(u64 tick) { + master_semaphore->Wait(tick); +} + +void VKScheduler::Flush(VkSemaphore semaphore) { SubmitExecution(semaphore); - if (release_fence) { - current_fence->Release(); - } AllocateNewContext(); } -void VKScheduler::Finish(bool release_fence, VkSemaphore semaphore) { +void VKScheduler::Finish(VkSemaphore semaphore) { + const u64 presubmit_tick = CurrentTick(); SubmitExecution(semaphore); - current_fence->Wait(); - if (release_fence) { - current_fence->Release(); - } + Wait(presubmit_tick); AllocateNewContext(); } @@ -100,16 +108,19 @@ void VKScheduler::RequestRenderpass(VkRenderPass renderpass, VkFramebuffer frame state.framebuffer = framebuffer; state.render_area = render_area; - VkRenderPassBeginInfo renderpass_bi; - renderpass_bi.sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO; - renderpass_bi.pNext = nullptr; - renderpass_bi.renderPass = renderpass; - renderpass_bi.framebuffer = framebuffer; - renderpass_bi.renderArea.offset.x = 0; - renderpass_bi.renderArea.offset.y = 0; - renderpass_bi.renderArea.extent = render_area; - renderpass_bi.clearValueCount = 0; - renderpass_bi.pClearValues = nullptr; + const VkRenderPassBeginInfo renderpass_bi{ + .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO, + .pNext = nullptr, + .renderPass = renderpass, + .framebuffer = framebuffer, + .renderArea = + { + .offset = {.x = 0, .y = 0}, + .extent = render_area, + }, + .clearValueCount = 0, + .pClearValues = nullptr, + }; Record([renderpass_bi, end_renderpass](vk::CommandBuffer cmdbuf) { if (end_renderpass) { @@ -134,6 +145,7 @@ void VKScheduler::BindGraphicsPipeline(VkPipeline pipeline) { } void VKScheduler::WorkerThread() { + Common::SetCurrentThreadPriority(Common::ThreadPriority::High); std::unique_lock lock{mutex}; do { cv.wait(lock, [this] { return !chunk_queue.Empty() || quit; }); @@ -156,35 +168,58 @@ void VKScheduler::SubmitExecution(VkSemaphore semaphore) { current_cmdbuf.End(); - VkSubmitInfo submit_info; - submit_info.sType = VK_STRUCTURE_TYPE_SUBMIT_INFO; - submit_info.pNext = nullptr; - submit_info.waitSemaphoreCount = 0; - submit_info.pWaitSemaphores = nullptr; - submit_info.pWaitDstStageMask = nullptr; - submit_info.commandBufferCount = 1; - submit_info.pCommandBuffers = current_cmdbuf.address(); - submit_info.signalSemaphoreCount = semaphore ? 1 : 0; - submit_info.pSignalSemaphores = &semaphore; - device.GetGraphicsQueue().Submit(submit_info, *current_fence); + const VkSemaphore timeline_semaphore = master_semaphore->Handle(); + const u32 num_signal_semaphores = semaphore ? 2U : 1U; + + const u64 signal_value = master_semaphore->CurrentTick(); + const u64 wait_value = signal_value - 1; + const VkPipelineStageFlags wait_stage_mask = VK_PIPELINE_STAGE_ALL_COMMANDS_BIT; + + master_semaphore->NextTick(); + + const std::array signal_values{signal_value, u64(0)}; + const std::array signal_semaphores{timeline_semaphore, semaphore}; + + const VkTimelineSemaphoreSubmitInfoKHR timeline_si{ + .sType = VK_STRUCTURE_TYPE_TIMELINE_SEMAPHORE_SUBMIT_INFO_KHR, + .pNext = nullptr, + .waitSemaphoreValueCount = 1, + .pWaitSemaphoreValues = &wait_value, + .signalSemaphoreValueCount = num_signal_semaphores, + .pSignalSemaphoreValues = signal_values.data(), + }; + const VkSubmitInfo submit_info{ + .sType = VK_STRUCTURE_TYPE_SUBMIT_INFO, + .pNext = &timeline_si, + .waitSemaphoreCount = 1, + .pWaitSemaphores = &timeline_semaphore, + .pWaitDstStageMask = &wait_stage_mask, + .commandBufferCount = 1, + .pCommandBuffers = current_cmdbuf.address(), + .signalSemaphoreCount = num_signal_semaphores, + .pSignalSemaphores = signal_semaphores.data(), + }; + switch (const VkResult result = device.GetGraphicsQueue().Submit(submit_info)) { + case VK_SUCCESS: + break; + case VK_ERROR_DEVICE_LOST: + device.ReportLoss(); + [[fallthrough]]; + default: + vk::Check(result); + } } void VKScheduler::AllocateNewContext() { - ++ticks; - - VkCommandBufferBeginInfo cmdbuf_bi; - cmdbuf_bi.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO; - cmdbuf_bi.pNext = nullptr; - cmdbuf_bi.flags = VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT; - cmdbuf_bi.pInheritanceInfo = nullptr; - std::unique_lock lock{mutex}; - current_fence = next_fence; - next_fence = &resource_manager.CommitFence(); - current_cmdbuf = vk::CommandBuffer(resource_manager.CommitCommandBuffer(*current_fence), - device.GetDispatchLoader()); - current_cmdbuf.Begin(cmdbuf_bi); + current_cmdbuf = vk::CommandBuffer(command_pool->Commit(), device.GetDispatchLoader()); + current_cmdbuf.Begin({ + .sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO, + .pNext = nullptr, + .flags = VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT, + .pInheritanceInfo = nullptr, + }); // Enable counters once again. These are disabled when a command buffer is finished. if (query_cache) { diff --git a/src/video_core/renderer_vulkan/vk_scheduler.h b/src/video_core/renderer_vulkan/vk_scheduler.h index 82a8adc69..7be8a19f0 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.h +++ b/src/video_core/renderer_vulkan/vk_scheduler.h @@ -7,7 +7,6 @@ #include <atomic> #include <condition_variable> #include <memory> -#include <optional> #include <stack> #include <thread> #include <utility> @@ -17,42 +16,33 @@ namespace Vulkan { +class CommandPool; +class MasterSemaphore; class StateTracker; class VKDevice; -class VKFence; class VKQueryCache; -class VKResourceManager; - -class VKFenceView { -public: - VKFenceView() = default; - VKFenceView(VKFence* const& fence) : fence{fence} {} - - VKFence* operator->() const noexcept { - return fence; - } - - operator VKFence&() const noexcept { - return *fence; - } - -private: - VKFence* const& fence; -}; /// The scheduler abstracts command buffer and fence management with an interface that's able to do /// OpenGL-like operations on Vulkan command buffers. class VKScheduler { public: - explicit VKScheduler(const VKDevice& device, VKResourceManager& resource_manager, - StateTracker& state_tracker); + explicit VKScheduler(const VKDevice& device, StateTracker& state_tracker); ~VKScheduler(); + /// Returns the current command buffer tick. + [[nodiscard]] u64 CurrentTick() const noexcept; + + /// Returns true when a tick has been triggered by the GPU. + [[nodiscard]] bool IsFree(u64 tick) const noexcept; + + /// Waits for the given tick to trigger on the GPU. + void Wait(u64 tick); + /// Sends the current execution context to the GPU. - void Flush(bool release_fence = true, VkSemaphore semaphore = nullptr); + void Flush(VkSemaphore semaphore = nullptr); /// Sends the current execution context to the GPU and waits for it to complete. - void Finish(bool release_fence = true, VkSemaphore semaphore = nullptr); + void Finish(VkSemaphore semaphore = nullptr); /// Waits for the worker thread to finish executing everything. After this function returns it's /// safe to touch worker resources. @@ -87,14 +77,9 @@ public: (void)chunk->Record(command); } - /// Gets a reference to the current fence. - VKFenceView GetFence() const { - return current_fence; - } - - /// Returns the current command buffer tick. - u64 Ticks() const { - return ticks; + /// Returns the master timeline semaphore. + [[nodiscard]] MasterSemaphore& GetMasterSemaphore() const noexcept { + return *master_semaphore; } private: @@ -172,6 +157,13 @@ private: std::array<u8, 0x8000> data{}; }; + struct State { + VkRenderPass renderpass = nullptr; + VkFramebuffer framebuffer = nullptr; + VkExtent2D render_area = {0, 0}; + VkPipeline graphics_pipeline = nullptr; + }; + void WorkerThread(); void SubmitExecution(VkSemaphore semaphore); @@ -187,30 +179,23 @@ private: void AcquireNewChunk(); const VKDevice& device; - VKResourceManager& resource_manager; StateTracker& state_tracker; + std::unique_ptr<MasterSemaphore> master_semaphore; + std::unique_ptr<CommandPool> command_pool; + VKQueryCache* query_cache = nullptr; vk::CommandBuffer current_cmdbuf; - VKFence* current_fence = nullptr; - VKFence* next_fence = nullptr; - - struct State { - VkRenderPass renderpass = nullptr; - VkFramebuffer framebuffer = nullptr; - VkExtent2D render_area = {0, 0}; - VkPipeline graphics_pipeline = nullptr; - } state; std::unique_ptr<CommandChunk> chunk; std::thread worker_thread; + State state; Common::SPSCQueue<std::unique_ptr<CommandChunk>> chunk_queue; Common::SPSCQueue<std::unique_ptr<CommandChunk>> chunk_reserve; std::mutex mutex; std::condition_variable cv; - std::atomic<u64> ticks = 0; bool quit = false; }; diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp index aaa138f52..a20452b87 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp @@ -103,8 +103,8 @@ struct GenericVaryingDescription { }; spv::Dim GetSamplerDim(const Sampler& sampler) { - ASSERT(!sampler.IsBuffer()); - switch (sampler.GetType()) { + ASSERT(!sampler.is_buffer); + switch (sampler.type) { case Tegra::Shader::TextureType::Texture1D: return spv::Dim::Dim1D; case Tegra::Shader::TextureType::Texture2D: @@ -114,13 +114,13 @@ spv::Dim GetSamplerDim(const Sampler& sampler) { case Tegra::Shader::TextureType::TextureCube: return spv::Dim::Cube; default: - UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<u32>(sampler.GetType())); + UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<int>(sampler.type)); return spv::Dim::Dim2D; } } std::pair<spv::Dim, bool> GetImageDim(const Image& image) { - switch (image.GetType()) { + switch (image.type) { case Tegra::Shader::ImageType::Texture1D: return {spv::Dim::Dim1D, false}; case Tegra::Shader::ImageType::TextureBuffer: @@ -134,7 +134,7 @@ std::pair<spv::Dim, bool> GetImageDim(const Image& image) { case Tegra::Shader::ImageType::Texture3D: return {spv::Dim::Dim3D, false}; default: - UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<u32>(image.GetType())); + UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<int>(image.type)); return {spv::Dim::Dim2D, false}; } } @@ -272,12 +272,19 @@ bool IsPrecise(Operation operand) { return false; } +u32 ShaderVersion(const VKDevice& device) { + if (device.InstanceApiVersion() < VK_API_VERSION_1_1) { + return 0x00010000; + } + return 0x00010300; +} + class SPIRVDecompiler final : public Sirit::Module { public: explicit SPIRVDecompiler(const VKDevice& device, const ShaderIR& ir, ShaderType stage, const Registry& registry, const Specialization& specialization) - : Module(0x00010300), device{device}, ir{ir}, stage{stage}, header{ir.GetHeader()}, - registry{registry}, specialization{specialization} { + : Module(ShaderVersion(device)), device{device}, ir{ir}, stage{stage}, + header{ir.GetHeader()}, registry{registry}, specialization{specialization} { if (stage != ShaderType::Compute) { transform_feedback = BuildTransformFeedback(registry.GetGraphicsInfo()); } @@ -293,6 +300,7 @@ public: AddCapability(spv::Capability::DrawParameters); AddCapability(spv::Capability::SubgroupBallotKHR); AddCapability(spv::Capability::SubgroupVoteKHR); + AddExtension("SPV_KHR_16bit_storage"); AddExtension("SPV_KHR_shader_ballot"); AddExtension("SPV_KHR_subgroup_vote"); AddExtension("SPV_KHR_storage_buffer_storage_class"); @@ -400,8 +408,9 @@ private: u32 binding = specialization.base_binding; binding = DeclareConstantBuffers(binding); binding = DeclareGlobalBuffers(binding); - binding = DeclareTexelBuffers(binding); + binding = DeclareUniformTexels(binding); binding = DeclareSamplers(binding); + binding = DeclareStorageTexels(binding); binding = DeclareImages(binding); const Id main = OpFunction(t_void, {}, TypeFunction(t_void)); @@ -515,6 +524,16 @@ private: void DeclareCommon() { thread_id = DeclareInputBuiltIn(spv::BuiltIn::SubgroupLocalInvocationId, t_in_uint, "thread_id"); + thread_masks[0] = + DeclareInputBuiltIn(spv::BuiltIn::SubgroupEqMask, t_in_uint4, "thread_eq_mask"); + thread_masks[1] = + DeclareInputBuiltIn(spv::BuiltIn::SubgroupGeMask, t_in_uint4, "thread_ge_mask"); + thread_masks[2] = + DeclareInputBuiltIn(spv::BuiltIn::SubgroupGtMask, t_in_uint4, "thread_gt_mask"); + thread_masks[3] = + DeclareInputBuiltIn(spv::BuiltIn::SubgroupLeMask, t_in_uint4, "thread_le_mask"); + thread_masks[4] = + DeclareInputBuiltIn(spv::BuiltIn::SubgroupLtMask, t_in_uint4, "thread_lt_mask"); } void DeclareVertex() { @@ -674,13 +693,19 @@ private: } t_smem_uint = TypePointer(spv::StorageClass::Workgroup, t_uint); - const u32 smem_size = specialization.shared_memory_size; + u32 smem_size = specialization.shared_memory_size * 4; if (smem_size == 0) { // Avoid declaring an empty array. return; } - const auto element_count = static_cast<u32>(Common::AlignUp(smem_size, 4) / 4); - const Id type_array = TypeArray(t_uint, Constant(t_uint, element_count)); + const u32 limit = device.GetMaxComputeSharedMemorySize(); + if (smem_size > limit) { + LOG_ERROR(Render_Vulkan, "Shared memory size {} is clamped to host's limit {}", + smem_size, limit); + smem_size = limit; + } + + const Id type_array = TypeArray(t_uint, Constant(t_uint, smem_size / 4)); const Id type_pointer = TypePointer(spv::StorageClass::Workgroup, type_array); Name(type_pointer, "SharedMemory"); @@ -689,9 +714,9 @@ private: } void DeclareInternalFlags() { - constexpr std::array names = {"zero", "sign", "carry", "overflow"}; + static constexpr std::array names{"zero", "sign", "carry", "overflow"}; + for (std::size_t flag = 0; flag < INTERNAL_FLAGS_COUNT; ++flag) { - const auto flag_code = static_cast<InternalFlag>(flag); const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false); internal_flags[flag] = AddGlobalVariable(Name(id, names[flag])); } @@ -731,8 +756,10 @@ private: if (!IsGenericAttribute(index)) { continue; } - const u32 location = GetGenericAttributeLocation(index); + if (!IsAttributeEnabled(location)) { + continue; + } const auto type_descriptor = GetAttributeType(location); Id type; if (IsInputAttributeArray()) { @@ -877,13 +904,13 @@ private: return binding; } - u32 DeclareTexelBuffers(u32 binding) { + u32 DeclareUniformTexels(u32 binding) { for (const auto& sampler : ir.GetSamplers()) { - if (!sampler.IsBuffer()) { + if (!sampler.is_buffer) { continue; } - ASSERT(!sampler.IsArray()); - ASSERT(!sampler.IsShadow()); + ASSERT(!sampler.is_array); + ASSERT(!sampler.is_shadow); constexpr auto dim = spv::Dim::Buffer; constexpr int depth = 0; @@ -894,23 +921,23 @@ private: const Id image_type = TypeImage(t_float, dim, depth, arrayed, ms, sampled, format); const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type); const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); - AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex()))); + AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.index))); Decorate(id, spv::Decoration::Binding, binding++); Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); - texel_buffers.emplace(sampler.GetIndex(), TexelBuffer{image_type, id}); + uniform_texels.emplace(sampler.index, TexelBuffer{image_type, id}); } return binding; } u32 DeclareSamplers(u32 binding) { for (const auto& sampler : ir.GetSamplers()) { - if (sampler.IsBuffer()) { + if (sampler.is_buffer) { continue; } const auto dim = GetSamplerDim(sampler); - const int depth = sampler.IsShadow() ? 1 : 0; - const int arrayed = sampler.IsArray() ? 1 : 0; + const int depth = sampler.is_shadow ? 1 : 0; + const int arrayed = sampler.is_array ? 1 : 0; constexpr bool ms = false; constexpr int sampled = 1; constexpr auto format = spv::ImageFormat::Unknown; @@ -918,46 +945,63 @@ private: const Id sampler_type = TypeSampledImage(image_type); const Id sampler_pointer_type = TypePointer(spv::StorageClass::UniformConstant, sampler_type); - const Id type = sampler.IsIndexed() - ? TypeArray(sampler_type, Constant(t_uint, sampler.Size())) + const Id type = sampler.is_indexed + ? TypeArray(sampler_type, Constant(t_uint, sampler.size)) : sampler_type; const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, type); const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); - AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex()))); + AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.index))); Decorate(id, spv::Decoration::Binding, binding++); Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); - sampled_images.emplace(sampler.GetIndex(), SampledImage{image_type, sampler_type, - sampler_pointer_type, id}); + sampled_images.emplace( + sampler.index, SampledImage{image_type, sampler_type, sampler_pointer_type, id}); } return binding; } - u32 DeclareImages(u32 binding) { + u32 DeclareStorageTexels(u32 binding) { for (const auto& image : ir.GetImages()) { - const auto [dim, arrayed] = GetImageDim(image); - constexpr int depth = 0; - constexpr bool ms = false; - constexpr int sampled = 2; // This won't be accessed with a sampler - constexpr auto format = spv::ImageFormat::Unknown; - const Id image_type = TypeImage(t_uint, dim, depth, arrayed, ms, sampled, format, {}); - const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type); - const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); - AddGlobalVariable(Name(id, fmt::format("image_{}", image.GetIndex()))); - - Decorate(id, spv::Decoration::Binding, binding++); - Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); - if (image.IsRead() && !image.IsWritten()) { - Decorate(id, spv::Decoration::NonWritable); - } else if (image.IsWritten() && !image.IsRead()) { - Decorate(id, spv::Decoration::NonReadable); + if (image.type != Tegra::Shader::ImageType::TextureBuffer) { + continue; } + DeclareImage(image, binding); + } + return binding; + } - images.emplace(static_cast<u32>(image.GetIndex()), StorageImage{image_type, id}); + u32 DeclareImages(u32 binding) { + for (const auto& image : ir.GetImages()) { + if (image.type == Tegra::Shader::ImageType::TextureBuffer) { + continue; + } + DeclareImage(image, binding); } return binding; } + void DeclareImage(const Image& image, u32& binding) { + const auto [dim, arrayed] = GetImageDim(image); + constexpr int depth = 0; + constexpr bool ms = false; + constexpr int sampled = 2; // This won't be accessed with a sampler + const auto format = image.is_atomic ? spv::ImageFormat::R32ui : spv::ImageFormat::Unknown; + const Id image_type = TypeImage(t_uint, dim, depth, arrayed, ms, sampled, format, {}); + const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type); + const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); + AddGlobalVariable(Name(id, fmt::format("image_{}", image.index))); + + Decorate(id, spv::Decoration::Binding, binding++); + Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); + if (image.is_read && !image.is_written) { + Decorate(id, spv::Decoration::NonWritable); + } else if (image.is_written && !image.is_read) { + Decorate(id, spv::Decoration::NonReadable); + } + + images.emplace(image.index, StorageImage{image_type, id}); + } + bool IsRenderTargetEnabled(u32 rt) const { for (u32 component = 0; component < 4; ++component) { if (header.ps.IsColorComponentOutputEnabled(rt, component)) { @@ -976,6 +1020,10 @@ private: return stage == ShaderType::TesselationControl; } + bool IsAttributeEnabled(u32 location) const { + return stage != ShaderType::Vertex || specialization.enabled_attributes[location]; + } + u32 GetNumInputVertices() const { switch (stage) { case ShaderType::Geometry: @@ -1071,8 +1119,7 @@ private: void VisitBasicBlock(const NodeBlock& bb) { for (const auto& node : bb) { - [[maybe_unused]] const Type type = Visit(node).type; - ASSERT(type == Type::Void); + Visit(node); } } @@ -1192,16 +1239,20 @@ private: UNIMPLEMENTED_MSG("Unmanaged FrontFacing element={}", element); return {v_float_zero, Type::Float}; default: - if (IsGenericAttribute(attribute)) { - const u32 location = GetGenericAttributeLocation(attribute); - const auto type_descriptor = GetAttributeType(location); - const Type type = type_descriptor.type; - const Id attribute_id = input_attributes.at(attribute); - const std::vector elements = {element}; - const Id pointer = ArrayPass(type_descriptor.scalar, attribute_id, elements); - return {OpLoad(GetTypeDefinition(type), pointer), type}; + if (!IsGenericAttribute(attribute)) { + break; } - break; + const u32 location = GetGenericAttributeLocation(attribute); + if (!IsAttributeEnabled(location)) { + // Disabled attributes (also known as constant attributes) always return zero. + return {v_float_zero, Type::Float}; + } + const auto type_descriptor = GetAttributeType(location); + const Type type = type_descriptor.type; + const Id attribute_id = input_attributes.at(attribute); + const std::vector elements = {element}; + const Id pointer = ArrayPass(type_descriptor.scalar, attribute_id, elements); + return {OpLoad(GetTypeDefinition(type), pointer), type}; } UNIMPLEMENTED_MSG("Unhandled input attribute: {}", static_cast<u32>(attribute)); return {v_float_zero, Type::Float}; @@ -1237,7 +1288,7 @@ private: } else { UNREACHABLE_MSG("Unmanaged offset node type"); } - pointer = OpAccessChain(t_cbuf_float, buffer_id, Constant(t_uint, 0), buffer_index, + pointer = OpAccessChain(t_cbuf_float, buffer_id, v_uint_zero, buffer_index, buffer_element); } return {OpLoad(t_float, pointer), Type::Float}; @@ -1362,7 +1413,9 @@ private: Expression target{}; if (const auto gpr = std::get_if<GprNode>(&*dest)) { if (gpr->GetIndex() == Register::ZeroIndex) { - // Writing to Register::ZeroIndex is a no op + // Writing to Register::ZeroIndex is a no op but we still have to visit its source + // because it might have side effects. + Visit(src); return {}; } target = {registers.at(gpr->GetIndex()), Type::Float}; @@ -1584,6 +1637,15 @@ private: return {OpCompositeConstruct(t_half, low, high), Type::HalfFloat}; } + Expression LogicalAddCarry(Operation operation) { + const Id op_a = AsUint(Visit(operation[0])); + const Id op_b = AsUint(Visit(operation[1])); + + const Id result = OpIAddCarry(TypeStruct({t_uint, t_uint}), op_a, op_b); + const Id carry = OpCompositeExtract(t_uint, result, 1); + return {OpINotEqual(t_bool, carry, v_uint_zero), Type::Bool}; + } + Expression LogicalAssign(Operation operation) { const Node& dest = operation[0]; const Node& src = operation[1]; @@ -1609,13 +1671,31 @@ private: return {}; } + Expression LogicalFOrdered(Operation operation) { + // Emulate SPIR-V's OpOrdered + const Id op_a = AsFloat(Visit(operation[0])); + const Id op_b = AsFloat(Visit(operation[1])); + const Id is_num_a = OpFOrdEqual(t_bool, op_a, op_a); + const Id is_num_b = OpFOrdEqual(t_bool, op_b, op_b); + return {OpLogicalAnd(t_bool, is_num_a, is_num_b), Type::Bool}; + } + + Expression LogicalFUnordered(Operation operation) { + // Emulate SPIR-V's OpUnordered + const Id op_a = AsFloat(Visit(operation[0])); + const Id op_b = AsFloat(Visit(operation[1])); + const Id is_nan_a = OpIsNan(t_bool, op_a); + const Id is_nan_b = OpIsNan(t_bool, op_b); + return {OpLogicalOr(t_bool, is_nan_a, is_nan_b), Type::Bool}; + } + Id GetTextureSampler(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); - ASSERT(!meta.sampler.IsBuffer()); + ASSERT(!meta.sampler.is_buffer); - const auto& entry = sampled_images.at(meta.sampler.GetIndex()); + const auto& entry = sampled_images.at(meta.sampler.index); Id sampler = entry.variable; - if (meta.sampler.IsIndexed()) { + if (meta.sampler.is_indexed) { const Id index = AsInt(Visit(meta.index)); sampler = OpAccessChain(entry.sampler_pointer_type, sampler, index); } @@ -1624,9 +1704,9 @@ private: Id GetTextureImage(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); - const u32 index = meta.sampler.GetIndex(); - if (meta.sampler.IsBuffer()) { - const auto& entry = texel_buffers.at(index); + const u32 index = meta.sampler.index; + if (meta.sampler.is_buffer) { + const auto& entry = uniform_texels.at(index); return OpLoad(entry.image_type, entry.image); } else { const auto& entry = sampled_images.at(index); @@ -1636,7 +1716,7 @@ private: Id GetImage(Operation operation) { const auto& meta = std::get<MetaImage>(operation.GetMeta()); - const auto entry = images.at(meta.image.GetIndex()); + const auto entry = images.at(meta.image.index); return OpLoad(entry.image_type, entry.image); } @@ -1652,7 +1732,7 @@ private: } if (const auto meta = std::get_if<MetaTexture>(&operation.GetMeta())) { // Add array coordinate for textures - if (meta->sampler.IsArray()) { + if (meta->sampler.is_array) { Id array = AsInt(Visit(meta->array)); if (type == Type::Float) { array = OpConvertSToF(t_float, array); @@ -1758,7 +1838,7 @@ private: operands.push_back(GetOffsetCoordinates(operation)); } - if (meta.sampler.IsShadow()) { + if (meta.sampler.is_shadow) { const Id dref = AsFloat(Visit(meta.depth_compare)); return {OpImageSampleDrefExplicitLod(t_float, sampler, coords, dref, mask, operands), Type::Float}; @@ -1773,7 +1853,7 @@ private: const Id coords = GetCoordinates(operation, Type::Float); Id texture{}; - if (meta.sampler.IsShadow()) { + if (meta.sampler.is_shadow) { texture = OpImageDrefGather(t_float4, GetTextureSampler(operation), coords, AsFloat(Visit(meta.depth_compare))); } else { @@ -1800,8 +1880,8 @@ private: } const Id lod = AsUint(Visit(operation[0])); - const std::size_t coords_count = [&]() { - switch (const auto type = meta.sampler.GetType(); type) { + const std::size_t coords_count = [&meta] { + switch (const auto type = meta.sampler.type) { case Tegra::Shader::TextureType::Texture1D: return 1; case Tegra::Shader::TextureType::Texture2D: @@ -1810,7 +1890,7 @@ private: case Tegra::Shader::TextureType::Texture3D: return 3; default: - UNREACHABLE_MSG("Invalid texture type={}", static_cast<u32>(type)); + UNREACHABLE_MSG("Invalid texture type={}", static_cast<int>(type)); return 2; } }(); @@ -1853,7 +1933,7 @@ private: const Id image = GetTextureImage(operation); const Id coords = GetCoordinates(operation, Type::Int); Id fetch; - if (meta.lod && !meta.sampler.IsBuffer()) { + if (meta.lod && !meta.sampler.is_buffer) { fetch = OpImageFetch(t_float4, image, coords, spv::ImageOperandsMask::Lod, AsInt(Visit(meta.lod))); } else { @@ -1903,39 +1983,20 @@ private: return {}; } - Expression AtomicImageAdd(Operation operation) { - UNIMPLEMENTED(); - return {}; - } - - Expression AtomicImageMin(Operation operation) { - UNIMPLEMENTED(); - return {}; - } - - Expression AtomicImageMax(Operation operation) { - UNIMPLEMENTED(); - return {}; - } - - Expression AtomicImageAnd(Operation operation) { - UNIMPLEMENTED(); - return {}; - } - - Expression AtomicImageOr(Operation operation) { - UNIMPLEMENTED(); - return {}; - } + template <Id (Module::*func)(Id, Id, Id, Id, Id)> + Expression AtomicImage(Operation operation) { + const auto& meta{std::get<MetaImage>(operation.GetMeta())}; + ASSERT(meta.values.size() == 1); - Expression AtomicImageXor(Operation operation) { - UNIMPLEMENTED(); - return {}; - } + const Id coordinate = GetCoordinates(operation, Type::Int); + const Id image = images.at(meta.image.index).image; + const Id sample = v_uint_zero; + const Id pointer = OpImageTexelPointer(t_image_uint, image, coordinate, sample); - Expression AtomicImageExchange(Operation operation) { - UNIMPLEMENTED(); - return {}; + const Id scope = Constant(t_uint, static_cast<u32>(spv::Scope::Device)); + const Id semantics = v_uint_zero; + const Id value = AsUint(Visit(meta.values[0])); + return {(this->*func)(t_uint, pointer, scope, semantics, value), Type::Uint}; } template <Id (Module::*func)(Id, Id, Id, Id, Id)> @@ -1950,7 +2011,7 @@ private: return {v_float_zero, Type::Float}; } const Id scope = Constant(t_uint, static_cast<u32>(spv::Scope::Device)); - const Id semantics = Constant(t_uint, 0); + const Id semantics = v_uint_zero; const Id value = AsUint(Visit(operation[1])); return {(this->*func)(t_uint, pointer, scope, semantics, value), Type::Uint}; @@ -2148,14 +2209,37 @@ private: return {OpLoad(t_uint, thread_id), Type::Uint}; } + template <std::size_t index> + Expression ThreadMask(Operation) { + // TODO(Rodrigo): Handle devices with different warp sizes + const Id mask = thread_masks[index]; + return {OpLoad(t_uint, AccessElement(t_in_uint, mask, 0)), Type::Uint}; + } + Expression ShuffleIndexed(Operation operation) { const Id value = AsFloat(Visit(operation[0])); const Id index = AsUint(Visit(operation[1])); return {OpSubgroupReadInvocationKHR(t_float, value, index), Type::Float}; } - Expression MemoryBarrierGL(Operation) { - const auto scope = spv::Scope::Device; + Expression Barrier(Operation) { + if (!ir.IsDecompiled()) { + LOG_ERROR(Render_Vulkan, "OpBarrier used by shader is not decompiled"); + return {}; + } + + const auto scope = spv::Scope::Workgroup; + const auto memory = spv::Scope::Workgroup; + const auto semantics = + spv::MemorySemanticsMask::WorkgroupMemory | spv::MemorySemanticsMask::AcquireRelease; + OpControlBarrier(Constant(t_uint, static_cast<u32>(scope)), + Constant(t_uint, static_cast<u32>(memory)), + Constant(t_uint, static_cast<u32>(semantics))); + return {}; + } + + template <spv::Scope scope> + Expression MemoryBarrier(Operation) { const auto semantics = spv::MemorySemanticsMask::AcquireRelease | spv::MemorySemanticsMask::UniformMemory | spv::MemorySemanticsMask::WorkgroupMemory | @@ -2502,7 +2586,14 @@ private: &SPIRVDecompiler::Binary<&Module::OpFOrdGreaterThan, Type::Bool, Type::Float>, &SPIRVDecompiler::Binary<&Module::OpFOrdNotEqual, Type::Bool, Type::Float>, &SPIRVDecompiler::Binary<&Module::OpFOrdGreaterThanEqual, Type::Bool, Type::Float>, - &SPIRVDecompiler::Unary<&Module::OpIsNan, Type::Bool, Type::Float>, + &SPIRVDecompiler::LogicalFOrdered, + &SPIRVDecompiler::LogicalFUnordered, + &SPIRVDecompiler::Binary<&Module::OpFUnordLessThan, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFUnordEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFUnordLessThanEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFUnordGreaterThan, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFUnordNotEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFUnordGreaterThanEqual, Type::Bool, Type::Float>, &SPIRVDecompiler::Binary<&Module::OpSLessThan, Type::Bool, Type::Int>, &SPIRVDecompiler::Binary<&Module::OpIEqual, Type::Bool, Type::Int>, @@ -2518,6 +2609,8 @@ private: &SPIRVDecompiler::Binary<&Module::OpINotEqual, Type::Bool, Type::Uint>, &SPIRVDecompiler::Binary<&Module::OpUGreaterThanEqual, Type::Bool, Type::Uint>, + &SPIRVDecompiler::LogicalAddCarry, + &SPIRVDecompiler::Binary<&Module::OpFOrdLessThan, Type::Bool2, Type::HalfFloat>, &SPIRVDecompiler::Binary<&Module::OpFOrdEqual, Type::Bool2, Type::HalfFloat>, &SPIRVDecompiler::Binary<&Module::OpFOrdLessThanEqual, Type::Bool2, Type::HalfFloat>, @@ -2542,11 +2635,11 @@ private: &SPIRVDecompiler::ImageLoad, &SPIRVDecompiler::ImageStore, - &SPIRVDecompiler::AtomicImageAdd, - &SPIRVDecompiler::AtomicImageAnd, - &SPIRVDecompiler::AtomicImageOr, - &SPIRVDecompiler::AtomicImageXor, - &SPIRVDecompiler::AtomicImageExchange, + &SPIRVDecompiler::AtomicImage<&Module::OpAtomicIAdd>, + &SPIRVDecompiler::AtomicImage<&Module::OpAtomicAnd>, + &SPIRVDecompiler::AtomicImage<&Module::OpAtomicOr>, + &SPIRVDecompiler::AtomicImage<&Module::OpAtomicXor>, + &SPIRVDecompiler::AtomicImage<&Module::OpAtomicExchange>, &SPIRVDecompiler::Atomic<&Module::OpAtomicExchange>, &SPIRVDecompiler::Atomic<&Module::OpAtomicIAdd>, @@ -2603,9 +2696,16 @@ private: &SPIRVDecompiler::Vote<&Module::OpSubgroupAllEqualKHR>, &SPIRVDecompiler::ThreadId, + &SPIRVDecompiler::ThreadMask<0>, // Eq + &SPIRVDecompiler::ThreadMask<1>, // Ge + &SPIRVDecompiler::ThreadMask<2>, // Gt + &SPIRVDecompiler::ThreadMask<3>, // Le + &SPIRVDecompiler::ThreadMask<4>, // Lt &SPIRVDecompiler::ShuffleIndexed, - &SPIRVDecompiler::MemoryBarrierGL, + &SPIRVDecompiler::Barrier, + &SPIRVDecompiler::MemoryBarrier<spv::Scope::Workgroup>, + &SPIRVDecompiler::MemoryBarrier<spv::Scope::Device>, }; static_assert(operation_decompilers.size() == static_cast<std::size_t>(OperationCode::Amount)); @@ -2681,8 +2781,11 @@ private: Decorate(TypeStruct(t_gmem_array), spv::Decoration::Block), 0, spv::Decoration::Offset, 0); const Id t_gmem_ssbo = TypePointer(spv::StorageClass::StorageBuffer, t_gmem_struct); + const Id t_image_uint = TypePointer(spv::StorageClass::Image, t_uint); + const Id v_float_zero = Constant(t_float, 0.0f); const Id v_float_one = Constant(t_float, 1.0f); + const Id v_uint_zero = Constant(t_uint, 0); // Nvidia uses these defaults for varyings (e.g. position and generic attributes) const Id v_varying_default = @@ -2707,15 +2810,15 @@ private: std::unordered_map<u8, GenericVaryingDescription> output_attributes; std::map<u32, Id> constant_buffers; std::map<GlobalMemoryBase, Id> global_buffers; - std::map<u32, TexelBuffer> texel_buffers; + std::map<u32, TexelBuffer> uniform_texels; std::map<u32, SampledImage> sampled_images; std::map<u32, StorageImage> images; + std::array<Id, Maxwell::NumRenderTargets> frag_colors{}; Id instance_index{}; Id vertex_index{}; Id base_instance{}; Id base_vertex{}; - std::array<Id, Maxwell::NumRenderTargets> frag_colors{}; Id frag_depth{}; Id frag_coord{}; Id front_facing{}; @@ -2727,6 +2830,7 @@ private: Id workgroup_id{}; Id local_invocation_id{}; Id thread_id{}; + std::array<Id, 5> thread_masks{}; // eq, ge, gt, le, lt VertexIndices in_indices; VertexIndices out_indices; @@ -2969,14 +3073,18 @@ ShaderEntries GenerateShaderEntries(const VideoCommon::Shader::ShaderIR& ir) { entries.global_buffers.emplace_back(base.cbuf_index, base.cbuf_offset, usage.is_written); } for (const auto& sampler : ir.GetSamplers()) { - if (sampler.IsBuffer()) { - entries.texel_buffers.emplace_back(sampler); + if (sampler.is_buffer) { + entries.uniform_texels.emplace_back(sampler); } else { entries.samplers.emplace_back(sampler); } } for (const auto& image : ir.GetImages()) { - entries.images.emplace_back(image); + if (image.type == Tegra::Shader::ImageType::TextureBuffer) { + entries.storage_texels.emplace_back(image); + } else { + entries.images.emplace_back(image); + } } for (const auto& attribute : ir.GetInputAttributes()) { if (IsGenericAttribute(attribute)) { diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.h b/src/video_core/renderer_vulkan/vk_shader_decompiler.h index ffea4709e..2b0e90396 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.h +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.h @@ -5,11 +5,7 @@ #pragma once #include <array> -#include <bitset> -#include <memory> #include <set> -#include <type_traits> -#include <utility> #include <vector> #include "common/common_types.h" @@ -25,8 +21,9 @@ class VKDevice; namespace Vulkan { using Maxwell = Tegra::Engines::Maxwell3D::Regs; -using TexelBufferEntry = VideoCommon::Shader::Sampler; +using UniformTexelEntry = VideoCommon::Shader::Sampler; using SamplerEntry = VideoCommon::Shader::Sampler; +using StorageTexelEntry = VideoCommon::Shader::Image; using ImageEntry = VideoCommon::Shader::Image; constexpr u32 DESCRIPTOR_SET = 0; @@ -70,13 +67,15 @@ private: struct ShaderEntries { u32 NumBindings() const { return static_cast<u32>(const_buffers.size() + global_buffers.size() + - texel_buffers.size() + samplers.size() + images.size()); + uniform_texels.size() + samplers.size() + storage_texels.size() + + images.size()); } std::vector<ConstBufferEntry> const_buffers; std::vector<GlobalBufferEntry> global_buffers; - std::vector<TexelBufferEntry> texel_buffers; + std::vector<UniformTexelEntry> uniform_texels; std::vector<SamplerEntry> samplers; + std::vector<StorageTexelEntry> storage_texels; std::vector<ImageEntry> images; std::set<u32> attributes; std::array<bool, Maxwell::NumClipDistances> clip_distances{}; @@ -92,7 +91,8 @@ struct Specialization final { u32 shared_memory_size{}; // Graphics specific - std::optional<float> point_size{}; + std::optional<float> point_size; + std::bitset<Maxwell::NumVertexAttributes> enabled_attributes; std::array<Maxwell::VertexAttribute::Type, Maxwell::NumVertexAttributes> attribute_types{}; bool ndc_minus_one_to_one{}; }; diff --git a/src/video_core/renderer_vulkan/vk_shader_util.cpp b/src/video_core/renderer_vulkan/vk_shader_util.cpp index 784839327..c1a218d76 100644 --- a/src/video_core/renderer_vulkan/vk_shader_util.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_util.cpp @@ -4,8 +4,7 @@ #include <cstring> #include <memory> -#include <vector> -#include "common/alignment.h" + #include "common/assert.h" #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_device.h" @@ -20,13 +19,13 @@ vk::ShaderModule BuildShader(const VKDevice& device, std::size_t code_size, cons const auto data = std::make_unique<u32[]>(code_size / sizeof(u32)); std::memcpy(data.get(), code_data, code_size); - VkShaderModuleCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.codeSize = code_size; - ci.pCode = data.get(); - return device.GetLogical().CreateShaderModule(ci); + return device.GetLogical().CreateShaderModule({ + .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .codeSize = code_size, + .pCode = data.get(), + }); } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_shader_util.h b/src/video_core/renderer_vulkan/vk_shader_util.h index be38d6697..d1d3f3cae 100644 --- a/src/video_core/renderer_vulkan/vk_shader_util.h +++ b/src/video_core/renderer_vulkan/vk_shader_util.h @@ -4,7 +4,6 @@ #pragma once -#include <vector> #include "common/common_types.h" #include "video_core/renderer_vulkan/wrapper.h" diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp index 94d954d7a..2fd3b7f39 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp @@ -10,37 +10,18 @@ #include "common/bit_util.h" #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { -VKStagingBufferPool::StagingBuffer::StagingBuffer(std::unique_ptr<VKBuffer> buffer, VKFence& fence, - u64 last_epoch) - : buffer{std::move(buffer)}, watch{fence}, last_epoch{last_epoch} {} +VKStagingBufferPool::StagingBuffer::StagingBuffer(std::unique_ptr<VKBuffer> buffer_) + : buffer{std::move(buffer_)} {} -VKStagingBufferPool::StagingBuffer::StagingBuffer(StagingBuffer&& rhs) noexcept { - buffer = std::move(rhs.buffer); - watch = std::move(rhs.watch); - last_epoch = rhs.last_epoch; -} - -VKStagingBufferPool::StagingBuffer::~StagingBuffer() = default; - -VKStagingBufferPool::StagingBuffer& VKStagingBufferPool::StagingBuffer::operator=( - StagingBuffer&& rhs) noexcept { - buffer = std::move(rhs.buffer); - watch = std::move(rhs.watch); - last_epoch = rhs.last_epoch; - return *this; -} - -VKStagingBufferPool::VKStagingBufferPool(const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler) - : device{device}, memory_manager{memory_manager}, scheduler{scheduler}, - is_device_integrated{device.IsIntegrated()} {} +VKStagingBufferPool::VKStagingBufferPool(const VKDevice& device_, VKMemoryManager& memory_manager_, + VKScheduler& scheduler_) + : device{device_}, memory_manager{memory_manager_}, scheduler{scheduler_} {} VKStagingBufferPool::~VKStagingBufferPool() = default; @@ -52,21 +33,19 @@ VKBuffer& VKStagingBufferPool::GetUnusedBuffer(std::size_t size, bool host_visib } void VKStagingBufferPool::TickFrame() { - ++epoch; current_delete_level = (current_delete_level + 1) % NumLevels; ReleaseCache(true); - if (!is_device_integrated) { - ReleaseCache(false); - } + ReleaseCache(false); } VKBuffer* VKStagingBufferPool::TryGetReservedBuffer(std::size_t size, bool host_visible) { - for (auto& entry : GetCache(host_visible)[Common::Log2Ceil64(size)].entries) { - if (entry.watch.TryWatch(scheduler.GetFence())) { - entry.last_epoch = epoch; - return &*entry.buffer; + for (StagingBuffer& entry : GetCache(host_visible)[Common::Log2Ceil64(size)].entries) { + if (!scheduler.IsFree(entry.tick)) { + continue; } + entry.tick = scheduler.CurrentTick(); + return &*entry.buffer; } return nullptr; } @@ -74,28 +53,29 @@ VKBuffer* VKStagingBufferPool::TryGetReservedBuffer(std::size_t size, bool host_ VKBuffer& VKStagingBufferPool::CreateStagingBuffer(std::size_t size, bool host_visible) { const u32 log2 = Common::Log2Ceil64(size); - VkBufferCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.size = 1ULL << log2; - ci.usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | - VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | - VK_BUFFER_USAGE_INDEX_BUFFER_BIT; - ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - ci.queueFamilyIndexCount = 0; - ci.pQueueFamilyIndices = nullptr; - auto buffer = std::make_unique<VKBuffer>(); - buffer->handle = device.GetLogical().CreateBuffer(ci); + buffer->handle = device.GetLogical().CreateBuffer({ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = 1ULL << log2, + .usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | + VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | + VK_BUFFER_USAGE_INDEX_BUFFER_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }); buffer->commit = memory_manager.Commit(buffer->handle, host_visible); - auto& entries = GetCache(host_visible)[log2].entries; - return *entries.emplace_back(std::move(buffer), scheduler.GetFence(), epoch).buffer; + std::vector<StagingBuffer>& entries = GetCache(host_visible)[log2].entries; + StagingBuffer& entry = entries.emplace_back(std::move(buffer)); + entry.tick = scheduler.CurrentTick(); + return *entry.buffer; } VKStagingBufferPool::StagingBuffersCache& VKStagingBufferPool::GetCache(bool host_visible) { - return is_device_integrated || host_visible ? host_staging_buffers : device_staging_buffers; + return host_visible ? host_staging_buffers : device_staging_buffers; } void VKStagingBufferPool::ReleaseCache(bool host_visible) { @@ -113,9 +93,8 @@ u64 VKStagingBufferPool::ReleaseLevel(StagingBuffersCache& cache, std::size_t lo auto& entries = staging.entries; const std::size_t old_size = entries.size(); - const auto is_deleteable = [this](const auto& entry) { - static constexpr u64 epochs_to_destroy = 180; - return entry.last_epoch + epochs_to_destroy < epoch && !entry.watch.IsUsed(); + const auto is_deleteable = [this](const StagingBuffer& entry) { + return scheduler.IsFree(entry.tick); }; const std::size_t begin_offset = staging.delete_index; const std::size_t end_offset = std::min(begin_offset + deletions_per_tick, old_size); diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h index a0840ff8c..2dd5049ac 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h @@ -5,20 +5,16 @@ #pragma once #include <climits> -#include <unordered_map> -#include <utility> #include <vector> #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/wrapper.h" namespace Vulkan { class VKDevice; -class VKFenceWatch; class VKScheduler; struct VKBuffer final { @@ -38,16 +34,10 @@ public: private: struct StagingBuffer final { - explicit StagingBuffer(std::unique_ptr<VKBuffer> buffer, VKFence& fence, u64 last_epoch); - StagingBuffer(StagingBuffer&& rhs) noexcept; - StagingBuffer(const StagingBuffer&) = delete; - ~StagingBuffer(); - - StagingBuffer& operator=(StagingBuffer&& rhs) noexcept; + explicit StagingBuffer(std::unique_ptr<VKBuffer> buffer); std::unique_ptr<VKBuffer> buffer; - VKFenceWatch watch; - u64 last_epoch = 0; + u64 tick = 0; }; struct StagingBuffers final { @@ -71,13 +61,10 @@ private: const VKDevice& device; VKMemoryManager& memory_manager; VKScheduler& scheduler; - const bool is_device_integrated; StagingBuffersCache host_staging_buffers; StagingBuffersCache device_staging_buffers; - u64 epoch = 0; - std::size_t current_delete_level = 0; }; diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.cpp b/src/video_core/renderer_vulkan/vk_state_tracker.cpp index 94a89e388..5d2c4a796 100644 --- a/src/video_core/renderer_vulkan/vk_state_tracker.cpp +++ b/src/video_core/renderer_vulkan/vk_state_tracker.cpp @@ -36,6 +36,14 @@ Flags MakeInvalidationFlags() { flags[BlendConstants] = true; flags[DepthBounds] = true; flags[StencilProperties] = true; + flags[CullMode] = true; + flags[DepthBoundsEnable] = true; + flags[DepthTestEnable] = true; + flags[DepthWriteEnable] = true; + flags[DepthCompareOp] = true; + flags[FrontFace] = true; + flags[StencilOp] = true; + flags[StencilTestEnable] = true; return flags; } @@ -75,14 +83,58 @@ void SetupDirtyStencilProperties(Tables& tables) { table[OFF(stencil_back_func_mask)] = StencilProperties; } -} // Anonymous namespace +void SetupDirtyCullMode(Tables& tables) { + auto& table = tables[0]; + table[OFF(cull_face)] = CullMode; + table[OFF(cull_test_enabled)] = CullMode; +} + +void SetupDirtyDepthBoundsEnable(Tables& tables) { + tables[0][OFF(depth_bounds_enable)] = DepthBoundsEnable; +} + +void SetupDirtyDepthTestEnable(Tables& tables) { + tables[0][OFF(depth_test_enable)] = DepthTestEnable; +} + +void SetupDirtyDepthWriteEnable(Tables& tables) { + tables[0][OFF(depth_write_enabled)] = DepthWriteEnable; +} + +void SetupDirtyDepthCompareOp(Tables& tables) { + tables[0][OFF(depth_test_func)] = DepthCompareOp; +} -StateTracker::StateTracker(Core::System& system) - : system{system}, invalidation_flags{MakeInvalidationFlags()} {} +void SetupDirtyFrontFace(Tables& tables) { + auto& table = tables[0]; + table[OFF(front_face)] = FrontFace; + table[OFF(screen_y_control)] = FrontFace; +} + +void SetupDirtyStencilOp(Tables& tables) { + auto& table = tables[0]; + table[OFF(stencil_front_op_fail)] = StencilOp; + table[OFF(stencil_front_op_zfail)] = StencilOp; + table[OFF(stencil_front_op_zpass)] = StencilOp; + table[OFF(stencil_front_func_func)] = StencilOp; + table[OFF(stencil_back_op_fail)] = StencilOp; + table[OFF(stencil_back_op_zfail)] = StencilOp; + table[OFF(stencil_back_op_zpass)] = StencilOp; + table[OFF(stencil_back_func_func)] = StencilOp; + + // Table 0 is used by StencilProperties + tables[1][OFF(stencil_two_side_enable)] = StencilOp; +} -void StateTracker::Initialize() { - auto& dirty = system.GPU().Maxwell3D().dirty; - auto& tables = dirty.tables; +void SetupDirtyStencilTestEnable(Tables& tables) { + tables[0][OFF(stencil_enable)] = StencilTestEnable; +} + +} // Anonymous namespace + +StateTracker::StateTracker(Tegra::GPU& gpu) + : flags{gpu.Maxwell3D().dirty.flags}, invalidation_flags{MakeInvalidationFlags()} { + auto& tables = gpu.Maxwell3D().dirty.tables; SetupDirtyRenderTargets(tables); SetupDirtyViewports(tables); SetupDirtyScissors(tables); @@ -90,10 +142,14 @@ void StateTracker::Initialize() { SetupDirtyBlendConstants(tables); SetupDirtyDepthBounds(tables); SetupDirtyStencilProperties(tables); -} - -void StateTracker::InvalidateCommandBufferState() { - system.GPU().Maxwell3D().dirty.flags |= invalidation_flags; + SetupDirtyCullMode(tables); + SetupDirtyDepthBoundsEnable(tables); + SetupDirtyDepthTestEnable(tables); + SetupDirtyDepthWriteEnable(tables); + SetupDirtyDepthCompareOp(tables); + SetupDirtyFrontFace(tables); + SetupDirtyStencilOp(tables); + SetupDirtyStencilTestEnable(tables); } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.h b/src/video_core/renderer_vulkan/vk_state_tracker.h index 03bc415b2..1de789e57 100644 --- a/src/video_core/renderer_vulkan/vk_state_tracker.h +++ b/src/video_core/renderer_vulkan/vk_state_tracker.h @@ -26,6 +26,15 @@ enum : u8 { DepthBounds, StencilProperties, + CullMode, + DepthBoundsEnable, + DepthTestEnable, + DepthWriteEnable, + DepthCompareOp, + FrontFace, + StencilOp, + StencilTestEnable, + Last }; static_assert(Last <= std::numeric_limits<u8>::max()); @@ -33,12 +42,15 @@ static_assert(Last <= std::numeric_limits<u8>::max()); } // namespace Dirty class StateTracker { -public: - explicit StateTracker(Core::System& system); + using Maxwell = Tegra::Engines::Maxwell3D::Regs; - void Initialize(); +public: + explicit StateTracker(Tegra::GPU& gpu); - void InvalidateCommandBufferState(); + void InvalidateCommandBufferState() { + flags |= invalidation_flags; + current_topology = INVALID_TOPOLOGY; + } bool TouchViewports() { return Exchange(Dirty::Viewports, false); @@ -64,16 +76,60 @@ public: return Exchange(Dirty::StencilProperties, false); } + bool TouchCullMode() { + return Exchange(Dirty::CullMode, false); + } + + bool TouchDepthBoundsTestEnable() { + return Exchange(Dirty::DepthBoundsEnable, false); + } + + bool TouchDepthTestEnable() { + return Exchange(Dirty::DepthTestEnable, false); + } + + bool TouchDepthBoundsEnable() { + return Exchange(Dirty::DepthBoundsEnable, false); + } + + bool TouchDepthWriteEnable() { + return Exchange(Dirty::DepthWriteEnable, false); + } + + bool TouchDepthCompareOp() { + return Exchange(Dirty::DepthCompareOp, false); + } + + bool TouchFrontFace() { + return Exchange(Dirty::FrontFace, false); + } + + bool TouchStencilOp() { + return Exchange(Dirty::StencilOp, false); + } + + bool TouchStencilTestEnable() { + return Exchange(Dirty::StencilTestEnable, false); + } + + bool ChangePrimitiveTopology(Maxwell::PrimitiveTopology new_topology) { + const bool has_changed = current_topology != new_topology; + current_topology = new_topology; + return has_changed; + } + private: + static constexpr auto INVALID_TOPOLOGY = static_cast<Maxwell::PrimitiveTopology>(~0u); + bool Exchange(std::size_t id, bool new_value) const noexcept { - auto& flags = system.GPU().Maxwell3D().dirty.flags; const bool is_dirty = flags[id]; flags[id] = new_value; return is_dirty; } - Core::System& system; + Tegra::Engines::Maxwell3D::DirtyState::Flags& flags; Tegra::Engines::Maxwell3D::DirtyState::Flags invalidation_flags; + Maxwell::PrimitiveTopology current_topology = INVALID_TOPOLOGY; }; } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp index 38a93a01a..1b59612b9 100644 --- a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp +++ b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <algorithm> +#include <limits> #include <optional> #include <tuple> #include <vector> @@ -10,7 +11,6 @@ #include "common/alignment.h" #include "common/assert.h" #include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -22,27 +22,43 @@ namespace { constexpr u64 WATCHES_INITIAL_RESERVE = 0x4000; constexpr u64 WATCHES_RESERVE_CHUNK = 0x1000; -constexpr u64 STREAM_BUFFER_SIZE = 256 * 1024 * 1024; +constexpr u64 PREFERRED_STREAM_BUFFER_SIZE = 256 * 1024 * 1024; -std::optional<u32> FindMemoryType(const VKDevice& device, u32 filter, - VkMemoryPropertyFlags wanted) { - const auto properties = device.GetPhysical().GetMemoryProperties(); - for (u32 i = 0; i < properties.memoryTypeCount; i++) { - if (!(filter & (1 << i))) { - continue; - } - if ((properties.memoryTypes[i].propertyFlags & wanted) == wanted) { +/// Find a memory type with the passed requirements +std::optional<u32> FindMemoryType(const VkPhysicalDeviceMemoryProperties& properties, + VkMemoryPropertyFlags wanted, + u32 filter = std::numeric_limits<u32>::max()) { + for (u32 i = 0; i < properties.memoryTypeCount; ++i) { + const auto flags = properties.memoryTypes[i].propertyFlags; + if ((flags & wanted) == wanted && (filter & (1U << i)) != 0) { return i; } } return std::nullopt; } +/// Get the preferred host visible memory type. +u32 GetMemoryType(const VkPhysicalDeviceMemoryProperties& properties, + u32 filter = std::numeric_limits<u32>::max()) { + // Prefer device local host visible allocations. Both AMD and Nvidia now provide one. + // Otherwise search for a host visible allocation. + static constexpr auto HOST_MEMORY = + VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT; + static constexpr auto DYNAMIC_MEMORY = HOST_MEMORY | VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT; + + std::optional preferred_type = FindMemoryType(properties, DYNAMIC_MEMORY); + if (!preferred_type) { + preferred_type = FindMemoryType(properties, HOST_MEMORY); + ASSERT_MSG(preferred_type, "No host visible and coherent memory type found"); + } + return preferred_type.value_or(0); +} + } // Anonymous namespace -VKStreamBuffer::VKStreamBuffer(const VKDevice& device, VKScheduler& scheduler, +VKStreamBuffer::VKStreamBuffer(const VKDevice& device_, VKScheduler& scheduler_, VkBufferUsageFlags usage) - : device{device}, scheduler{scheduler} { + : device{device_}, scheduler{scheduler_} { CreateBuffers(usage); ReserveWatches(current_watches, WATCHES_INITIAL_RESERVE); ReserveWatches(previous_watches, WATCHES_INITIAL_RESERVE); @@ -51,7 +67,7 @@ VKStreamBuffer::VKStreamBuffer(const VKDevice& device, VKScheduler& scheduler, VKStreamBuffer::~VKStreamBuffer() = default; std::tuple<u8*, u64, bool> VKStreamBuffer::Map(u64 size, u64 alignment) { - ASSERT(size <= STREAM_BUFFER_SIZE); + ASSERT(size <= stream_buffer_size); mapped_size = size; if (alignment > 0) { @@ -61,7 +77,7 @@ std::tuple<u8*, u64, bool> VKStreamBuffer::Map(u64 size, u64 alignment) { WaitPendingOperations(offset); bool invalidated = false; - if (offset + size > STREAM_BUFFER_SIZE) { + if (offset + size > stream_buffer_size) { // The buffer would overflow, save the amount of used watches and reset the state. invalidation_mark = current_watch_cursor; current_watch_cursor = 0; @@ -94,44 +110,39 @@ void VKStreamBuffer::Unmap(u64 size) { } auto& watch = current_watches[current_watch_cursor++]; watch.upper_bound = offset; - watch.fence.Watch(scheduler.GetFence()); + watch.tick = scheduler.CurrentTick(); } void VKStreamBuffer::CreateBuffers(VkBufferUsageFlags usage) { - VkBufferCreateInfo buffer_ci; - buffer_ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO; - buffer_ci.pNext = nullptr; - buffer_ci.flags = 0; - buffer_ci.size = STREAM_BUFFER_SIZE; - buffer_ci.usage = usage; - buffer_ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - buffer_ci.queueFamilyIndexCount = 0; - buffer_ci.pQueueFamilyIndices = nullptr; - - const auto& dev = device.GetLogical(); - buffer = dev.CreateBuffer(buffer_ci); - - const auto& dld = device.GetDispatchLoader(); - const auto requirements = dev.GetBufferMemoryRequirements(*buffer); - // Prefer device local host visible allocations (this should hit AMD's pinned memory). - auto type = - FindMemoryType(device, requirements.memoryTypeBits, - VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT | - VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT); - if (!type) { - // Otherwise search for a host visible allocation. - type = FindMemoryType(device, requirements.memoryTypeBits, - VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | - VK_MEMORY_PROPERTY_HOST_COHERENT_BIT); - ASSERT_MSG(type, "No host visible and coherent memory type found"); - } - VkMemoryAllocateInfo memory_ai; - memory_ai.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO; - memory_ai.pNext = nullptr; - memory_ai.allocationSize = requirements.size; - memory_ai.memoryTypeIndex = *type; - - memory = dev.AllocateMemory(memory_ai); + const auto memory_properties = device.GetPhysical().GetMemoryProperties(); + const u32 preferred_type = GetMemoryType(memory_properties); + const u32 preferred_heap = memory_properties.memoryTypes[preferred_type].heapIndex; + + // Substract from the preferred heap size some bytes to avoid getting out of memory. + const VkDeviceSize heap_size = memory_properties.memoryHeaps[preferred_heap].size; + // As per DXVK's example, using `heap_size / 2` + const VkDeviceSize allocable_size = heap_size / 2; + buffer = device.GetLogical().CreateBuffer({ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = std::min(PREFERRED_STREAM_BUFFER_SIZE, allocable_size), + .usage = usage, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }); + + const auto requirements = device.GetLogical().GetBufferMemoryRequirements(*buffer); + const u32 required_flags = requirements.memoryTypeBits; + stream_buffer_size = static_cast<u64>(requirements.size); + + memory = device.GetLogical().AllocateMemory({ + .sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO, + .pNext = nullptr, + .allocationSize = requirements.size, + .memoryTypeIndex = GetMemoryType(memory_properties, required_flags), + }); buffer.BindMemory(*memory, 0); } @@ -146,7 +157,7 @@ void VKStreamBuffer::WaitPendingOperations(u64 requested_upper_bound) { while (requested_upper_bound < wait_bound && wait_cursor < *invalidation_mark) { auto& watch = previous_watches[wait_cursor]; wait_bound = watch.upper_bound; - watch.fence.Wait(); + scheduler.Wait(watch.tick); ++wait_cursor; } } diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.h b/src/video_core/renderer_vulkan/vk_stream_buffer.h index 58ce8b973..5e15ad78f 100644 --- a/src/video_core/renderer_vulkan/vk_stream_buffer.h +++ b/src/video_core/renderer_vulkan/vk_stream_buffer.h @@ -14,7 +14,6 @@ namespace Vulkan { class VKDevice; -class VKFence; class VKFenceWatch; class VKScheduler; @@ -35,13 +34,17 @@ public: /// Ensures that "size" bytes of memory are available to the GPU, potentially recording a copy. void Unmap(u64 size); - VkBuffer GetHandle() const { + VkBuffer Handle() const noexcept { return *buffer; } + u64 Address() const noexcept { + return 0; + } + private: - struct Watch final { - VKFenceWatch fence; + struct Watch { + u64 tick{}; u64 upper_bound{}; }; @@ -56,8 +59,9 @@ private: const VKDevice& device; ///< Vulkan device manager. VKScheduler& scheduler; ///< Command scheduler. - vk::Buffer buffer; ///< Mapped buffer. - vk::DeviceMemory memory; ///< Memory allocation. + vk::Buffer buffer; ///< Mapped buffer. + vk::DeviceMemory memory; ///< Memory allocation. + u64 stream_buffer_size{}; ///< Stream buffer size. u64 offset{}; ///< Buffer iterator. u64 mapped_size{}; ///< Size reserved for the current copy. diff --git a/src/video_core/renderer_vulkan/vk_swapchain.cpp b/src/video_core/renderer_vulkan/vk_swapchain.cpp index bffd8f32a..9636a7c65 100644 --- a/src/video_core/renderer_vulkan/vk_swapchain.cpp +++ b/src/video_core/renderer_vulkan/vk_swapchain.cpp @@ -12,7 +12,7 @@ #include "core/core.h" #include "core/frontend/framebuffer_layout.h" #include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_resource_manager.h" +#include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_swapchain.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -56,8 +56,8 @@ VkExtent2D ChooseSwapExtent(const VkSurfaceCapabilitiesKHR& capabilities, u32 wi } // Anonymous namespace -VKSwapchain::VKSwapchain(VkSurfaceKHR surface, const VKDevice& device) - : surface{surface}, device{device} {} +VKSwapchain::VKSwapchain(VkSurfaceKHR surface_, const VKDevice& device_, VKScheduler& scheduler_) + : surface{surface_}, device{device_}, scheduler{scheduler_} {} VKSwapchain::~VKSwapchain() = default; @@ -75,35 +75,33 @@ void VKSwapchain::Create(u32 width, u32 height, bool srgb) { CreateSemaphores(); CreateImageViews(); - fences.resize(image_count, nullptr); + resource_ticks.clear(); + resource_ticks.resize(image_count); } void VKSwapchain::AcquireNextImage() { device.GetLogical().AcquireNextImageKHR(*swapchain, std::numeric_limits<u64>::max(), *present_semaphores[frame_index], {}, &image_index); - if (auto& fence = fences[image_index]; fence) { - fence->Wait(); - fence->Release(); - fence = nullptr; - } + scheduler.Wait(resource_ticks[image_index]); } -bool VKSwapchain::Present(VkSemaphore render_semaphore, VKFence& fence) { +bool VKSwapchain::Present(VkSemaphore render_semaphore) { const VkSemaphore present_semaphore{*present_semaphores[frame_index]}; const std::array<VkSemaphore, 2> semaphores{present_semaphore, render_semaphore}; const auto present_queue{device.GetPresentQueue()}; bool recreated = false; - VkPresentInfoKHR present_info; - present_info.sType = VK_STRUCTURE_TYPE_PRESENT_INFO_KHR; - present_info.pNext = nullptr; - present_info.waitSemaphoreCount = render_semaphore ? 2U : 1U; - present_info.pWaitSemaphores = semaphores.data(); - present_info.swapchainCount = 1; - present_info.pSwapchains = swapchain.address(); - present_info.pImageIndices = &image_index; - present_info.pResults = nullptr; + const VkPresentInfoKHR present_info{ + .sType = VK_STRUCTURE_TYPE_PRESENT_INFO_KHR, + .pNext = nullptr, + .waitSemaphoreCount = render_semaphore ? 2U : 1U, + .pWaitSemaphores = semaphores.data(), + .swapchainCount = 1, + .pSwapchains = swapchain.address(), + .pImageIndices = &image_index, + .pResults = nullptr, + }; switch (const VkResult result = present_queue.Present(present_info)) { case VK_SUCCESS: @@ -122,8 +120,7 @@ bool VKSwapchain::Present(VkSemaphore render_semaphore, VKFence& fence) { break; } - ASSERT(fences[image_index] == nullptr); - fences[image_index] = &fence; + resource_ticks[image_index] = scheduler.CurrentTick(); frame_index = (frame_index + 1) % static_cast<u32>(image_count); return recreated; } @@ -147,24 +144,26 @@ void VKSwapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, requested_image_count = capabilities.maxImageCount; } - VkSwapchainCreateInfoKHR swapchain_ci; - swapchain_ci.sType = VK_STRUCTURE_TYPE_SWAPCHAIN_CREATE_INFO_KHR; - swapchain_ci.pNext = nullptr; - swapchain_ci.flags = 0; - swapchain_ci.surface = surface; - swapchain_ci.minImageCount = requested_image_count; - swapchain_ci.imageFormat = surface_format.format; - swapchain_ci.imageColorSpace = surface_format.colorSpace; - swapchain_ci.imageArrayLayers = 1; - swapchain_ci.imageUsage = VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT; - swapchain_ci.imageSharingMode = VK_SHARING_MODE_EXCLUSIVE; - swapchain_ci.queueFamilyIndexCount = 0; - swapchain_ci.pQueueFamilyIndices = nullptr; - swapchain_ci.preTransform = capabilities.currentTransform; - swapchain_ci.compositeAlpha = VK_COMPOSITE_ALPHA_OPAQUE_BIT_KHR; - swapchain_ci.presentMode = present_mode; - swapchain_ci.clipped = VK_FALSE; - swapchain_ci.oldSwapchain = nullptr; + VkSwapchainCreateInfoKHR swapchain_ci{ + .sType = VK_STRUCTURE_TYPE_SWAPCHAIN_CREATE_INFO_KHR, + .pNext = nullptr, + .flags = 0, + .surface = surface, + .minImageCount = requested_image_count, + .imageFormat = surface_format.format, + .imageColorSpace = surface_format.colorSpace, + .imageExtent = {}, + .imageArrayLayers = 1, + .imageUsage = VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT, + .imageSharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + .preTransform = capabilities.currentTransform, + .compositeAlpha = VK_COMPOSITE_ALPHA_OPAQUE_BIT_KHR, + .presentMode = present_mode, + .clipped = VK_FALSE, + .oldSwapchain = nullptr, + }; const u32 graphics_family{device.GetGraphicsFamily()}; const u32 present_family{device.GetPresentFamily()}; @@ -173,8 +172,6 @@ void VKSwapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities, swapchain_ci.imageSharingMode = VK_SHARING_MODE_CONCURRENT; swapchain_ci.queueFamilyIndexCount = static_cast<u32>(queue_indices.size()); swapchain_ci.pQueueFamilyIndices = queue_indices.data(); - } else { - swapchain_ci.imageSharingMode = VK_SHARING_MODE_EXCLUSIVE; } // Request the size again to reduce the possibility of a TOCTOU race condition. @@ -200,20 +197,29 @@ void VKSwapchain::CreateSemaphores() { } void VKSwapchain::CreateImageViews() { - VkImageViewCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - // ci.image - ci.viewType = VK_IMAGE_VIEW_TYPE_2D; - ci.format = image_format; - ci.components = {VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY, - VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY}; - ci.subresourceRange.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT; - ci.subresourceRange.baseMipLevel = 0; - ci.subresourceRange.levelCount = 1; - ci.subresourceRange.baseArrayLayer = 0; - ci.subresourceRange.layerCount = 1; + VkImageViewCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = {}, + .viewType = VK_IMAGE_VIEW_TYPE_2D, + .format = image_format, + .components = + { + .r = VK_COMPONENT_SWIZZLE_IDENTITY, + .g = VK_COMPONENT_SWIZZLE_IDENTITY, + .b = VK_COMPONENT_SWIZZLE_IDENTITY, + .a = VK_COMPONENT_SWIZZLE_IDENTITY, + }, + .subresourceRange = + { + .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT, + .baseMipLevel = 0, + .levelCount = 1, + .baseArrayLayer = 0, + .layerCount = 1, + }, + }; image_views.resize(image_count); for (std::size_t i = 0; i < image_count; i++) { diff --git a/src/video_core/renderer_vulkan/vk_swapchain.h b/src/video_core/renderer_vulkan/vk_swapchain.h index a35d61345..6b39befdf 100644 --- a/src/video_core/renderer_vulkan/vk_swapchain.h +++ b/src/video_core/renderer_vulkan/vk_swapchain.h @@ -16,11 +16,11 @@ struct FramebufferLayout; namespace Vulkan { class VKDevice; -class VKFence; +class VKScheduler; class VKSwapchain { public: - explicit VKSwapchain(VkSurfaceKHR surface, const VKDevice& device); + explicit VKSwapchain(VkSurfaceKHR surface, const VKDevice& device, VKScheduler& scheduler); ~VKSwapchain(); /// Creates (or recreates) the swapchain with a given size. @@ -31,7 +31,7 @@ public: /// Presents the rendered image to the swapchain. Returns true when the swapchains had to be /// recreated. Takes responsability for the ownership of fence. - bool Present(VkSemaphore render_semaphore, VKFence& fence); + bool Present(VkSemaphore render_semaphore); /// Returns true when the framebuffer layout has changed. bool HasFramebufferChanged(const Layout::FramebufferLayout& framebuffer) const; @@ -74,6 +74,7 @@ private: const VkSurfaceKHR surface; const VKDevice& device; + VKScheduler& scheduler; vk::SwapchainKHR swapchain; @@ -81,7 +82,7 @@ private: std::vector<VkImage> images; std::vector<vk::ImageView> image_views; std::vector<vk::Framebuffer> framebuffers; - std::vector<VKFence*> fences; + std::vector<u64> resource_ticks; std::vector<vk::Semaphore> present_semaphores; u32 image_index{}; diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index de4c23120..f2c8f2ae1 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -10,11 +10,9 @@ #include <variant> #include <vector> -#include "common/alignment.h" #include "common/assert.h" #include "common/common_types.h" #include "core/core.h" -#include "core/memory.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/morton.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" @@ -26,7 +24,6 @@ #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/surface.h" -#include "video_core/textures/convert.h" namespace Vulkan { @@ -98,17 +95,18 @@ VkImageViewType GetImageViewType(SurfaceTarget target) { vk::Buffer CreateBuffer(const VKDevice& device, const SurfaceParams& params, std::size_t host_memory_size) { // TODO(Rodrigo): Move texture buffer creation to the buffer cache - VkBufferCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.size = static_cast<VkDeviceSize>(host_memory_size); - ci.usage = VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | - VK_BUFFER_USAGE_TRANSFER_DST_BIT; - ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - ci.queueFamilyIndexCount = 0; - ci.pQueueFamilyIndices = nullptr; - return device.GetLogical().CreateBuffer(ci); + return device.GetLogical().CreateBuffer({ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = static_cast<VkDeviceSize>(host_memory_size), + .usage = VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT | + VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | + VK_BUFFER_USAGE_TRANSFER_DST_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }); } VkBufferViewCreateInfo GenerateBufferViewCreateInfo(const VKDevice& device, @@ -116,15 +114,16 @@ VkBufferViewCreateInfo GenerateBufferViewCreateInfo(const VKDevice& device, std::size_t host_memory_size) { ASSERT(params.IsBuffer()); - VkBufferViewCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_BUFFER_VIEW_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.buffer = buffer; - ci.format = MaxwellToVK::SurfaceFormat(device, FormatType::Buffer, params.pixel_format).format; - ci.offset = 0; - ci.range = static_cast<VkDeviceSize>(host_memory_size); - return ci; + return { + .sType = VK_STRUCTURE_TYPE_BUFFER_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .buffer = buffer, + .format = + MaxwellToVK::SurfaceFormat(device, FormatType::Buffer, params.pixel_format).format, + .offset = 0, + .range = static_cast<VkDeviceSize>(host_memory_size), + }; } VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceParams& params) { @@ -133,23 +132,24 @@ VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceP const auto [format, attachable, storage] = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, params.pixel_format); - VkImageCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.imageType = SurfaceTargetToImage(params.target); - ci.format = format; - ci.mipLevels = params.num_levels; - ci.arrayLayers = static_cast<u32>(params.GetNumLayers()); - ci.samples = VK_SAMPLE_COUNT_1_BIT; - ci.tiling = VK_IMAGE_TILING_OPTIMAL; - ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE; - ci.queueFamilyIndexCount = 0; - ci.pQueueFamilyIndices = nullptr; - ci.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED; - - ci.usage = VK_IMAGE_USAGE_SAMPLED_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT | - VK_IMAGE_USAGE_TRANSFER_SRC_BIT; + VkImageCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .imageType = SurfaceTargetToImage(params.target), + .format = format, + .extent = {}, + .mipLevels = params.num_levels, + .arrayLayers = static_cast<u32>(params.GetNumLayers()), + .samples = VK_SAMPLE_COUNT_1_BIT, + .tiling = VK_IMAGE_TILING_OPTIMAL, + .usage = VK_IMAGE_USAGE_SAMPLED_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT | + VK_IMAGE_USAGE_TRANSFER_SRC_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, + }; if (attachable) { ci.usage |= params.IsPixelFormatZeta() ? VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT : VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT; @@ -170,6 +170,7 @@ VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceP ci.extent = {params.width, params.height, 1}; break; case SurfaceTarget::Texture3D: + ci.flags |= VK_IMAGE_CREATE_2D_ARRAY_COMPATIBLE_BIT; ci.extent = {params.width, params.height, params.depth}; break; case SurfaceTarget::TextureBuffer: @@ -179,14 +180,18 @@ VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceP return ci; } +u32 EncodeSwizzle(Tegra::Texture::SwizzleSource x_source, Tegra::Texture::SwizzleSource y_source, + Tegra::Texture::SwizzleSource z_source, Tegra::Texture::SwizzleSource w_source) { + return (static_cast<u32>(x_source) << 24) | (static_cast<u32>(y_source) << 16) | + (static_cast<u32>(z_source) << 8) | static_cast<u32>(w_source); +} + } // Anonymous namespace -CachedSurface::CachedSurface(Core::System& system, const VKDevice& device, - VKResourceManager& resource_manager, VKMemoryManager& memory_manager, +CachedSurface::CachedSurface(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler, VKStagingBufferPool& staging_pool, GPUVAddr gpu_addr, const SurfaceParams& params) - : SurfaceBase<View>{gpu_addr, params, device.IsOptimalAstcSupported()}, system{system}, - device{device}, resource_manager{resource_manager}, + : SurfaceBase<View>{gpu_addr, params, device.IsOptimalAstcSupported()}, device{device}, memory_manager{memory_manager}, scheduler{scheduler}, staging_pool{staging_pool} { if (params.IsBuffer()) { buffer = CreateBuffer(device, params, host_memory_size); @@ -206,9 +211,11 @@ CachedSurface::CachedSurface(Core::System& system, const VKDevice& device, } // TODO(Rodrigo): Move this to a virtual function. - main_view = CreateViewInner( - ViewParams(params.target, 0, static_cast<u32>(params.GetNumLayers()), 0, params.num_levels), - true); + u32 num_layers = 1; + if (params.is_layered || params.target == SurfaceTarget::Texture3D) { + num_layers = params.depth; + } + main_view = CreateView(ViewParams(params.target, 0, num_layers, 0, params.num_levels)); } CachedSurface::~CachedSurface() = default; @@ -227,7 +234,7 @@ void CachedSurface::UploadTexture(const std::vector<u8>& staging_buffer) { void CachedSurface::DownloadTexture(std::vector<u8>& staging_buffer) { UNIMPLEMENTED_IF(params.IsBuffer()); - if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5U) { + if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5_UNORM) { LOG_WARNING(Render_Vulkan, "A1B5G5R5 flushing is stubbed"); } @@ -256,12 +263,8 @@ void CachedSurface::DecorateSurfaceName() { } View CachedSurface::CreateView(const ViewParams& params) { - return CreateViewInner(params, false); -} - -View CachedSurface::CreateViewInner(const ViewParams& params, bool is_proxy) { // TODO(Rodrigo): Add name decorations - return views[params] = std::make_shared<CachedSurfaceView>(device, *this, params, is_proxy); + return views[params] = std::make_shared<CachedSurfaceView>(device, *this, params); } void CachedSurface::UploadBuffer(const std::vector<u8>& staging_buffer) { @@ -279,12 +282,10 @@ void CachedSurface::UploadBuffer(const std::vector<u8>& staging_buffer) { VkBufferMemoryBarrier barrier; barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; barrier.pNext = nullptr; - barrier.srcAccessMask = VK_PIPELINE_STAGE_TRANSFER_BIT; - barrier.dstAccessMask = VK_PIPELINE_STAGE_VERTEX_SHADER_BIT; - barrier.srcQueueFamilyIndex = VK_ACCESS_TRANSFER_WRITE_BIT; - barrier.dstQueueFamilyIndex = VK_ACCESS_SHADER_READ_BIT; - barrier.srcQueueFamilyIndex = 0; - barrier.dstQueueFamilyIndex = 0; + barrier.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT; + barrier.dstAccessMask = VK_ACCESS_SHADER_READ_BIT; + barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; // They'll be ignored anyway + barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; barrier.buffer = dst_buffer; barrier.offset = 0; barrier.size = size; @@ -321,22 +322,25 @@ void CachedSurface::UploadImage(const std::vector<u8>& staging_buffer) { } VkBufferImageCopy CachedSurface::GetBufferImageCopy(u32 level) const { - VkBufferImageCopy copy; - copy.bufferOffset = params.GetHostMipmapLevelOffset(level, is_converted); - copy.bufferRowLength = 0; - copy.bufferImageHeight = 0; - copy.imageSubresource.aspectMask = image->GetAspectMask(); - copy.imageSubresource.mipLevel = level; - copy.imageSubresource.baseArrayLayer = 0; - copy.imageSubresource.layerCount = static_cast<u32>(params.GetNumLayers()); - copy.imageOffset.x = 0; - copy.imageOffset.y = 0; - copy.imageOffset.z = 0; - copy.imageExtent.width = params.GetMipWidth(level); - copy.imageExtent.height = params.GetMipHeight(level); - copy.imageExtent.depth = - params.target == SurfaceTarget::Texture3D ? params.GetMipDepth(level) : 1; - return copy; + return { + .bufferOffset = params.GetHostMipmapLevelOffset(level, is_converted), + .bufferRowLength = 0, + .bufferImageHeight = 0, + .imageSubresource = + { + .aspectMask = image->GetAspectMask(), + .mipLevel = level, + .baseArrayLayer = 0, + .layerCount = static_cast<u32>(params.GetNumLayers()), + }, + .imageOffset = {.x = 0, .y = 0, .z = 0}, + .imageExtent = + { + .width = params.GetMipWidth(level), + .height = params.GetMipHeight(level), + .depth = params.target == SurfaceTarget::Texture3D ? params.GetMipDepth(level) : 1U, + }, + }; } VkImageSubresourceRange CachedSurface::GetImageSubresourceRange() const { @@ -345,38 +349,44 @@ VkImageSubresourceRange CachedSurface::GetImageSubresourceRange() const { } CachedSurfaceView::CachedSurfaceView(const VKDevice& device, CachedSurface& surface, - const ViewParams& params, bool is_proxy) + const ViewParams& params) : VideoCommon::ViewBase{params}, params{surface.GetSurfaceParams()}, image{surface.GetImageHandle()}, buffer_view{surface.GetBufferViewHandle()}, aspect_mask{surface.GetAspectMask()}, device{device}, surface{surface}, - base_layer{params.base_layer}, num_layers{params.num_layers}, base_level{params.base_level}, - num_levels{params.num_levels}, image_view_type{image ? GetImageViewType(params.target) - : VK_IMAGE_VIEW_TYPE_1D} {} + base_level{params.base_level}, num_levels{params.num_levels}, + image_view_type{image ? GetImageViewType(params.target) : VK_IMAGE_VIEW_TYPE_1D} { + if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { + base_layer = 0; + num_layers = 1; + base_slice = params.base_layer; + num_slices = params.num_layers; + } else { + base_layer = params.base_layer; + num_layers = params.num_layers; + } +} CachedSurfaceView::~CachedSurfaceView() = default; -VkImageView CachedSurfaceView::GetHandle(SwizzleSource x_source, SwizzleSource y_source, - SwizzleSource z_source, SwizzleSource w_source) { - const u32 swizzle = EncodeSwizzle(x_source, y_source, z_source, w_source); - if (last_image_view && last_swizzle == swizzle) { +VkImageView CachedSurfaceView::GetImageView(SwizzleSource x_source, SwizzleSource y_source, + SwizzleSource z_source, SwizzleSource w_source) { + const u32 new_swizzle = EncodeSwizzle(x_source, y_source, z_source, w_source); + if (last_image_view && last_swizzle == new_swizzle) { return last_image_view; } - last_swizzle = swizzle; + last_swizzle = new_swizzle; - const auto [entry, is_cache_miss] = view_cache.try_emplace(swizzle); + const auto [entry, is_cache_miss] = view_cache.try_emplace(new_swizzle); auto& image_view = entry->second; if (!is_cache_miss) { return last_image_view = *image_view; } - auto swizzle_x = MaxwellToVK::SwizzleSource(x_source); - auto swizzle_y = MaxwellToVK::SwizzleSource(y_source); - auto swizzle_z = MaxwellToVK::SwizzleSource(z_source); - auto swizzle_w = MaxwellToVK::SwizzleSource(w_source); - - if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5U) { + std::array swizzle{MaxwellToVK::SwizzleSource(x_source), MaxwellToVK::SwizzleSource(y_source), + MaxwellToVK::SwizzleSource(z_source), MaxwellToVK::SwizzleSource(w_source)}; + if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5_UNORM) { // A1B5G5R5 is implemented as A1R5G5B5, we have to change the swizzle here. - std::swap(swizzle_x, swizzle_z); + std::swap(swizzle[0], swizzle[2]); } // Games can sample depth or stencil values on textures. This is decided by the swizzle value on @@ -386,11 +396,11 @@ VkImageView CachedSurfaceView::GetHandle(SwizzleSource x_source, SwizzleSource y UNIMPLEMENTED_IF(x_source != SwizzleSource::R && x_source != SwizzleSource::G); const bool is_first = x_source == SwizzleSource::R; switch (params.pixel_format) { - case VideoCore::Surface::PixelFormat::Z24S8: - case VideoCore::Surface::PixelFormat::Z32FS8: + case VideoCore::Surface::PixelFormat::D24_UNORM_S8_UINT: + case VideoCore::Surface::PixelFormat::D32_FLOAT_S8_UINT: aspect = is_first ? VK_IMAGE_ASPECT_DEPTH_BIT : VK_IMAGE_ASPECT_STENCIL_BIT; break; - case VideoCore::Surface::PixelFormat::S8Z24: + case VideoCore::Surface::PixelFormat::S8_UINT_D24_UNORM: aspect = is_first ? VK_IMAGE_ASPECT_STENCIL_BIT : VK_IMAGE_ASPECT_DEPTH_BIT; break; default: @@ -398,44 +408,100 @@ VkImageView CachedSurfaceView::GetHandle(SwizzleSource x_source, SwizzleSource y UNIMPLEMENTED(); } - // Vulkan doesn't seem to understand swizzling of a depth stencil image, use identity - swizzle_x = VK_COMPONENT_SWIZZLE_R; - swizzle_y = VK_COMPONENT_SWIZZLE_G; - swizzle_z = VK_COMPONENT_SWIZZLE_B; - swizzle_w = VK_COMPONENT_SWIZZLE_A; + // Make sure we sample the first component + std::transform( + swizzle.begin(), swizzle.end(), swizzle.begin(), [](VkComponentSwizzle component) { + return component == VK_COMPONENT_SWIZZLE_G ? VK_COMPONENT_SWIZZLE_R : component; + }); } - VkImageViewCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.image = surface.GetImageHandle(); - ci.viewType = image_view_type; - ci.format = surface.GetImage().GetFormat(); - ci.components = {swizzle_x, swizzle_y, swizzle_z, swizzle_w}; - ci.subresourceRange.aspectMask = aspect; - ci.subresourceRange.baseMipLevel = base_level; - ci.subresourceRange.levelCount = num_levels; - ci.subresourceRange.baseArrayLayer = base_layer; - ci.subresourceRange.layerCount = num_layers; - image_view = device.GetLogical().CreateImageView(ci); + if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { + ASSERT(base_slice == 0); + ASSERT(num_slices == params.depth); + } + + image_view = device.GetLogical().CreateImageView({ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = surface.GetImageHandle(), + .viewType = image_view_type, + .format = surface.GetImage().GetFormat(), + .components = + { + .r = swizzle[0], + .g = swizzle[1], + .b = swizzle[2], + .a = swizzle[3], + }, + .subresourceRange = + { + .aspectMask = aspect, + .baseMipLevel = base_level, + .levelCount = num_levels, + .baseArrayLayer = base_layer, + .layerCount = num_layers, + }, + }); return last_image_view = *image_view; } -VKTextureCache::VKTextureCache(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - const VKDevice& device, VKResourceManager& resource_manager, - VKMemoryManager& memory_manager, VKScheduler& scheduler, - VKStagingBufferPool& staging_pool) - : TextureCache(system, rasterizer, device.IsOptimalAstcSupported()), device{device}, - resource_manager{resource_manager}, memory_manager{memory_manager}, scheduler{scheduler}, - staging_pool{staging_pool} {} +VkImageView CachedSurfaceView::GetAttachment() { + if (render_target) { + return *render_target; + } + + VkImageViewCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = surface.GetImageHandle(), + .viewType = VK_IMAGE_VIEW_TYPE_1D, + .format = surface.GetImage().GetFormat(), + .components = + { + .r = VK_COMPONENT_SWIZZLE_IDENTITY, + .g = VK_COMPONENT_SWIZZLE_IDENTITY, + .b = VK_COMPONENT_SWIZZLE_IDENTITY, + .a = VK_COMPONENT_SWIZZLE_IDENTITY, + }, + .subresourceRange = + { + .aspectMask = aspect_mask, + .baseMipLevel = base_level, + .levelCount = num_levels, + .baseArrayLayer = 0, + .layerCount = 0, + }, + }; + if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { + ci.viewType = num_slices > 1 ? VK_IMAGE_VIEW_TYPE_2D_ARRAY : VK_IMAGE_VIEW_TYPE_2D; + ci.subresourceRange.baseArrayLayer = base_slice; + ci.subresourceRange.layerCount = num_slices; + } else { + ci.viewType = image_view_type; + ci.subresourceRange.baseArrayLayer = base_layer; + ci.subresourceRange.layerCount = num_layers; + } + render_target = device.GetLogical().CreateImageView(ci); + return *render_target; +} + +VKTextureCache::VKTextureCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::MemoryManager& gpu_memory, const VKDevice& device_, + VKMemoryManager& memory_manager_, VKScheduler& scheduler_, + VKStagingBufferPool& staging_pool_) + : TextureCache(rasterizer, maxwell3d, gpu_memory, device_.IsOptimalAstcSupported()), + device{device_}, memory_manager{memory_manager_}, scheduler{scheduler_}, staging_pool{ + staging_pool_} {} VKTextureCache::~VKTextureCache() = default; Surface VKTextureCache::CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) { - return std::make_shared<CachedSurface>(system, device, resource_manager, memory_manager, - scheduler, staging_pool, gpu_addr, params); + return std::make_shared<CachedSurface>(device, memory_manager, scheduler, staging_pool, + gpu_addr, params); } void VKTextureCache::ImageCopy(Surface& src_surface, Surface& dst_surface, @@ -462,24 +528,40 @@ void VKTextureCache::ImageCopy(Surface& src_surface, Surface& dst_surface, VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_WRITE_BIT, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL); - VkImageCopy copy; - copy.srcSubresource.aspectMask = src_surface->GetAspectMask(); - copy.srcSubresource.mipLevel = copy_params.source_level; - copy.srcSubresource.baseArrayLayer = copy_params.source_z; - copy.srcSubresource.layerCount = num_layers; - copy.srcOffset.x = copy_params.source_x; - copy.srcOffset.y = copy_params.source_y; - copy.srcOffset.z = 0; - copy.dstSubresource.aspectMask = dst_surface->GetAspectMask(); - copy.dstSubresource.mipLevel = copy_params.dest_level; - copy.dstSubresource.baseArrayLayer = dst_base_layer; - copy.dstSubresource.layerCount = num_layers; - copy.dstOffset.x = copy_params.dest_x; - copy.dstOffset.y = copy_params.dest_y; - copy.dstOffset.z = dst_offset_z; - copy.extent.width = copy_params.width; - copy.extent.height = copy_params.height; - copy.extent.depth = extent_z; + const VkImageCopy copy{ + .srcSubresource = + { + .aspectMask = src_surface->GetAspectMask(), + .mipLevel = copy_params.source_level, + .baseArrayLayer = copy_params.source_z, + .layerCount = num_layers, + }, + .srcOffset = + { + .x = static_cast<s32>(copy_params.source_x), + .y = static_cast<s32>(copy_params.source_y), + .z = 0, + }, + .dstSubresource = + { + .aspectMask = dst_surface->GetAspectMask(), + .mipLevel = copy_params.dest_level, + .baseArrayLayer = dst_base_layer, + .layerCount = num_layers, + }, + .dstOffset = + { + .x = static_cast<s32>(copy_params.dest_x), + .y = static_cast<s32>(copy_params.dest_y), + .z = static_cast<s32>(dst_offset_z), + }, + .extent = + { + .width = copy_params.width, + .height = copy_params.height, + .depth = extent_z, + }, + }; const VkImage src_image = src_surface->GetImageHandle(); const VkImage dst_image = dst_surface->GetImageHandle(); diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index 115595f28..39202feba 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -7,23 +7,13 @@ #include <memory> #include <unordered_map> -#include "common/assert.h" #include "common/common_types.h" -#include "common/logging/log.h" -#include "common/math_util.h" -#include "video_core/gpu.h" -#include "video_core/rasterizer_cache.h" #include "video_core/renderer_vulkan/vk_image.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/wrapper.h" #include "video_core/texture_cache/surface_base.h" #include "video_core/texture_cache/texture_cache.h" -#include "video_core/textures/decoders.h" - -namespace Core { -class System; -} namespace VideoCore { class RasterizerInterface; @@ -33,7 +23,6 @@ namespace Vulkan { class RasterizerVulkan; class VKDevice; -class VKResourceManager; class VKScheduler; class VKStagingBufferPool; @@ -51,8 +40,7 @@ class CachedSurface final : public VideoCommon::SurfaceBase<View> { friend CachedSurfaceView; public: - explicit CachedSurface(Core::System& system, const VKDevice& device, - VKResourceManager& resource_manager, VKMemoryManager& memory_manager, + explicit CachedSurface(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler, VKStagingBufferPool& staging_pool, GPUVAddr gpu_addr, const SurfaceParams& params); ~CachedSurface(); @@ -97,7 +85,6 @@ protected: void DecorateSurfaceName(); View CreateView(const ViewParams& params) override; - View CreateViewInner(const ViewParams& params, bool is_proxy); private: void UploadBuffer(const std::vector<u8>& staging_buffer); @@ -108,9 +95,7 @@ private: VkImageSubresourceRange GetImageSubresourceRange() const; - Core::System& system; const VKDevice& device; - VKResourceManager& resource_manager; VKMemoryManager& memory_manager; VKScheduler& scheduler; VKStagingBufferPool& staging_pool; @@ -126,23 +111,20 @@ private: class CachedSurfaceView final : public VideoCommon::ViewBase { public: explicit CachedSurfaceView(const VKDevice& device, CachedSurface& surface, - const ViewParams& params, bool is_proxy); + const ViewParams& params); ~CachedSurfaceView(); - VkImageView GetHandle(Tegra::Texture::SwizzleSource x_source, - Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, - Tegra::Texture::SwizzleSource w_source); + VkImageView GetImageView(Tegra::Texture::SwizzleSource x_source, + Tegra::Texture::SwizzleSource y_source, + Tegra::Texture::SwizzleSource z_source, + Tegra::Texture::SwizzleSource w_source); + + VkImageView GetAttachment(); bool IsSameSurface(const CachedSurfaceView& rhs) const { return &surface == &rhs.surface; } - VkImageView GetHandle() { - return GetHandle(Tegra::Texture::SwizzleSource::R, Tegra::Texture::SwizzleSource::G, - Tegra::Texture::SwizzleSource::B, Tegra::Texture::SwizzleSource::A); - } - u32 GetWidth() const { return params.GetMipWidth(base_level); } @@ -186,14 +168,6 @@ public: } private: - static u32 EncodeSwizzle(Tegra::Texture::SwizzleSource x_source, - Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, - Tegra::Texture::SwizzleSource w_source) { - return (static_cast<u32>(x_source) << 24) | (static_cast<u32>(y_source) << 16) | - (static_cast<u32>(z_source) << 8) | static_cast<u32>(w_source); - } - // Store a copy of these values to avoid double dereference when reading them const SurfaceParams params; const VkImage image; @@ -202,24 +176,27 @@ private: const VKDevice& device; CachedSurface& surface; - const u32 base_layer; - const u32 num_layers; const u32 base_level; const u32 num_levels; const VkImageViewType image_view_type; + u32 base_layer = 0; + u32 num_layers = 0; + u32 base_slice = 0; + u32 num_slices = 0; VkImageView last_image_view = nullptr; u32 last_swizzle = 0; + vk::ImageView render_target; std::unordered_map<u32, vk::ImageView> view_cache; }; class VKTextureCache final : public TextureCacheBase { public: - explicit VKTextureCache(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - const VKDevice& device, VKResourceManager& resource_manager, - VKMemoryManager& memory_manager, VKScheduler& scheduler, - VKStagingBufferPool& staging_pool); + explicit VKTextureCache(VideoCore::RasterizerInterface& rasterizer, + Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, + const VKDevice& device, VKMemoryManager& memory_manager, + VKScheduler& scheduler, VKStagingBufferPool& staging_pool); ~VKTextureCache(); private: @@ -234,7 +211,6 @@ private: void BufferCopy(Surface& src_surface, Surface& dst_surface) override; const VKDevice& device; - VKResourceManager& resource_manager; VKMemoryManager& memory_manager; VKScheduler& scheduler; VKStagingBufferPool& staging_pool; diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp index 4bfec0077..351c048d2 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp @@ -24,34 +24,25 @@ void VKUpdateDescriptorQueue::TickFrame() { } void VKUpdateDescriptorQueue::Acquire() { - entries.clear(); -} + // Minimum number of entries required. + // This is the maximum number of entries a single draw call migth use. + static constexpr std::size_t MIN_ENTRIES = 0x400; -void VKUpdateDescriptorQueue::Send(VkDescriptorUpdateTemplateKHR update_template, - VkDescriptorSet set) { - if (payload.size() + entries.size() >= payload.max_size()) { + if (payload.size() + MIN_ENTRIES >= payload.max_size()) { LOG_WARNING(Render_Vulkan, "Payload overflow, waiting for worker thread"); scheduler.WaitWorker(); payload.clear(); } + upload_start = &*payload.end(); +} - const auto payload_start = payload.data() + payload.size(); - for (const auto& entry : entries) { - if (const auto image = std::get_if<VkDescriptorImageInfo>(&entry)) { - payload.push_back(*image); - } else if (const auto buffer = std::get_if<Buffer>(&entry)) { - payload.emplace_back(*buffer->buffer, buffer->offset, buffer->size); - } else if (const auto texel = std::get_if<VkBufferView>(&entry)) { - payload.push_back(*texel); - } else { - UNREACHABLE(); - } - } - - scheduler.Record( - [payload_start, set, update_template, logical = &device.GetLogical()](vk::CommandBuffer) { - logical->UpdateDescriptorSet(set, update_template, payload_start); - }); +void VKUpdateDescriptorQueue::Send(VkDescriptorUpdateTemplateKHR update_template, + VkDescriptorSet set) { + const void* const data = upload_start; + const vk::Device* const logical = &device.GetLogical(); + scheduler.Record([data, logical, set, update_template](vk::CommandBuffer) { + logical->UpdateDescriptorSet(set, update_template, data); + }); } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.h b/src/video_core/renderer_vulkan/vk_update_descriptor.h index a9e3d5dba..945320c72 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.h +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.h @@ -4,7 +4,6 @@ #pragma once -#include <type_traits> #include <variant> #include <boost/container/static_vector.hpp> @@ -16,18 +15,13 @@ namespace Vulkan { class VKDevice; class VKScheduler; -class DescriptorUpdateEntry { -public: - explicit DescriptorUpdateEntry() : image{} {} - - DescriptorUpdateEntry(VkDescriptorImageInfo image) : image{image} {} +struct DescriptorUpdateEntry { + DescriptorUpdateEntry(VkDescriptorImageInfo image_) : image{image_} {} - DescriptorUpdateEntry(VkBuffer buffer, VkDeviceSize offset, VkDeviceSize size) - : buffer{buffer, offset, size} {} + DescriptorUpdateEntry(VkDescriptorBufferInfo buffer_) : buffer{buffer_} {} - DescriptorUpdateEntry(VkBufferView texel_buffer) : texel_buffer{texel_buffer} {} + DescriptorUpdateEntry(VkBufferView texel_buffer_) : texel_buffer{texel_buffer_} {} -private: union { VkDescriptorImageInfo image; VkDescriptorBufferInfo buffer; @@ -47,37 +41,34 @@ public: void Send(VkDescriptorUpdateTemplateKHR update_template, VkDescriptorSet set); void AddSampledImage(VkSampler sampler, VkImageView image_view) { - entries.emplace_back(VkDescriptorImageInfo{sampler, image_view, {}}); + payload.emplace_back(VkDescriptorImageInfo{sampler, image_view, {}}); } void AddImage(VkImageView image_view) { - entries.emplace_back(VkDescriptorImageInfo{{}, image_view, {}}); + payload.emplace_back(VkDescriptorImageInfo{{}, image_view, {}}); } - void AddBuffer(const VkBuffer* buffer, u64 offset, std::size_t size) { - entries.push_back(Buffer{buffer, offset, size}); + void AddBuffer(VkBuffer buffer, u64 offset, std::size_t size) { + payload.emplace_back(VkDescriptorBufferInfo{buffer, offset, size}); } void AddTexelBuffer(VkBufferView texel_buffer) { - entries.emplace_back(texel_buffer); + payload.emplace_back(texel_buffer); } - VkImageLayout* GetLastImageLayout() { - return &std::get<VkDescriptorImageInfo>(entries.back()).imageLayout; + VkImageLayout* LastImageLayout() { + return &payload.back().image.imageLayout; } -private: - struct Buffer { - const VkBuffer* buffer = nullptr; - u64 offset = 0; - std::size_t size = 0; - }; - using Variant = std::variant<VkDescriptorImageInfo, Buffer, VkBufferView>; + const VkImageLayout* LastImageLayout() const { + return &payload.back().image.imageLayout; + } +private: const VKDevice& device; VKScheduler& scheduler; - boost::container::static_vector<Variant, 0x400> entries; + const DescriptorUpdateEntry* upload_start = nullptr; boost::container::static_vector<DescriptorUpdateEntry, 0x10000> payload; }; diff --git a/src/video_core/renderer_vulkan/wrapper.cpp b/src/video_core/renderer_vulkan/wrapper.cpp index 9b94dfff1..4e83303d8 100644 --- a/src/video_core/renderer_vulkan/wrapper.cpp +++ b/src/video_core/renderer_vulkan/wrapper.cpp @@ -2,13 +2,16 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> #include <exception> #include <memory> #include <optional> +#include <string_view> #include <utility> #include <vector> #include "common/common_types.h" +#include "common/logging/log.h" #include "video_core/renderer_vulkan/wrapper.h" @@ -16,6 +19,44 @@ namespace Vulkan::vk { namespace { +template <typename Func> +void SortPhysicalDevices(std::vector<VkPhysicalDevice>& devices, const InstanceDispatch& dld, + Func&& func) { + // Calling GetProperties calls Vulkan more than needed. But they are supposed to be cheap + // functions. + std::stable_sort(devices.begin(), devices.end(), + [&dld, &func](VkPhysicalDevice lhs, VkPhysicalDevice rhs) { + return func(vk::PhysicalDevice(lhs, dld).GetProperties(), + vk::PhysicalDevice(rhs, dld).GetProperties()); + }); +} + +void SortPhysicalDevicesPerVendor(std::vector<VkPhysicalDevice>& devices, + const InstanceDispatch& dld, + std::initializer_list<u32> vendor_ids) { + for (auto it = vendor_ids.end(); it != vendor_ids.begin();) { + --it; + SortPhysicalDevices(devices, dld, [id = *it](const auto& lhs, const auto& rhs) { + return lhs.vendorID == id && rhs.vendorID != id; + }); + } +} + +void SortPhysicalDevices(std::vector<VkPhysicalDevice>& devices, const InstanceDispatch& dld) { + // Sort by name, this will set a base and make GPUs with higher numbers appear first + // (e.g. GTX 1650 will intentionally be listed before a GTX 1080). + SortPhysicalDevices(devices, dld, [](const auto& lhs, const auto& rhs) { + return std::string_view{lhs.deviceName} > std::string_view{rhs.deviceName}; + }); + // Prefer discrete over non-discrete + SortPhysicalDevices(devices, dld, [](const auto& lhs, const auto& rhs) { + return lhs.deviceType == VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU && + rhs.deviceType != VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU; + }); + // Prefer Nvidia over AMD, AMD over Intel, Intel over the rest. + SortPhysicalDevicesPerVendor(devices, dld, {0x10DE, 0x1002, 0x8086}); +} + template <typename T> bool Proc(T& result, const InstanceDispatch& dld, const char* proc_name, VkInstance instance = nullptr) noexcept { @@ -61,14 +102,25 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkCmdPipelineBarrier); X(vkCmdPushConstants); X(vkCmdSetBlendConstants); - X(vkCmdSetCheckpointNV); X(vkCmdSetDepthBias); X(vkCmdSetDepthBounds); + X(vkCmdSetEvent); X(vkCmdSetScissor); X(vkCmdSetStencilCompareMask); X(vkCmdSetStencilReference); X(vkCmdSetStencilWriteMask); X(vkCmdSetViewport); + X(vkCmdWaitEvents); + X(vkCmdBindVertexBuffers2EXT); + X(vkCmdSetCullModeEXT); + X(vkCmdSetDepthBoundsTestEnableEXT); + X(vkCmdSetDepthCompareOpEXT); + X(vkCmdSetDepthTestEnableEXT); + X(vkCmdSetDepthWriteEnableEXT); + X(vkCmdSetFrontFaceEXT); + X(vkCmdSetPrimitiveTopologyEXT); + X(vkCmdSetStencilOpEXT); + X(vkCmdSetStencilTestEnableEXT); X(vkCreateBuffer); X(vkCreateBufferView); X(vkCreateCommandPool); @@ -76,6 +128,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkCreateDescriptorPool); X(vkCreateDescriptorSetLayout); X(vkCreateDescriptorUpdateTemplateKHR); + X(vkCreateEvent); X(vkCreateFence); X(vkCreateFramebuffer); X(vkCreateGraphicsPipelines); @@ -94,6 +147,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkDestroyDescriptorPool); X(vkDestroyDescriptorSetLayout); X(vkDestroyDescriptorUpdateTemplateKHR); + X(vkDestroyEvent); X(vkDestroyFence); X(vkDestroyFramebuffer); X(vkDestroyImage); @@ -113,10 +167,11 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkFreeMemory); X(vkGetBufferMemoryRequirements); X(vkGetDeviceQueue); + X(vkGetEventStatus); X(vkGetFenceStatus); X(vkGetImageMemoryRequirements); X(vkGetQueryPoolResults); - X(vkGetQueueCheckpointDataNV); + X(vkGetSemaphoreCounterValueKHR); X(vkMapMemory); X(vkQueueSubmit); X(vkResetFences); @@ -125,6 +180,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkUpdateDescriptorSetWithTemplateKHR); X(vkUpdateDescriptorSets); X(vkWaitForFences); + X(vkWaitSemaphoresKHR); #undef X } @@ -132,7 +188,8 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { bool Load(InstanceDispatch& dld) noexcept { #define X(name) Proc(dld.name, dld, #name) - return X(vkCreateInstance) && X(vkEnumerateInstanceExtensionProperties); + return X(vkCreateInstance) && X(vkEnumerateInstanceExtensionProperties) && + X(vkEnumerateInstanceLayerProperties); #undef X } @@ -230,6 +287,22 @@ const char* ToString(VkResult result) noexcept { return "VK_ERROR_INVALID_DEVICE_ADDRESS_EXT"; case VkResult::VK_ERROR_FULL_SCREEN_EXCLUSIVE_MODE_LOST_EXT: return "VK_ERROR_FULL_SCREEN_EXCLUSIVE_MODE_LOST_EXT"; + case VkResult::VK_ERROR_UNKNOWN: + return "VK_ERROR_UNKNOWN"; + case VkResult::VK_ERROR_INCOMPATIBLE_VERSION_KHR: + return "VK_ERROR_INCOMPATIBLE_VERSION_KHR"; + case VkResult::VK_THREAD_IDLE_KHR: + return "VK_THREAD_IDLE_KHR"; + case VkResult::VK_THREAD_DONE_KHR: + return "VK_THREAD_DONE_KHR"; + case VkResult::VK_OPERATION_DEFERRED_KHR: + return "VK_OPERATION_DEFERRED_KHR"; + case VkResult::VK_OPERATION_NOT_DEFERRED_KHR: + return "VK_OPERATION_NOT_DEFERRED_KHR"; + case VkResult::VK_PIPELINE_COMPILE_REQUIRED_EXT: + return "VK_PIPELINE_COMPILE_REQUIRED_EXT"; + case VkResult::VK_RESULT_MAX_ENUM: + return "VK_RESULT_MAX_ENUM"; } return "Unknown"; } @@ -271,6 +344,10 @@ void Destroy(VkDevice device, VkDeviceMemory handle, const DeviceDispatch& dld) dld.vkFreeMemory(device, handle, nullptr); } +void Destroy(VkDevice device, VkEvent handle, const DeviceDispatch& dld) noexcept { + dld.vkDestroyEvent(device, handle, nullptr); +} + void Destroy(VkDevice device, VkFence handle, const DeviceDispatch& dld) noexcept { dld.vkDestroyFence(device, handle, nullptr); } @@ -339,26 +416,27 @@ VkResult Free(VkDevice device, VkCommandPool handle, Span<VkCommandBuffer> buffe return VK_SUCCESS; } -Instance Instance::Create(Span<const char*> layers, Span<const char*> extensions, +Instance Instance::Create(u32 version, Span<const char*> layers, Span<const char*> extensions, InstanceDispatch& dld) noexcept { - VkApplicationInfo application_info; - application_info.sType = VK_STRUCTURE_TYPE_APPLICATION_INFO; - application_info.pNext = nullptr; - application_info.pApplicationName = "yuzu Emulator"; - application_info.applicationVersion = VK_MAKE_VERSION(0, 1, 0); - application_info.pEngineName = "yuzu Emulator"; - application_info.engineVersion = VK_MAKE_VERSION(0, 1, 0); - application_info.apiVersion = VK_API_VERSION_1_1; - - VkInstanceCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; - ci.pApplicationInfo = &application_info; - ci.enabledLayerCount = layers.size(); - ci.ppEnabledLayerNames = layers.data(); - ci.enabledExtensionCount = extensions.size(); - ci.ppEnabledExtensionNames = extensions.data(); + const VkApplicationInfo application_info{ + .sType = VK_STRUCTURE_TYPE_APPLICATION_INFO, + .pNext = nullptr, + .pApplicationName = "yuzu Emulator", + .applicationVersion = VK_MAKE_VERSION(0, 1, 0), + .pEngineName = "yuzu Emulator", + .engineVersion = VK_MAKE_VERSION(0, 1, 0), + .apiVersion = version, + }; + const VkInstanceCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .pApplicationInfo = &application_info, + .enabledLayerCount = layers.size(), + .ppEnabledLayerNames = layers.data(), + .enabledExtensionCount = extensions.size(), + .ppEnabledExtensionNames = extensions.data(), + }; VkInstance instance; if (dld.vkCreateInstance(&ci, nullptr, &instance) != VK_SUCCESS) { @@ -383,24 +461,26 @@ std::optional<std::vector<VkPhysicalDevice>> Instance::EnumeratePhysicalDevices( if (dld->vkEnumeratePhysicalDevices(handle, &num, physical_devices.data()) != VK_SUCCESS) { return std::nullopt; } - return physical_devices; + SortPhysicalDevices(physical_devices, *dld); + return std::make_optional(std::move(physical_devices)); } DebugCallback Instance::TryCreateDebugCallback( PFN_vkDebugUtilsMessengerCallbackEXT callback) noexcept { - VkDebugUtilsMessengerCreateInfoEXT ci; - ci.sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT; - ci.pNext = nullptr; - ci.flags = 0; - ci.messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT; - ci.messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT; - ci.pfnUserCallback = callback; - ci.pUserData = nullptr; + const VkDebugUtilsMessengerCreateInfoEXT ci{ + .sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT, + .pNext = nullptr, + .flags = 0, + .messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT, + .messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT, + .pfnUserCallback = callback, + .pUserData = nullptr, + }; VkDebugUtilsMessengerEXT messenger; if (dld->vkCreateDebugUtilsMessengerEXT(handle, &ci, nullptr, &messenger) != VK_SUCCESS) { @@ -409,17 +489,6 @@ DebugCallback Instance::TryCreateDebugCallback( return DebugCallback(messenger, handle, *dld); } -std::vector<VkCheckpointDataNV> Queue::GetCheckpointDataNV(const DeviceDispatch& dld) const { - if (!dld.vkGetQueueCheckpointDataNV) { - return {}; - } - u32 num; - dld.vkGetQueueCheckpointDataNV(queue, &num, nullptr); - std::vector<VkCheckpointDataNV> checkpoints(num); - dld.vkGetQueueCheckpointDataNV(queue, &num, checkpoints.data()); - return checkpoints; -} - void Buffer::BindMemory(VkDeviceMemory memory, VkDeviceSize offset) const { Check(dld->vkBindBufferMemory(owner, handle, memory, offset)); } @@ -442,12 +511,13 @@ DescriptorSets DescriptorPool::Allocate(const VkDescriptorSetAllocateInfo& ai) c } CommandBuffers CommandPool::Allocate(std::size_t num_buffers, VkCommandBufferLevel level) const { - VkCommandBufferAllocateInfo ai; - ai.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO; - ai.pNext = nullptr; - ai.commandPool = handle; - ai.level = level; - ai.commandBufferCount = static_cast<u32>(num_buffers); + const VkCommandBufferAllocateInfo ai{ + .sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO, + .pNext = nullptr, + .commandPool = handle, + .level = level, + .commandBufferCount = static_cast<u32>(num_buffers), + }; std::unique_ptr buffers = std::make_unique<VkCommandBuffer[]>(num_buffers); switch (const VkResult result = dld->vkAllocateCommandBuffers(owner, &ai, buffers.get())) { @@ -469,20 +539,20 @@ std::vector<VkImage> SwapchainKHR::GetImages() const { } Device Device::Create(VkPhysicalDevice physical_device, Span<VkDeviceQueueCreateInfo> queues_ci, - Span<const char*> enabled_extensions, - const VkPhysicalDeviceFeatures2& enabled_features, + Span<const char*> enabled_extensions, const void* next, DeviceDispatch& dld) noexcept { - VkDeviceCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO; - ci.pNext = &enabled_features; - ci.flags = 0; - ci.queueCreateInfoCount = queues_ci.size(); - ci.pQueueCreateInfos = queues_ci.data(); - ci.enabledLayerCount = 0; - ci.ppEnabledLayerNames = nullptr; - ci.enabledExtensionCount = enabled_extensions.size(); - ci.ppEnabledExtensionNames = enabled_extensions.data(); - ci.pEnabledFeatures = nullptr; + const VkDeviceCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO, + .pNext = next, + .flags = 0, + .queueCreateInfoCount = queues_ci.size(), + .pQueueCreateInfos = queues_ci.data(), + .enabledLayerCount = 0, + .ppEnabledLayerNames = nullptr, + .enabledExtensionCount = enabled_extensions.size(), + .ppEnabledExtensionNames = enabled_extensions.data(), + .pEnabledFeatures = nullptr, + }; VkDevice device; if (dld.vkCreateDevice(physical_device, &ci, nullptr, &device) != VK_SUCCESS) { @@ -523,11 +593,15 @@ ImageView Device::CreateImageView(const VkImageViewCreateInfo& ci) const { } Semaphore Device::CreateSemaphore() const { - VkSemaphoreCreateInfo ci; - ci.sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO; - ci.pNext = nullptr; - ci.flags = 0; + static constexpr VkSemaphoreCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + }; + return CreateSemaphore(ci); +} +Semaphore Device::CreateSemaphore(const VkSemaphoreCreateInfo& ci) const { VkSemaphore object; Check(dld->vkCreateSemaphore(handle, &ci, nullptr, &object)); return Semaphore(object, handle, *dld); @@ -613,6 +687,18 @@ ShaderModule Device::CreateShaderModule(const VkShaderModuleCreateInfo& ci) cons return ShaderModule(object, handle, *dld); } +Event Device::CreateEvent() const { + static constexpr VkEventCreateInfo ci{ + .sType = VK_STRUCTURE_TYPE_EVENT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + }; + + VkEvent object; + Check(dld->vkCreateEvent(handle, &ci, nullptr, &object)); + return Event(object, handle, *dld); +} + SwapchainKHR Device::CreateSwapchainKHR(const VkSwapchainCreateInfoKHR& ci) const { VkSwapchainKHR object; Check(dld->vkCreateSwapchainKHR(handle, &ci, nullptr, &object)); @@ -701,8 +787,7 @@ bool PhysicalDevice::GetSurfaceSupportKHR(u32 queue_family_index, VkSurfaceKHR s return supported == VK_TRUE; } -VkSurfaceCapabilitiesKHR PhysicalDevice::GetSurfaceCapabilitiesKHR(VkSurfaceKHR surface) const - noexcept { +VkSurfaceCapabilitiesKHR PhysicalDevice::GetSurfaceCapabilitiesKHR(VkSurfaceKHR surface) const { VkSurfaceCapabilitiesKHR capabilities; Check(dld->vkGetPhysicalDeviceSurfaceCapabilitiesKHR(physical_device, surface, &capabilities)); return capabilities; @@ -733,6 +818,21 @@ VkPhysicalDeviceMemoryProperties PhysicalDevice::GetMemoryProperties() const noe return properties; } +u32 AvailableVersion(const InstanceDispatch& dld) noexcept { + PFN_vkEnumerateInstanceVersion vkEnumerateInstanceVersion; + if (!Proc(vkEnumerateInstanceVersion, dld, "vkEnumerateInstanceVersion")) { + // If the procedure is not found, Vulkan 1.0 is assumed + return VK_API_VERSION_1_0; + } + u32 version; + if (const VkResult result = vkEnumerateInstanceVersion(&version); result != VK_SUCCESS) { + LOG_ERROR(Render_Vulkan, "vkEnumerateInstanceVersion returned {}, assuming Vulkan 1.1", + ToString(result)); + return VK_API_VERSION_1_1; + } + return version; +} + std::optional<std::vector<VkExtensionProperties>> EnumerateInstanceExtensionProperties( const InstanceDispatch& dld) { u32 num; @@ -747,4 +847,17 @@ std::optional<std::vector<VkExtensionProperties>> EnumerateInstanceExtensionProp return properties; } +std::optional<std::vector<VkLayerProperties>> EnumerateInstanceLayerProperties( + const InstanceDispatch& dld) { + u32 num; + if (dld.vkEnumerateInstanceLayerProperties(&num, nullptr) != VK_SUCCESS) { + return std::nullopt; + } + std::vector<VkLayerProperties> properties(num); + if (dld.vkEnumerateInstanceLayerProperties(&num, properties.data()) != VK_SUCCESS) { + return std::nullopt; + } + return properties; +} + } // namespace Vulkan::vk diff --git a/src/video_core/renderer_vulkan/wrapper.h b/src/video_core/renderer_vulkan/wrapper.h index fb3657819..f64919623 100644 --- a/src/video_core/renderer_vulkan/wrapper.h +++ b/src/video_core/renderer_vulkan/wrapper.h @@ -141,6 +141,7 @@ struct InstanceDispatch { PFN_vkCreateInstance vkCreateInstance; PFN_vkDestroyInstance vkDestroyInstance; PFN_vkEnumerateInstanceExtensionProperties vkEnumerateInstanceExtensionProperties; + PFN_vkEnumerateInstanceLayerProperties vkEnumerateInstanceLayerProperties; PFN_vkCreateDebugUtilsMessengerEXT vkCreateDebugUtilsMessengerEXT; PFN_vkCreateDevice vkCreateDevice; @@ -197,14 +198,25 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkCmdPipelineBarrier vkCmdPipelineBarrier; PFN_vkCmdPushConstants vkCmdPushConstants; PFN_vkCmdSetBlendConstants vkCmdSetBlendConstants; - PFN_vkCmdSetCheckpointNV vkCmdSetCheckpointNV; PFN_vkCmdSetDepthBias vkCmdSetDepthBias; PFN_vkCmdSetDepthBounds vkCmdSetDepthBounds; + PFN_vkCmdSetEvent vkCmdSetEvent; PFN_vkCmdSetScissor vkCmdSetScissor; PFN_vkCmdSetStencilCompareMask vkCmdSetStencilCompareMask; PFN_vkCmdSetStencilReference vkCmdSetStencilReference; PFN_vkCmdSetStencilWriteMask vkCmdSetStencilWriteMask; PFN_vkCmdSetViewport vkCmdSetViewport; + PFN_vkCmdWaitEvents vkCmdWaitEvents; + PFN_vkCmdBindVertexBuffers2EXT vkCmdBindVertexBuffers2EXT; + PFN_vkCmdSetCullModeEXT vkCmdSetCullModeEXT; + PFN_vkCmdSetDepthBoundsTestEnableEXT vkCmdSetDepthBoundsTestEnableEXT; + PFN_vkCmdSetDepthCompareOpEXT vkCmdSetDepthCompareOpEXT; + PFN_vkCmdSetDepthTestEnableEXT vkCmdSetDepthTestEnableEXT; + PFN_vkCmdSetDepthWriteEnableEXT vkCmdSetDepthWriteEnableEXT; + PFN_vkCmdSetFrontFaceEXT vkCmdSetFrontFaceEXT; + PFN_vkCmdSetPrimitiveTopologyEXT vkCmdSetPrimitiveTopologyEXT; + PFN_vkCmdSetStencilOpEXT vkCmdSetStencilOpEXT; + PFN_vkCmdSetStencilTestEnableEXT vkCmdSetStencilTestEnableEXT; PFN_vkCreateBuffer vkCreateBuffer; PFN_vkCreateBufferView vkCreateBufferView; PFN_vkCreateCommandPool vkCreateCommandPool; @@ -212,6 +224,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkCreateDescriptorPool vkCreateDescriptorPool; PFN_vkCreateDescriptorSetLayout vkCreateDescriptorSetLayout; PFN_vkCreateDescriptorUpdateTemplateKHR vkCreateDescriptorUpdateTemplateKHR; + PFN_vkCreateEvent vkCreateEvent; PFN_vkCreateFence vkCreateFence; PFN_vkCreateFramebuffer vkCreateFramebuffer; PFN_vkCreateGraphicsPipelines vkCreateGraphicsPipelines; @@ -230,6 +243,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkDestroyDescriptorPool vkDestroyDescriptorPool; PFN_vkDestroyDescriptorSetLayout vkDestroyDescriptorSetLayout; PFN_vkDestroyDescriptorUpdateTemplateKHR vkDestroyDescriptorUpdateTemplateKHR; + PFN_vkDestroyEvent vkDestroyEvent; PFN_vkDestroyFence vkDestroyFence; PFN_vkDestroyFramebuffer vkDestroyFramebuffer; PFN_vkDestroyImage vkDestroyImage; @@ -249,10 +263,11 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkFreeMemory vkFreeMemory; PFN_vkGetBufferMemoryRequirements vkGetBufferMemoryRequirements; PFN_vkGetDeviceQueue vkGetDeviceQueue; + PFN_vkGetEventStatus vkGetEventStatus; PFN_vkGetFenceStatus vkGetFenceStatus; PFN_vkGetImageMemoryRequirements vkGetImageMemoryRequirements; PFN_vkGetQueryPoolResults vkGetQueryPoolResults; - PFN_vkGetQueueCheckpointDataNV vkGetQueueCheckpointDataNV; + PFN_vkGetSemaphoreCounterValueKHR vkGetSemaphoreCounterValueKHR; PFN_vkMapMemory vkMapMemory; PFN_vkQueueSubmit vkQueueSubmit; PFN_vkResetFences vkResetFences; @@ -261,6 +276,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkUpdateDescriptorSetWithTemplateKHR vkUpdateDescriptorSetWithTemplateKHR; PFN_vkUpdateDescriptorSets vkUpdateDescriptorSets; PFN_vkWaitForFences vkWaitForFences; + PFN_vkWaitSemaphoresKHR vkWaitSemaphoresKHR; }; /// Loads instance agnostic function pointers. @@ -281,6 +297,7 @@ void Destroy(VkDevice, VkDescriptorPool, const DeviceDispatch&) noexcept; void Destroy(VkDevice, VkDescriptorSetLayout, const DeviceDispatch&) noexcept; void Destroy(VkDevice, VkDescriptorUpdateTemplateKHR, const DeviceDispatch&) noexcept; void Destroy(VkDevice, VkDeviceMemory, const DeviceDispatch&) noexcept; +void Destroy(VkDevice, VkEvent, const DeviceDispatch&) noexcept; void Destroy(VkDevice, VkFence, const DeviceDispatch&) noexcept; void Destroy(VkDevice, VkFramebuffer, const DeviceDispatch&) noexcept; void Destroy(VkDevice, VkImage, const DeviceDispatch&) noexcept; @@ -535,7 +552,6 @@ using PipelineLayout = Handle<VkPipelineLayout, VkDevice, DeviceDispatch>; using QueryPool = Handle<VkQueryPool, VkDevice, DeviceDispatch>; using RenderPass = Handle<VkRenderPass, VkDevice, DeviceDispatch>; using Sampler = Handle<VkSampler, VkDevice, DeviceDispatch>; -using Semaphore = Handle<VkSemaphore, VkDevice, DeviceDispatch>; using ShaderModule = Handle<VkShaderModule, VkDevice, DeviceDispatch>; using SurfaceKHR = Handle<VkSurfaceKHR, VkInstance, InstanceDispatch>; @@ -548,7 +564,7 @@ class Instance : public Handle<VkInstance, NoOwner, InstanceDispatch> { public: /// Creates a Vulkan instance. Use "operator bool" for error handling. - static Instance Create(Span<const char*> layers, Span<const char*> extensions, + static Instance Create(u32 version, Span<const char*> layers, Span<const char*> extensions, InstanceDispatch& dld) noexcept; /// Enumerates physical devices. @@ -567,12 +583,9 @@ public: /// Construct a queue handle. constexpr Queue(VkQueue queue, const DeviceDispatch& dld) noexcept : queue{queue}, dld{&dld} {} - /// Returns the checkpoint data. - /// @note Returns an empty vector when the function pointer is not present. - std::vector<VkCheckpointDataNV> GetCheckpointDataNV(const DeviceDispatch& dld) const; - - void Submit(Span<VkSubmitInfo> submit_infos, VkFence fence) const { - Check(dld->vkQueueSubmit(queue, submit_infos.size(), submit_infos.data(), fence)); + VkResult Submit(Span<VkSubmitInfo> submit_infos, + VkFence fence = VK_NULL_HANDLE) const noexcept { + return dld->vkQueueSubmit(queue, submit_infos.size(), submit_infos.data(), fence); } VkResult Present(const VkPresentInfoKHR& present_info) const noexcept { @@ -654,13 +667,59 @@ public: std::vector<VkImage> GetImages() const; }; +class Event : public Handle<VkEvent, VkDevice, DeviceDispatch> { + using Handle<VkEvent, VkDevice, DeviceDispatch>::Handle; + +public: + VkResult GetStatus() const noexcept { + return dld->vkGetEventStatus(owner, handle); + } +}; + +class Semaphore : public Handle<VkSemaphore, VkDevice, DeviceDispatch> { + using Handle<VkSemaphore, VkDevice, DeviceDispatch>::Handle; + +public: + [[nodiscard]] u64 GetCounter() const { + u64 value; + Check(dld->vkGetSemaphoreCounterValueKHR(owner, handle, &value)); + return value; + } + + /** + * Waits for a timeline semaphore on the host. + * + * @param value Value to wait + * @param timeout Time in nanoseconds to timeout + * @return True on successful wait, false on timeout + */ + bool Wait(u64 value, u64 timeout = std::numeric_limits<u64>::max()) const { + const VkSemaphoreWaitInfoKHR wait_info{ + .sType = VK_STRUCTURE_TYPE_SEMAPHORE_WAIT_INFO_KHR, + .pNext = nullptr, + .flags = 0, + .semaphoreCount = 1, + .pSemaphores = &handle, + .pValues = &value, + }; + const VkResult result = dld->vkWaitSemaphoresKHR(owner, &wait_info, timeout); + switch (result) { + case VK_SUCCESS: + return true; + case VK_TIMEOUT: + return false; + default: + throw Exception(result); + } + } +}; + class Device : public Handle<VkDevice, NoOwner, DeviceDispatch> { using Handle<VkDevice, NoOwner, DeviceDispatch>::Handle; public: static Device Create(VkPhysicalDevice physical_device, Span<VkDeviceQueueCreateInfo> queues_ci, - Span<const char*> enabled_extensions, - const VkPhysicalDeviceFeatures2& enabled_features, + Span<const char*> enabled_extensions, const void* next, DeviceDispatch& dld) noexcept; Queue GetQueue(u32 family_index) const noexcept; @@ -675,6 +734,8 @@ public: Semaphore CreateSemaphore() const; + Semaphore CreateSemaphore(const VkSemaphoreCreateInfo& ci) const; + Fence CreateFence(const VkFenceCreateInfo& ci) const; DescriptorPool CreateDescriptorPool(const VkDescriptorPoolCreateInfo& ci) const; @@ -702,6 +763,8 @@ public: ShaderModule CreateShaderModule(const VkShaderModuleCreateInfo& ci) const; + Event CreateEvent() const; + SwapchainKHR CreateSwapchainKHR(const VkSwapchainCreateInfoKHR& ci) const; DeviceMemory TryAllocateMemory(const VkMemoryAllocateInfo& ai) const noexcept; @@ -734,18 +797,11 @@ public: dld->vkResetQueryPoolEXT(handle, query_pool, first, count); } - void GetQueryResults(VkQueryPool query_pool, u32 first, u32 count, std::size_t data_size, - void* data, VkDeviceSize stride, VkQueryResultFlags flags) const { - Check(dld->vkGetQueryPoolResults(handle, query_pool, first, count, data_size, data, stride, - flags)); - } - - template <typename T> - T GetQueryResult(VkQueryPool query_pool, u32 first, VkQueryResultFlags flags) const { - static_assert(std::is_trivially_copyable_v<T>); - T value; - GetQueryResults(query_pool, first, 1, sizeof(T), &value, sizeof(T), flags); - return value; + VkResult GetQueryResults(VkQueryPool query_pool, u32 first, u32 count, std::size_t data_size, + void* data, VkDeviceSize stride, + VkQueryResultFlags flags) const noexcept { + return dld->vkGetQueryPoolResults(handle, query_pool, first, count, data_size, data, stride, + flags); } }; @@ -776,7 +832,7 @@ public: bool GetSurfaceSupportKHR(u32 queue_family_index, VkSurfaceKHR) const; - VkSurfaceCapabilitiesKHR GetSurfaceCapabilitiesKHR(VkSurfaceKHR) const noexcept; + VkSurfaceCapabilitiesKHR GetSurfaceCapabilitiesKHR(VkSurfaceKHR) const; std::vector<VkSurfaceFormatKHR> GetSurfaceFormatsKHR(VkSurfaceKHR) const; @@ -835,8 +891,8 @@ public: dld->vkCmdBindPipeline(handle, bind_point, pipeline); } - void BindIndexBuffer(VkBuffer buffer, VkDeviceSize offset, VkIndexType index_type) const - noexcept { + void BindIndexBuffer(VkBuffer buffer, VkDeviceSize offset, + VkIndexType index_type) const noexcept { dld->vkCmdBindIndexBuffer(handle, buffer, offset, index_type); } @@ -849,8 +905,8 @@ public: BindVertexBuffers(binding, 1, &buffer, &offset); } - void Draw(u32 vertex_count, u32 instance_count, u32 first_vertex, u32 first_instance) const - noexcept { + void Draw(u32 vertex_count, u32 instance_count, u32 first_vertex, + u32 first_instance) const noexcept { dld->vkCmdDraw(handle, vertex_count, instance_count, first_vertex, first_instance); } @@ -860,15 +916,15 @@ public: first_instance); } - void ClearAttachments(Span<VkClearAttachment> attachments, Span<VkClearRect> rects) const - noexcept { + void ClearAttachments(Span<VkClearAttachment> attachments, + Span<VkClearRect> rects) const noexcept { dld->vkCmdClearAttachments(handle, attachments.size(), attachments.data(), rects.size(), rects.data()); } void BlitImage(VkImage src_image, VkImageLayout src_layout, VkImage dst_image, - VkImageLayout dst_layout, Span<VkImageBlit> regions, VkFilter filter) const - noexcept { + VkImageLayout dst_layout, Span<VkImageBlit> regions, + VkFilter filter) const noexcept { dld->vkCmdBlitImage(handle, src_image, src_layout, dst_image, dst_layout, regions.size(), regions.data(), filter); } @@ -893,8 +949,8 @@ public: regions.data()); } - void CopyBuffer(VkBuffer src_buffer, VkBuffer dst_buffer, Span<VkBufferCopy> regions) const - noexcept { + void CopyBuffer(VkBuffer src_buffer, VkBuffer dst_buffer, + Span<VkBufferCopy> regions) const noexcept { dld->vkCmdCopyBuffer(handle, src_buffer, dst_buffer, regions.size(), regions.data()); } @@ -910,8 +966,8 @@ public: regions.data()); } - void FillBuffer(VkBuffer dst_buffer, VkDeviceSize dst_offset, VkDeviceSize size, u32 data) const - noexcept { + void FillBuffer(VkBuffer dst_buffer, VkDeviceSize dst_offset, VkDeviceSize size, + u32 data) const noexcept { dld->vkCmdFillBuffer(handle, dst_buffer, dst_offset, size, data); } @@ -920,10 +976,6 @@ public: dld->vkCmdPushConstants(handle, layout, flags, offset, size, values); } - void SetCheckpointNV(const void* checkpoint_marker) const noexcept { - dld->vkCmdSetCheckpointNV(handle, checkpoint_marker); - } - void SetViewport(u32 first, Span<VkViewport> viewports) const noexcept { dld->vkCmdSetViewport(handle, first, viewports.size(), viewports.data()); } @@ -956,6 +1008,63 @@ public: dld->vkCmdSetDepthBounds(handle, min_depth_bounds, max_depth_bounds); } + void SetEvent(VkEvent event, VkPipelineStageFlags stage_flags) const noexcept { + dld->vkCmdSetEvent(handle, event, stage_flags); + } + + void WaitEvents(Span<VkEvent> events, VkPipelineStageFlags src_stage_mask, + VkPipelineStageFlags dst_stage_mask, Span<VkMemoryBarrier> memory_barriers, + Span<VkBufferMemoryBarrier> buffer_barriers, + Span<VkImageMemoryBarrier> image_barriers) const noexcept { + dld->vkCmdWaitEvents(handle, events.size(), events.data(), src_stage_mask, dst_stage_mask, + memory_barriers.size(), memory_barriers.data(), buffer_barriers.size(), + buffer_barriers.data(), image_barriers.size(), image_barriers.data()); + } + + void BindVertexBuffers2EXT(u32 first_binding, u32 binding_count, const VkBuffer* buffers, + const VkDeviceSize* offsets, const VkDeviceSize* sizes, + const VkDeviceSize* strides) const noexcept { + dld->vkCmdBindVertexBuffers2EXT(handle, first_binding, binding_count, buffers, offsets, + sizes, strides); + } + + void SetCullModeEXT(VkCullModeFlags cull_mode) const noexcept { + dld->vkCmdSetCullModeEXT(handle, cull_mode); + } + + void SetDepthBoundsTestEnableEXT(bool enable) const noexcept { + dld->vkCmdSetDepthBoundsTestEnableEXT(handle, enable ? VK_TRUE : VK_FALSE); + } + + void SetDepthCompareOpEXT(VkCompareOp compare_op) const noexcept { + dld->vkCmdSetDepthCompareOpEXT(handle, compare_op); + } + + void SetDepthTestEnableEXT(bool enable) const noexcept { + dld->vkCmdSetDepthTestEnableEXT(handle, enable ? VK_TRUE : VK_FALSE); + } + + void SetDepthWriteEnableEXT(bool enable) const noexcept { + dld->vkCmdSetDepthWriteEnableEXT(handle, enable ? VK_TRUE : VK_FALSE); + } + + void SetFrontFaceEXT(VkFrontFace front_face) const noexcept { + dld->vkCmdSetFrontFaceEXT(handle, front_face); + } + + void SetPrimitiveTopologyEXT(VkPrimitiveTopology primitive_topology) const noexcept { + dld->vkCmdSetPrimitiveTopologyEXT(handle, primitive_topology); + } + + void SetStencilOpEXT(VkStencilFaceFlags face_mask, VkStencilOp fail_op, VkStencilOp pass_op, + VkStencilOp depth_fail_op, VkCompareOp compare_op) const noexcept { + dld->vkCmdSetStencilOpEXT(handle, face_mask, fail_op, pass_op, depth_fail_op, compare_op); + } + + void SetStencilTestEnableEXT(bool enable) const noexcept { + dld->vkCmdSetStencilTestEnableEXT(handle, enable ? VK_TRUE : VK_FALSE); + } + void BindTransformFeedbackBuffersEXT(u32 first, u32 count, const VkBuffer* buffers, const VkDeviceSize* offsets, const VkDeviceSize* sizes) const noexcept { @@ -981,7 +1090,12 @@ private: const DeviceDispatch* dld; }; +u32 AvailableVersion(const InstanceDispatch& dld) noexcept; + std::optional<std::vector<VkExtensionProperties>> EnumerateInstanceExtensionProperties( const InstanceDispatch& dld); +std::optional<std::vector<VkLayerProperties>> EnumerateInstanceLayerProperties( + const InstanceDispatch& dld); + } // namespace Vulkan::vk diff --git a/src/video_core/shader/ast.h b/src/video_core/shader/ast.h index cca13bcde..8e5a22ab3 100644 --- a/src/video_core/shader/ast.h +++ b/src/video_core/shader/ast.h @@ -199,55 +199,48 @@ public: } std::optional<u32> GetGotoLabel() const { - auto inner = std::get_if<ASTGoto>(&data); - if (inner) { + if (const auto* inner = std::get_if<ASTGoto>(&data)) { return {inner->label}; } - return {}; + return std::nullopt; } Expr GetGotoCondition() const { - auto inner = std::get_if<ASTGoto>(&data); - if (inner) { + if (const auto* inner = std::get_if<ASTGoto>(&data)) { return inner->condition; } return nullptr; } void MarkLabelUnused() { - auto inner = std::get_if<ASTLabel>(&data); - if (inner) { + if (auto* inner = std::get_if<ASTLabel>(&data)) { inner->unused = true; } } bool IsLabelUnused() const { - auto inner = std::get_if<ASTLabel>(&data); - if (inner) { + if (const auto* inner = std::get_if<ASTLabel>(&data)) { return inner->unused; } return true; } std::optional<u32> GetLabelIndex() const { - auto inner = std::get_if<ASTLabel>(&data); - if (inner) { + if (const auto* inner = std::get_if<ASTLabel>(&data)) { return {inner->index}; } - return {}; + return std::nullopt; } Expr GetIfCondition() const { - auto inner = std::get_if<ASTIfThen>(&data); - if (inner) { + if (const auto* inner = std::get_if<ASTIfThen>(&data)) { return inner->condition; } return nullptr; } void SetGotoCondition(Expr new_condition) { - auto inner = std::get_if<ASTGoto>(&data); - if (inner) { + if (auto* inner = std::get_if<ASTGoto>(&data)) { inner->condition = std::move(new_condition); } } diff --git a/src/video_core/shader/async_shaders.cpp b/src/video_core/shader/async_shaders.cpp new file mode 100644 index 000000000..6920afdf2 --- /dev/null +++ b/src/video_core/shader/async_shaders.cpp @@ -0,0 +1,216 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <condition_variable> +#include <mutex> +#include <thread> +#include <vector> +#include "video_core/engines/maxwell_3d.h" +#include "video_core/renderer_base.h" +#include "video_core/renderer_opengl/gl_shader_cache.h" +#include "video_core/shader/async_shaders.h" + +namespace VideoCommon::Shader { + +AsyncShaders::AsyncShaders(Core::Frontend::EmuWindow& emu_window) : emu_window(emu_window) {} + +AsyncShaders::~AsyncShaders() { + KillWorkers(); +} + +void AsyncShaders::AllocateWorkers() { + // Use at least one thread + u32 num_workers = 1; + + // Deduce how many more threads we can use + const u32 thread_count = std::thread::hardware_concurrency(); + if (thread_count >= 8) { + // Increase async workers by 1 for every 2 threads >= 8 + num_workers += 1 + (thread_count - 8) / 2; + } + + // If we already have workers queued, ignore + if (num_workers == worker_threads.size()) { + return; + } + + // If workers already exist, clear them + if (!worker_threads.empty()) { + FreeWorkers(); + } + + // Create workers + for (std::size_t i = 0; i < num_workers; i++) { + context_list.push_back(emu_window.CreateSharedContext()); + worker_threads.emplace_back(&AsyncShaders::ShaderCompilerThread, this, + context_list[i].get()); + } +} + +void AsyncShaders::FreeWorkers() { + // Mark all threads to quit + is_thread_exiting.store(true); + cv.notify_all(); + for (auto& thread : worker_threads) { + thread.join(); + } + // Clear our shared contexts + context_list.clear(); + + // Clear our worker threads + worker_threads.clear(); +} + +void AsyncShaders::KillWorkers() { + is_thread_exiting.store(true); + for (auto& thread : worker_threads) { + thread.detach(); + } + // Clear our shared contexts + context_list.clear(); + + // Clear our worker threads + worker_threads.clear(); +} + +bool AsyncShaders::HasWorkQueued() const { + return !pending_queue.empty(); +} + +bool AsyncShaders::HasCompletedWork() const { + std::shared_lock lock{completed_mutex}; + return !finished_work.empty(); +} + +bool AsyncShaders::IsShaderAsync(const Tegra::GPU& gpu) const { + const auto& regs = gpu.Maxwell3D().regs; + + // If something is using depth, we can assume that games are not rendering anything which will + // be used one time. + if (regs.zeta_enable) { + return true; + } + + // If games are using a small index count, we can assume these are full screen quads. Usually + // these shaders are only used once for building textures so we can assume they can't be built + // async + if (regs.index_array.count <= 6 || regs.vertex_buffer.count <= 6) { + return false; + } + + return true; +} + +std::vector<AsyncShaders::Result> AsyncShaders::GetCompletedWork() { + std::vector<Result> results; + { + std::unique_lock lock{completed_mutex}; + results = std::move(finished_work); + finished_work.clear(); + } + return results; +} + +void AsyncShaders::QueueOpenGLShader(const OpenGL::Device& device, + Tegra::Engines::ShaderType shader_type, u64 uid, + std::vector<u64> code, std::vector<u64> code_b, + u32 main_offset, CompilerSettings compiler_settings, + const Registry& registry, VAddr cpu_addr) { + std::unique_lock lock(queue_mutex); + pending_queue.push({ + .backend = device.UseAssemblyShaders() ? Backend::GLASM : Backend::OpenGL, + .device = &device, + .shader_type = shader_type, + .uid = uid, + .code = std::move(code), + .code_b = std::move(code_b), + .main_offset = main_offset, + .compiler_settings = compiler_settings, + .registry = registry, + .cpu_address = cpu_addr, + }); + cv.notify_one(); +} + +void AsyncShaders::QueueVulkanShader(Vulkan::VKPipelineCache* pp_cache, + const Vulkan::VKDevice& device, Vulkan::VKScheduler& scheduler, + Vulkan::VKDescriptorPool& descriptor_pool, + Vulkan::VKUpdateDescriptorQueue& update_descriptor_queue, + Vulkan::VKRenderPassCache& renderpass_cache, + std::vector<VkDescriptorSetLayoutBinding> bindings, + Vulkan::SPIRVProgram program, + Vulkan::GraphicsPipelineCacheKey key) { + std::unique_lock lock(queue_mutex); + pending_queue.push({ + .backend = Backend::Vulkan, + .pp_cache = pp_cache, + .vk_device = &device, + .scheduler = &scheduler, + .descriptor_pool = &descriptor_pool, + .update_descriptor_queue = &update_descriptor_queue, + .renderpass_cache = &renderpass_cache, + .bindings = std::move(bindings), + .program = std::move(program), + .key = key, + }); + cv.notify_one(); +} + +void AsyncShaders::ShaderCompilerThread(Core::Frontend::GraphicsContext* context) { + while (!is_thread_exiting.load(std::memory_order_relaxed)) { + std::unique_lock lock{queue_mutex}; + cv.wait(lock, [this] { return HasWorkQueued() || is_thread_exiting; }); + if (is_thread_exiting) { + return; + } + + // Partial lock to allow all threads to read at the same time + if (!HasWorkQueued()) { + continue; + } + // Another thread beat us, just unlock and wait for the next load + if (pending_queue.empty()) { + continue; + } + + // Pull work from queue + WorkerParams work = std::move(pending_queue.front()); + pending_queue.pop(); + lock.unlock(); + + if (work.backend == Backend::OpenGL || work.backend == Backend::GLASM) { + const ShaderIR ir(work.code, work.main_offset, work.compiler_settings, *work.registry); + const auto scope = context->Acquire(); + auto program = + OpenGL::BuildShader(*work.device, work.shader_type, work.uid, ir, *work.registry); + Result result{}; + result.backend = work.backend; + result.cpu_address = work.cpu_address; + result.uid = work.uid; + result.code = std::move(work.code); + result.code_b = std::move(work.code_b); + result.shader_type = work.shader_type; + + if (work.backend == Backend::OpenGL) { + result.program.opengl = std::move(program->source_program); + } else if (work.backend == Backend::GLASM) { + result.program.glasm = std::move(program->assembly_program); + } + + { + std::unique_lock complete_lock(completed_mutex); + finished_work.push_back(std::move(result)); + } + } else if (work.backend == Backend::Vulkan) { + auto pipeline = std::make_unique<Vulkan::VKGraphicsPipeline>( + *work.vk_device, *work.scheduler, *work.descriptor_pool, + *work.update_descriptor_queue, *work.renderpass_cache, work.key, work.bindings, + work.program); + + work.pp_cache->EmplacePipeline(std::move(pipeline)); + } + } +} + +} // namespace VideoCommon::Shader diff --git a/src/video_core/shader/async_shaders.h b/src/video_core/shader/async_shaders.h new file mode 100644 index 000000000..7a99e1dc5 --- /dev/null +++ b/src/video_core/shader/async_shaders.h @@ -0,0 +1,147 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <condition_variable> +#include <memory> +#include <shared_mutex> +#include <thread> + +// This header includes both Vulkan and OpenGL headers, this has to be fixed +// Unfortunately, including OpenGL will include Windows.h that defines macros that can cause issues. +// Forcefully include glad early and undefine macros +#include <glad/glad.h> +#ifdef CreateEvent +#undef CreateEvent +#endif +#ifdef CreateSemaphore +#undef CreateSemaphore +#endif + +#include "common/common_types.h" +#include "video_core/renderer_opengl/gl_device.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/renderer_opengl/gl_shader_decompiler.h" +#include "video_core/renderer_vulkan/vk_device.h" +#include "video_core/renderer_vulkan/vk_pipeline_cache.h" +#include "video_core/renderer_vulkan/vk_scheduler.h" + +namespace Core::Frontend { +class EmuWindow; +class GraphicsContext; +} // namespace Core::Frontend + +namespace Tegra { +class GPU; +} + +namespace Vulkan { +class VKPipelineCache; +} + +namespace VideoCommon::Shader { + +class AsyncShaders { +public: + enum class Backend { + OpenGL, + GLASM, + Vulkan, + }; + + struct ResultPrograms { + OpenGL::OGLProgram opengl; + OpenGL::OGLAssemblyProgram glasm; + }; + + struct Result { + u64 uid; + VAddr cpu_address; + Backend backend; + ResultPrograms program; + std::vector<u64> code; + std::vector<u64> code_b; + Tegra::Engines::ShaderType shader_type; + }; + + explicit AsyncShaders(Core::Frontend::EmuWindow& emu_window); + ~AsyncShaders(); + + /// Start up shader worker threads + void AllocateWorkers(); + + /// Clear the shader queue and kill all worker threads + void FreeWorkers(); + + // Force end all threads + void KillWorkers(); + + /// Check to see if any shaders have actually been compiled + [[nodiscard]] bool HasCompletedWork() const; + + /// Deduce if a shader can be build on another thread of MUST be built in sync. We cannot build + /// every shader async as some shaders are only built and executed once. We try to "guess" which + /// shader would be used only once + [[nodiscard]] bool IsShaderAsync(const Tegra::GPU& gpu) const; + + /// Pulls completed compiled shaders + [[nodiscard]] std::vector<Result> GetCompletedWork(); + + void QueueOpenGLShader(const OpenGL::Device& device, Tegra::Engines::ShaderType shader_type, + u64 uid, std::vector<u64> code, std::vector<u64> code_b, u32 main_offset, + CompilerSettings compiler_settings, const Registry& registry, + VAddr cpu_addr); + + void QueueVulkanShader(Vulkan::VKPipelineCache* pp_cache, const Vulkan::VKDevice& device, + Vulkan::VKScheduler& scheduler, + Vulkan::VKDescriptorPool& descriptor_pool, + Vulkan::VKUpdateDescriptorQueue& update_descriptor_queue, + Vulkan::VKRenderPassCache& renderpass_cache, + std::vector<VkDescriptorSetLayoutBinding> bindings, + Vulkan::SPIRVProgram program, Vulkan::GraphicsPipelineCacheKey key); + +private: + void ShaderCompilerThread(Core::Frontend::GraphicsContext* context); + + /// Check our worker queue to see if we have any work queued already + [[nodiscard]] bool HasWorkQueued() const; + + struct WorkerParams { + Backend backend; + // For OGL + const OpenGL::Device* device; + Tegra::Engines::ShaderType shader_type; + u64 uid; + std::vector<u64> code; + std::vector<u64> code_b; + u32 main_offset; + CompilerSettings compiler_settings; + std::optional<Registry> registry; + VAddr cpu_address; + + // For Vulkan + Vulkan::VKPipelineCache* pp_cache; + const Vulkan::VKDevice* vk_device; + Vulkan::VKScheduler* scheduler; + Vulkan::VKDescriptorPool* descriptor_pool; + Vulkan::VKUpdateDescriptorQueue* update_descriptor_queue; + Vulkan::VKRenderPassCache* renderpass_cache; + std::vector<VkDescriptorSetLayoutBinding> bindings; + Vulkan::SPIRVProgram program; + Vulkan::GraphicsPipelineCacheKey key; + }; + + std::condition_variable cv; + mutable std::mutex queue_mutex; + mutable std::shared_mutex completed_mutex; + std::atomic<bool> is_thread_exiting{}; + std::vector<std::unique_ptr<Core::Frontend::GraphicsContext>> context_list; + std::vector<std::thread> worker_threads; + std::queue<WorkerParams> pending_queue; + std::vector<Result> finished_work; + Core::Frontend::EmuWindow& emu_window; +}; + +} // namespace VideoCommon::Shader diff --git a/src/video_core/shader/control_flow.cpp b/src/video_core/shader/control_flow.cpp index 2e2711350..4c8971615 100644 --- a/src/video_core/shader/control_flow.cpp +++ b/src/video_core/shader/control_flow.cpp @@ -13,6 +13,7 @@ #include "common/common_types.h" #include "video_core/shader/ast.h" #include "video_core/shader/control_flow.h" +#include "video_core/shader/memory_util.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" @@ -115,17 +116,6 @@ Pred GetPredicate(u32 index, bool negated) { return static_cast<Pred>(static_cast<u64>(index) + (negated ? 8ULL : 0ULL)); } -/** - * Returns whether the instruction at the specified offset is a 'sched' instruction. - * Sched instructions always appear before a sequence of 3 instructions. - */ -constexpr bool IsSchedInstruction(u32 offset, u32 main_offset) { - constexpr u32 SchedPeriod = 4; - u32 absolute_offset = offset - main_offset; - - return (absolute_offset % SchedPeriod) == 0; -} - enum class ParseResult : u32 { ControlCaught, BlockEnd, @@ -197,24 +187,26 @@ std::optional<std::pair<BufferInfo, u64>> TrackLDC(const CFGRebuildState& state, std::optional<u64> TrackSHLRegister(const CFGRebuildState& state, u32& pos, u64 ldc_tracked_register) { - return TrackInstruction<u64>(state, pos, - [ldc_tracked_register](auto instr, const auto& opcode) { - return opcode.GetId() == OpCode::Id::SHL_IMM && - instr.gpr0.Value() == ldc_tracked_register; - }, - [](auto instr, const auto&) { return instr.gpr8.Value(); }); + return TrackInstruction<u64>( + state, pos, + [ldc_tracked_register](auto instr, const auto& opcode) { + return opcode.GetId() == OpCode::Id::SHL_IMM && + instr.gpr0.Value() == ldc_tracked_register; + }, + [](auto instr, const auto&) { return instr.gpr8.Value(); }); } std::optional<u32> TrackIMNMXValue(const CFGRebuildState& state, u32& pos, u64 shl_tracked_register) { - return TrackInstruction<u32>(state, pos, - [shl_tracked_register](auto instr, const auto& opcode) { - return opcode.GetId() == OpCode::Id::IMNMX_IMM && - instr.gpr0.Value() == shl_tracked_register; - }, - [](auto instr, const auto&) { - return static_cast<u32>(instr.alu.GetSignedImm20_20() + 1); - }); + return TrackInstruction<u32>( + state, pos, + [shl_tracked_register](auto instr, const auto& opcode) { + return opcode.GetId() == OpCode::Id::IMNMX_IMM && + instr.gpr0.Value() == shl_tracked_register; + }, + [](auto instr, const auto&) { + return static_cast<u32>(instr.alu.GetSignedImm20_20() + 1); + }); } std::optional<BranchIndirectInfo> TrackBranchIndirectInfo(const CFGRebuildState& state, u32 pos) { @@ -484,17 +476,17 @@ bool TryInspectAddress(CFGRebuildState& state) { } case BlockCollision::Inside: { // This case is the tricky one: - // We need to Split the block in 2 sepparate blocks + // We need to split the block into 2 separate blocks const u32 end = state.block_info[block_index].end; BlockInfo& new_block = CreateBlockInfo(state, address, end); BlockInfo& current_block = state.block_info[block_index]; current_block.end = address - 1; - new_block.branch = current_block.branch; + new_block.branch = std::move(current_block.branch); BlockBranchInfo forward_branch = MakeBranchInfo<SingleBranch>(); const auto branch = std::get_if<SingleBranch>(forward_branch.get()); branch->address = address; branch->ignore = true; - current_block.branch = forward_branch; + current_block.branch = std::move(forward_branch); return true; } default: @@ -555,13 +547,13 @@ bool TryQuery(CFGRebuildState& state) { gather_labels(q2.ssy_stack, state.ssy_labels, block); gather_labels(q2.pbk_stack, state.pbk_labels, block); if (std::holds_alternative<SingleBranch>(*block.branch)) { - const auto branch = std::get_if<SingleBranch>(block.branch.get()); + auto* branch = std::get_if<SingleBranch>(block.branch.get()); if (!branch->condition.IsUnconditional()) { q2.address = block.end + 1; state.queries.push_back(q2); } - Query conditional_query{q2}; + auto& conditional_query = state.queries.emplace_back(q2); if (branch->is_sync) { if (branch->address == unassigned_branch) { branch->address = conditional_query.ssy_stack.top(); @@ -575,23 +567,21 @@ bool TryQuery(CFGRebuildState& state) { conditional_query.pbk_stack.pop(); } conditional_query.address = branch->address; - state.queries.push_back(std::move(conditional_query)); return true; } - const auto multi_branch = std::get_if<MultiBranch>(block.branch.get()); + + const auto* multi_branch = std::get_if<MultiBranch>(block.branch.get()); for (const auto& branch_case : multi_branch->branches) { - Query conditional_query{q2}; + auto& conditional_query = state.queries.emplace_back(q2); conditional_query.address = branch_case.address; - state.queries.push_back(std::move(conditional_query)); } + return true; } -} // Anonymous namespace - void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) { - const auto get_expr = ([&](const Condition& cond) -> Expr { - Expr result{}; + const auto get_expr = [](const Condition& cond) -> Expr { + Expr result; if (cond.cc != ConditionCode::T) { result = MakeExpr<ExprCondCode>(cond.cc); } @@ -604,10 +594,10 @@ void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) { } Expr extra = MakeExpr<ExprPredicate>(pred); if (negate) { - extra = MakeExpr<ExprNot>(extra); + extra = MakeExpr<ExprNot>(std::move(extra)); } if (result) { - return MakeExpr<ExprAnd>(extra, result); + return MakeExpr<ExprAnd>(std::move(extra), std::move(result)); } return extra; } @@ -615,9 +605,10 @@ void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) { return result; } return MakeExpr<ExprBoolean>(true); - }); + }; + if (std::holds_alternative<SingleBranch>(*branch_info)) { - const auto branch = std::get_if<SingleBranch>(branch_info.get()); + const auto* branch = std::get_if<SingleBranch>(branch_info.get()); if (branch->address < 0) { if (branch->kill) { mm.InsertReturn(get_expr(branch->condition), true); @@ -629,7 +620,7 @@ void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) { mm.InsertGoto(get_expr(branch->condition), branch->address); return; } - const auto multi_branch = std::get_if<MultiBranch>(branch_info.get()); + const auto* multi_branch = std::get_if<MultiBranch>(branch_info.get()); for (const auto& branch_case : multi_branch->branches) { mm.InsertGoto(MakeExpr<ExprGprEqual>(multi_branch->gpr, branch_case.cmp_value), branch_case.address); @@ -655,6 +646,8 @@ void DecompileShader(CFGRebuildState& state) { state.manager->Decompile(); } +} // Anonymous namespace + std::unique_ptr<ShaderCharacteristics> ScanFlow(const ProgramCode& program_code, u32 start_address, const CompilerSettings& settings, Registry& registry) { diff --git a/src/video_core/shader/decode.cpp b/src/video_core/shader/decode.cpp index 87ac9ac6c..eeac328a6 100644 --- a/src/video_core/shader/decode.cpp +++ b/src/video_core/shader/decode.cpp @@ -13,6 +13,7 @@ #include "video_core/engines/shader_bytecode.h" #include "video_core/engines/shader_header.h" #include "video_core/shader/control_flow.h" +#include "video_core/shader/memory_util.h" #include "video_core/shader/node_helper.h" #include "video_core/shader/shader_ir.h" @@ -23,17 +24,6 @@ using Tegra::Shader::OpCode; namespace { -/** - * Returns whether the instruction at the specified offset is a 'sched' instruction. - * Sched instructions always appear before a sequence of 3 instructions. - */ -constexpr bool IsSchedInstruction(u32 offset, u32 main_offset) { - constexpr u32 SchedPeriod = 4; - u32 absolute_offset = offset - main_offset; - - return (absolute_offset % SchedPeriod) == 0; -} - void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, const std::list<Sampler>& used_samplers) { if (gpu_driver.IsTextureHandlerSizeKnown() || used_samplers.size() <= 1) { @@ -42,11 +32,11 @@ void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, u32 count{}; std::vector<u32> bound_offsets; for (const auto& sampler : used_samplers) { - if (sampler.IsBindless()) { + if (sampler.is_bindless) { continue; } ++count; - bound_offsets.emplace_back(sampler.GetOffset()); + bound_offsets.emplace_back(sampler.offset); } if (count > 1) { gpu_driver.DeduceTextureHandlerSize(std::move(bound_offsets)); @@ -56,14 +46,14 @@ void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, std::optional<u32> TryDeduceSamplerSize(const Sampler& sampler_to_deduce, VideoCore::GuestDriverProfile& gpu_driver, const std::list<Sampler>& used_samplers) { - const u32 base_offset = sampler_to_deduce.GetOffset(); + const u32 base_offset = sampler_to_deduce.offset; u32 max_offset{std::numeric_limits<u32>::max()}; for (const auto& sampler : used_samplers) { - if (sampler.IsBindless()) { + if (sampler.is_bindless) { continue; } - if (sampler.GetOffset() > base_offset) { - max_offset = std::min(sampler.GetOffset(), max_offset); + if (sampler.offset > base_offset) { + max_offset = std::min(sampler.offset, max_offset); } } if (max_offset == std::numeric_limits<u32>::max()) { @@ -265,7 +255,7 @@ void ShaderIR::InsertControlFlow(NodeBlock& bb, const ShaderBlock& block) { Node n = Operation(OperationCode::Branch, Immediate(branch_case.address)); Node op_b = Immediate(branch_case.cmp_value); Node condition = - GetPredicateComparisonInteger(Tegra::Shader::PredCondition::Equal, false, op_a, op_b); + GetPredicateComparisonInteger(Tegra::Shader::PredCondition::EQ, false, op_a, op_b); auto result = Conditional(condition, {n}); bb.push_back(result); global_code.push_back(result); @@ -363,14 +353,14 @@ void ShaderIR::PostDecode() { return; } for (auto& sampler : used_samplers) { - if (!sampler.IsIndexed()) { + if (!sampler.is_indexed) { continue; } if (const auto size = TryDeduceSamplerSize(sampler, gpu_driver, used_samplers)) { - sampler.SetSize(*size); + sampler.size = *size; } else { LOG_CRITICAL(HW_GPU, "Failed to deduce size of indexed sampler"); - sampler.SetSize(1); + sampler.size = 1; } } } diff --git a/src/video_core/shader/decode/arithmetic.cpp b/src/video_core/shader/decode/arithmetic.cpp index 4db329fa5..afef5948d 100644 --- a/src/video_core/shader/decode/arithmetic.cpp +++ b/src/video_core/shader/decode/arithmetic.cpp @@ -137,7 +137,8 @@ u32 ShaderIR::DecodeArithmetic(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::FCMP_RR: - case OpCode::Id::FCMP_RC: { + case OpCode::Id::FCMP_RC: + case OpCode::Id::FCMP_IMMR: { UNIMPLEMENTED_IF(instr.fcmp.ftz == 0); Node op_c = GetRegister(instr.gpr39); Node comp = GetPredicateComparisonFloat(instr.fcmp.cond, std::move(op_c), Immediate(0.0f)); diff --git a/src/video_core/shader/decode/arithmetic_half.cpp b/src/video_core/shader/decode/arithmetic_half.cpp index ee7d9a29d..88103fede 100644 --- a/src/video_core/shader/decode/arithmetic_half.cpp +++ b/src/video_core/shader/decode/arithmetic_half.cpp @@ -19,22 +19,49 @@ u32 ShaderIR::DecodeArithmeticHalf(NodeBlock& bb, u32 pc) { const Instruction instr = {program_code[pc]}; const auto opcode = OpCode::Decode(instr); - if (opcode->get().GetId() == OpCode::Id::HADD2_C || - opcode->get().GetId() == OpCode::Id::HADD2_R) { + bool negate_a = false; + bool negate_b = false; + bool absolute_a = false; + bool absolute_b = false; + + switch (opcode->get().GetId()) { + case OpCode::Id::HADD2_R: if (instr.alu_half.ftz == 0) { LOG_DEBUG(HW_GPU, "{} without FTZ is not implemented", opcode->get().GetName()); } + negate_a = ((instr.value >> 43) & 1) != 0; + negate_b = ((instr.value >> 31) & 1) != 0; + absolute_a = ((instr.value >> 44) & 1) != 0; + absolute_b = ((instr.value >> 30) & 1) != 0; + break; + case OpCode::Id::HADD2_C: + if (instr.alu_half.ftz == 0) { + LOG_DEBUG(HW_GPU, "{} without FTZ is not implemented", opcode->get().GetName()); + } + negate_a = ((instr.value >> 43) & 1) != 0; + negate_b = ((instr.value >> 56) & 1) != 0; + absolute_a = ((instr.value >> 44) & 1) != 0; + absolute_b = ((instr.value >> 54) & 1) != 0; + break; + case OpCode::Id::HMUL2_R: + negate_a = ((instr.value >> 43) & 1) != 0; + absolute_a = ((instr.value >> 44) & 1) != 0; + absolute_b = ((instr.value >> 30) & 1) != 0; + break; + case OpCode::Id::HMUL2_C: + negate_b = ((instr.value >> 31) & 1) != 0; + absolute_a = ((instr.value >> 44) & 1) != 0; + absolute_b = ((instr.value >> 54) & 1) != 0; + break; + default: + UNREACHABLE(); + break; } - const bool negate_a = - opcode->get().GetId() != OpCode::Id::HMUL2_R && instr.alu_half.negate_a != 0; - const bool negate_b = - opcode->get().GetId() != OpCode::Id::HMUL2_C && instr.alu_half.negate_b != 0; - Node op_a = UnpackHalfFloat(GetRegister(instr.gpr8), instr.alu_half.type_a); - op_a = GetOperandAbsNegHalf(op_a, instr.alu_half.abs_a, negate_a); + op_a = GetOperandAbsNegHalf(op_a, absolute_a, negate_a); - auto [type_b, op_b] = [&]() -> std::tuple<HalfType, Node> { + auto [type_b, op_b] = [this, instr, opcode]() -> std::pair<HalfType, Node> { switch (opcode->get().GetId()) { case OpCode::Id::HADD2_C: case OpCode::Id::HMUL2_C: @@ -48,17 +75,16 @@ u32 ShaderIR::DecodeArithmeticHalf(NodeBlock& bb, u32 pc) { } }(); op_b = UnpackHalfFloat(op_b, type_b); - // redeclaration to avoid a bug in clang with reusing local bindings in lambdas - Node op_b_alt = GetOperandAbsNegHalf(op_b, instr.alu_half.abs_b, negate_b); + op_b = GetOperandAbsNegHalf(op_b, absolute_b, negate_b); - Node value = [&]() { + Node value = [this, opcode, op_a, op_b = op_b] { switch (opcode->get().GetId()) { case OpCode::Id::HADD2_C: case OpCode::Id::HADD2_R: - return Operation(OperationCode::HAdd, PRECISE, op_a, op_b_alt); + return Operation(OperationCode::HAdd, PRECISE, op_a, op_b); case OpCode::Id::HMUL2_C: case OpCode::Id::HMUL2_R: - return Operation(OperationCode::HMul, PRECISE, op_a, op_b_alt); + return Operation(OperationCode::HMul, PRECISE, op_a, op_b); default: UNIMPLEMENTED_MSG("Unhandled half float instruction: {}", opcode->get().GetName()); return Immediate(0); diff --git a/src/video_core/shader/decode/arithmetic_integer.cpp b/src/video_core/shader/decode/arithmetic_integer.cpp index 0f4c3103a..73155966f 100644 --- a/src/video_core/shader/decode/arithmetic_integer.cpp +++ b/src/video_core/shader/decode/arithmetic_integer.cpp @@ -35,15 +35,38 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { case OpCode::Id::IADD_C: case OpCode::Id::IADD_R: case OpCode::Id::IADD_IMM: { - UNIMPLEMENTED_IF_MSG(instr.alu.saturate_d, "IADD saturation not implemented"); + UNIMPLEMENTED_IF_MSG(instr.alu.saturate_d, "IADD.SAT"); + UNIMPLEMENTED_IF_MSG(instr.iadd.x && instr.generates_cc, "IADD.X Rd.CC"); op_a = GetOperandAbsNegInteger(op_a, false, instr.alu_integer.negate_a, true); op_b = GetOperandAbsNegInteger(op_b, false, instr.alu_integer.negate_b, true); - const Node value = Operation(OperationCode::IAdd, PRECISE, op_a, op_b); + Node value = Operation(OperationCode::UAdd, op_a, op_b); - SetInternalFlagsFromInteger(bb, value, instr.generates_cc); - SetRegister(bb, instr.gpr0, value); + if (instr.iadd.x) { + Node carry = GetInternalFlag(InternalFlag::Carry); + Node x = Operation(OperationCode::Select, std::move(carry), Immediate(1), Immediate(0)); + value = Operation(OperationCode::UAdd, std::move(value), std::move(x)); + } + + if (instr.generates_cc) { + const Node i0 = Immediate(0); + + Node zero = Operation(OperationCode::LogicalIEqual, value, i0); + Node sign = Operation(OperationCode::LogicalILessThan, value, i0); + Node carry = Operation(OperationCode::LogicalAddCarry, op_a, op_b); + + Node pos_a = Operation(OperationCode::LogicalIGreaterThan, op_a, i0); + Node pos_b = Operation(OperationCode::LogicalIGreaterThan, op_b, i0); + Node pos = Operation(OperationCode::LogicalAnd, std::move(pos_a), std::move(pos_b)); + Node overflow = Operation(OperationCode::LogicalAnd, pos, sign); + + SetInternalFlag(bb, InternalFlag::Zero, std::move(zero)); + SetInternalFlag(bb, InternalFlag::Sign, std::move(sign)); + SetInternalFlag(bb, InternalFlag::Carry, std::move(carry)); + SetInternalFlag(bb, InternalFlag::Overflow, std::move(overflow)); + } + SetRegister(bb, instr.gpr0, std::move(value)); break; } case OpCode::Id::IADD3_C: @@ -75,12 +98,12 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { op_b = GetOperandAbsNegInteger(op_b, false, instr.iadd3.neg_b, true); op_c = GetOperandAbsNegInteger(op_c, false, instr.iadd3.neg_c, true); - const Node value = [&]() { - const Node add_ab = Operation(OperationCode::IAdd, NO_PRECISE, op_a, op_b); + const Node value = [&] { + Node add_ab = Operation(OperationCode::IAdd, NO_PRECISE, op_a, op_b); if (opcode->get().GetId() != OpCode::Id::IADD3_R) { return Operation(OperationCode::IAdd, NO_PRECISE, add_ab, op_c); } - const Node shifted = [&]() { + const Node shifted = [&] { switch (instr.iadd3.mode) { case Tegra::Shader::IAdd3Mode::RightShift: // TODO(tech4me): According to @@ -249,8 +272,8 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { } case OpCode::Id::LEA_IMM: { const bool neg = instr.lea.imm.neg != 0; - return {Immediate(static_cast<u32>(instr.lea.imm.entry_a)), - GetOperandAbsNegInteger(GetRegister(instr.gpr8), false, neg, true), + return {GetOperandAbsNegInteger(GetRegister(instr.gpr8), false, neg, true), + Immediate(static_cast<u32>(instr.lea.imm.entry_a)), Immediate(static_cast<u32>(instr.lea.imm.entry_b))}; } case OpCode::Id::LEA_RZ: { diff --git a/src/video_core/shader/decode/arithmetic_integer_immediate.cpp b/src/video_core/shader/decode/arithmetic_integer_immediate.cpp index 73880db0e..2a30aab2b 100644 --- a/src/video_core/shader/decode/arithmetic_integer_immediate.cpp +++ b/src/video_core/shader/decode/arithmetic_integer_immediate.cpp @@ -28,23 +28,26 @@ u32 ShaderIR::DecodeArithmeticIntegerImmediate(NodeBlock& bb, u32 pc) { case OpCode::Id::IADD32I: { UNIMPLEMENTED_IF_MSG(instr.iadd32i.saturate, "IADD32I saturation is not implemented"); - op_a = GetOperandAbsNegInteger(op_a, false, instr.iadd32i.negate_a, true); + op_a = GetOperandAbsNegInteger(std::move(op_a), false, instr.iadd32i.negate_a != 0, true); - const Node value = Operation(OperationCode::IAdd, PRECISE, op_a, op_b); + Node value = Operation(OperationCode::IAdd, PRECISE, std::move(op_a), std::move(op_b)); - SetInternalFlagsFromInteger(bb, value, instr.op_32.generates_cc); - SetRegister(bb, instr.gpr0, value); + SetInternalFlagsFromInteger(bb, value, instr.op_32.generates_cc != 0); + SetRegister(bb, instr.gpr0, std::move(value)); break; } case OpCode::Id::LOP32I: { - if (instr.alu.lop32i.invert_a) - op_a = Operation(OperationCode::IBitwiseNot, NO_PRECISE, op_a); + if (instr.alu.lop32i.invert_a) { + op_a = Operation(OperationCode::IBitwiseNot, NO_PRECISE, std::move(op_a)); + } - if (instr.alu.lop32i.invert_b) - op_b = Operation(OperationCode::IBitwiseNot, NO_PRECISE, op_b); + if (instr.alu.lop32i.invert_b) { + op_b = Operation(OperationCode::IBitwiseNot, NO_PRECISE, std::move(op_b)); + } - WriteLogicOperation(bb, instr.gpr0, instr.alu.lop32i.operation, op_a, op_b, - PredicateResultMode::None, Pred::UnusedIndex, instr.op_32.generates_cc); + WriteLogicOperation(bb, instr.gpr0, instr.alu.lop32i.operation, std::move(op_a), + std::move(op_b), PredicateResultMode::None, Pred::UnusedIndex, + instr.op_32.generates_cc != 0); break; } default: @@ -58,14 +61,14 @@ u32 ShaderIR::DecodeArithmeticIntegerImmediate(NodeBlock& bb, u32 pc) { void ShaderIR::WriteLogicOperation(NodeBlock& bb, Register dest, LogicOperation logic_op, Node op_a, Node op_b, PredicateResultMode predicate_mode, Pred predicate, bool sets_cc) { - const Node result = [&]() { + Node result = [&] { switch (logic_op) { case LogicOperation::And: - return Operation(OperationCode::IBitwiseAnd, PRECISE, op_a, op_b); + return Operation(OperationCode::IBitwiseAnd, PRECISE, std::move(op_a), std::move(op_b)); case LogicOperation::Or: - return Operation(OperationCode::IBitwiseOr, PRECISE, op_a, op_b); + return Operation(OperationCode::IBitwiseOr, PRECISE, std::move(op_a), std::move(op_b)); case LogicOperation::Xor: - return Operation(OperationCode::IBitwiseXor, PRECISE, op_a, op_b); + return Operation(OperationCode::IBitwiseXor, PRECISE, std::move(op_a), std::move(op_b)); case LogicOperation::PassB: return op_b; default: @@ -84,8 +87,8 @@ void ShaderIR::WriteLogicOperation(NodeBlock& bb, Register dest, LogicOperation return; case PredicateResultMode::NotZero: { // Set the predicate to true if the result is not zero. - const Node compare = Operation(OperationCode::LogicalINotEqual, result, Immediate(0)); - SetPredicate(bb, static_cast<u64>(predicate), compare); + Node compare = Operation(OperationCode::LogicalINotEqual, std::move(result), Immediate(0)); + SetPredicate(bb, static_cast<u64>(predicate), std::move(compare)); break; } default: diff --git a/src/video_core/shader/decode/half_set.cpp b/src/video_core/shader/decode/half_set.cpp index 848e46874..b2e88fa20 100644 --- a/src/video_core/shader/decode/half_set.cpp +++ b/src/video_core/shader/decode/half_set.cpp @@ -13,55 +13,101 @@ namespace VideoCommon::Shader { +using std::move; using Tegra::Shader::Instruction; using Tegra::Shader::OpCode; +using Tegra::Shader::PredCondition; u32 ShaderIR::DecodeHalfSet(NodeBlock& bb, u32 pc) { const Instruction instr = {program_code[pc]}; const auto opcode = OpCode::Decode(instr); - if (instr.hset2.ftz == 0) { - LOG_DEBUG(HW_GPU, "{} without FTZ is not implemented", opcode->get().GetName()); + PredCondition cond; + bool bf; + bool ftz; + bool neg_a; + bool abs_a; + bool neg_b; + bool abs_b; + switch (opcode->get().GetId()) { + case OpCode::Id::HSET2_C: + case OpCode::Id::HSET2_IMM: + cond = instr.hsetp2.cbuf_and_imm.cond; + bf = instr.Bit(53); + ftz = instr.Bit(54); + neg_a = instr.Bit(43); + abs_a = instr.Bit(44); + neg_b = instr.Bit(56); + abs_b = instr.Bit(54); + break; + case OpCode::Id::HSET2_R: + cond = instr.hsetp2.reg.cond; + bf = instr.Bit(49); + ftz = instr.Bit(50); + neg_a = instr.Bit(43); + abs_a = instr.Bit(44); + neg_b = instr.Bit(31); + abs_b = instr.Bit(30); + break; + default: + UNREACHABLE(); } - Node op_a = UnpackHalfFloat(GetRegister(instr.gpr8), instr.hset2.type_a); - op_a = GetOperandAbsNegHalf(op_a, instr.hset2.abs_a, instr.hset2.negate_a); - - Node op_b = [&]() { + Node op_b = [this, instr, opcode] { switch (opcode->get().GetId()) { + case OpCode::Id::HSET2_C: + // Inform as unimplemented as this is not tested. + UNIMPLEMENTED_MSG("HSET2_C is not implemented"); + return GetConstBuffer(instr.cbuf34.index, instr.cbuf34.GetOffset()); case OpCode::Id::HSET2_R: return GetRegister(instr.gpr20); + case OpCode::Id::HSET2_IMM: + return UnpackHalfImmediate(instr, true); default: UNREACHABLE(); - return Immediate(0); + return Node{}; } }(); - op_b = UnpackHalfFloat(op_b, instr.hset2.type_b); - op_b = GetOperandAbsNegHalf(op_b, instr.hset2.abs_b, instr.hset2.negate_b); - const Node second_pred = GetPredicate(instr.hset2.pred39, instr.hset2.neg_pred); + if (!ftz) { + LOG_DEBUG(HW_GPU, "{} without FTZ is not implemented", opcode->get().GetName()); + } + + Node op_a = UnpackHalfFloat(GetRegister(instr.gpr8), instr.hset2.type_a); + op_a = GetOperandAbsNegHalf(op_a, abs_a, neg_a); + + switch (opcode->get().GetId()) { + case OpCode::Id::HSET2_R: + op_b = GetOperandAbsNegHalf(move(op_b), abs_b, neg_b); + [[fallthrough]]; + case OpCode::Id::HSET2_C: + op_b = UnpackHalfFloat(move(op_b), instr.hset2.type_b); + break; + default: + break; + } - const Node comparison_pair = GetPredicateComparisonHalf(instr.hset2.cond, op_a, op_b); + Node second_pred = GetPredicate(instr.hset2.pred39, instr.hset2.neg_pred); + + Node comparison_pair = GetPredicateComparisonHalf(cond, op_a, op_b); const OperationCode combiner = GetPredicateCombiner(instr.hset2.op); // HSET2 operates on each half float in the pack. std::array<Node, 2> values; for (u32 i = 0; i < 2; ++i) { - const u32 raw_value = instr.hset2.bf ? 0x3c00 : 0xffff; - const Node true_value = Immediate(raw_value << (i * 16)); - const Node false_value = Immediate(0); - - const Node comparison = - Operation(OperationCode::LogicalPick2, comparison_pair, Immediate(i)); - const Node predicate = Operation(combiner, comparison, second_pred); + const u32 raw_value = bf ? 0x3c00 : 0xffff; + Node true_value = Immediate(raw_value << (i * 16)); + Node false_value = Immediate(0); + Node comparison = Operation(OperationCode::LogicalPick2, comparison_pair, Immediate(i)); + Node predicate = Operation(combiner, comparison, second_pred); values[i] = - Operation(OperationCode::Select, NO_PRECISE, predicate, true_value, false_value); + Operation(OperationCode::Select, predicate, move(true_value), move(false_value)); } - const Node value = Operation(OperationCode::UBitwiseOr, NO_PRECISE, values[0], values[1]); - SetRegister(bb, instr.gpr0, value); + Node value = Operation(OperationCode::UBitwiseOr, values[0], values[1]); + SetRegister(bb, instr.gpr0, move(value)); return pc; } diff --git a/src/video_core/shader/decode/image.cpp b/src/video_core/shader/decode/image.cpp index 08ebca38b..1ed4212ee 100644 --- a/src/video_core/shader/decode/image.cpp +++ b/src/video_core/shader/decode/image.cpp @@ -31,11 +31,11 @@ ComponentType GetComponentType(Tegra::Engines::SamplerDescriptor descriptor, std::size_t component) { const TextureFormat format{descriptor.format}; switch (format) { - case TextureFormat::R16_G16_B16_A16: - case TextureFormat::R32_G32_B32_A32: - case TextureFormat::R32_G32_B32: - case TextureFormat::R32_G32: - case TextureFormat::R16_G16: + case TextureFormat::R16G16B16A16: + case TextureFormat::R32G32B32A32: + case TextureFormat::R32G32B32: + case TextureFormat::R32G32: + case TextureFormat::R16G16: case TextureFormat::R32: case TextureFormat::R16: case TextureFormat::R8: @@ -97,6 +97,7 @@ ComponentType GetComponentType(Tegra::Engines::SamplerDescriptor descriptor, break; case TextureFormat::B5G6R5: case TextureFormat::B6G5R5: + case TextureFormat::B10G11R11: if (component == 0) { return descriptor.b_type; } @@ -107,9 +108,9 @@ ComponentType GetComponentType(Tegra::Engines::SamplerDescriptor descriptor, return descriptor.r_type; } break; - case TextureFormat::G8R24: - case TextureFormat::G24R8: - case TextureFormat::G8R8: + case TextureFormat::R24G8: + case TextureFormat::R8G24: + case TextureFormat::R8G8: case TextureFormat::G4R4: if (component == 0) { return descriptor.g_type; @@ -118,6 +119,8 @@ ComponentType GetComponentType(Tegra::Engines::SamplerDescriptor descriptor, return descriptor.r_type; } break; + default: + break; } UNIMPLEMENTED_MSG("Texture format not implemented={}", format); return ComponentType::FLOAT; @@ -136,15 +139,15 @@ bool IsComponentEnabled(std::size_t component_mask, std::size_t component) { u32 GetComponentSize(TextureFormat format, std::size_t component) { switch (format) { - case TextureFormat::R32_G32_B32_A32: + case TextureFormat::R32G32B32A32: return 32; - case TextureFormat::R16_G16_B16_A16: + case TextureFormat::R16G16B16A16: return 16; - case TextureFormat::R32_G32_B32: + case TextureFormat::R32G32B32: return component <= 2 ? 32 : 0; - case TextureFormat::R32_G32: + case TextureFormat::R32G32: return component <= 1 ? 32 : 0; - case TextureFormat::R16_G16: + case TextureFormat::R16G16: return component <= 1 ? 16 : 0; case TextureFormat::R32: return component == 0 ? 32 : 0; @@ -191,7 +194,15 @@ u32 GetComponentSize(TextureFormat format, std::size_t component) { return 6; } return 0; - case TextureFormat::G8R24: + case TextureFormat::B10G11R11: + if (component == 1 || component == 2) { + return 11; + } + if (component == 0) { + return 10; + } + return 0; + case TextureFormat::R24G8: if (component == 0) { return 8; } @@ -199,7 +210,7 @@ u32 GetComponentSize(TextureFormat format, std::size_t component) { return 24; } return 0; - case TextureFormat::G24R8: + case TextureFormat::R8G24: if (component == 0) { return 24; } @@ -207,7 +218,7 @@ u32 GetComponentSize(TextureFormat format, std::size_t component) { return 8; } return 0; - case TextureFormat::G8R8: + case TextureFormat::R8G8: return (component == 0 || component == 1) ? 8 : 0; case TextureFormat::G4R4: return (component == 0 || component == 1) ? 4 : 0; @@ -223,24 +234,25 @@ std::size_t GetImageComponentMask(TextureFormat format) { constexpr u8 B = 0b0100; constexpr u8 A = 0b1000; switch (format) { - case TextureFormat::R32_G32_B32_A32: - case TextureFormat::R16_G16_B16_A16: + case TextureFormat::R32G32B32A32: + case TextureFormat::R16G16B16A16: case TextureFormat::A8R8G8B8: case TextureFormat::A2B10G10R10: case TextureFormat::A4B4G4R4: case TextureFormat::A5B5G5R1: case TextureFormat::A1B5G5R5: return std::size_t{R | G | B | A}; - case TextureFormat::R32_G32_B32: + case TextureFormat::R32G32B32: case TextureFormat::R32_B24G8: case TextureFormat::B5G6R5: case TextureFormat::B6G5R5: + case TextureFormat::B10G11R11: return std::size_t{R | G | B}; - case TextureFormat::R32_G32: - case TextureFormat::R16_G16: - case TextureFormat::G8R24: - case TextureFormat::G24R8: - case TextureFormat::G8R8: + case TextureFormat::R32G32: + case TextureFormat::R16G16: + case TextureFormat::R24G8: + case TextureFormat::R8G24: + case TextureFormat::R8G8: case TextureFormat::G4R4: return std::size_t{R | G}; case TextureFormat::R32: @@ -299,7 +311,7 @@ std::pair<Node, bool> ShaderIR::GetComponentValue(ComponentType component_type, return {std::move(original_value), true}; } default: - UNIMPLEMENTED_MSG("Unimplement component type={}", component_type); + UNIMPLEMENTED_MSG("Unimplemented component type={}", component_type); return {std::move(original_value), true}; } } @@ -352,8 +364,10 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { registry.ObtainBoundSampler(static_cast<u32>(instr.image.index.Value())); } else { const Node image_register = GetRegister(instr.gpr39); - const auto [base_image, buffer, offset] = TrackCbuf( - image_register, global_code, static_cast<s64>(global_code.size())); + const auto result = TrackCbuf(image_register, global_code, + static_cast<s64>(global_code.size())); + const auto buffer = std::get<1>(result); + const auto offset = std::get<2>(result); descriptor = registry.ObtainBindlessSampler(buffer, offset); } if (!descriptor) { @@ -453,11 +467,14 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { return OperationCode::AtomicImageXor; case Tegra::Shader::ImageAtomicOperation::Exch: return OperationCode::AtomicImageExchange; + default: + break; } + break; default: break; } - UNIMPLEMENTED_MSG("Unimplemented operation={} type={}", + UNIMPLEMENTED_MSG("Unimplemented operation={}, type={}", static_cast<u64>(instr.suatom_d.operation.Value()), static_cast<u64>(instr.suatom_d.operation_type.Value())); return OperationCode::AtomicImageAdd; @@ -483,11 +500,10 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { Image& ShaderIR::GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type) { const auto offset = static_cast<u32>(image.index.Value()); - const auto it = - std::find_if(std::begin(used_images), std::end(used_images), - [offset](const Image& entry) { return entry.GetOffset() == offset; }); + const auto it = std::find_if(std::begin(used_images), std::end(used_images), + [offset](const Image& entry) { return entry.offset == offset; }); if (it != std::end(used_images)) { - ASSERT(!it->IsBindless() && it->GetType() == it->GetType()); + ASSERT(!it->is_bindless && it->type == type); return *it; } @@ -497,16 +513,18 @@ Image& ShaderIR::GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType t Image& ShaderIR::GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::ImageType type) { const Node image_register = GetRegister(reg); - const auto [base_image, buffer, offset] = + const auto result = TrackCbuf(image_register, global_code, static_cast<s64>(global_code.size())); - const auto it = - std::find_if(std::begin(used_images), std::end(used_images), - [buffer = buffer, offset = offset](const Image& entry) { - return entry.GetBuffer() == buffer && entry.GetOffset() == offset; - }); + const auto buffer = std::get<1>(result); + const auto offset = std::get<2>(result); + + const auto it = std::find_if(std::begin(used_images), std::end(used_images), + [buffer, offset](const Image& entry) { + return entry.buffer == buffer && entry.offset == offset; + }); if (it != std::end(used_images)) { - ASSERT(it->IsBindless() && it->GetType() == it->GetType()); + ASSERT(it->is_bindless && it->type == type); return *it; } diff --git a/src/video_core/shader/decode/memory.cpp b/src/video_core/shader/decode/memory.cpp index 8112ead3e..e2bba88dd 100644 --- a/src/video_core/shader/decode/memory.cpp +++ b/src/video_core/shader/decode/memory.cpp @@ -386,8 +386,8 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::RED: { - UNIMPLEMENTED_IF_MSG(instr.red.type != GlobalAtomicType::U32); - UNIMPLEMENTED_IF_MSG(instr.red.operation != AtomicOp::Add); + UNIMPLEMENTED_IF_MSG(instr.red.type != GlobalAtomicType::U32, "type={}", + static_cast<int>(instr.red.type.Value())); const auto [real_address, base_address, descriptor] = TrackGlobalMemory(bb, instr, true, true); if (!real_address || !base_address) { @@ -396,7 +396,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { } Node gmem = MakeNode<GmemNode>(real_address, base_address, descriptor); Node value = GetRegister(instr.gpr0); - bb.push_back(Operation(OperationCode::ReduceIAdd, move(gmem), move(value))); + bb.push_back(Operation(GetAtomOperation(instr.red.operation), move(gmem), move(value))); break; } case OpCode::Id::ATOM: { @@ -472,14 +472,14 @@ std::tuple<Node, Node, GlobalMemoryBase> ShaderIR::TrackGlobalMemory(NodeBlock& const auto [base_address, index, offset] = TrackCbuf(addr_register, global_code, static_cast<s64>(global_code.size())); - ASSERT_OR_EXECUTE_MSG(base_address != nullptr, - { return std::make_tuple(nullptr, nullptr, GlobalMemoryBase{}); }, - "Global memory tracking failed"); + ASSERT_OR_EXECUTE_MSG( + base_address != nullptr, { return std::make_tuple(nullptr, nullptr, GlobalMemoryBase{}); }, + "Global memory tracking failed"); bb.push_back(Comment(fmt::format("Base address is c[0x{:x}][0x{:x}]", index, offset))); const GlobalMemoryBase descriptor{index, offset}; - const auto& [entry, is_new] = used_global_memory.try_emplace(descriptor); + const auto& entry = used_global_memory.try_emplace(descriptor).first; auto& usage = entry->second; usage.is_written |= is_write; usage.is_read |= is_read; diff --git a/src/video_core/shader/decode/other.cpp b/src/video_core/shader/decode/other.cpp index d4f95b18c..29a7cfbfe 100644 --- a/src/video_core/shader/decode/other.cpp +++ b/src/video_core/shader/decode/other.cpp @@ -75,15 +75,14 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { const Node value = [this, instr] { switch (instr.sys20) { case SystemVariable::LaneId: - LOG_WARNING(HW_GPU, "S2R instruction with LaneId is incomplete"); - return Immediate(0U); + return Operation(OperationCode::ThreadId); case SystemVariable::InvocationId: return Operation(OperationCode::InvocationId); case SystemVariable::Ydirection: return Operation(OperationCode::YNegate); case SystemVariable::InvocationInfo: LOG_WARNING(HW_GPU, "S2R instruction with InvocationInfo is incomplete"); - return Immediate(0U); + return Immediate(0x00ff'0000U); case SystemVariable::WscaleFactorXY: UNIMPLEMENTED_MSG("S2R WscaleFactorXY is not implemented"); return Immediate(0U); @@ -109,6 +108,27 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { return Operation(OperationCode::WorkGroupIdY); case SystemVariable::CtaIdZ: return Operation(OperationCode::WorkGroupIdZ); + case SystemVariable::EqMask: + case SystemVariable::LtMask: + case SystemVariable::LeMask: + case SystemVariable::GtMask: + case SystemVariable::GeMask: + uses_warps = true; + switch (instr.sys20) { + case SystemVariable::EqMask: + return Operation(OperationCode::ThreadEqMask); + case SystemVariable::LtMask: + return Operation(OperationCode::ThreadLtMask); + case SystemVariable::LeMask: + return Operation(OperationCode::ThreadLeMask); + case SystemVariable::GtMask: + return Operation(OperationCode::ThreadGtMask); + case SystemVariable::GeMask: + return Operation(OperationCode::ThreadGeMask); + default: + UNREACHABLE(); + return Immediate(0u); + } default: UNIMPLEMENTED_MSG("Unhandled system move: {}", static_cast<u32>(instr.sys20.Value())); @@ -272,10 +292,25 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { SetRegister(bb, instr.gpr0, GetRegister(instr.gpr8)); break; } + case OpCode::Id::BAR: { + UNIMPLEMENTED_IF_MSG(instr.value != 0xF0A81B8000070000ULL, "BAR is not BAR.SYNC 0x0"); + bb.push_back(Operation(OperationCode::Barrier)); + break; + } case OpCode::Id::MEMBAR: { - UNIMPLEMENTED_IF(instr.membar.type != Tegra::Shader::MembarType::GL); UNIMPLEMENTED_IF(instr.membar.unknown != Tegra::Shader::MembarUnknown::Default); - bb.push_back(Operation(OperationCode::MemoryBarrierGL)); + const OperationCode type = [instr] { + switch (instr.membar.type) { + case Tegra::Shader::MembarType::CTA: + return OperationCode::MemoryBarrierGroup; + case Tegra::Shader::MembarType::GL: + return OperationCode::MemoryBarrierGlobal; + default: + UNIMPLEMENTED_MSG("MEMBAR type={}", static_cast<int>(instr.membar.type.Value())); + return OperationCode::MemoryBarrierGlobal; + } + }(); + bb.push_back(Operation(type)); break; } case OpCode::Id::DEPBAR: { diff --git a/src/video_core/shader/decode/register_set_predicate.cpp b/src/video_core/shader/decode/register_set_predicate.cpp index 8d54cce34..6116c31aa 100644 --- a/src/video_core/shader/decode/register_set_predicate.cpp +++ b/src/video_core/shader/decode/register_set_predicate.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <utility> + #include "common/assert.h" #include "common/common_types.h" #include "video_core/engines/shader_bytecode.h" @@ -10,20 +12,20 @@ namespace VideoCommon::Shader { +using std::move; using Tegra::Shader::Instruction; using Tegra::Shader::OpCode; namespace { -constexpr u64 NUM_PROGRAMMABLE_PREDICATES = 7; -} +constexpr u64 NUM_CONDITION_CODES = 4; +constexpr u64 NUM_PREDICATES = 7; +} // namespace u32 ShaderIR::DecodeRegisterSetPredicate(NodeBlock& bb, u32 pc) { const Instruction instr = {program_code[pc]}; const auto opcode = OpCode::Decode(instr); - UNIMPLEMENTED_IF(instr.p2r_r2p.mode != Tegra::Shader::R2pMode::Pr); - - const Node apply_mask = [&] { + Node apply_mask = [this, opcode, instr] { switch (opcode->get().GetId()) { case OpCode::Id::R2P_IMM: case OpCode::Id::P2R_IMM: @@ -34,39 +36,43 @@ u32 ShaderIR::DecodeRegisterSetPredicate(NodeBlock& bb, u32 pc) { } }(); - const auto offset = static_cast<u32>(instr.p2r_r2p.byte) * 8; + const u32 offset = static_cast<u32>(instr.p2r_r2p.byte) * 8; + + const bool cc = instr.p2r_r2p.mode == Tegra::Shader::R2pMode::Cc; + const u64 num_entries = cc ? NUM_CONDITION_CODES : NUM_PREDICATES; + const auto get_entry = [this, cc](u64 entry) { + return cc ? GetInternalFlag(static_cast<InternalFlag>(entry)) : GetPredicate(entry); + }; switch (opcode->get().GetId()) { case OpCode::Id::R2P_IMM: { - const Node mask = GetRegister(instr.gpr8); + Node mask = GetRegister(instr.gpr8); - for (u64 pred = 0; pred < NUM_PROGRAMMABLE_PREDICATES; ++pred) { - const auto shift = static_cast<u32>(pred); + for (u64 entry = 0; entry < num_entries; ++entry) { + const u32 shift = static_cast<u32>(entry); - const Node apply_compare = BitfieldExtract(apply_mask, shift, 1); - const Node condition = - Operation(OperationCode::LogicalUNotEqual, apply_compare, Immediate(0)); + Node apply = BitfieldExtract(apply_mask, shift, 1); + Node condition = Operation(OperationCode::LogicalUNotEqual, apply, Immediate(0)); - const Node value_compare = BitfieldExtract(mask, offset + shift, 1); - const Node value = - Operation(OperationCode::LogicalUNotEqual, value_compare, Immediate(0)); + Node compare = BitfieldExtract(mask, offset + shift, 1); + Node value = Operation(OperationCode::LogicalUNotEqual, move(compare), Immediate(0)); - const Node code = Operation(OperationCode::LogicalAssign, GetPredicate(pred), value); - bb.push_back(Conditional(condition, {code})); + Node code = Operation(OperationCode::LogicalAssign, get_entry(entry), move(value)); + bb.push_back(Conditional(condition, {move(code)})); } break; } case OpCode::Id::P2R_IMM: { Node value = Immediate(0); - for (u64 pred = 0; pred < NUM_PROGRAMMABLE_PREDICATES; ++pred) { - Node bit = Operation(OperationCode::Select, GetPredicate(pred), Immediate(1U << pred), + for (u64 entry = 0; entry < num_entries; ++entry) { + Node bit = Operation(OperationCode::Select, get_entry(entry), Immediate(1U << entry), Immediate(0)); - value = Operation(OperationCode::UBitwiseOr, std::move(value), std::move(bit)); + value = Operation(OperationCode::UBitwiseOr, move(value), move(bit)); } - value = Operation(OperationCode::UBitwiseAnd, std::move(value), apply_mask); - value = BitfieldInsert(GetRegister(instr.gpr8), std::move(value), offset, 8); + value = Operation(OperationCode::UBitwiseAnd, move(value), apply_mask); + value = BitfieldInsert(GetRegister(instr.gpr8), move(value), offset, 8); - SetRegister(bb, instr.gpr0, std::move(value)); + SetRegister(bb, instr.gpr0, move(value)); break; } default: diff --git a/src/video_core/shader/decode/shift.cpp b/src/video_core/shader/decode/shift.cpp index 3b391d3e6..d4ffa8014 100644 --- a/src/video_core/shader/decode/shift.cpp +++ b/src/video_core/shader/decode/shift.cpp @@ -23,7 +23,6 @@ Node IsFull(Node shift) { } Node Shift(OperationCode opcode, Node value, Node shift) { - Node is_full = Operation(OperationCode::LogicalIEqual, shift, Immediate(32)); Node shifted = Operation(opcode, move(value), shift); return Operation(OperationCode::Select, IsFull(move(shift)), Immediate(0), move(shifted)); } diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp index 6c4a1358b..02fdccd86 100644 --- a/src/video_core/shader/decode/texture.cpp +++ b/src/video_core/shader/decode/texture.cpp @@ -139,15 +139,15 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { } const Node component = Immediate(static_cast<u32>(instr.tld4s.component)); - const SamplerInfo info{TextureType::Texture2D, false, is_depth_compare}; - const Sampler& sampler = *GetSampler(instr.sampler, info); + SamplerInfo info; + info.is_shadow = is_depth_compare; + const std::optional<Sampler> sampler = GetSampler(instr.sampler, info); Node4 values; for (u32 element = 0; element < values.size(); ++element) { - auto coords_copy = coords; - MetaTexture meta{sampler, {}, depth_compare, aoffi, {}, {}, - {}, {}, component, element, {}}; - values[element] = Operation(OperationCode::TextureGather, meta, std::move(coords_copy)); + MetaTexture meta{*sampler, {}, depth_compare, aoffi, {}, {}, + {}, {}, component, element, {}}; + values[element] = Operation(OperationCode::TextureGather, meta, coords); } if (instr.tld4s.fp16_flag) { @@ -165,19 +165,20 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { "AOFFI is not implemented"); const bool is_array = instr.txd.is_array != 0; - u64 base_reg = instr.gpr8.Value(); const auto derivate_reg = instr.gpr20.Value(); const auto texture_type = instr.txd.texture_type.Value(); const auto coord_count = GetCoordCount(texture_type); - Node index_var{}; - const Sampler* sampler = - is_bindless ? GetBindlessSampler(base_reg, index_var, {{texture_type, is_array, false}}) - : GetSampler(instr.sampler, {{texture_type, is_array, false}}); + u64 base_reg = instr.gpr8.Value(); + Node index_var; + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + const std::optional<Sampler> sampler = is_bindless + ? GetBindlessSampler(base_reg, info, index_var) + : GetSampler(instr.sampler, info); Node4 values; - if (sampler == nullptr) { - for (u32 element = 0; element < values.size(); ++element) { - values[element] = Immediate(0); - } + if (!sampler) { + std::generate(values.begin(), values.end(), [this] { return Immediate(0); }); WriteTexInstructionFloat(bb, instr, values); break; } @@ -215,14 +216,12 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { is_bindless = true; [[fallthrough]]; case OpCode::Id::TXQ: { - // TODO: The new commits on the texture refactor, change the way samplers work. - // Sadly, not all texture instructions specify the type of texture their sampler - // uses. This must be fixed at a later instance. - Node index_var{}; - const Sampler* sampler = - is_bindless ? GetBindlessSampler(instr.gpr8, index_var) : GetSampler(instr.sampler); - - if (sampler == nullptr) { + Node index_var; + const std::optional<Sampler> sampler = is_bindless + ? GetBindlessSampler(instr.gpr8, {}, index_var) + : GetSampler(instr.sampler, {}); + + if (!sampler) { u32 indexer = 0; for (u32 element = 0; element < 4; ++element) { if (!instr.txq.IsComponentEnabled(element)) { @@ -268,13 +267,17 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { UNIMPLEMENTED_IF_MSG(instr.tmml.UsesMiscMode(Tegra::Shader::TextureMiscMode::NDV), "NDV is not implemented"); - auto texture_type = instr.tmml.texture_type.Value(); + const auto texture_type = instr.tmml.texture_type.Value(); const bool is_array = instr.tmml.array != 0; - Node index_var{}; - const Sampler* sampler = - is_bindless ? GetBindlessSampler(instr.gpr20, index_var) : GetSampler(instr.sampler); - - if (sampler == nullptr) { + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + Node index_var; + const std::optional<Sampler> sampler = + is_bindless ? GetBindlessSampler(instr.gpr20, info, index_var) + : GetSampler(instr.sampler, info); + + if (!sampler) { u32 indexer = 0; for (u32 element = 0; element < 2; ++element) { if (!instr.tmml.IsComponentEnabled(element)) { @@ -289,34 +292,36 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { break; } - std::vector<Node> coords; - - // TODO: Add coordinates for different samplers once other texture types are implemented. - switch (texture_type) { - case TextureType::Texture1D: - coords.push_back(GetRegister(instr.gpr8)); - break; - case TextureType::Texture2D: - coords.push_back(GetRegister(instr.gpr8.Value() + 0)); - coords.push_back(GetRegister(instr.gpr8.Value() + 1)); - break; - default: - UNIMPLEMENTED_MSG("Unhandled texture type {}", static_cast<u32>(texture_type)); + const u64 base_index = is_array ? 1 : 0; + const u64 num_components = [texture_type] { + switch (texture_type) { + case TextureType::Texture1D: + return 1; + case TextureType::Texture2D: + return 2; + case TextureType::TextureCube: + return 3; + default: + UNIMPLEMENTED_MSG("Unhandled texture type {}", static_cast<int>(texture_type)); + return 2; + } + }(); + // TODO: What's the array component used for? - // Fallback to interpreting as a 2D texture for now - coords.push_back(GetRegister(instr.gpr8.Value() + 0)); - coords.push_back(GetRegister(instr.gpr8.Value() + 1)); - texture_type = TextureType::Texture2D; + std::vector<Node> coords; + coords.reserve(num_components); + for (u64 component = 0; component < num_components; ++component) { + coords.push_back(GetRegister(instr.gpr8.Value() + base_index + component)); } + u32 indexer = 0; for (u32 element = 0; element < 2; ++element) { if (!instr.tmml.IsComponentEnabled(element)) { continue; } - auto params = coords; MetaTexture meta{*sampler, {}, {}, {}, {}, {}, {}, {}, {}, element, index_var}; - const Node value = Operation(OperationCode::TextureQueryLod, meta, std::move(params)); - SetTemporary(bb, indexer++, value); + Node value = Operation(OperationCode::TextureQueryLod, meta, coords); + SetTemporary(bb, indexer++, std::move(value)); } for (u32 i = 0; i < indexer; ++i) { SetRegister(bb, instr.gpr0.Value() + i, GetTemporary(i)); @@ -355,98 +360,122 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { return pc; } -ShaderIR::SamplerInfo ShaderIR::GetSamplerInfo(std::optional<SamplerInfo> sampler_info, u32 offset, - std::optional<u32> buffer) { - if (sampler_info) { - return *sampler_info; +ShaderIR::SamplerInfo ShaderIR::GetSamplerInfo( + SamplerInfo info, std::optional<Tegra::Engines::SamplerDescriptor> sampler) { + if (info.IsComplete()) { + return info; } - const auto sampler = buffer ? registry.ObtainBindlessSampler(*buffer, offset) - : registry.ObtainBoundSampler(offset); if (!sampler) { LOG_WARNING(HW_GPU, "Unknown sampler info"); - return SamplerInfo{TextureType::Texture2D, false, false, false}; - } - return SamplerInfo{sampler->texture_type, sampler->is_array != 0, sampler->is_shadow != 0, - sampler->is_buffer != 0}; + info.type = info.type.value_or(Tegra::Shader::TextureType::Texture2D); + info.is_array = info.is_array.value_or(false); + info.is_shadow = info.is_shadow.value_or(false); + info.is_buffer = info.is_buffer.value_or(false); + return info; + } + info.type = info.type.value_or(sampler->texture_type); + info.is_array = info.is_array.value_or(sampler->is_array != 0); + info.is_shadow = info.is_shadow.value_or(sampler->is_shadow != 0); + info.is_buffer = info.is_buffer.value_or(sampler->is_buffer != 0); + return info; } -const Sampler* ShaderIR::GetSampler(const Tegra::Shader::Sampler& sampler, - std::optional<SamplerInfo> sampler_info) { - const auto offset = static_cast<u32>(sampler.index.Value()); - const auto info = GetSamplerInfo(sampler_info, offset); +std::optional<Sampler> ShaderIR::GetSampler(Tegra::Shader::Sampler sampler, + SamplerInfo sampler_info) { + const u32 offset = static_cast<u32>(sampler.index.Value()); + const auto info = GetSamplerInfo(sampler_info, registry.ObtainBoundSampler(offset)); // If this sampler has already been used, return the existing mapping. - const auto it = - std::find_if(used_samplers.begin(), used_samplers.end(), - [offset](const Sampler& entry) { return entry.GetOffset() == offset; }); + const auto it = std::find_if(used_samplers.begin(), used_samplers.end(), + [offset](const Sampler& entry) { return entry.offset == offset; }); if (it != used_samplers.end()) { - ASSERT(!it->IsBindless() && it->GetType() == info.type && it->IsArray() == info.is_array && - it->IsShadow() == info.is_shadow && it->IsBuffer() == info.is_buffer); - return &*it; + ASSERT(!it->is_bindless && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer); + return *it; } // Otherwise create a new mapping for this sampler const auto next_index = static_cast<u32>(used_samplers.size()); - return &used_samplers.emplace_back(next_index, offset, info.type, info.is_array, info.is_shadow, - info.is_buffer, false); + return used_samplers.emplace_back(next_index, offset, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer, false); } -const Sampler* ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, Node& index_var, - std::optional<SamplerInfo> sampler_info) { +std::optional<Sampler> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, + Node& index_var) { const Node sampler_register = GetRegister(reg); const auto [base_node, tracked_sampler_info] = TrackBindlessSampler(sampler_register, global_code, static_cast<s64>(global_code.size())); - ASSERT(base_node != nullptr); - if (base_node == nullptr) { - return nullptr; + if (!base_node) { + UNREACHABLE(); + return std::nullopt; } - if (const auto bindless_sampler_info = - std::get_if<BindlessSamplerNode>(&*tracked_sampler_info)) { - const u32 buffer = bindless_sampler_info->GetIndex(); - const u32 offset = bindless_sampler_info->GetOffset(); - const auto info = GetSamplerInfo(sampler_info, offset, buffer); + if (const auto sampler_info = std::get_if<BindlessSamplerNode>(&*tracked_sampler_info)) { + const u32 buffer = sampler_info->index; + const u32 offset = sampler_info->offset; + info = GetSamplerInfo(info, registry.ObtainBindlessSampler(buffer, offset)); // If this sampler has already been used, return the existing mapping. - const auto it = - std::find_if(used_samplers.begin(), used_samplers.end(), - [buffer = buffer, offset = offset](const Sampler& entry) { - return entry.GetBuffer() == buffer && entry.GetOffset() == offset; - }); + const auto it = std::find_if(used_samplers.begin(), used_samplers.end(), + [buffer, offset](const Sampler& entry) { + return entry.buffer == buffer && entry.offset == offset; + }); if (it != used_samplers.end()) { - ASSERT(it->IsBindless() && it->GetType() == info.type && - it->IsArray() == info.is_array && it->IsShadow() == info.is_shadow); - return &*it; + ASSERT(it->is_bindless && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow); + return *it; } // Otherwise create a new mapping for this sampler const auto next_index = static_cast<u32>(used_samplers.size()); - return &used_samplers.emplace_back(next_index, offset, buffer, info.type, info.is_array, - info.is_shadow, info.is_buffer, false); - } else if (const auto array_sampler_info = - std::get_if<ArraySamplerNode>(&*tracked_sampler_info)) { - const u32 base_offset = array_sampler_info->GetBaseOffset() / 4; - index_var = GetCustomVariable(array_sampler_info->GetIndexVar()); - const auto info = GetSamplerInfo(sampler_info, base_offset); + return used_samplers.emplace_back(next_index, offset, buffer, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer, false); + } + if (const auto sampler_info = std::get_if<SeparateSamplerNode>(&*tracked_sampler_info)) { + const std::pair indices = sampler_info->indices; + const std::pair offsets = sampler_info->offsets; + info = GetSamplerInfo(info, registry.ObtainSeparateSampler(indices, offsets)); + + // Try to use an already created sampler if it exists + const auto it = std::find_if( + used_samplers.begin(), used_samplers.end(), [indices, offsets](const Sampler& entry) { + return offsets == std::pair{entry.offset, entry.secondary_offset} && + indices == std::pair{entry.buffer, entry.secondary_buffer}; + }); + if (it != used_samplers.end()) { + ASSERT(it->is_separated && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer); + return *it; + } + + // Otherwise create a new mapping for this sampler + const u32 next_index = static_cast<u32>(used_samplers.size()); + return used_samplers.emplace_back(next_index, offsets, indices, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer); + } + if (const auto sampler_info = std::get_if<ArraySamplerNode>(&*tracked_sampler_info)) { + const u32 base_offset = sampler_info->base_offset / 4; + index_var = GetCustomVariable(sampler_info->bindless_var); + info = GetSamplerInfo(info, registry.ObtainBoundSampler(base_offset)); // If this sampler has already been used, return the existing mapping. const auto it = std::find_if( used_samplers.begin(), used_samplers.end(), - [base_offset](const Sampler& entry) { return entry.GetOffset() == base_offset; }); + [base_offset](const Sampler& entry) { return entry.offset == base_offset; }); if (it != used_samplers.end()) { - ASSERT(!it->IsBindless() && it->GetType() == info.type && - it->IsArray() == info.is_array && it->IsShadow() == info.is_shadow && - it->IsBuffer() == info.is_buffer && it->IsIndexed()); - return &*it; + ASSERT(!it->is_bindless && it->type == info.type && it->is_array == info.is_array && + it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer && + it->is_indexed); + return *it; } uses_indexed_samplers = true; // Otherwise create a new mapping for this sampler const auto next_index = static_cast<u32>(used_samplers.size()); - return &used_samplers.emplace_back(next_index, base_offset, info.type, info.is_array, - info.is_shadow, info.is_buffer, true); + return used_samplers.emplace_back(next_index, base_offset, *info.type, *info.is_array, + *info.is_shadow, *info.is_buffer, true); } - return nullptr; + return std::nullopt; } void ShaderIR::WriteTexInstructionFloat(NodeBlock& bb, Instruction instr, const Node4& components) { @@ -527,14 +556,19 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, const bool is_shadow = depth_compare != nullptr; const bool is_bindless = bindless_reg.has_value(); - UNIMPLEMENTED_IF(texture_type == TextureType::TextureCube && is_array && is_shadow); ASSERT_MSG(texture_type != TextureType::Texture3D || !is_array || !is_shadow, "Illegal texture type"); - const SamplerInfo info{texture_type, is_array, is_shadow, false}; + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + info.is_shadow = is_shadow; + info.is_buffer = false; + Node index_var; - const Sampler* sampler = is_bindless ? GetBindlessSampler(*bindless_reg, index_var, info) - : GetSampler(instr.sampler, info); + const std::optional<Sampler> sampler = is_bindless + ? GetBindlessSampler(*bindless_reg, info, index_var) + : GetSampler(instr.sampler, info); if (!sampler) { return {Immediate(0), Immediate(0), Immediate(0), Immediate(0)}; } @@ -593,8 +627,9 @@ Node4 ShaderIR::GetTexCode(Instruction instr, TextureType texture_type, ++parameter_register; } - const auto [coord_count, total_coord_count] = ValidateAndGetCoordinateElement( - texture_type, depth_compare, is_array, lod_bias_enabled, 4, 5); + const auto coord_counts = ValidateAndGetCoordinateElement(texture_type, depth_compare, is_array, + lod_bias_enabled, 4, 5); + const auto coord_count = std::get<0>(coord_counts); // If enabled arrays index is always stored in the gpr8 field const u64 array_register = instr.gpr8.Value(); // First coordinate index is the gpr8 or gpr8 + 1 when arrays are used @@ -632,8 +667,10 @@ Node4 ShaderIR::GetTexsCode(Instruction instr, TextureType texture_type, const bool lod_bias_enabled = (process_mode != TextureProcessMode::None && process_mode != TextureProcessMode::LZ); - const auto [coord_count, total_coord_count] = ValidateAndGetCoordinateElement( - texture_type, depth_compare, is_array, lod_bias_enabled, 4, 4); + const auto coord_counts = ValidateAndGetCoordinateElement(texture_type, depth_compare, is_array, + lod_bias_enabled, 4, 4); + const auto coord_count = std::get<0>(coord_counts); + // If enabled arrays index is always stored in the gpr8 field const u64 array_register = instr.gpr8.Value(); // First coordinate index is stored in gpr8 field or (gpr8 + 1) when arrays are used @@ -682,12 +719,17 @@ Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool de u64 parameter_register = instr.gpr20.Value(); - const SamplerInfo info{texture_type, is_array, depth_compare, false}; - Node index_var{}; - const Sampler* sampler = is_bindless ? GetBindlessSampler(parameter_register++, index_var, info) - : GetSampler(instr.sampler, info); + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + info.is_shadow = depth_compare; + + Node index_var; + const std::optional<Sampler> sampler = + is_bindless ? GetBindlessSampler(parameter_register++, info, index_var) + : GetSampler(instr.sampler, info); Node4 values; - if (sampler == nullptr) { + if (!sampler) { for (u32 element = 0; element < values.size(); ++element) { values[element] = Immediate(0); } @@ -723,7 +765,7 @@ Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool de Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) { const auto texture_type{instr.tld.texture_type}; - const bool is_array{instr.tld.is_array}; + const bool is_array{instr.tld.is_array != 0}; const bool lod_enabled{instr.tld.GetTextureProcessMode() == TextureProcessMode::LL}; const std::size_t coord_count{GetCoordCount(texture_type)}; @@ -742,12 +784,12 @@ Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) { // const Node aoffi_register{is_aoffi ? GetRegister(gpr20_cursor++) : nullptr}; // const Node multisample{is_multisample ? GetRegister(gpr20_cursor++) : nullptr}; - const auto& sampler = *GetSampler(instr.sampler); + const std::optional<Sampler> sampler = GetSampler(instr.sampler, {}); Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, array_register, {}, {}, {}, {}, {}, lod, {}, element, {}}; + MetaTexture meta{*sampler, array_register, {}, {}, {}, {}, {}, lod, {}, element, {}}; values[element] = Operation(OperationCode::TexelFetch, meta, std::move(coords_copy)); } @@ -755,7 +797,11 @@ Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) { } Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is_array) { - const Sampler& sampler = *GetSampler(instr.sampler); + SamplerInfo info; + info.type = texture_type; + info.is_array = is_array; + info.is_shadow = false; + const std::optional<Sampler> sampler = GetSampler(instr.sampler, info); const std::size_t type_coord_count = GetCoordCount(texture_type); const bool lod_enabled = instr.tlds.GetTextureProcessMode() == TextureProcessMode::LL; @@ -783,7 +829,7 @@ Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, array, {}, {}, {}, {}, {}, lod, {}, element, {}}; + MetaTexture meta{*sampler, array, {}, {}, {}, {}, {}, lod, {}, element, {}}; values[element] = Operation(OperationCode::TexelFetch, meta, std::move(coords_copy)); } return values; diff --git a/src/video_core/shader/decode/video.cpp b/src/video_core/shader/decode/video.cpp index 64ba60ea2..1c0957277 100644 --- a/src/video_core/shader/decode/video.cpp +++ b/src/video_core/shader/decode/video.cpp @@ -91,29 +91,28 @@ u32 ShaderIR::DecodeVideo(NodeBlock& bb, u32 pc) { return pc; } -Node ShaderIR::GetVideoOperand(Node op, bool is_chunk, bool is_signed, - Tegra::Shader::VideoType type, u64 byte_height) { +Node ShaderIR::GetVideoOperand(Node op, bool is_chunk, bool is_signed, VideoType type, + u64 byte_height) { if (!is_chunk) { return BitfieldExtract(op, static_cast<u32>(byte_height * 8), 8); } - const Node zero = Immediate(0); switch (type) { - case Tegra::Shader::VideoType::Size16_Low: + case VideoType::Size16_Low: return BitfieldExtract(op, 0, 16); - case Tegra::Shader::VideoType::Size16_High: + case VideoType::Size16_High: return BitfieldExtract(op, 16, 16); - case Tegra::Shader::VideoType::Size32: + case VideoType::Size32: // TODO(Rodrigo): From my hardware tests it becomes a bit "mad" when this type is used // (1 * 1 + 0 == 0x5b800000). Until a better explanation is found: abort. UNIMPLEMENTED(); - return zero; - case Tegra::Shader::VideoType::Invalid: + return Immediate(0); + case VideoType::Invalid: UNREACHABLE_MSG("Invalid instruction encoding"); - return zero; + return Immediate(0); default: UNREACHABLE(); - return zero; + return Immediate(0); } } diff --git a/src/video_core/shader/decode/xmad.cpp b/src/video_core/shader/decode/xmad.cpp index 6191ffba1..233b8fa42 100644 --- a/src/video_core/shader/decode/xmad.cpp +++ b/src/video_core/shader/decode/xmad.cpp @@ -81,35 +81,36 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) { SetTemporary(bb, 0, product); product = GetTemporary(0); - const Node original_c = op_c; + Node original_c = op_c; const Tegra::Shader::XmadMode set_mode = mode; // Workaround to clang compile error - op_c = [&]() { + op_c = [&] { switch (set_mode) { case Tegra::Shader::XmadMode::None: return original_c; case Tegra::Shader::XmadMode::CLo: - return BitfieldExtract(original_c, 0, 16); + return BitfieldExtract(std::move(original_c), 0, 16); case Tegra::Shader::XmadMode::CHi: - return BitfieldExtract(original_c, 16, 16); + return BitfieldExtract(std::move(original_c), 16, 16); case Tegra::Shader::XmadMode::CBcc: { - const Node shifted_b = SignedOperation(OperationCode::ILogicalShiftLeft, is_signed_b, - original_b, Immediate(16)); - return SignedOperation(OperationCode::IAdd, is_signed_c, original_c, shifted_b); + Node shifted_b = SignedOperation(OperationCode::ILogicalShiftLeft, is_signed_b, + original_b, Immediate(16)); + return SignedOperation(OperationCode::IAdd, is_signed_c, std::move(original_c), + std::move(shifted_b)); } case Tegra::Shader::XmadMode::CSfu: { - const Node comp_a = GetPredicateComparisonInteger(PredCondition::Equal, is_signed_a, - op_a, Immediate(0)); - const Node comp_b = GetPredicateComparisonInteger(PredCondition::Equal, is_signed_b, - op_b, Immediate(0)); + const Node comp_a = + GetPredicateComparisonInteger(PredCondition::EQ, is_signed_a, op_a, Immediate(0)); + const Node comp_b = + GetPredicateComparisonInteger(PredCondition::EQ, is_signed_b, op_b, Immediate(0)); const Node comp = Operation(OperationCode::LogicalOr, comp_a, comp_b); const Node comp_minus_a = GetPredicateComparisonInteger( - PredCondition::NotEqual, is_signed_a, + PredCondition::NE, is_signed_a, SignedOperation(OperationCode::IBitwiseAnd, is_signed_a, op_a, Immediate(0x80000000)), Immediate(0)); const Node comp_minus_b = GetPredicateComparisonInteger( - PredCondition::NotEqual, is_signed_b, + PredCondition::NE, is_signed_b, SignedOperation(OperationCode::IBitwiseAnd, is_signed_b, op_b, Immediate(0x80000000)), Immediate(0)); diff --git a/src/video_core/shader/memory_util.cpp b/src/video_core/shader/memory_util.cpp new file mode 100644 index 000000000..e18ccba8e --- /dev/null +++ b/src/video_core/shader/memory_util.cpp @@ -0,0 +1,76 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <cstddef> + +#include <boost/container_hash/hash.hpp> + +#include "common/common_types.h" +#include "core/core.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/memory_manager.h" +#include "video_core/shader/memory_util.h" +#include "video_core/shader/shader_ir.h" + +namespace VideoCommon::Shader { + +GPUVAddr GetShaderAddress(Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::Engines::Maxwell3D::Regs::ShaderProgram program) { + const auto& shader_config{maxwell3d.regs.shader_config[static_cast<std::size_t>(program)]}; + return maxwell3d.regs.code_address.CodeAddress() + shader_config.offset; +} + +bool IsSchedInstruction(std::size_t offset, std::size_t main_offset) { + // Sched instructions appear once every 4 instructions. + constexpr std::size_t SchedPeriod = 4; + const std::size_t absolute_offset = offset - main_offset; + return (absolute_offset % SchedPeriod) == 0; +} + +std::size_t CalculateProgramSize(const ProgramCode& program, bool is_compute) { + // This is the encoded version of BRA that jumps to itself. All Nvidia + // shaders end with one. + static constexpr u64 SELF_JUMPING_BRANCH = 0xE2400FFFFF07000FULL; + static constexpr u64 MASK = 0xFFFFFFFFFF7FFFFFULL; + + const std::size_t start_offset = is_compute ? KERNEL_MAIN_OFFSET : STAGE_MAIN_OFFSET; + std::size_t offset = start_offset; + while (offset < program.size()) { + const u64 instruction = program[offset]; + if (!IsSchedInstruction(offset, start_offset)) { + if ((instruction & MASK) == SELF_JUMPING_BRANCH) { + // End on Maxwell's "nop" instruction + break; + } + if (instruction == 0) { + break; + } + } + ++offset; + } + // The last instruction is included in the program size + return std::min(offset + 1, program.size()); +} + +ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, GPUVAddr gpu_addr, + const u8* host_ptr, bool is_compute) { + ProgramCode code(VideoCommon::Shader::MAX_PROGRAM_LENGTH); + ASSERT_OR_EXECUTE(host_ptr != nullptr, { return code; }); + memory_manager.ReadBlockUnsafe(gpu_addr, code.data(), code.size() * sizeof(u64)); + code.resize(CalculateProgramSize(code, is_compute)); + return code; +} + +u64 GetUniqueIdentifier(Tegra::Engines::ShaderType shader_type, bool is_a, const ProgramCode& code, + const ProgramCode& code_b) { + size_t unique_identifier = boost::hash_value(code); + if (is_a) { + // VertexA programs include two programs + boost::hash_combine(unique_identifier, boost::hash_value(code_b)); + } + return static_cast<u64>(unique_identifier); +} + +} // namespace VideoCommon::Shader diff --git a/src/video_core/shader/memory_util.h b/src/video_core/shader/memory_util.h new file mode 100644 index 000000000..4624d38e6 --- /dev/null +++ b/src/video_core/shader/memory_util.h @@ -0,0 +1,43 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstddef> +#include <vector> + +#include "common/common_types.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/engines/shader_type.h" + +namespace Tegra { +class MemoryManager; +} + +namespace VideoCommon::Shader { + +using ProgramCode = std::vector<u64>; + +constexpr u32 STAGE_MAIN_OFFSET = 10; +constexpr u32 KERNEL_MAIN_OFFSET = 0; + +/// Gets the address for the specified shader stage program +GPUVAddr GetShaderAddress(Tegra::Engines::Maxwell3D& maxwell3d, + Tegra::Engines::Maxwell3D::Regs::ShaderProgram program); + +/// Gets if the current instruction offset is a scheduler instruction +bool IsSchedInstruction(std::size_t offset, std::size_t main_offset); + +/// Calculates the size of a program stream +std::size_t CalculateProgramSize(const ProgramCode& program, bool is_compute); + +/// Gets the shader program code from memory for the specified address +ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, GPUVAddr gpu_addr, + const u8* host_ptr, bool is_compute); + +/// Hashes one (or two) program streams +u64 GetUniqueIdentifier(Tegra::Engines::ShaderType shader_type, bool is_a, const ProgramCode& code, + const ProgramCode& code_b = {}); + +} // namespace VideoCommon::Shader diff --git a/src/video_core/shader/node.h b/src/video_core/shader/node.h index 3eee961f5..8f230d57a 100644 --- a/src/video_core/shader/node.h +++ b/src/video_core/shader/node.h @@ -110,13 +110,20 @@ enum class OperationCode { LogicalPick2, /// (bool2 pair, uint index) -> bool LogicalAnd2, /// (bool2 a) -> bool - LogicalFLessThan, /// (float a, float b) -> bool - LogicalFEqual, /// (float a, float b) -> bool - LogicalFLessEqual, /// (float a, float b) -> bool - LogicalFGreaterThan, /// (float a, float b) -> bool - LogicalFNotEqual, /// (float a, float b) -> bool - LogicalFGreaterEqual, /// (float a, float b) -> bool - LogicalFIsNan, /// (float a) -> bool + LogicalFOrdLessThan, /// (float a, float b) -> bool + LogicalFOrdEqual, /// (float a, float b) -> bool + LogicalFOrdLessEqual, /// (float a, float b) -> bool + LogicalFOrdGreaterThan, /// (float a, float b) -> bool + LogicalFOrdNotEqual, /// (float a, float b) -> bool + LogicalFOrdGreaterEqual, /// (float a, float b) -> bool + LogicalFOrdered, /// (float a, float b) -> bool + LogicalFUnordered, /// (float a, float b) -> bool + LogicalFUnordLessThan, /// (float a, float b) -> bool + LogicalFUnordEqual, /// (float a, float b) -> bool + LogicalFUnordLessEqual, /// (float a, float b) -> bool + LogicalFUnordGreaterThan, /// (float a, float b) -> bool + LogicalFUnordNotEqual, /// (float a, float b) -> bool + LogicalFUnordGreaterEqual, /// (float a, float b) -> bool LogicalILessThan, /// (int a, int b) -> bool LogicalIEqual, /// (int a, int b) -> bool @@ -132,6 +139,8 @@ enum class OperationCode { LogicalUNotEqual, /// (uint a, uint b) -> bool LogicalUGreaterEqual, /// (uint a, uint b) -> bool + LogicalAddCarry, /// (uint a, uint b) -> bool + Logical2HLessThan, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2 Logical2HEqual, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2 Logical2HLessEqual, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2 @@ -217,9 +226,16 @@ enum class OperationCode { VoteEqual, /// (bool) -> bool ThreadId, /// () -> uint + ThreadEqMask, /// () -> uint + ThreadGeMask, /// () -> uint + ThreadGtMask, /// () -> uint + ThreadLeMask, /// () -> uint + ThreadLtMask, /// () -> uint ShuffleIndexed, /// (uint value, uint index) -> uint - MemoryBarrierGL, /// () -> void + Barrier, /// () -> void + MemoryBarrierGroup, /// () -> void + MemoryBarrierGlobal, /// () -> void Amount, }; @@ -259,133 +275,76 @@ using Node = std::shared_ptr<NodeData>; using Node4 = std::array<Node, 4>; using NodeBlock = std::vector<Node>; -class BindlessSamplerNode; -class ArraySamplerNode; +struct ArraySamplerNode; +struct BindlessSamplerNode; +struct SeparateSamplerNode; -using TrackSamplerData = std::variant<BindlessSamplerNode, ArraySamplerNode>; +using TrackSamplerData = std::variant<BindlessSamplerNode, SeparateSamplerNode, ArraySamplerNode>; using TrackSampler = std::shared_ptr<TrackSamplerData>; -class Sampler { -public: - /// This constructor is for bound samplers +struct Sampler { + /// Bound samplers constructor constexpr explicit Sampler(u32 index, u32 offset, Tegra::Shader::TextureType type, bool is_array, bool is_shadow, bool is_buffer, bool is_indexed) : index{index}, offset{offset}, type{type}, is_array{is_array}, is_shadow{is_shadow}, is_buffer{is_buffer}, is_indexed{is_indexed} {} - /// This constructor is for bindless samplers + /// Separate sampler constructor + constexpr explicit Sampler(u32 index, std::pair<u32, u32> offsets, std::pair<u32, u32> buffers, + Tegra::Shader::TextureType type, bool is_array, bool is_shadow, + bool is_buffer) + : index{index}, offset{offsets.first}, secondary_offset{offsets.second}, + buffer{buffers.first}, secondary_buffer{buffers.second}, type{type}, is_array{is_array}, + is_shadow{is_shadow}, is_buffer{is_buffer}, is_separated{true} {} + + /// Bindless samplers constructor constexpr explicit Sampler(u32 index, u32 offset, u32 buffer, Tegra::Shader::TextureType type, bool is_array, bool is_shadow, bool is_buffer, bool is_indexed) : index{index}, offset{offset}, buffer{buffer}, type{type}, is_array{is_array}, is_shadow{is_shadow}, is_buffer{is_buffer}, is_bindless{true}, is_indexed{is_indexed} {} - constexpr u32 GetIndex() const { - return index; - } - - constexpr u32 GetOffset() const { - return offset; - } - - constexpr u32 GetBuffer() const { - return buffer; - } - - constexpr Tegra::Shader::TextureType GetType() const { - return type; - } - - constexpr bool IsArray() const { - return is_array; - } - - constexpr bool IsShadow() const { - return is_shadow; - } - - constexpr bool IsBuffer() const { - return is_buffer; - } - - constexpr bool IsBindless() const { - return is_bindless; - } - - constexpr bool IsIndexed() const { - return is_indexed; - } - - constexpr u32 Size() const { - return size; - } - - constexpr void SetSize(u32 new_size) { - size = new_size; - } - -private: - u32 index{}; ///< Emulated index given for the this sampler. - u32 offset{}; ///< Offset in the const buffer from where the sampler is being read. - u32 buffer{}; ///< Buffer where the bindless sampler is being read (unused on bound samplers). - u32 size{1}; ///< Size of the sampler. + u32 index = 0; ///< Emulated index given for the this sampler. + u32 offset = 0; ///< Offset in the const buffer from where the sampler is being read. + u32 secondary_offset = 0; ///< Secondary offset in the const buffer. + u32 buffer = 0; ///< Buffer where the bindless sampler is read. + u32 secondary_buffer = 0; ///< Secondary buffer where the bindless sampler is read. + u32 size = 1; ///< Size of the sampler. Tegra::Shader::TextureType type{}; ///< The type used to sample this texture (Texture2D, etc) - bool is_array{}; ///< Whether the texture is being sampled as an array texture or not. - bool is_shadow{}; ///< Whether the texture is being sampled as a depth texture or not. - bool is_buffer{}; ///< Whether the texture is a texture buffer without sampler. - bool is_bindless{}; ///< Whether this sampler belongs to a bindless texture or not. - bool is_indexed{}; ///< Whether this sampler is an indexed array of textures. + bool is_array = false; ///< Whether the texture is being sampled as an array texture or not. + bool is_shadow = false; ///< Whether the texture is being sampled as a depth texture or not. + bool is_buffer = false; ///< Whether the texture is a texture buffer without sampler. + bool is_bindless = false; ///< Whether this sampler belongs to a bindless texture or not. + bool is_indexed = false; ///< Whether this sampler is an indexed array of textures. + bool is_separated = false; ///< Whether the image and sampler is separated or not. }; /// Represents a tracked bindless sampler into a direct const buffer -class ArraySamplerNode final { -public: - explicit ArraySamplerNode(u32 index, u32 base_offset, u32 bindless_var) - : index{index}, base_offset{base_offset}, bindless_var{bindless_var} {} - - constexpr u32 GetIndex() const { - return index; - } - - constexpr u32 GetBaseOffset() const { - return base_offset; - } - - constexpr u32 GetIndexVar() const { - return bindless_var; - } - -private: +struct ArraySamplerNode { u32 index; u32 base_offset; u32 bindless_var; }; -/// Represents a tracked bindless sampler into a direct const buffer -class BindlessSamplerNode final { -public: - explicit BindlessSamplerNode(u32 index, u32 offset) : index{index}, offset{offset} {} - - constexpr u32 GetIndex() const { - return index; - } - - constexpr u32 GetOffset() const { - return offset; - } +/// Represents a tracked separate sampler image pair that was folded statically +struct SeparateSamplerNode { + std::pair<u32, u32> indices; + std::pair<u32, u32> offsets; +}; -private: +/// Represents a tracked bindless sampler into a direct const buffer +struct BindlessSamplerNode { u32 index; u32 offset; }; -class Image final { +struct Image { public: - /// This constructor is for bound images + /// Bound images constructor constexpr explicit Image(u32 index, u32 offset, Tegra::Shader::ImageType type) : index{index}, offset{offset}, type{type} {} - /// This constructor is for bindless samplers + /// Bindless samplers constructor constexpr explicit Image(u32 index, u32 offset, u32 buffer, Tegra::Shader::ImageType type) : index{index}, offset{offset}, buffer{buffer}, type{type}, is_bindless{true} {} @@ -403,53 +362,20 @@ public: is_atomic = true; } - constexpr u32 GetIndex() const { - return index; - } - - constexpr u32 GetOffset() const { - return offset; - } - - constexpr u32 GetBuffer() const { - return buffer; - } - - constexpr Tegra::Shader::ImageType GetType() const { - return type; - } - - constexpr bool IsBindless() const { - return is_bindless; - } - - constexpr bool IsWritten() const { - return is_written; - } - - constexpr bool IsRead() const { - return is_read; - } - - constexpr bool IsAtomic() const { - return is_atomic; - } - -private: - u32 index{}; - u32 offset{}; - u32 buffer{}; + u32 index = 0; + u32 offset = 0; + u32 buffer = 0; Tegra::Shader::ImageType type{}; - bool is_bindless{}; - bool is_written{}; - bool is_read{}; - bool is_atomic{}; + bool is_bindless = false; + bool is_written = false; + bool is_read = false; + bool is_atomic = false; }; struct GlobalMemoryBase { - u32 cbuf_index{}; - u32 cbuf_offset{}; + u32 cbuf_index = 0; + u32 cbuf_offset = 0; bool operator<(const GlobalMemoryBase& rhs) const { return std::tie(cbuf_index, cbuf_offset) < std::tie(rhs.cbuf_index, rhs.cbuf_offset); @@ -463,7 +389,7 @@ struct MetaArithmetic { /// Parameters describing a texture sampler struct MetaTexture { - const Sampler& sampler; + Sampler sampler; Node array; Node depth_compare; std::vector<Node> aoffi; diff --git a/src/video_core/shader/node_helper.h b/src/video_core/shader/node_helper.h index 11231bbea..1e0886185 100644 --- a/src/video_core/shader/node_helper.h +++ b/src/video_core/shader/node_helper.h @@ -48,7 +48,7 @@ Node MakeNode(Args&&... args) { template <typename T, typename... Args> TrackSampler MakeTrackSampler(Args&&... args) { static_assert(std::is_convertible_v<T, TrackSamplerData>); - return std::make_shared<TrackSamplerData>(T(std::forward<Args>(args)...)); + return std::make_shared<TrackSamplerData>(T{std::forward<Args>(args)...}); } template <typename... Args> diff --git a/src/video_core/shader/registry.cpp b/src/video_core/shader/registry.cpp index af70b3f35..148d91fcb 100644 --- a/src/video_core/shader/registry.cpp +++ b/src/video_core/shader/registry.cpp @@ -24,44 +24,45 @@ GraphicsInfo MakeGraphicsInfo(ShaderType shader_stage, ConstBufferEngineInterfac if (shader_stage == ShaderType::Compute) { return {}; } - auto& graphics = static_cast<Tegra::Engines::Maxwell3D&>(engine); - - GraphicsInfo info; - info.tfb_layouts = graphics.regs.tfb_layouts; - info.tfb_varying_locs = graphics.regs.tfb_varying_locs; - info.primitive_topology = graphics.regs.draw.topology; - info.tessellation_primitive = graphics.regs.tess_mode.prim; - info.tessellation_spacing = graphics.regs.tess_mode.spacing; - info.tfb_enabled = graphics.regs.tfb_enabled; - info.tessellation_clockwise = graphics.regs.tess_mode.cw; - return info; + + auto& graphics = dynamic_cast<Tegra::Engines::Maxwell3D&>(engine); + + return { + .tfb_layouts = graphics.regs.tfb_layouts, + .tfb_varying_locs = graphics.regs.tfb_varying_locs, + .primitive_topology = graphics.regs.draw.topology, + .tessellation_primitive = graphics.regs.tess_mode.prim, + .tessellation_spacing = graphics.regs.tess_mode.spacing, + .tfb_enabled = graphics.regs.tfb_enabled != 0, + .tessellation_clockwise = graphics.regs.tess_mode.cw.Value() != 0, + }; } ComputeInfo MakeComputeInfo(ShaderType shader_stage, ConstBufferEngineInterface& engine) { if (shader_stage != ShaderType::Compute) { return {}; } - auto& compute = static_cast<Tegra::Engines::KeplerCompute&>(engine); + + auto& compute = dynamic_cast<Tegra::Engines::KeplerCompute&>(engine); const auto& launch = compute.launch_description; - ComputeInfo info; - info.workgroup_size = {launch.block_dim_x, launch.block_dim_y, launch.block_dim_z}; - info.local_memory_size_in_words = launch.local_pos_alloc; - info.shared_memory_size_in_words = launch.shared_alloc; - return info; + return { + .workgroup_size = {launch.block_dim_x, launch.block_dim_y, launch.block_dim_z}, + .shared_memory_size_in_words = launch.shared_alloc, + .local_memory_size_in_words = launch.local_pos_alloc, + }; } } // Anonymous namespace -Registry::Registry(Tegra::Engines::ShaderType shader_stage, const SerializedRegistryInfo& info) +Registry::Registry(ShaderType shader_stage, const SerializedRegistryInfo& info) : stage{shader_stage}, stored_guest_driver_profile{info.guest_driver_profile}, bound_buffer{info.bound_buffer}, graphics_info{info.graphics}, compute_info{info.compute} {} -Registry::Registry(Tegra::Engines::ShaderType shader_stage, - Tegra::Engines::ConstBufferEngineInterface& engine) - : stage{shader_stage}, engine{&engine}, bound_buffer{engine.GetBoundBuffer()}, - graphics_info{MakeGraphicsInfo(shader_stage, engine)}, compute_info{MakeComputeInfo( - shader_stage, engine)} {} +Registry::Registry(ShaderType shader_stage, ConstBufferEngineInterface& engine_) + : stage{shader_stage}, engine{&engine_}, bound_buffer{engine_.GetBoundBuffer()}, + graphics_info{MakeGraphicsInfo(shader_stage, engine_)}, compute_info{MakeComputeInfo( + shader_stage, engine_)} {} Registry::~Registry() = default; @@ -93,8 +94,27 @@ std::optional<SamplerDescriptor> Registry::ObtainBoundSampler(u32 offset) { return value; } -std::optional<Tegra::Engines::SamplerDescriptor> Registry::ObtainBindlessSampler(u32 buffer, - u32 offset) { +std::optional<Tegra::Engines::SamplerDescriptor> Registry::ObtainSeparateSampler( + std::pair<u32, u32> buffers, std::pair<u32, u32> offsets) { + SeparateSamplerKey key; + key.buffers = buffers; + key.offsets = offsets; + const auto iter = separate_samplers.find(key); + if (iter != separate_samplers.end()) { + return iter->second; + } + if (!engine) { + return std::nullopt; + } + + const u32 handle_1 = engine->AccessConstBuffer32(stage, key.buffers.first, key.offsets.first); + const u32 handle_2 = engine->AccessConstBuffer32(stage, key.buffers.second, key.offsets.second); + const SamplerDescriptor value = engine->AccessSampler(handle_1 | handle_2); + separate_samplers.emplace(key, value); + return value; +} + +std::optional<SamplerDescriptor> Registry::ObtainBindlessSampler(u32 buffer, u32 offset) { const std::pair key = {buffer, offset}; const auto iter = bindless_samplers.find(key); if (iter != bindless_samplers.end()) { diff --git a/src/video_core/shader/registry.h b/src/video_core/shader/registry.h index 0c80d35fd..4bebefdde 100644 --- a/src/video_core/shader/registry.h +++ b/src/video_core/shader/registry.h @@ -19,8 +19,39 @@ namespace VideoCommon::Shader { +struct SeparateSamplerKey { + std::pair<u32, u32> buffers; + std::pair<u32, u32> offsets; +}; + +} // namespace VideoCommon::Shader + +namespace std { + +template <> +struct hash<VideoCommon::Shader::SeparateSamplerKey> { + std::size_t operator()(const VideoCommon::Shader::SeparateSamplerKey& key) const noexcept { + return std::hash<u32>{}(key.buffers.first ^ key.buffers.second ^ key.offsets.first ^ + key.offsets.second); + } +}; + +template <> +struct equal_to<VideoCommon::Shader::SeparateSamplerKey> { + bool operator()(const VideoCommon::Shader::SeparateSamplerKey& lhs, + const VideoCommon::Shader::SeparateSamplerKey& rhs) const noexcept { + return lhs.buffers == rhs.buffers && lhs.offsets == rhs.offsets; + } +}; + +} // namespace std + +namespace VideoCommon::Shader { + using KeyMap = std::unordered_map<std::pair<u32, u32>, u32, Common::PairHash>; using BoundSamplerMap = std::unordered_map<u32, Tegra::Engines::SamplerDescriptor>; +using SeparateSamplerMap = + std::unordered_map<SeparateSamplerKey, Tegra::Engines::SamplerDescriptor>; using BindlessSamplerMap = std::unordered_map<std::pair<u32, u32>, Tegra::Engines::SamplerDescriptor, Common::PairHash>; @@ -63,7 +94,7 @@ public: explicit Registry(Tegra::Engines::ShaderType shader_stage, const SerializedRegistryInfo& info); explicit Registry(Tegra::Engines::ShaderType shader_stage, - Tegra::Engines::ConstBufferEngineInterface& engine); + Tegra::Engines::ConstBufferEngineInterface& engine_); ~Registry(); @@ -73,6 +104,9 @@ public: std::optional<Tegra::Engines::SamplerDescriptor> ObtainBoundSampler(u32 offset); + std::optional<Tegra::Engines::SamplerDescriptor> ObtainSeparateSampler( + std::pair<u32, u32> buffers, std::pair<u32, u32> offsets); + std::optional<Tegra::Engines::SamplerDescriptor> ObtainBindlessSampler(u32 buffer, u32 offset); /// Inserts a key. @@ -128,6 +162,7 @@ private: Tegra::Engines::ConstBufferEngineInterface* engine = nullptr; KeyMap keys; BoundSamplerMap bound_samplers; + SeparateSamplerMap separate_samplers; BindlessSamplerMap bindless_samplers; u32 bound_buffer; GraphicsInfo graphics_info; diff --git a/src/video_core/shader/shader_ir.cpp b/src/video_core/shader/shader_ir.cpp index 8852c8a1b..29d794b34 100644 --- a/src/video_core/shader/shader_ir.cpp +++ b/src/video_core/shader/shader_ir.cpp @@ -10,6 +10,7 @@ #include "common/common_types.h" #include "common/logging/log.h" #include "video_core/engines/shader_bytecode.h" +#include "video_core/shader/node.h" #include "video_core/shader/node_helper.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" @@ -56,8 +57,7 @@ Node ShaderIR::GetConstBuffer(u64 index_, u64 offset_) { const auto index = static_cast<u32>(index_); const auto offset = static_cast<u32>(offset_); - const auto [entry, is_new] = used_cbufs.try_emplace(index); - entry->second.MarkAsUsed(offset); + used_cbufs.try_emplace(index).first->second.MarkAsUsed(offset); return MakeNode<CbufNode>(index, Immediate(offset)); } @@ -66,8 +66,7 @@ Node ShaderIR::GetConstBufferIndirect(u64 index_, u64 offset_, Node node) { const auto index = static_cast<u32>(index_); const auto offset = static_cast<u32>(offset_); - const auto [entry, is_new] = used_cbufs.try_emplace(index); - entry->second.MarkAsUsedIndirect(); + used_cbufs.try_emplace(index).first->second.MarkAsUsedIndirect(); Node final_offset = [&] { // Attempt to inline constant buffer without a variable offset. This is done to allow @@ -113,9 +112,9 @@ Node ShaderIR::GetOutputAttribute(Attribute::Index index, u64 element, Node buff } Node ShaderIR::GetInternalFlag(InternalFlag flag, bool negated) const { - const Node node = MakeNode<InternalFlagNode>(flag); + Node node = MakeNode<InternalFlagNode>(flag); if (negated) { - return Operation(OperationCode::LogicalNegate, node); + return Operation(OperationCode::LogicalNegate, std::move(node)); } return node; } @@ -166,6 +165,7 @@ Node ShaderIR::ConvertIntegerSize(Node value, Register::Size size, bool is_signe std::move(value), Immediate(16)); value = SignedOperation(OperationCode::IArithmeticShiftRight, is_signed, NO_PRECISE, std::move(value), Immediate(16)); + return value; case Register::Size::Word: // Default - do nothing return value; @@ -244,56 +244,44 @@ Node ShaderIR::GetSaturatedHalfFloat(Node value, bool saturate) { } Node ShaderIR::GetPredicateComparisonFloat(PredCondition condition, Node op_a, Node op_b) { + if (condition == PredCondition::T) { + return GetPredicate(true); + } else if (condition == PredCondition::F) { + return GetPredicate(false); + } + static constexpr std::array comparison_table{ - std::pair{PredCondition::LessThan, OperationCode::LogicalFLessThan}, - std::pair{PredCondition::Equal, OperationCode::LogicalFEqual}, - std::pair{PredCondition::LessEqual, OperationCode::LogicalFLessEqual}, - std::pair{PredCondition::GreaterThan, OperationCode::LogicalFGreaterThan}, - std::pair{PredCondition::NotEqual, OperationCode::LogicalFNotEqual}, - std::pair{PredCondition::GreaterEqual, OperationCode::LogicalFGreaterEqual}, - std::pair{PredCondition::LessThanWithNan, OperationCode::LogicalFLessThan}, - std::pair{PredCondition::NotEqualWithNan, OperationCode::LogicalFNotEqual}, - std::pair{PredCondition::LessEqualWithNan, OperationCode::LogicalFLessEqual}, - std::pair{PredCondition::GreaterThanWithNan, OperationCode::LogicalFGreaterThan}, - std::pair{PredCondition::GreaterEqualWithNan, OperationCode::LogicalFGreaterEqual}, + OperationCode(0), + OperationCode::LogicalFOrdLessThan, // LT + OperationCode::LogicalFOrdEqual, // EQ + OperationCode::LogicalFOrdLessEqual, // LE + OperationCode::LogicalFOrdGreaterThan, // GT + OperationCode::LogicalFOrdNotEqual, // NE + OperationCode::LogicalFOrdGreaterEqual, // GE + OperationCode::LogicalFOrdered, // NUM + OperationCode::LogicalFUnordered, // NAN + OperationCode::LogicalFUnordLessThan, // LTU + OperationCode::LogicalFUnordEqual, // EQU + OperationCode::LogicalFUnordLessEqual, // LEU + OperationCode::LogicalFUnordGreaterThan, // GTU + OperationCode::LogicalFUnordNotEqual, // NEU + OperationCode::LogicalFUnordGreaterEqual, // GEU }; + const std::size_t index = static_cast<std::size_t>(condition); + ASSERT_MSG(index < std::size(comparison_table), "Invalid condition={}", index); - const auto comparison = - std::find_if(comparison_table.cbegin(), comparison_table.cend(), - [condition](const auto entry) { return condition == entry.first; }); - UNIMPLEMENTED_IF_MSG(comparison == comparison_table.cend(), - "Unknown predicate comparison operation"); - - Node predicate = Operation(comparison->second, NO_PRECISE, op_a, op_b); - - if (condition == PredCondition::LessThanWithNan || - condition == PredCondition::NotEqualWithNan || - condition == PredCondition::LessEqualWithNan || - condition == PredCondition::GreaterThanWithNan || - condition == PredCondition::GreaterEqualWithNan) { - predicate = Operation(OperationCode::LogicalOr, predicate, - Operation(OperationCode::LogicalFIsNan, op_a)); - predicate = Operation(OperationCode::LogicalOr, predicate, - Operation(OperationCode::LogicalFIsNan, op_b)); - } - - return predicate; + return Operation(comparison_table[index], op_a, op_b); } Node ShaderIR::GetPredicateComparisonInteger(PredCondition condition, bool is_signed, Node op_a, Node op_b) { static constexpr std::array comparison_table{ - std::pair{PredCondition::LessThan, OperationCode::LogicalILessThan}, - std::pair{PredCondition::Equal, OperationCode::LogicalIEqual}, - std::pair{PredCondition::LessEqual, OperationCode::LogicalILessEqual}, - std::pair{PredCondition::GreaterThan, OperationCode::LogicalIGreaterThan}, - std::pair{PredCondition::NotEqual, OperationCode::LogicalINotEqual}, - std::pair{PredCondition::GreaterEqual, OperationCode::LogicalIGreaterEqual}, - std::pair{PredCondition::LessThanWithNan, OperationCode::LogicalILessThan}, - std::pair{PredCondition::NotEqualWithNan, OperationCode::LogicalINotEqual}, - std::pair{PredCondition::LessEqualWithNan, OperationCode::LogicalILessEqual}, - std::pair{PredCondition::GreaterThanWithNan, OperationCode::LogicalIGreaterThan}, - std::pair{PredCondition::GreaterEqualWithNan, OperationCode::LogicalIGreaterEqual}, + std::pair{PredCondition::LT, OperationCode::LogicalILessThan}, + std::pair{PredCondition::EQ, OperationCode::LogicalIEqual}, + std::pair{PredCondition::LE, OperationCode::LogicalILessEqual}, + std::pair{PredCondition::GT, OperationCode::LogicalIGreaterThan}, + std::pair{PredCondition::NE, OperationCode::LogicalINotEqual}, + std::pair{PredCondition::GE, OperationCode::LogicalIGreaterEqual}, }; const auto comparison = @@ -302,32 +290,24 @@ Node ShaderIR::GetPredicateComparisonInteger(PredCondition condition, bool is_si UNIMPLEMENTED_IF_MSG(comparison == comparison_table.cend(), "Unknown predicate comparison operation"); - Node predicate = SignedOperation(comparison->second, is_signed, NO_PRECISE, std::move(op_a), - std::move(op_b)); - - UNIMPLEMENTED_IF_MSG(condition == PredCondition::LessThanWithNan || - condition == PredCondition::NotEqualWithNan || - condition == PredCondition::LessEqualWithNan || - condition == PredCondition::GreaterThanWithNan || - condition == PredCondition::GreaterEqualWithNan, - "NaN comparisons for integers are not implemented"); - return predicate; + return SignedOperation(comparison->second, is_signed, NO_PRECISE, std::move(op_a), + std::move(op_b)); } Node ShaderIR::GetPredicateComparisonHalf(Tegra::Shader::PredCondition condition, Node op_a, Node op_b) { static constexpr std::array comparison_table{ - std::pair{PredCondition::LessThan, OperationCode::Logical2HLessThan}, - std::pair{PredCondition::Equal, OperationCode::Logical2HEqual}, - std::pair{PredCondition::LessEqual, OperationCode::Logical2HLessEqual}, - std::pair{PredCondition::GreaterThan, OperationCode::Logical2HGreaterThan}, - std::pair{PredCondition::NotEqual, OperationCode::Logical2HNotEqual}, - std::pair{PredCondition::GreaterEqual, OperationCode::Logical2HGreaterEqual}, - std::pair{PredCondition::LessThanWithNan, OperationCode::Logical2HLessThanWithNan}, - std::pair{PredCondition::NotEqualWithNan, OperationCode::Logical2HNotEqualWithNan}, - std::pair{PredCondition::LessEqualWithNan, OperationCode::Logical2HLessEqualWithNan}, - std::pair{PredCondition::GreaterThanWithNan, OperationCode::Logical2HGreaterThanWithNan}, - std::pair{PredCondition::GreaterEqualWithNan, OperationCode::Logical2HGreaterEqualWithNan}, + std::pair{PredCondition::LT, OperationCode::Logical2HLessThan}, + std::pair{PredCondition::EQ, OperationCode::Logical2HEqual}, + std::pair{PredCondition::LE, OperationCode::Logical2HLessEqual}, + std::pair{PredCondition::GT, OperationCode::Logical2HGreaterThan}, + std::pair{PredCondition::NE, OperationCode::Logical2HNotEqual}, + std::pair{PredCondition::GE, OperationCode::Logical2HGreaterEqual}, + std::pair{PredCondition::LTU, OperationCode::Logical2HLessThanWithNan}, + std::pair{PredCondition::LEU, OperationCode::Logical2HLessEqualWithNan}, + std::pair{PredCondition::GTU, OperationCode::Logical2HGreaterThanWithNan}, + std::pair{PredCondition::NEU, OperationCode::Logical2HNotEqualWithNan}, + std::pair{PredCondition::GEU, OperationCode::Logical2HGreaterEqualWithNan}, }; const auto comparison = @@ -398,7 +378,7 @@ void ShaderIR::SetInternalFlagsFromFloat(NodeBlock& bb, Node value, bool sets_cc if (!sets_cc) { return; } - Node zerop = Operation(OperationCode::LogicalFEqual, std::move(value), Immediate(0.0f)); + Node zerop = Operation(OperationCode::LogicalFOrdEqual, std::move(value), Immediate(0.0f)); SetInternalFlag(bb, InternalFlag::Zero, std::move(zerop)); LOG_WARNING(HW_GPU, "Condition codes implementation is incomplete"); } diff --git a/src/video_core/shader/shader_ir.h b/src/video_core/shader/shader_ir.h index c6e7bdf50..3a98b2104 100644 --- a/src/video_core/shader/shader_ir.h +++ b/src/video_core/shader/shader_ir.h @@ -18,6 +18,7 @@ #include "video_core/engines/shader_header.h" #include "video_core/shader/ast.h" #include "video_core/shader/compiler_settings.h" +#include "video_core/shader/memory_util.h" #include "video_core/shader/node.h" #include "video_core/shader/registry.h" @@ -25,16 +26,13 @@ namespace VideoCommon::Shader { struct ShaderBlock; -using ProgramCode = std::vector<u64>; - constexpr u32 MAX_PROGRAM_LENGTH = 0x1000; -class ConstBuffer { -public: - explicit ConstBuffer(u32 max_offset, bool is_indirect) +struct ConstBuffer { + constexpr explicit ConstBuffer(u32 max_offset, bool is_indirect) : max_offset{max_offset}, is_indirect{is_indirect} {} - ConstBuffer() = default; + constexpr ConstBuffer() = default; void MarkAsUsed(u64 offset) { max_offset = std::max(max_offset, static_cast<u32>(offset)); @@ -57,8 +55,8 @@ public: } private: - u32 max_offset{}; - bool is_indirect{}; + u32 max_offset = 0; + bool is_indirect = false; }; struct GlobalMemoryUsage { @@ -192,10 +190,14 @@ private: friend class ASTDecoder; struct SamplerInfo { - Tegra::Shader::TextureType type; - bool is_array; - bool is_shadow; - bool is_buffer; + std::optional<Tegra::Shader::TextureType> type; + std::optional<bool> is_array; + std::optional<bool> is_shadow; + std::optional<bool> is_buffer; + + constexpr bool IsComplete() const noexcept { + return type && is_array && is_shadow && is_buffer; + } }; void Decode(); @@ -328,16 +330,15 @@ private: OperationCode GetPredicateCombiner(Tegra::Shader::PredOperation operation); /// Queries the missing sampler info from the execution context. - SamplerInfo GetSamplerInfo(std::optional<SamplerInfo> sampler_info, u32 offset, - std::optional<u32> buffer = std::nullopt); + SamplerInfo GetSamplerInfo(SamplerInfo info, + std::optional<Tegra::Engines::SamplerDescriptor> sampler); - /// Accesses a texture sampler - const Sampler* GetSampler(const Tegra::Shader::Sampler& sampler, - std::optional<SamplerInfo> sampler_info = std::nullopt); + /// Accesses a texture sampler. + std::optional<Sampler> GetSampler(Tegra::Shader::Sampler sampler, SamplerInfo info); /// Accesses a texture sampler for a bindless texture. - const Sampler* GetBindlessSampler(Tegra::Shader::Register reg, Node& index_var, - std::optional<SamplerInfo> sampler_info = std::nullopt); + std::optional<Sampler> GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, + Node& index_var); /// Accesses an image. Image& GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type); @@ -408,8 +409,14 @@ private: std::tuple<Node, u32, u32> TrackCbuf(Node tracked, const NodeBlock& code, s64 cursor) const; - std::tuple<Node, TrackSampler> TrackBindlessSampler(Node tracked, const NodeBlock& code, - s64 cursor); + std::pair<Node, TrackSampler> TrackBindlessSampler(Node tracked, const NodeBlock& code, + s64 cursor); + + std::pair<Node, TrackSampler> HandleBindlessIndirectRead(const CbufNode& cbuf, + const OperationNode& operation, + Node gpr, Node base_offset, + Node tracked, const NodeBlock& code, + s64 cursor); std::optional<u32> TrackImmediate(Node tracked, const NodeBlock& code, s64 cursor) const; diff --git a/src/video_core/shader/track.cpp b/src/video_core/shader/track.cpp index 10739b37d..6be3ea92b 100644 --- a/src/video_core/shader/track.cpp +++ b/src/video_core/shader/track.cpp @@ -14,6 +14,7 @@ namespace VideoCommon::Shader { namespace { + std::pair<Node, s64> FindOperation(const NodeBlock& code, s64 cursor, OperationCode operation_code) { for (; cursor >= 0; --cursor) { @@ -27,8 +28,9 @@ std::pair<Node, s64> FindOperation(const NodeBlock& code, s64 cursor, if (const auto conditional = std::get_if<ConditionalNode>(&*node)) { const auto& conditional_code = conditional->GetCode(); - auto [found, internal_cursor] = FindOperation( + auto result = FindOperation( conditional_code, static_cast<s64>(conditional_code.size() - 1), operation_code); + auto& found = result.first; if (found) { return {std::move(found), cursor}; } @@ -62,7 +64,8 @@ bool AmendNodeCv(std::size_t amend_index, Node node) { if (const auto operation = std::get_if<OperationNode>(&*node)) { operation->SetAmendIndex(amend_index); return true; - } else if (const auto conditional = std::get_if<ConditionalNode>(&*node)) { + } + if (const auto conditional = std::get_if<ConditionalNode>(&*node)) { conditional->SetAmendIndex(amend_index); return true; } @@ -71,39 +74,27 @@ bool AmendNodeCv(std::size_t amend_index, Node node) { } // Anonymous namespace -std::tuple<Node, TrackSampler> ShaderIR::TrackBindlessSampler(Node tracked, const NodeBlock& code, - s64 cursor) { +std::pair<Node, TrackSampler> ShaderIR::TrackBindlessSampler(Node tracked, const NodeBlock& code, + s64 cursor) { if (const auto cbuf = std::get_if<CbufNode>(&*tracked)) { + const u32 cbuf_index = cbuf->GetIndex(); + // Constant buffer found, test if it's an immediate - const auto offset = cbuf->GetOffset(); + const auto& offset = cbuf->GetOffset(); if (const auto immediate = std::get_if<ImmediateNode>(&*offset)) { - auto track = - MakeTrackSampler<BindlessSamplerNode>(cbuf->GetIndex(), immediate->GetValue()); + auto track = MakeTrackSampler<BindlessSamplerNode>(cbuf_index, immediate->GetValue()); return {tracked, track}; - } else if (const auto operation = std::get_if<OperationNode>(&*offset)) { + } + if (const auto operation = std::get_if<OperationNode>(&*offset)) { const u32 bound_buffer = registry.GetBoundBuffer(); - if (bound_buffer != cbuf->GetIndex()) { + if (bound_buffer != cbuf_index) { return {}; } - const auto pair = DecoupleIndirectRead(*operation); - if (!pair) { - return {}; + if (const std::optional pair = DecoupleIndirectRead(*operation)) { + auto [gpr, base_offset] = *pair; + return HandleBindlessIndirectRead(*cbuf, *operation, gpr, base_offset, tracked, + code, cursor); } - auto [gpr, base_offset] = *pair; - const auto offset_inm = std::get_if<ImmediateNode>(&*base_offset); - const auto& gpu_driver = registry.AccessGuestDriverProfile(); - const u32 bindless_cv = NewCustomVariable(); - const Node op = - Operation(OperationCode::UDiv, gpr, Immediate(gpu_driver.GetTextureHandlerSize())); - - const Node cv_node = GetCustomVariable(bindless_cv); - Node amend_op = Operation(OperationCode::Assign, cv_node, std::move(op)); - const std::size_t amend_index = DeclareAmend(amend_op); - AmendNodeCv(amend_index, code[cursor]); - // TODO Implement Bindless Index custom variable - auto track = MakeTrackSampler<ArraySamplerNode>(cbuf->GetIndex(), - offset_inm->GetValue(), bindless_cv); - return {tracked, track}; } return {}; } @@ -120,10 +111,23 @@ std::tuple<Node, TrackSampler> ShaderIR::TrackBindlessSampler(Node tracked, cons return TrackBindlessSampler(source, code, new_cursor); } if (const auto operation = std::get_if<OperationNode>(&*tracked)) { - for (std::size_t i = operation->GetOperandsCount(); i > 0; --i) { - if (auto found = TrackBindlessSampler((*operation)[i - 1], code, cursor); - std::get<0>(found)) { - // Cbuf found in operand. + const OperationNode& op = *operation; + + const OperationCode opcode = operation->GetCode(); + if (opcode == OperationCode::IBitwiseOr || opcode == OperationCode::UBitwiseOr) { + ASSERT(op.GetOperandsCount() == 2); + auto [node_a, index_a, offset_a] = TrackCbuf(op[0], code, cursor); + auto [node_b, index_b, offset_b] = TrackCbuf(op[1], code, cursor); + if (node_a && node_b) { + auto track = MakeTrackSampler<SeparateSamplerNode>(std::pair{index_a, index_b}, + std::pair{offset_a, offset_b}); + return {tracked, std::move(track)}; + } + } + std::size_t i = op.GetOperandsCount(); + while (i--) { + if (auto found = TrackBindlessSampler(op[i - 1], code, cursor); std::get<0>(found)) { + // Constant buffer found in operand. return found; } } @@ -137,11 +141,31 @@ std::tuple<Node, TrackSampler> ShaderIR::TrackBindlessSampler(Node tracked, cons return {}; } +std::pair<Node, TrackSampler> ShaderIR::HandleBindlessIndirectRead( + const CbufNode& cbuf, const OperationNode& operation, Node gpr, Node base_offset, Node tracked, + const NodeBlock& code, s64 cursor) { + const auto offset_imm = std::get<ImmediateNode>(*base_offset); + const auto& gpu_driver = registry.AccessGuestDriverProfile(); + const u32 bindless_cv = NewCustomVariable(); + const u32 texture_handler_size = gpu_driver.GetTextureHandlerSize(); + Node op = Operation(OperationCode::UDiv, gpr, Immediate(texture_handler_size)); + + Node cv_node = GetCustomVariable(bindless_cv); + Node amend_op = Operation(OperationCode::Assign, std::move(cv_node), std::move(op)); + const std::size_t amend_index = DeclareAmend(std::move(amend_op)); + AmendNodeCv(amend_index, code[cursor]); + + // TODO: Implement bindless index custom variable + auto track = + MakeTrackSampler<ArraySamplerNode>(cbuf.GetIndex(), offset_imm.GetValue(), bindless_cv); + return {tracked, track}; +} + std::tuple<Node, u32, u32> ShaderIR::TrackCbuf(Node tracked, const NodeBlock& code, s64 cursor) const { if (const auto cbuf = std::get_if<CbufNode>(&*tracked)) { // Constant buffer found, test if it's an immediate - const auto offset = cbuf->GetOffset(); + const auto& offset = cbuf->GetOffset(); if (const auto immediate = std::get_if<ImmediateNode>(&*offset)) { return {tracked, cbuf->GetIndex(), immediate->GetValue()}; } @@ -151,21 +175,13 @@ std::tuple<Node, u32, u32> ShaderIR::TrackCbuf(Node tracked, const NodeBlock& co if (gpr->GetIndex() == Tegra::Shader::Register::ZeroIndex) { return {}; } - s64 current_cursor = cursor; - while (current_cursor > 0) { - // Reduce the cursor in one to avoid infinite loops when the instruction sets the same - // register that it uses as operand - const auto [source, new_cursor] = TrackRegister(gpr, code, current_cursor - 1); - current_cursor = new_cursor; - if (!source) { - continue; - } - const auto [base_address, index, offset] = TrackCbuf(source, code, current_cursor); - if (base_address != nullptr) { - return {base_address, index, offset}; - } + // Reduce the cursor in one to avoid infinite loops when the instruction sets the same + // register that it uses as operand + const auto [source, new_cursor] = TrackRegister(gpr, code, cursor - 1); + if (!source) { + return {}; } - return {}; + return TrackCbuf(source, code, new_cursor); } if (const auto operation = std::get_if<OperationNode>(&*tracked)) { for (std::size_t i = operation->GetOperandsCount(); i > 0; --i) { @@ -186,15 +202,15 @@ std::tuple<Node, u32, u32> ShaderIR::TrackCbuf(Node tracked, const NodeBlock& co std::optional<u32> ShaderIR::TrackImmediate(Node tracked, const NodeBlock& code, s64 cursor) const { // Reduce the cursor in one to avoid infinite loops when the instruction sets the same register // that it uses as operand - const auto [found, found_cursor] = - TrackRegister(&std::get<GprNode>(*tracked), code, cursor - 1); + const auto result = TrackRegister(&std::get<GprNode>(*tracked), code, cursor - 1); + const auto& found = result.first; if (!found) { - return {}; + return std::nullopt; } if (const auto immediate = std::get_if<ImmediateNode>(&*found)) { return immediate->GetValue(); } - return {}; + return std::nullopt; } std::pair<Node, s64> ShaderIR::TrackRegister(const GprNode* tracked, const NodeBlock& code, diff --git a/src/video_core/shader_cache.h b/src/video_core/shader_cache.h new file mode 100644 index 000000000..015a789d6 --- /dev/null +++ b/src/video_core/shader_cache.h @@ -0,0 +1,240 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <algorithm> +#include <memory> +#include <mutex> +#include <unordered_map> +#include <utility> +#include <vector> + +#include "common/assert.h" +#include "common/common_types.h" +#include "video_core/rasterizer_interface.h" + +namespace VideoCommon { + +template <class T> +class ShaderCache { + static constexpr u64 PAGE_BITS = 14; + static constexpr u64 PAGE_SIZE = u64(1) << PAGE_BITS; + + struct Entry { + VAddr addr_start; + VAddr addr_end; + T* data; + + bool is_memory_marked = true; + + constexpr bool Overlaps(VAddr start, VAddr end) const noexcept { + return start < addr_end && addr_start < end; + } + }; + +public: + virtual ~ShaderCache() = default; + + /// @brief Removes shaders inside a given region + /// @note Checks for ranges + /// @param addr Start address of the invalidation + /// @param size Number of bytes of the invalidation + void InvalidateRegion(VAddr addr, std::size_t size) { + std::scoped_lock lock{invalidation_mutex}; + InvalidatePagesInRegion(addr, size); + RemovePendingShaders(); + } + + /// @brief Unmarks a memory region as cached and marks it for removal + /// @param addr Start address of the CPU write operation + /// @param size Number of bytes of the CPU write operation + void OnCPUWrite(VAddr addr, std::size_t size) { + std::lock_guard lock{invalidation_mutex}; + InvalidatePagesInRegion(addr, size); + } + + /// @brief Flushes delayed removal operations + void SyncGuestHost() { + std::scoped_lock lock{invalidation_mutex}; + RemovePendingShaders(); + } + + /// @brief Tries to obtain a cached shader starting in a given address + /// @note Doesn't check for ranges, the given address has to be the start of the shader + /// @param addr Start address of the shader, this doesn't cache for region + /// @return Pointer to a valid shader, nullptr when nothing is found + T* TryGet(VAddr addr) const { + std::scoped_lock lock{lookup_mutex}; + + const auto it = lookup_cache.find(addr); + if (it == lookup_cache.end()) { + return nullptr; + } + return it->second->data; + } + +protected: + explicit ShaderCache(VideoCore::RasterizerInterface& rasterizer_) : rasterizer{rasterizer_} {} + + /// @brief Register in the cache a given entry + /// @param data Shader to store in the cache + /// @param addr Start address of the shader that will be registered + /// @param size Size in bytes of the shader + void Register(std::unique_ptr<T> data, VAddr addr, std::size_t size) { + std::scoped_lock lock{invalidation_mutex, lookup_mutex}; + + const VAddr addr_end = addr + size; + Entry* const entry = NewEntry(addr, addr_end, data.get()); + + const u64 page_end = (addr_end + PAGE_SIZE - 1) >> PAGE_BITS; + for (u64 page = addr >> PAGE_BITS; page < page_end; ++page) { + invalidation_cache[page].push_back(entry); + } + + storage.push_back(std::move(data)); + + rasterizer.UpdatePagesCachedCount(addr, size, 1); + } + + /// @brief Called when a shader is going to be removed + /// @param shader Shader that will be removed + /// @pre invalidation_cache is locked + /// @pre lookup_mutex is locked + virtual void OnShaderRemoval([[maybe_unused]] T* shader) {} + +private: + /// @brief Invalidate pages in a given region + /// @pre invalidation_mutex is locked + void InvalidatePagesInRegion(VAddr addr, std::size_t size) { + const VAddr addr_end = addr + size; + const u64 page_end = (addr_end + PAGE_SIZE - 1) >> PAGE_BITS; + for (u64 page = addr >> PAGE_BITS; page < page_end; ++page) { + auto it = invalidation_cache.find(page); + if (it == invalidation_cache.end()) { + continue; + } + InvalidatePageEntries(it->second, addr, addr_end); + } + } + + /// @brief Remove shaders marked for deletion + /// @pre invalidation_mutex is locked + void RemovePendingShaders() { + if (marked_for_removal.empty()) { + return; + } + // Remove duplicates + std::sort(marked_for_removal.begin(), marked_for_removal.end()); + marked_for_removal.erase(std::unique(marked_for_removal.begin(), marked_for_removal.end()), + marked_for_removal.end()); + + std::vector<T*> removed_shaders; + removed_shaders.reserve(marked_for_removal.size()); + + std::scoped_lock lock{lookup_mutex}; + + for (Entry* const entry : marked_for_removal) { + removed_shaders.push_back(entry->data); + + const auto it = lookup_cache.find(entry->addr_start); + ASSERT(it != lookup_cache.end()); + lookup_cache.erase(it); + } + marked_for_removal.clear(); + + if (!removed_shaders.empty()) { + RemoveShadersFromStorage(std::move(removed_shaders)); + } + } + + /// @brief Invalidates entries in a given range for the passed page + /// @param entries Vector of entries in the page, it will be modified on overlaps + /// @param addr Start address of the invalidation + /// @param addr_end Non-inclusive end address of the invalidation + /// @pre invalidation_mutex is locked + void InvalidatePageEntries(std::vector<Entry*>& entries, VAddr addr, VAddr addr_end) { + std::size_t index = 0; + while (index < entries.size()) { + Entry* const entry = entries[index]; + if (!entry->Overlaps(addr, addr_end)) { + ++index; + continue; + } + + UnmarkMemory(entry); + RemoveEntryFromInvalidationCache(entry); + marked_for_removal.push_back(entry); + } + } + + /// @brief Removes all references to an entry in the invalidation cache + /// @param entry Entry to remove from the invalidation cache + /// @pre invalidation_mutex is locked + void RemoveEntryFromInvalidationCache(const Entry* entry) { + const u64 page_end = (entry->addr_end + PAGE_SIZE - 1) >> PAGE_BITS; + for (u64 page = entry->addr_start >> PAGE_BITS; page < page_end; ++page) { + const auto entries_it = invalidation_cache.find(page); + ASSERT(entries_it != invalidation_cache.end()); + std::vector<Entry*>& entries = entries_it->second; + + const auto entry_it = std::find(entries.begin(), entries.end(), entry); + ASSERT(entry_it != entries.end()); + entries.erase(entry_it); + } + } + + /// @brief Unmarks an entry from the rasterizer cache + /// @param entry Entry to unmark from memory + void UnmarkMemory(Entry* entry) { + if (!entry->is_memory_marked) { + return; + } + entry->is_memory_marked = false; + + const VAddr addr = entry->addr_start; + const std::size_t size = entry->addr_end - addr; + rasterizer.UpdatePagesCachedCount(addr, size, -1); + } + + /// @brief Removes a vector of shaders from a list + /// @param removed_shaders Shaders to be removed from the storage + /// @pre invalidation_mutex is locked + /// @pre lookup_mutex is locked + void RemoveShadersFromStorage(std::vector<T*> removed_shaders) { + // Notify removals + for (T* const shader : removed_shaders) { + OnShaderRemoval(shader); + } + + // Remove them from the cache + const auto is_removed = [&removed_shaders](const std::unique_ptr<T>& shader) { + return std::find(removed_shaders.begin(), removed_shaders.end(), shader.get()) != + removed_shaders.end(); + }; + std::erase_if(storage, is_removed); + } + + /// @brief Creates a new entry in the lookup cache and returns its pointer + /// @pre lookup_mutex is locked + Entry* NewEntry(VAddr addr, VAddr addr_end, T* data) { + auto entry = std::make_unique<Entry>(Entry{addr, addr_end, data}); + Entry* const entry_pointer = entry.get(); + + lookup_cache.emplace(addr, std::move(entry)); + return entry_pointer; + } + + VideoCore::RasterizerInterface& rasterizer; + + mutable std::mutex lookup_mutex; + std::mutex invalidation_mutex; + + std::unordered_map<u64, std::unique_ptr<Entry>> lookup_cache; + std::unordered_map<u64, std::vector<Entry*>> invalidation_cache; + std::vector<std::unique_ptr<T>> storage; + std::vector<Entry*> marked_for_removal; +}; + +} // namespace VideoCommon diff --git a/src/video_core/shader_notify.cpp b/src/video_core/shader_notify.cpp new file mode 100644 index 000000000..c3c71657d --- /dev/null +++ b/src/video_core/shader_notify.cpp @@ -0,0 +1,42 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "video_core/shader_notify.h" + +using namespace std::chrono_literals; + +namespace VideoCore { +namespace { +constexpr auto UPDATE_TICK = 32ms; +} + +ShaderNotify::ShaderNotify() = default; +ShaderNotify::~ShaderNotify() = default; + +std::size_t ShaderNotify::GetShadersBuilding() { + const auto now = std::chrono::high_resolution_clock::now(); + const auto diff = now - last_update; + if (diff > UPDATE_TICK) { + std::shared_lock lock(mutex); + last_updated_count = accurate_count; + } + return last_updated_count; +} + +std::size_t ShaderNotify::GetShadersBuildingAccurate() { + std::shared_lock lock{mutex}; + return accurate_count; +} + +void ShaderNotify::MarkShaderComplete() { + std::unique_lock lock{mutex}; + accurate_count--; +} + +void ShaderNotify::MarkSharderBuilding() { + std::unique_lock lock{mutex}; + accurate_count++; +} + +} // namespace VideoCore diff --git a/src/video_core/shader_notify.h b/src/video_core/shader_notify.h new file mode 100644 index 000000000..a9c92d179 --- /dev/null +++ b/src/video_core/shader_notify.h @@ -0,0 +1,29 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <chrono> +#include <shared_mutex> +#include "common/common_types.h" + +namespace VideoCore { +class ShaderNotify { +public: + ShaderNotify(); + ~ShaderNotify(); + + std::size_t GetShadersBuilding(); + std::size_t GetShadersBuildingAccurate(); + + void MarkShaderComplete(); + void MarkSharderBuilding(); + +private: + std::size_t last_updated_count{}; + std::size_t accurate_count{}; + std::shared_mutex mutex; + std::chrono::high_resolution_clock::time_point last_update{}; +}; +} // namespace VideoCore diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index cc7181229..1688267bb 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -74,115 +74,131 @@ bool SurfaceTargetIsArray(SurfaceTarget target) { PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) { switch (format) { - case Tegra::DepthFormat::S8_Z24_UNORM: - return PixelFormat::S8Z24; - case Tegra::DepthFormat::Z24_S8_UNORM: - return PixelFormat::Z24S8; - case Tegra::DepthFormat::Z32_FLOAT: - return PixelFormat::Z32F; - case Tegra::DepthFormat::Z16_UNORM: - return PixelFormat::Z16; - case Tegra::DepthFormat::Z32_S8_X24_FLOAT: - return PixelFormat::Z32FS8; + case Tegra::DepthFormat::S8_UINT_Z24_UNORM: + return PixelFormat::S8_UINT_D24_UNORM; + case Tegra::DepthFormat::D24S8_UNORM: + return PixelFormat::D24_UNORM_S8_UINT; + case Tegra::DepthFormat::D32_FLOAT: + return PixelFormat::D32_FLOAT; + case Tegra::DepthFormat::D16_UNORM: + return PixelFormat::D16_UNORM; + case Tegra::DepthFormat::D32_FLOAT_S8X24_UINT: + return PixelFormat::D32_FLOAT_S8_UINT; default: - LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format)); - UNREACHABLE(); - return PixelFormat::S8Z24; + UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<u32>(format)); + return PixelFormat::S8_UINT_D24_UNORM; } } PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) { switch (format) { - case Tegra::RenderTargetFormat::RGBA8_SRGB: - return PixelFormat::RGBA8_SRGB; - case Tegra::RenderTargetFormat::RGBA8_UNORM: - return PixelFormat::ABGR8U; - case Tegra::RenderTargetFormat::RGBA8_SNORM: - return PixelFormat::ABGR8S; - case Tegra::RenderTargetFormat::RGBA8_UINT: - return PixelFormat::ABGR8UI; - case Tegra::RenderTargetFormat::BGRA8_SRGB: - return PixelFormat::BGRA8_SRGB; - case Tegra::RenderTargetFormat::BGRA8_UNORM: - return PixelFormat::BGRA8; - case Tegra::RenderTargetFormat::RGB10_A2_UNORM: - return PixelFormat::A2B10G10R10U; - case Tegra::RenderTargetFormat::RGBA16_FLOAT: - return PixelFormat::RGBA16F; - case Tegra::RenderTargetFormat::RGBA16_UNORM: - return PixelFormat::RGBA16U; - case Tegra::RenderTargetFormat::RGBA16_SNORM: - return PixelFormat::RGBA16S; - case Tegra::RenderTargetFormat::RGBA16_UINT: - return PixelFormat::RGBA16UI; - case Tegra::RenderTargetFormat::RGBA32_FLOAT: - return PixelFormat::RGBA32F; - case Tegra::RenderTargetFormat::RG32_FLOAT: - return PixelFormat::RG32F; - case Tegra::RenderTargetFormat::R11G11B10_FLOAT: - return PixelFormat::R11FG11FB10F; - case Tegra::RenderTargetFormat::B5G6R5_UNORM: - return PixelFormat::B5G6R5U; - case Tegra::RenderTargetFormat::BGR5A1_UNORM: - return PixelFormat::A1B5G5R5U; - case Tegra::RenderTargetFormat::RGBA32_UINT: - return PixelFormat::RGBA32UI; - case Tegra::RenderTargetFormat::R8_UNORM: - return PixelFormat::R8U; - case Tegra::RenderTargetFormat::R8_UINT: - return PixelFormat::R8UI; - case Tegra::RenderTargetFormat::RG16_FLOAT: - return PixelFormat::RG16F; - case Tegra::RenderTargetFormat::RG16_UINT: - return PixelFormat::RG16UI; - case Tegra::RenderTargetFormat::RG16_SINT: - return PixelFormat::RG16I; - case Tegra::RenderTargetFormat::RG16_UNORM: - return PixelFormat::RG16; - case Tegra::RenderTargetFormat::RG16_SNORM: - return PixelFormat::RG16S; - case Tegra::RenderTargetFormat::RG8_UNORM: - return PixelFormat::RG8U; - case Tegra::RenderTargetFormat::RG8_SNORM: - return PixelFormat::RG8S; - case Tegra::RenderTargetFormat::R16_FLOAT: - return PixelFormat::R16F; + case Tegra::RenderTargetFormat::R32B32G32A32_FLOAT: + return PixelFormat::R32G32B32A32_FLOAT; + case Tegra::RenderTargetFormat::R32G32B32A32_SINT: + return PixelFormat::R32G32B32A32_SINT; + case Tegra::RenderTargetFormat::R32G32B32A32_UINT: + return PixelFormat::R32G32B32A32_UINT; + case Tegra::RenderTargetFormat::R16G16B16A16_UNORM: + return PixelFormat::R16G16B16A16_UNORM; + case Tegra::RenderTargetFormat::R16G16B16A16_SNORM: + return PixelFormat::R16G16B16A16_SNORM; + case Tegra::RenderTargetFormat::R16G16B16A16_SINT: + return PixelFormat::R16G16B16A16_SINT; + case Tegra::RenderTargetFormat::R16G16B16A16_UINT: + return PixelFormat::R16G16B16A16_UINT; + case Tegra::RenderTargetFormat::R16G16B16A16_FLOAT: + return PixelFormat::R16G16B16A16_FLOAT; + case Tegra::RenderTargetFormat::R32G32_FLOAT: + return PixelFormat::R32G32_FLOAT; + case Tegra::RenderTargetFormat::R32G32_SINT: + return PixelFormat::R32G32_SINT; + case Tegra::RenderTargetFormat::R32G32_UINT: + return PixelFormat::R32G32_UINT; + case Tegra::RenderTargetFormat::R16G16B16X16_FLOAT: + return PixelFormat::R16G16B16X16_FLOAT; + case Tegra::RenderTargetFormat::B8G8R8A8_UNORM: + return PixelFormat::B8G8R8A8_UNORM; + case Tegra::RenderTargetFormat::B8G8R8A8_SRGB: + return PixelFormat::B8G8R8A8_SRGB; + case Tegra::RenderTargetFormat::A2B10G10R10_UNORM: + return PixelFormat::A2B10G10R10_UNORM; + case Tegra::RenderTargetFormat::A2B10G10R10_UINT: + return PixelFormat::A2B10G10R10_UINT; + case Tegra::RenderTargetFormat::A8B8G8R8_UNORM: + return PixelFormat::A8B8G8R8_UNORM; + case Tegra::RenderTargetFormat::A8B8G8R8_SRGB: + return PixelFormat::A8B8G8R8_SRGB; + case Tegra::RenderTargetFormat::A8B8G8R8_SNORM: + return PixelFormat::A8B8G8R8_SNORM; + case Tegra::RenderTargetFormat::A8B8G8R8_SINT: + return PixelFormat::A8B8G8R8_SINT; + case Tegra::RenderTargetFormat::A8B8G8R8_UINT: + return PixelFormat::A8B8G8R8_UINT; + case Tegra::RenderTargetFormat::R16G16_UNORM: + return PixelFormat::R16G16_UNORM; + case Tegra::RenderTargetFormat::R16G16_SNORM: + return PixelFormat::R16G16_SNORM; + case Tegra::RenderTargetFormat::R16G16_SINT: + return PixelFormat::R16G16_SINT; + case Tegra::RenderTargetFormat::R16G16_UINT: + return PixelFormat::R16G16_UINT; + case Tegra::RenderTargetFormat::R16G16_FLOAT: + return PixelFormat::R16G16_FLOAT; + case Tegra::RenderTargetFormat::B10G11R11_FLOAT: + return PixelFormat::B10G11R11_FLOAT; + case Tegra::RenderTargetFormat::R32_SINT: + return PixelFormat::R32_SINT; + case Tegra::RenderTargetFormat::R32_UINT: + return PixelFormat::R32_UINT; + case Tegra::RenderTargetFormat::R32_FLOAT: + return PixelFormat::R32_FLOAT; + case Tegra::RenderTargetFormat::R5G6B5_UNORM: + return PixelFormat::R5G6B5_UNORM; + case Tegra::RenderTargetFormat::A1R5G5B5_UNORM: + return PixelFormat::A1R5G5B5_UNORM; + case Tegra::RenderTargetFormat::R8G8_UNORM: + return PixelFormat::R8G8_UNORM; + case Tegra::RenderTargetFormat::R8G8_SNORM: + return PixelFormat::R8G8_SNORM; + case Tegra::RenderTargetFormat::R8G8_SINT: + return PixelFormat::R8G8_SINT; + case Tegra::RenderTargetFormat::R8G8_UINT: + return PixelFormat::R8G8_UINT; case Tegra::RenderTargetFormat::R16_UNORM: - return PixelFormat::R16U; + return PixelFormat::R16_UNORM; case Tegra::RenderTargetFormat::R16_SNORM: - return PixelFormat::R16S; - case Tegra::RenderTargetFormat::R16_UINT: - return PixelFormat::R16UI; + return PixelFormat::R16_SNORM; case Tegra::RenderTargetFormat::R16_SINT: - return PixelFormat::R16I; - case Tegra::RenderTargetFormat::R32_FLOAT: - return PixelFormat::R32F; - case Tegra::RenderTargetFormat::R32_SINT: - return PixelFormat::R32I; - case Tegra::RenderTargetFormat::R32_UINT: - return PixelFormat::R32UI; - case Tegra::RenderTargetFormat::RG32_UINT: - return PixelFormat::RG32UI; - case Tegra::RenderTargetFormat::RGBX16_FLOAT: - return PixelFormat::RGBX16F; + return PixelFormat::R16_SINT; + case Tegra::RenderTargetFormat::R16_UINT: + return PixelFormat::R16_UINT; + case Tegra::RenderTargetFormat::R16_FLOAT: + return PixelFormat::R16_FLOAT; + case Tegra::RenderTargetFormat::R8_UNORM: + return PixelFormat::R8_UNORM; + case Tegra::RenderTargetFormat::R8_SNORM: + return PixelFormat::R8_SNORM; + case Tegra::RenderTargetFormat::R8_SINT: + return PixelFormat::R8_SINT; + case Tegra::RenderTargetFormat::R8_UINT: + return PixelFormat::R8_UINT; default: - LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format)); - UNREACHABLE(); - return PixelFormat::RGBA8_SRGB; + UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<int>(format)); + return PixelFormat::A8B8G8R8_UNORM; } } PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) { switch (format) { - case Tegra::FramebufferConfig::PixelFormat::ABGR8: - return PixelFormat::ABGR8U; - case Tegra::FramebufferConfig::PixelFormat::RGB565: - return PixelFormat::B5G6R5U; - case Tegra::FramebufferConfig::PixelFormat::BGRA8: - return PixelFormat::BGRA8; + case Tegra::FramebufferConfig::PixelFormat::A8B8G8R8_UNORM: + return PixelFormat::A8B8G8R8_UNORM; + case Tegra::FramebufferConfig::PixelFormat::RGB565_UNORM: + return PixelFormat::R5G6B5_UNORM; + case Tegra::FramebufferConfig::PixelFormat::B8G8R8A8_UNORM: + return PixelFormat::B8G8R8A8_UNORM; default: UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<u32>(format)); - return PixelFormat::ABGR8U; + return PixelFormat::A8B8G8R8_UNORM; } } @@ -210,27 +226,27 @@ SurfaceType GetFormatType(PixelFormat pixel_format) { bool IsPixelFormatASTC(PixelFormat format) { switch (format) { - case PixelFormat::ASTC_2D_4X4: - case PixelFormat::ASTC_2D_5X4: - case PixelFormat::ASTC_2D_5X5: - case PixelFormat::ASTC_2D_8X8: - case PixelFormat::ASTC_2D_8X5: + case PixelFormat::ASTC_2D_4X4_UNORM: + case PixelFormat::ASTC_2D_5X4_UNORM: + case PixelFormat::ASTC_2D_5X5_UNORM: + case PixelFormat::ASTC_2D_8X8_UNORM: + case PixelFormat::ASTC_2D_8X5_UNORM: case PixelFormat::ASTC_2D_4X4_SRGB: case PixelFormat::ASTC_2D_5X4_SRGB: case PixelFormat::ASTC_2D_5X5_SRGB: case PixelFormat::ASTC_2D_8X8_SRGB: case PixelFormat::ASTC_2D_8X5_SRGB: - case PixelFormat::ASTC_2D_10X8: + case PixelFormat::ASTC_2D_10X8_UNORM: case PixelFormat::ASTC_2D_10X8_SRGB: - case PixelFormat::ASTC_2D_6X6: + case PixelFormat::ASTC_2D_6X6_UNORM: case PixelFormat::ASTC_2D_6X6_SRGB: - case PixelFormat::ASTC_2D_10X10: + case PixelFormat::ASTC_2D_10X10_UNORM: case PixelFormat::ASTC_2D_10X10_SRGB: - case PixelFormat::ASTC_2D_12X12: + case PixelFormat::ASTC_2D_12X12_UNORM: case PixelFormat::ASTC_2D_12X12_SRGB: - case PixelFormat::ASTC_2D_8X6: + case PixelFormat::ASTC_2D_8X6_UNORM: case PixelFormat::ASTC_2D_8X6_SRGB: - case PixelFormat::ASTC_2D_6X5: + case PixelFormat::ASTC_2D_6X5_UNORM: case PixelFormat::ASTC_2D_6X5_SRGB: return true; default: @@ -240,12 +256,12 @@ bool IsPixelFormatASTC(PixelFormat format) { bool IsPixelFormatSRGB(PixelFormat format) { switch (format) { - case PixelFormat::RGBA8_SRGB: - case PixelFormat::BGRA8_SRGB: - case PixelFormat::DXT1_SRGB: - case PixelFormat::DXT23_SRGB: - case PixelFormat::DXT45_SRGB: - case PixelFormat::BC7U_SRGB: + case PixelFormat::A8B8G8R8_SRGB: + case PixelFormat::B8G8R8A8_SRGB: + case PixelFormat::BC1_RGBA_SRGB: + case PixelFormat::BC2_SRGB: + case PixelFormat::BC3_SRGB: + case PixelFormat::BC7_SRGB: case PixelFormat::ASTC_2D_4X4_SRGB: case PixelFormat::ASTC_2D_8X8_SRGB: case PixelFormat::ASTC_2D_8X5_SRGB: @@ -267,25 +283,4 @@ std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) { return {GetDefaultBlockWidth(format), GetDefaultBlockHeight(format)}; } -bool IsFormatBCn(PixelFormat format) { - switch (format) { - case PixelFormat::DXT1: - case PixelFormat::DXT23: - case PixelFormat::DXT45: - case PixelFormat::DXN1: - case PixelFormat::DXN2SNORM: - case PixelFormat::DXN2UNORM: - case PixelFormat::BC7U: - case PixelFormat::BC6H_UF16: - case PixelFormat::BC6H_SF16: - case PixelFormat::DXT1_SRGB: - case PixelFormat::DXT23_SRGB: - case PixelFormat::DXT45_SRGB: - case PixelFormat::BC7U_SRGB: - return true; - default: - return false; - } -} - } // namespace VideoCore::Surface diff --git a/src/video_core/surface.h b/src/video_core/surface.h index e0acd44d3..cfd12fa61 100644 --- a/src/video_core/surface.h +++ b/src/video_core/surface.h @@ -15,93 +15,105 @@ namespace VideoCore::Surface { enum class PixelFormat { - ABGR8U = 0, - ABGR8S = 1, - ABGR8UI = 2, - B5G6R5U = 3, - A2B10G10R10U = 4, - A1B5G5R5U = 5, - R8U = 6, - R8UI = 7, - RGBA16F = 8, - RGBA16U = 9, - RGBA16S = 10, - RGBA16UI = 11, - R11FG11FB10F = 12, - RGBA32UI = 13, - DXT1 = 14, - DXT23 = 15, - DXT45 = 16, - DXN1 = 17, // This is also known as BC4 - DXN2UNORM = 18, - DXN2SNORM = 19, - BC7U = 20, - BC6H_UF16 = 21, - BC6H_SF16 = 22, - ASTC_2D_4X4 = 23, - BGRA8 = 24, - RGBA32F = 25, - RG32F = 26, - R32F = 27, - R16F = 28, - R16U = 29, - R16S = 30, - R16UI = 31, - R16I = 32, - RG16 = 33, - RG16F = 34, - RG16UI = 35, - RG16I = 36, - RG16S = 37, - RGB32F = 38, - RGBA8_SRGB = 39, - RG8U = 40, - RG8S = 41, - RG32UI = 42, - RGBX16F = 43, - R32UI = 44, - R32I = 45, - ASTC_2D_8X8 = 46, - ASTC_2D_8X5 = 47, - ASTC_2D_5X4 = 48, - BGRA8_SRGB = 49, - DXT1_SRGB = 50, - DXT23_SRGB = 51, - DXT45_SRGB = 52, - BC7U_SRGB = 53, - R4G4B4A4U = 54, - ASTC_2D_4X4_SRGB = 55, - ASTC_2D_8X8_SRGB = 56, - ASTC_2D_8X5_SRGB = 57, - ASTC_2D_5X4_SRGB = 58, - ASTC_2D_5X5 = 59, - ASTC_2D_5X5_SRGB = 60, - ASTC_2D_10X8 = 61, - ASTC_2D_10X8_SRGB = 62, - ASTC_2D_6X6 = 63, - ASTC_2D_6X6_SRGB = 64, - ASTC_2D_10X10 = 65, - ASTC_2D_10X10_SRGB = 66, - ASTC_2D_12X12 = 67, - ASTC_2D_12X12_SRGB = 68, - ASTC_2D_8X6 = 69, - ASTC_2D_8X6_SRGB = 70, - ASTC_2D_6X5 = 71, - ASTC_2D_6X5_SRGB = 72, - E5B9G9R9F = 73, + A8B8G8R8_UNORM, + A8B8G8R8_SNORM, + A8B8G8R8_SINT, + A8B8G8R8_UINT, + R5G6B5_UNORM, + B5G6R5_UNORM, + A1R5G5B5_UNORM, + A2B10G10R10_UNORM, + A2B10G10R10_UINT, + A1B5G5R5_UNORM, + R8_UNORM, + R8_SNORM, + R8_SINT, + R8_UINT, + R16G16B16A16_FLOAT, + R16G16B16A16_UNORM, + R16G16B16A16_SNORM, + R16G16B16A16_SINT, + R16G16B16A16_UINT, + B10G11R11_FLOAT, + R32G32B32A32_UINT, + BC1_RGBA_UNORM, + BC2_UNORM, + BC3_UNORM, + BC4_UNORM, + BC4_SNORM, + BC5_UNORM, + BC5_SNORM, + BC7_UNORM, + BC6H_UFLOAT, + BC6H_SFLOAT, + ASTC_2D_4X4_UNORM, + B8G8R8A8_UNORM, + R32G32B32A32_FLOAT, + R32G32B32A32_SINT, + R32G32_FLOAT, + R32G32_SINT, + R32_FLOAT, + R16_FLOAT, + R16_UNORM, + R16_SNORM, + R16_UINT, + R16_SINT, + R16G16_UNORM, + R16G16_FLOAT, + R16G16_UINT, + R16G16_SINT, + R16G16_SNORM, + R32G32B32_FLOAT, + A8B8G8R8_SRGB, + R8G8_UNORM, + R8G8_SNORM, + R8G8_SINT, + R8G8_UINT, + R32G32_UINT, + R16G16B16X16_FLOAT, + R32_UINT, + R32_SINT, + ASTC_2D_8X8_UNORM, + ASTC_2D_8X5_UNORM, + ASTC_2D_5X4_UNORM, + B8G8R8A8_SRGB, + BC1_RGBA_SRGB, + BC2_SRGB, + BC3_SRGB, + BC7_SRGB, + A4B4G4R4_UNORM, + ASTC_2D_4X4_SRGB, + ASTC_2D_8X8_SRGB, + ASTC_2D_8X5_SRGB, + ASTC_2D_5X4_SRGB, + ASTC_2D_5X5_UNORM, + ASTC_2D_5X5_SRGB, + ASTC_2D_10X8_UNORM, + ASTC_2D_10X8_SRGB, + ASTC_2D_6X6_UNORM, + ASTC_2D_6X6_SRGB, + ASTC_2D_10X10_UNORM, + ASTC_2D_10X10_SRGB, + ASTC_2D_12X12_UNORM, + ASTC_2D_12X12_SRGB, + ASTC_2D_8X6_UNORM, + ASTC_2D_8X6_SRGB, + ASTC_2D_6X5_UNORM, + ASTC_2D_6X5_SRGB, + E5B9G9R9_FLOAT, MaxColorFormat, // Depth formats - Z32F = 74, - Z16 = 75, + D32_FLOAT = MaxColorFormat, + D16_UNORM, MaxDepthFormat, // DepthStencil formats - Z24S8 = 76, - S8Z24 = 77, - Z32FS8 = 78, + D24_UNORM_S8_UINT = MaxDepthFormat, + S8_UINT_D24_UNORM, + D32_FLOAT_S8_UINT, MaxDepthStencilFormat, @@ -129,85 +141,97 @@ enum class SurfaceTarget { }; constexpr std::array<u32, MaxPixelFormat> compression_factor_shift_table = {{ - 0, // ABGR8U - 0, // ABGR8S - 0, // ABGR8UI - 0, // B5G6R5U - 0, // A2B10G10R10U - 0, // A1B5G5R5U - 0, // R8U - 0, // R8UI - 0, // RGBA16F - 0, // RGBA16U - 0, // RGBA16S - 0, // RGBA16UI - 0, // R11FG11FB10F - 0, // RGBA32UI - 2, // DXT1 - 2, // DXT23 - 2, // DXT45 - 2, // DXN1 - 2, // DXN2UNORM - 2, // DXN2SNORM - 2, // BC7U - 2, // BC6H_UF16 - 2, // BC6H_SF16 - 2, // ASTC_2D_4X4 - 0, // BGRA8 - 0, // RGBA32F - 0, // RG32F - 0, // R32F - 0, // R16F - 0, // R16U - 0, // R16S - 0, // R16UI - 0, // R16I - 0, // RG16 - 0, // RG16F - 0, // RG16UI - 0, // RG16I - 0, // RG16S - 0, // RGB32F - 0, // RGBA8_SRGB - 0, // RG8U - 0, // RG8S - 0, // RG32UI - 0, // RGBX16F - 0, // R32UI - 0, // R32I - 2, // ASTC_2D_8X8 - 2, // ASTC_2D_8X5 - 2, // ASTC_2D_5X4 - 0, // BGRA8_SRGB - 2, // DXT1_SRGB - 2, // DXT23_SRGB - 2, // DXT45_SRGB - 2, // BC7U_SRGB - 0, // R4G4B4A4U + 0, // A8B8G8R8_UNORM + 0, // A8B8G8R8_SNORM + 0, // A8B8G8R8_SINT + 0, // A8B8G8R8_UINT + 0, // R5G6B5_UNORM + 0, // B5G6R5_UNORM + 0, // A1R5G5B5_UNORM + 0, // A2B10G10R10_UNORM + 0, // A2B10G10R10_UINT + 0, // A1B5G5R5_UNORM + 0, // R8_UNORM + 0, // R8_SNORM + 0, // R8_SINT + 0, // R8_UINT + 0, // R16G16B16A16_FLOAT + 0, // R16G16B16A16_UNORM + 0, // R16G16B16A16_SNORM + 0, // R16G16B16A16_SINT + 0, // R16G16B16A16_UINT + 0, // B10G11R11_FLOAT + 0, // R32G32B32A32_UINT + 2, // BC1_RGBA_UNORM + 2, // BC2_UNORM + 2, // BC3_UNORM + 2, // BC4_UNORM + 2, // BC4_SNORM + 2, // BC5_UNORM + 2, // BC5_SNORM + 2, // BC7_UNORM + 2, // BC6H_UFLOAT + 2, // BC6H_SFLOAT + 2, // ASTC_2D_4X4_UNORM + 0, // B8G8R8A8_UNORM + 0, // R32G32B32A32_FLOAT + 0, // R32G32B32A32_SINT + 0, // R32G32_FLOAT + 0, // R32G32_SINT + 0, // R32_FLOAT + 0, // R16_FLOAT + 0, // R16_UNORM + 0, // R16_SNORM + 0, // R16_UINT + 0, // R16_SINT + 0, // R16G16_UNORM + 0, // R16G16_FLOAT + 0, // R16G16_UINT + 0, // R16G16_SINT + 0, // R16G16_SNORM + 0, // R32G32B32_FLOAT + 0, // A8B8G8R8_SRGB + 0, // R8G8_UNORM + 0, // R8G8_SNORM + 0, // R8G8_SINT + 0, // R8G8_UINT + 0, // R32G32_UINT + 0, // R16G16B16X16_FLOAT + 0, // R32_UINT + 0, // R32_SINT + 2, // ASTC_2D_8X8_UNORM + 2, // ASTC_2D_8X5_UNORM + 2, // ASTC_2D_5X4_UNORM + 0, // B8G8R8A8_SRGB + 2, // BC1_RGBA_SRGB + 2, // BC2_SRGB + 2, // BC3_SRGB + 2, // BC7_SRGB + 0, // A4B4G4R4_UNORM 2, // ASTC_2D_4X4_SRGB 2, // ASTC_2D_8X8_SRGB 2, // ASTC_2D_8X5_SRGB 2, // ASTC_2D_5X4_SRGB - 2, // ASTC_2D_5X5 + 2, // ASTC_2D_5X5_UNORM 2, // ASTC_2D_5X5_SRGB - 2, // ASTC_2D_10X8 + 2, // ASTC_2D_10X8_UNORM 2, // ASTC_2D_10X8_SRGB - 2, // ASTC_2D_6X6 + 2, // ASTC_2D_6X6_UNORM 2, // ASTC_2D_6X6_SRGB - 2, // ASTC_2D_10X10 + 2, // ASTC_2D_10X10_UNORM 2, // ASTC_2D_10X10_SRGB - 2, // ASTC_2D_12X12 + 2, // ASTC_2D_12X12_UNORM 2, // ASTC_2D_12X12_SRGB - 2, // ASTC_2D_8X6 + 2, // ASTC_2D_8X6_UNORM 2, // ASTC_2D_8X6_SRGB - 2, // ASTC_2D_6X5 + 2, // ASTC_2D_6X5_UNORM 2, // ASTC_2D_6X5_SRGB - 0, // E5B9G9R9F - 0, // Z32F - 0, // Z16 - 0, // Z24S8 - 0, // S8Z24 - 0, // Z32FS8 + 0, // E5B9G9R9_FLOAT + 0, // D32_FLOAT + 0, // D16_UNORM + 0, // D24_UNORM_S8_UINT + 0, // S8_UINT_D24_UNORM + 0, // D32_FLOAT_S8_UINT }}; /** @@ -227,85 +251,97 @@ inline constexpr u32 GetCompressionFactor(PixelFormat format) { } constexpr std::array<u32, MaxPixelFormat> block_width_table = {{ - 1, // ABGR8U - 1, // ABGR8S - 1, // ABGR8UI - 1, // B5G6R5U - 1, // A2B10G10R10U - 1, // A1B5G5R5U - 1, // R8U - 1, // R8UI - 1, // RGBA16F - 1, // RGBA16U - 1, // RGBA16S - 1, // RGBA16UI - 1, // R11FG11FB10F - 1, // RGBA32UI - 4, // DXT1 - 4, // DXT23 - 4, // DXT45 - 4, // DXN1 - 4, // DXN2UNORM - 4, // DXN2SNORM - 4, // BC7U - 4, // BC6H_UF16 - 4, // BC6H_SF16 - 4, // ASTC_2D_4X4 - 1, // BGRA8 - 1, // RGBA32F - 1, // RG32F - 1, // R32F - 1, // R16F - 1, // R16U - 1, // R16S - 1, // R16UI - 1, // R16I - 1, // RG16 - 1, // RG16F - 1, // RG16UI - 1, // RG16I - 1, // RG16S - 1, // RGB32F - 1, // RGBA8_SRGB - 1, // RG8U - 1, // RG8S - 1, // RG32UI - 1, // RGBX16F - 1, // R32UI - 1, // R32I - 8, // ASTC_2D_8X8 - 8, // ASTC_2D_8X5 - 5, // ASTC_2D_5X4 - 1, // BGRA8_SRGB - 4, // DXT1_SRGB - 4, // DXT23_SRGB - 4, // DXT45_SRGB - 4, // BC7U_SRGB - 1, // R4G4B4A4U + 1, // A8B8G8R8_UNORM + 1, // A8B8G8R8_SNORM + 1, // A8B8G8R8_SINT + 1, // A8B8G8R8_UINT + 1, // R5G6B5_UNORM + 1, // B5G6R5_UNORM + 1, // A1R5G5B5_UNORM + 1, // A2B10G10R10_UNORM + 1, // A2B10G10R10_UINT + 1, // A1B5G5R5_UNORM + 1, // R8_UNORM + 1, // R8_SNORM + 1, // R8_SINT + 1, // R8_UINT + 1, // R16G16B16A16_FLOAT + 1, // R16G16B16A16_UNORM + 1, // R16G16B16A16_SNORM + 1, // R16G16B16A16_SINT + 1, // R16G16B16A16_UINT + 1, // B10G11R11_FLOAT + 1, // R32G32B32A32_UINT + 4, // BC1_RGBA_UNORM + 4, // BC2_UNORM + 4, // BC3_UNORM + 4, // BC4_UNORM + 4, // BC4_SNORM + 4, // BC5_UNORM + 4, // BC5_SNORM + 4, // BC7_UNORM + 4, // BC6H_UFLOAT + 4, // BC6H_SFLOAT + 4, // ASTC_2D_4X4_UNORM + 1, // B8G8R8A8_UNORM + 1, // R32G32B32A32_FLOAT + 1, // R32G32B32A32_SINT + 1, // R32G32_FLOAT + 1, // R32G32_SINT + 1, // R32_FLOAT + 1, // R16_FLOAT + 1, // R16_UNORM + 1, // R16_SNORM + 1, // R16_UINT + 1, // R16_SINT + 1, // R16G16_UNORM + 1, // R16G16_FLOAT + 1, // R16G16_UINT + 1, // R16G16_SINT + 1, // R16G16_SNORM + 1, // R32G32B32_FLOAT + 1, // A8B8G8R8_SRGB + 1, // R8G8_UNORM + 1, // R8G8_SNORM + 1, // R8G8_SINT + 1, // R8G8_UINT + 1, // R32G32_UINT + 1, // R16G16B16X16_FLOAT + 1, // R32_UINT + 1, // R32_SINT + 8, // ASTC_2D_8X8_UNORM + 8, // ASTC_2D_8X5_UNORM + 5, // ASTC_2D_5X4_UNORM + 1, // B8G8R8A8_SRGB + 4, // BC1_RGBA_SRGB + 4, // BC2_SRGB + 4, // BC3_SRGB + 4, // BC7_SRGB + 1, // A4B4G4R4_UNORM 4, // ASTC_2D_4X4_SRGB 8, // ASTC_2D_8X8_SRGB 8, // ASTC_2D_8X5_SRGB 5, // ASTC_2D_5X4_SRGB - 5, // ASTC_2D_5X5 + 5, // ASTC_2D_5X5_UNORM 5, // ASTC_2D_5X5_SRGB - 10, // ASTC_2D_10X8 + 10, // ASTC_2D_10X8_UNORM 10, // ASTC_2D_10X8_SRGB - 6, // ASTC_2D_6X6 + 6, // ASTC_2D_6X6_UNORM 6, // ASTC_2D_6X6_SRGB - 10, // ASTC_2D_10X10 + 10, // ASTC_2D_10X10_UNORM 10, // ASTC_2D_10X10_SRGB - 12, // ASTC_2D_12X12 + 12, // ASTC_2D_12X12_UNORM 12, // ASTC_2D_12X12_SRGB - 8, // ASTC_2D_8X6 + 8, // ASTC_2D_8X6_UNORM 8, // ASTC_2D_8X6_SRGB - 6, // ASTC_2D_6X5 + 6, // ASTC_2D_6X5_UNORM 6, // ASTC_2D_6X5_SRGB - 1, // E5B9G9R9F - 1, // Z32F - 1, // Z16 - 1, // Z24S8 - 1, // S8Z24 - 1, // Z32FS8 + 1, // E5B9G9R9_FLOAT + 1, // D32_FLOAT + 1, // D16_UNORM + 1, // D24_UNORM_S8_UINT + 1, // S8_UINT_D24_UNORM + 1, // D32_FLOAT_S8_UINT }}; static constexpr u32 GetDefaultBlockWidth(PixelFormat format) { @@ -317,85 +353,97 @@ static constexpr u32 GetDefaultBlockWidth(PixelFormat format) { } constexpr std::array<u32, MaxPixelFormat> block_height_table = {{ - 1, // ABGR8U - 1, // ABGR8S - 1, // ABGR8UI - 1, // B5G6R5U - 1, // A2B10G10R10U - 1, // A1B5G5R5U - 1, // R8U - 1, // R8UI - 1, // RGBA16F - 1, // RGBA16U - 1, // RGBA16S - 1, // RGBA16UI - 1, // R11FG11FB10F - 1, // RGBA32UI - 4, // DXT1 - 4, // DXT23 - 4, // DXT45 - 4, // DXN1 - 4, // DXN2UNORM - 4, // DXN2SNORM - 4, // BC7U - 4, // BC6H_UF16 - 4, // BC6H_SF16 - 4, // ASTC_2D_4X4 - 1, // BGRA8 - 1, // RGBA32F - 1, // RG32F - 1, // R32F - 1, // R16F - 1, // R16U - 1, // R16S - 1, // R16UI - 1, // R16I - 1, // RG16 - 1, // RG16F - 1, // RG16UI - 1, // RG16I - 1, // RG16S - 1, // RGB32F - 1, // RGBA8_SRGB - 1, // RG8U - 1, // RG8S - 1, // RG32UI - 1, // RGBX16F - 1, // R32UI - 1, // R32I - 8, // ASTC_2D_8X8 - 5, // ASTC_2D_8X5 - 4, // ASTC_2D_5X4 - 1, // BGRA8_SRGB - 4, // DXT1_SRGB - 4, // DXT23_SRGB - 4, // DXT45_SRGB - 4, // BC7U_SRGB - 1, // R4G4B4A4U + 1, // A8B8G8R8_UNORM + 1, // A8B8G8R8_SNORM + 1, // A8B8G8R8_SINT + 1, // A8B8G8R8_UINT + 1, // R5G6B5_UNORM + 1, // B5G6R5_UNORM + 1, // A1R5G5B5_UNORM + 1, // A2B10G10R10_UNORM + 1, // A2B10G10R10_UINT + 1, // A1B5G5R5_UNORM + 1, // R8_UNORM + 1, // R8_SNORM + 1, // R8_SINT + 1, // R8_UINT + 1, // R16G16B16A16_FLOAT + 1, // R16G16B16A16_UNORM + 1, // R16G16B16A16_SNORM + 1, // R16G16B16A16_SINT + 1, // R16G16B16A16_UINT + 1, // B10G11R11_FLOAT + 1, // R32G32B32A32_UINT + 4, // BC1_RGBA_UNORM + 4, // BC2_UNORM + 4, // BC3_UNORM + 4, // BC4_UNORM + 4, // BC4_SNORM + 4, // BC5_UNORM + 4, // BC5_SNORM + 4, // BC7_UNORM + 4, // BC6H_UFLOAT + 4, // BC6H_SFLOAT + 4, // ASTC_2D_4X4_UNORM + 1, // B8G8R8A8_UNORM + 1, // R32G32B32A32_FLOAT + 1, // R32G32B32A32_SINT + 1, // R32G32_FLOAT + 1, // R32G32_SINT + 1, // R32_FLOAT + 1, // R16_FLOAT + 1, // R16_UNORM + 1, // R16_SNORM + 1, // R16_UINT + 1, // R16_SINT + 1, // R16G16_UNORM + 1, // R16G16_FLOAT + 1, // R16G16_UINT + 1, // R16G16_SINT + 1, // R16G16_SNORM + 1, // R32G32B32_FLOAT + 1, // A8B8G8R8_SRGB + 1, // R8G8_UNORM + 1, // R8G8_SNORM + 1, // R8G8_SINT + 1, // R8G8_UINT + 1, // R32G32_UINT + 1, // R16G16B16X16_FLOAT + 1, // R32_UINT + 1, // R32_SINT + 8, // ASTC_2D_8X8_UNORM + 5, // ASTC_2D_8X5_UNORM + 4, // ASTC_2D_5X4_UNORM + 1, // B8G8R8A8_SRGB + 4, // BC1_RGBA_SRGB + 4, // BC2_SRGB + 4, // BC3_SRGB + 4, // BC7_SRGB + 1, // A4B4G4R4_UNORM 4, // ASTC_2D_4X4_SRGB 8, // ASTC_2D_8X8_SRGB 5, // ASTC_2D_8X5_SRGB 4, // ASTC_2D_5X4_SRGB - 5, // ASTC_2D_5X5 + 5, // ASTC_2D_5X5_UNORM 5, // ASTC_2D_5X5_SRGB - 8, // ASTC_2D_10X8 + 8, // ASTC_2D_10X8_UNORM 8, // ASTC_2D_10X8_SRGB - 6, // ASTC_2D_6X6 + 6, // ASTC_2D_6X6_UNORM 6, // ASTC_2D_6X6_SRGB - 10, // ASTC_2D_10X10 + 10, // ASTC_2D_10X10_UNORM 10, // ASTC_2D_10X10_SRGB - 12, // ASTC_2D_12X12 + 12, // ASTC_2D_12X12_UNORM 12, // ASTC_2D_12X12_SRGB - 6, // ASTC_2D_8X6 + 6, // ASTC_2D_8X6_UNORM 6, // ASTC_2D_8X6_SRGB - 5, // ASTC_2D_6X5 + 5, // ASTC_2D_6X5_UNORM 5, // ASTC_2D_6X5_SRGB - 1, // E5B9G9R9F - 1, // Z32F - 1, // Z16 - 1, // Z24S8 - 1, // S8Z24 - 1, // Z32FS8 + 1, // E5B9G9R9_FLOAT + 1, // D32_FLOAT + 1, // D16_UNORM + 1, // D24_UNORM_S8_UINT + 1, // S8_UINT_D24_UNORM + 1, // D32_FLOAT_S8_UINT }}; static constexpr u32 GetDefaultBlockHeight(PixelFormat format) { @@ -407,85 +455,97 @@ static constexpr u32 GetDefaultBlockHeight(PixelFormat format) { } constexpr std::array<u32, MaxPixelFormat> bpp_table = {{ - 32, // ABGR8U - 32, // ABGR8S - 32, // ABGR8UI - 16, // B5G6R5U - 32, // A2B10G10R10U - 16, // A1B5G5R5U - 8, // R8U - 8, // R8UI - 64, // RGBA16F - 64, // RGBA16U - 64, // RGBA16S - 64, // RGBA16UI - 32, // R11FG11FB10F - 128, // RGBA32UI - 64, // DXT1 - 128, // DXT23 - 128, // DXT45 - 64, // DXN1 - 128, // DXN2UNORM - 128, // DXN2SNORM - 128, // BC7U - 128, // BC6H_UF16 - 128, // BC6H_SF16 - 128, // ASTC_2D_4X4 - 32, // BGRA8 - 128, // RGBA32F - 64, // RG32F - 32, // R32F - 16, // R16F - 16, // R16U - 16, // R16S - 16, // R16UI - 16, // R16I - 32, // RG16 - 32, // RG16F - 32, // RG16UI - 32, // RG16I - 32, // RG16S - 96, // RGB32F - 32, // RGBA8_SRGB - 16, // RG8U - 16, // RG8S - 64, // RG32UI - 64, // RGBX16F - 32, // R32UI - 32, // R32I - 128, // ASTC_2D_8X8 - 128, // ASTC_2D_8X5 - 128, // ASTC_2D_5X4 - 32, // BGRA8_SRGB - 64, // DXT1_SRGB - 128, // DXT23_SRGB - 128, // DXT45_SRGB - 128, // BC7U - 16, // R4G4B4A4U + 32, // A8B8G8R8_UNORM + 32, // A8B8G8R8_SNORM + 32, // A8B8G8R8_SINT + 32, // A8B8G8R8_UINT + 16, // R5G6B5_UNORM + 16, // B5G6R5_UNORM + 16, // A1R5G5B5_UNORM + 32, // A2B10G10R10_UNORM + 32, // A2B10G10R10_UINT + 16, // A1B5G5R5_UNORM + 8, // R8_UNORM + 8, // R8_SNORM + 8, // R8_SINT + 8, // R8_UINT + 64, // R16G16B16A16_FLOAT + 64, // R16G16B16A16_UNORM + 64, // R16G16B16A16_SNORM + 64, // R16G16B16A16_SINT + 64, // R16G16B16A16_UINT + 32, // B10G11R11_FLOAT + 128, // R32G32B32A32_UINT + 64, // BC1_RGBA_UNORM + 128, // BC2_UNORM + 128, // BC3_UNORM + 64, // BC4_UNORM + 64, // BC4_SNORM + 128, // BC5_UNORM + 128, // BC5_SNORM + 128, // BC7_UNORM + 128, // BC6H_UFLOAT + 128, // BC6H_SFLOAT + 128, // ASTC_2D_4X4_UNORM + 32, // B8G8R8A8_UNORM + 128, // R32G32B32A32_FLOAT + 128, // R32G32B32A32_SINT + 64, // R32G32_FLOAT + 64, // R32G32_SINT + 32, // R32_FLOAT + 16, // R16_FLOAT + 16, // R16_UNORM + 16, // R16_SNORM + 16, // R16_UINT + 16, // R16_SINT + 32, // R16G16_UNORM + 32, // R16G16_FLOAT + 32, // R16G16_UINT + 32, // R16G16_SINT + 32, // R16G16_SNORM + 96, // R32G32B32_FLOAT + 32, // A8B8G8R8_SRGB + 16, // R8G8_UNORM + 16, // R8G8_SNORM + 16, // R8G8_SINT + 16, // R8G8_UINT + 64, // R32G32_UINT + 64, // R16G16B16X16_FLOAT + 32, // R32_UINT + 32, // R32_SINT + 128, // ASTC_2D_8X8_UNORM + 128, // ASTC_2D_8X5_UNORM + 128, // ASTC_2D_5X4_UNORM + 32, // B8G8R8A8_SRGB + 64, // BC1_RGBA_SRGB + 128, // BC2_SRGB + 128, // BC3_SRGB + 128, // BC7_UNORM + 16, // A4B4G4R4_UNORM 128, // ASTC_2D_4X4_SRGB 128, // ASTC_2D_8X8_SRGB 128, // ASTC_2D_8X5_SRGB 128, // ASTC_2D_5X4_SRGB - 128, // ASTC_2D_5X5 + 128, // ASTC_2D_5X5_UNORM 128, // ASTC_2D_5X5_SRGB - 128, // ASTC_2D_10X8 + 128, // ASTC_2D_10X8_UNORM 128, // ASTC_2D_10X8_SRGB - 128, // ASTC_2D_6X6 + 128, // ASTC_2D_6X6_UNORM 128, // ASTC_2D_6X6_SRGB - 128, // ASTC_2D_10X10 + 128, // ASTC_2D_10X10_UNORM 128, // ASTC_2D_10X10_SRGB - 128, // ASTC_2D_12X12 + 128, // ASTC_2D_12X12_UNORM 128, // ASTC_2D_12X12_SRGB - 128, // ASTC_2D_8X6 + 128, // ASTC_2D_8X6_UNORM 128, // ASTC_2D_8X6_SRGB - 128, // ASTC_2D_6X5 + 128, // ASTC_2D_6X5_UNORM 128, // ASTC_2D_6X5_SRGB - 32, // E5B9G9R9F - 32, // Z32F - 16, // Z16 - 32, // Z24S8 - 32, // S8Z24 - 64, // Z32FS8 + 32, // E5B9G9R9_FLOAT + 32, // D32_FLOAT + 16, // D16_UNORM + 32, // D24_UNORM_S8_UINT + 32, // S8_UINT_D24_UNORM + 64, // D32_FLOAT_S8_UINT }}; static constexpr u32 GetFormatBpp(PixelFormat format) { @@ -524,7 +584,4 @@ bool IsPixelFormatSRGB(PixelFormat format); std::pair<u32, u32> GetASTCBlockSize(PixelFormat format); -/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN -bool IsFormatBCn(PixelFormat format); - } // namespace VideoCore::Surface diff --git a/src/video_core/texture_cache/format_lookup_table.cpp b/src/video_core/texture_cache/format_lookup_table.cpp index e151c26c4..7d5a75648 100644 --- a/src/video_core/texture_cache/format_lookup_table.cpp +++ b/src/video_core/texture_cache/format_lookup_table.cpp @@ -19,8 +19,6 @@ constexpr auto SNORM = ComponentType::SNORM; constexpr auto UNORM = ComponentType::UNORM; constexpr auto SINT = ComponentType::SINT; constexpr auto UINT = ComponentType::UINT; -constexpr auto SNORM_FORCE_FP16 = ComponentType::SNORM_FORCE_FP16; -constexpr auto UNORM_FORCE_FP16 = ComponentType::UNORM_FORCE_FP16; constexpr auto FLOAT = ComponentType::FLOAT; constexpr bool C = false; // Normal color constexpr bool S = true; // Srgb @@ -41,117 +39,126 @@ struct Table { ComponentType alpha_component; bool is_srgb; }; -constexpr std::array<Table, 76> DefinitionTable = {{ - {TextureFormat::A8R8G8B8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ABGR8U}, - {TextureFormat::A8R8G8B8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::ABGR8S}, - {TextureFormat::A8R8G8B8, C, UINT, UINT, UINT, UINT, PixelFormat::ABGR8UI}, - {TextureFormat::A8R8G8B8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::RGBA8_SRGB}, +constexpr std::array<Table, 86> DefinitionTable = {{ + {TextureFormat::A8R8G8B8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A8B8G8R8_UNORM}, + {TextureFormat::A8R8G8B8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::A8B8G8R8_SNORM}, + {TextureFormat::A8R8G8B8, C, UINT, UINT, UINT, UINT, PixelFormat::A8B8G8R8_UINT}, + {TextureFormat::A8R8G8B8, C, SINT, SINT, SINT, SINT, PixelFormat::A8B8G8R8_SINT}, + {TextureFormat::A8R8G8B8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::A8B8G8R8_SRGB}, - {TextureFormat::B5G6R5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::B5G6R5U}, + {TextureFormat::B5G6R5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::B5G6R5_UNORM}, - {TextureFormat::A2B10G10R10, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A2B10G10R10U}, + {TextureFormat::A2B10G10R10, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A2B10G10R10_UNORM}, + {TextureFormat::A2B10G10R10, C, UINT, UINT, UINT, UINT, PixelFormat::A2B10G10R10_UINT}, - {TextureFormat::A1B5G5R5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A1B5G5R5U}, + {TextureFormat::A1B5G5R5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A1B5G5R5_UNORM}, - {TextureFormat::A4B4G4R4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R4G4B4A4U}, + {TextureFormat::A4B4G4R4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A4B4G4R4_UNORM}, - {TextureFormat::R8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R8U}, - {TextureFormat::R8, C, UINT, UINT, UINT, UINT, PixelFormat::R8UI}, + {TextureFormat::R8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R8_UNORM}, + {TextureFormat::R8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R8_SNORM}, + {TextureFormat::R8, C, UINT, UINT, UINT, UINT, PixelFormat::R8_UINT}, + {TextureFormat::R8, C, SINT, SINT, SINT, SINT, PixelFormat::R8_SINT}, - {TextureFormat::G8R8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::RG8U}, - {TextureFormat::G8R8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::RG8S}, + {TextureFormat::R8G8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R8G8_UNORM}, + {TextureFormat::R8G8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R8G8_SNORM}, + {TextureFormat::R8G8, C, UINT, UINT, UINT, UINT, PixelFormat::R8G8_UINT}, + {TextureFormat::R8G8, C, SINT, SINT, SINT, SINT, PixelFormat::R8G8_SINT}, - {TextureFormat::R16_G16_B16_A16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::RGBA16S}, - {TextureFormat::R16_G16_B16_A16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::RGBA16U}, - {TextureFormat::R16_G16_B16_A16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::RGBA16F}, - {TextureFormat::R16_G16_B16_A16, C, UINT, UINT, UINT, UINT, PixelFormat::RGBA16UI}, + {TextureFormat::R16G16B16A16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16G16B16A16_SNORM}, + {TextureFormat::R16G16B16A16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16G16B16A16_UNORM}, + {TextureFormat::R16G16B16A16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16G16B16A16_FLOAT}, + {TextureFormat::R16G16B16A16, C, UINT, UINT, UINT, UINT, PixelFormat::R16G16B16A16_UINT}, + {TextureFormat::R16G16B16A16, C, SINT, SINT, SINT, SINT, PixelFormat::R16G16B16A16_SINT}, - {TextureFormat::R16_G16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::RG16F}, - {TextureFormat::R16_G16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::RG16}, - {TextureFormat::R16_G16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::RG16S}, - {TextureFormat::R16_G16, C, UINT, UINT, UINT, UINT, PixelFormat::RG16UI}, - {TextureFormat::R16_G16, C, SINT, SINT, SINT, SINT, PixelFormat::RG16I}, + {TextureFormat::R16G16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16G16_FLOAT}, + {TextureFormat::R16G16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16G16_UNORM}, + {TextureFormat::R16G16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16G16_SNORM}, + {TextureFormat::R16G16, C, UINT, UINT, UINT, UINT, PixelFormat::R16G16_UINT}, + {TextureFormat::R16G16, C, SINT, SINT, SINT, SINT, PixelFormat::R16G16_SINT}, - {TextureFormat::R16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16F}, - {TextureFormat::R16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16U}, - {TextureFormat::R16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16S}, - {TextureFormat::R16, C, UINT, UINT, UINT, UINT, PixelFormat::R16UI}, - {TextureFormat::R16, C, SINT, SINT, SINT, SINT, PixelFormat::R16I}, + {TextureFormat::R16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16_FLOAT}, + {TextureFormat::R16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16_UNORM}, + {TextureFormat::R16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16_SNORM}, + {TextureFormat::R16, C, UINT, UINT, UINT, UINT, PixelFormat::R16_UINT}, + {TextureFormat::R16, C, SINT, SINT, SINT, SINT, PixelFormat::R16_SINT}, - {TextureFormat::BF10GF11RF11, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R11FG11FB10F}, + {TextureFormat::B10G11R11, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::B10G11R11_FLOAT}, - {TextureFormat::R32_G32_B32_A32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::RGBA32F}, - {TextureFormat::R32_G32_B32_A32, C, UINT, UINT, UINT, UINT, PixelFormat::RGBA32UI}, + {TextureFormat::R32G32B32A32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32G32B32A32_FLOAT}, + {TextureFormat::R32G32B32A32, C, UINT, UINT, UINT, UINT, PixelFormat::R32G32B32A32_UINT}, + {TextureFormat::R32G32B32A32, C, SINT, SINT, SINT, SINT, PixelFormat::R32G32B32A32_SINT}, - {TextureFormat::R32_G32_B32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::RGB32F}, + {TextureFormat::R32G32B32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32G32B32_FLOAT}, - {TextureFormat::R32_G32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::RG32F}, - {TextureFormat::R32_G32, C, UINT, UINT, UINT, UINT, PixelFormat::RG32UI}, + {TextureFormat::R32G32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32G32_FLOAT}, + {TextureFormat::R32G32, C, UINT, UINT, UINT, UINT, PixelFormat::R32G32_UINT}, + {TextureFormat::R32G32, C, SINT, SINT, SINT, SINT, PixelFormat::R32G32_SINT}, - {TextureFormat::R32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32F}, - {TextureFormat::R32, C, UINT, UINT, UINT, UINT, PixelFormat::R32UI}, - {TextureFormat::R32, C, SINT, SINT, SINT, SINT, PixelFormat::R32I}, + {TextureFormat::R32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32_FLOAT}, + {TextureFormat::R32, C, UINT, UINT, UINT, UINT, PixelFormat::R32_UINT}, + {TextureFormat::R32, C, SINT, SINT, SINT, SINT, PixelFormat::R32_SINT}, - {TextureFormat::E5B9G9R9_SHAREDEXP, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::E5B9G9R9F}, + {TextureFormat::E5B9G9R9, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::E5B9G9R9_FLOAT}, - {TextureFormat::ZF32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::Z32F}, - {TextureFormat::Z16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::Z16}, - {TextureFormat::S8Z24, C, UINT, UNORM, UNORM, UNORM, PixelFormat::S8Z24}, - {TextureFormat::ZF32_X24S8, C, FLOAT, UINT, UNORM, UNORM, PixelFormat::Z32FS8}, + {TextureFormat::D32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::D32_FLOAT}, + {TextureFormat::D16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::D16_UNORM}, + {TextureFormat::S8D24, C, UINT, UNORM, UNORM, UNORM, PixelFormat::S8_UINT_D24_UNORM}, + {TextureFormat::R8G24, C, UINT, UNORM, UNORM, UNORM, PixelFormat::S8_UINT_D24_UNORM}, + {TextureFormat::D32S8, C, FLOAT, UINT, UNORM, UNORM, PixelFormat::D32_FLOAT_S8_UINT}, - {TextureFormat::DXT1, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXT1}, - {TextureFormat::DXT1, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXT1_SRGB}, + {TextureFormat::BC1_RGBA, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC1_RGBA_UNORM}, + {TextureFormat::BC1_RGBA, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC1_RGBA_SRGB}, - {TextureFormat::DXT23, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXT23}, - {TextureFormat::DXT23, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXT23_SRGB}, + {TextureFormat::BC2, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC2_UNORM}, + {TextureFormat::BC2, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC2_SRGB}, - {TextureFormat::DXT45, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXT45}, - {TextureFormat::DXT45, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXT45_SRGB}, + {TextureFormat::BC3, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC3_UNORM}, + {TextureFormat::BC3, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC3_SRGB}, - // TODO: Use a different pixel format for SNORM - {TextureFormat::DXN1, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXN1}, - {TextureFormat::DXN1, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::DXN1}, + {TextureFormat::BC4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC4_UNORM}, + {TextureFormat::BC4, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::BC4_SNORM}, - {TextureFormat::DXN2, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::DXN2UNORM}, - {TextureFormat::DXN2, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::DXN2SNORM}, + {TextureFormat::BC5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC5_UNORM}, + {TextureFormat::BC5, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::BC5_SNORM}, - {TextureFormat::BC7U, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC7U}, - {TextureFormat::BC7U, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC7U_SRGB}, + {TextureFormat::BC7, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC7_UNORM}, + {TextureFormat::BC7, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC7_SRGB}, - {TextureFormat::BC6H_SF16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::BC6H_SF16}, - {TextureFormat::BC6H_UF16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::BC6H_UF16}, + {TextureFormat::BC6H_SFLOAT, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::BC6H_SFLOAT}, + {TextureFormat::BC6H_UFLOAT, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::BC6H_UFLOAT}, - {TextureFormat::ASTC_2D_4X4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_4X4}, + {TextureFormat::ASTC_2D_4X4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_4X4_UNORM}, {TextureFormat::ASTC_2D_4X4, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_4X4_SRGB}, - {TextureFormat::ASTC_2D_5X4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X4}, + {TextureFormat::ASTC_2D_5X4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X4_UNORM}, {TextureFormat::ASTC_2D_5X4, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X4_SRGB}, - {TextureFormat::ASTC_2D_5X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X5}, + {TextureFormat::ASTC_2D_5X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X5_UNORM}, {TextureFormat::ASTC_2D_5X5, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X5_SRGB}, - {TextureFormat::ASTC_2D_8X8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X8}, + {TextureFormat::ASTC_2D_8X8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X8_UNORM}, {TextureFormat::ASTC_2D_8X8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X8_SRGB}, - {TextureFormat::ASTC_2D_8X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X5}, + {TextureFormat::ASTC_2D_8X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X5_UNORM}, {TextureFormat::ASTC_2D_8X5, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X5_SRGB}, - {TextureFormat::ASTC_2D_10X8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X8}, + {TextureFormat::ASTC_2D_10X8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X8_UNORM}, {TextureFormat::ASTC_2D_10X8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X8_SRGB}, - {TextureFormat::ASTC_2D_6X6, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X6}, + {TextureFormat::ASTC_2D_6X6, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X6_UNORM}, {TextureFormat::ASTC_2D_6X6, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X6_SRGB}, - {TextureFormat::ASTC_2D_10X10, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X10}, + {TextureFormat::ASTC_2D_10X10, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X10_UNORM}, {TextureFormat::ASTC_2D_10X10, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X10_SRGB}, - {TextureFormat::ASTC_2D_12X12, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_12X12}, + {TextureFormat::ASTC_2D_12X12, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_12X12_UNORM}, {TextureFormat::ASTC_2D_12X12, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_12X12_SRGB}, - {TextureFormat::ASTC_2D_8X6, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X6}, + {TextureFormat::ASTC_2D_8X6, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X6_UNORM}, {TextureFormat::ASTC_2D_8X6, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X6_SRGB}, - {TextureFormat::ASTC_2D_6X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X5}, + {TextureFormat::ASTC_2D_6X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X5_UNORM}, {TextureFormat::ASTC_2D_6X5, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X5_SRGB}, }}; @@ -182,7 +189,7 @@ PixelFormat FormatLookupTable::GetPixelFormat(TextureFormat format, bool is_srgb static_cast<int>(format), is_srgb, static_cast<int>(red_component), static_cast<int>(green_component), static_cast<int>(blue_component), static_cast<int>(alpha_component)); - return PixelFormat::ABGR8U; + return PixelFormat::A8B8G8R8_UNORM; } void FormatLookupTable::Set(TextureFormat format, bool is_srgb, ComponentType red_component, @@ -196,9 +203,9 @@ std::size_t FormatLookupTable::CalculateIndex(TextureFormat format, bool is_srgb ComponentType alpha_component) noexcept { const auto format_index = static_cast<std::size_t>(format); const auto red_index = static_cast<std::size_t>(red_component); - const auto green_index = static_cast<std::size_t>(red_component); - const auto blue_index = static_cast<std::size_t>(red_component); - const auto alpha_index = static_cast<std::size_t>(red_component); + const auto green_index = static_cast<std::size_t>(green_component); + const auto blue_index = static_cast<std::size_t>(blue_component); + const auto alpha_index = static_cast<std::size_t>(alpha_component); const std::size_t srgb_index = is_srgb ? 1 : 0; return format_index * PerFormat + diff --git a/src/video_core/texture_cache/surface_base.cpp b/src/video_core/texture_cache/surface_base.cpp index 7af0e792c..b44c09d71 100644 --- a/src/video_core/texture_cache/surface_base.cpp +++ b/src/video_core/texture_cache/surface_base.cpp @@ -115,17 +115,24 @@ std::optional<std::pair<u32, u32>> SurfaceBaseImpl::GetLayerMipmap( if (gpu_addr == candidate_gpu_addr) { return {{0, 0}}; } + if (candidate_gpu_addr < gpu_addr) { - return {}; + return std::nullopt; } + const auto relative_address{static_cast<GPUVAddr>(candidate_gpu_addr - gpu_addr)}; const auto layer{static_cast<u32>(relative_address / layer_size)}; + if (layer >= params.depth) { + return std::nullopt; + } + const GPUVAddr mipmap_address = relative_address - layer_size * layer; const auto mipmap_it = Common::BinaryFind(mipmap_offsets.begin(), mipmap_offsets.end(), mipmap_address); if (mipmap_it == mipmap_offsets.end()) { - return {}; + return std::nullopt; } + const auto level{static_cast<u32>(std::distance(mipmap_offsets.begin(), mipmap_it))}; return std::make_pair(layer, level); } @@ -225,7 +232,7 @@ void SurfaceBaseImpl::LoadBuffer(Tegra::MemoryManager& memory_manager, } } - if (!is_converted && params.pixel_format != PixelFormat::S8Z24) { + if (!is_converted && params.pixel_format != PixelFormat::S8_UINT_D24_UNORM) { return; } @@ -251,6 +258,11 @@ void SurfaceBaseImpl::FlushBuffer(Tegra::MemoryManager& memory_manager, tmp_buffer.resize(guest_memory_size); host_ptr = tmp_buffer.data(); + if (params.target == SurfaceTarget::Texture3D) { + // Special case for 3D texture segments + memory_manager.ReadBlockUnsafe(gpu_addr, host_ptr, guest_memory_size); + } + if (params.is_tiled) { ASSERT_MSG(params.block_width == 0, "Block width is defined as {}", params.block_width); for (u32 level = 0; level < params.num_levels; ++level) { diff --git a/src/video_core/texture_cache/surface_base.h b/src/video_core/texture_cache/surface_base.h index a39a8661b..173f2edba 100644 --- a/src/video_core/texture_cache/surface_base.h +++ b/src/video_core/texture_cache/surface_base.h @@ -72,9 +72,9 @@ public: return (cpu_addr < end) && (cpu_addr_end > start); } - bool IsInside(const GPUVAddr other_start, const GPUVAddr other_end) { + bool IsInside(const GPUVAddr other_start, const GPUVAddr other_end) const { const GPUVAddr gpu_addr_end = gpu_addr + guest_memory_size; - return (gpu_addr <= other_start && other_end <= gpu_addr_end); + return gpu_addr <= other_start && other_end <= gpu_addr_end; } // Use only when recycling a surface @@ -192,6 +192,22 @@ public: index = index_; } + void SetMemoryMarked(bool is_memory_marked_) { + is_memory_marked = is_memory_marked_; + } + + bool IsMemoryMarked() const { + return is_memory_marked; + } + + void SetSyncPending(bool is_sync_pending_) { + is_sync_pending = is_sync_pending_; + } + + bool IsSyncPending() const { + return is_sync_pending; + } + void MarkAsPicked(bool is_picked_) { is_picked = is_picked_; } @@ -201,8 +217,8 @@ public: } bool IsProtected() const { - // Only 3D Slices are to be protected - return is_target && params.block_depth > 0; + // Only 3D slices are to be protected + return is_target && params.target == SurfaceTarget::Texture3D; } bool IsRenderTarget() const { @@ -234,6 +250,11 @@ public: return GetView(ViewParams(overview_params.target, 0, num_layers, 0, params.num_levels)); } + TView Emplace3DView(u32 slice, u32 depth, u32 base_level, u32 num_levels) { + return GetView(ViewParams(VideoCore::Surface::SurfaceTarget::Texture3D, slice, depth, + base_level, num_levels)); + } + std::optional<TView> EmplaceIrregularView(const SurfaceParams& view_params, const GPUVAddr view_addr, const std::size_t candidate_size, const u32 mipmap, @@ -256,8 +277,8 @@ public: std::optional<TView> EmplaceView(const SurfaceParams& view_params, const GPUVAddr view_addr, const std::size_t candidate_size) { if (params.target == SurfaceTarget::Texture3D || - (params.num_levels == 1 && !params.is_layered) || - view_params.target == SurfaceTarget::Texture3D) { + view_params.target == SurfaceTarget::Texture3D || + (params.num_levels == 1 && !params.is_layered)) { return {}; } const auto layer_mipmap{GetLayerMipmap(view_addr)}; @@ -303,6 +324,8 @@ private: bool is_target{}; bool is_registered{}; bool is_picked{}; + bool is_memory_marked{}; + bool is_sync_pending{}; u32 index{NO_RT}; u64 modification_tick{}; }; diff --git a/src/video_core/texture_cache/surface_params.cpp b/src/video_core/texture_cache/surface_params.cpp index 6f3ef45be..13dd16356 100644 --- a/src/video_core/texture_cache/surface_params.cpp +++ b/src/video_core/texture_cache/surface_params.cpp @@ -74,21 +74,21 @@ SurfaceParams SurfaceParams::CreateForTexture(const FormatLookupTable& lookup_ta SurfaceParams params; params.is_tiled = tic.IsTiled(); params.srgb_conversion = tic.IsSrgbConversionEnabled(); - params.block_width = params.is_tiled ? tic.BlockWidth() : 0, - params.block_height = params.is_tiled ? tic.BlockHeight() : 0, - params.block_depth = params.is_tiled ? tic.BlockDepth() : 0, + params.block_width = params.is_tiled ? tic.BlockWidth() : 0; + params.block_height = params.is_tiled ? tic.BlockHeight() : 0; + params.block_depth = params.is_tiled ? tic.BlockDepth() : 0; params.tile_width_spacing = params.is_tiled ? (1 << tic.tile_width_spacing.Value()) : 1; params.pixel_format = lookup_table.GetPixelFormat( tic.format, params.srgb_conversion, tic.r_type, tic.g_type, tic.b_type, tic.a_type); params.type = GetFormatType(params.pixel_format); - if (entry.IsShadow() && params.type == SurfaceType::ColorTexture) { + if (entry.is_shadow && params.type == SurfaceType::ColorTexture) { switch (params.pixel_format) { - case PixelFormat::R16U: - case PixelFormat::R16F: - params.pixel_format = PixelFormat::Z16; + case PixelFormat::R16_UNORM: + case PixelFormat::R16_FLOAT: + params.pixel_format = PixelFormat::D16_UNORM; break; - case PixelFormat::R32F: - params.pixel_format = PixelFormat::Z32F; + case PixelFormat::R32_FLOAT: + params.pixel_format = PixelFormat::D32_FLOAT; break; default: UNIMPLEMENTED_MSG("Unimplemented shadow convert format: {}", @@ -96,7 +96,6 @@ SurfaceParams SurfaceParams::CreateForTexture(const FormatLookupTable& lookup_ta } params.type = GetFormatType(params.pixel_format); } - params.type = GetFormatType(params.pixel_format); // TODO: on 1DBuffer we should use the tic info. if (tic.IsBuffer()) { params.target = SurfaceTarget::TextureBuffer; @@ -108,7 +107,7 @@ SurfaceParams SurfaceParams::CreateForTexture(const FormatLookupTable& lookup_ta params.emulated_levels = 1; params.is_layered = false; } else { - params.target = TextureTypeToSurfaceTarget(entry.GetType(), entry.IsArray()); + params.target = TextureTypeToSurfaceTarget(entry.type, entry.is_array); params.width = tic.Width(); params.height = tic.Height(); params.depth = tic.Depth(); @@ -130,15 +129,14 @@ SurfaceParams SurfaceParams::CreateForImage(const FormatLookupTable& lookup_tabl SurfaceParams params; params.is_tiled = tic.IsTiled(); params.srgb_conversion = tic.IsSrgbConversionEnabled(); - params.block_width = params.is_tiled ? tic.BlockWidth() : 0, - params.block_height = params.is_tiled ? tic.BlockHeight() : 0, - params.block_depth = params.is_tiled ? tic.BlockDepth() : 0, + params.block_width = params.is_tiled ? tic.BlockWidth() : 0; + params.block_height = params.is_tiled ? tic.BlockHeight() : 0; + params.block_depth = params.is_tiled ? tic.BlockDepth() : 0; params.tile_width_spacing = params.is_tiled ? (1 << tic.tile_width_spacing.Value()) : 1; params.pixel_format = lookup_table.GetPixelFormat( tic.format, params.srgb_conversion, tic.r_type, tic.g_type, tic.b_type, tic.a_type); params.type = GetFormatType(params.pixel_format); - params.type = GetFormatType(params.pixel_format); - params.target = ImageTypeToSurfaceTarget(entry.GetType()); + params.target = ImageTypeToSurfaceTarget(entry.type); // TODO: on 1DBuffer we should use the tic info. if (tic.IsBuffer()) { params.target = SurfaceTarget::TextureBuffer; @@ -165,39 +163,40 @@ SurfaceParams SurfaceParams::CreateForImage(const FormatLookupTable& lookup_tabl return params; } -SurfaceParams SurfaceParams::CreateForDepthBuffer(Core::System& system) { - const auto& regs = system.GPU().Maxwell3D().regs; - regs.zeta_width, regs.zeta_height, regs.zeta.format, regs.zeta.memory_layout.type; - SurfaceParams params; - params.is_tiled = regs.zeta.memory_layout.type == - Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear; - params.srgb_conversion = false; - params.block_width = std::min(regs.zeta.memory_layout.block_width.Value(), 5U); - params.block_height = std::min(regs.zeta.memory_layout.block_height.Value(), 5U); - params.block_depth = std::min(regs.zeta.memory_layout.block_depth.Value(), 5U); - params.tile_width_spacing = 1; - params.pixel_format = PixelFormatFromDepthFormat(regs.zeta.format); - params.type = GetFormatType(params.pixel_format); - params.width = regs.zeta_width; - params.height = regs.zeta_height; - params.pitch = 0; - params.num_levels = 1; - params.emulated_levels = 1; - - const bool is_layered = regs.zeta_layers > 1 && params.block_depth == 0; - params.is_layered = is_layered; - params.target = is_layered ? SurfaceTarget::Texture2DArray : SurfaceTarget::Texture2D; - params.depth = is_layered ? regs.zeta_layers.Value() : 1U; - return params; +SurfaceParams SurfaceParams::CreateForDepthBuffer(Tegra::Engines::Maxwell3D& maxwell3d) { + const auto& regs = maxwell3d.regs; + const auto block_depth = std::min(regs.zeta.memory_layout.block_depth.Value(), 5U); + const bool is_layered = regs.zeta_layers > 1 && block_depth == 0; + const auto pixel_format = PixelFormatFromDepthFormat(regs.zeta.format); + return { + .is_tiled = regs.zeta.memory_layout.type == + Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear, + .srgb_conversion = false, + .is_layered = is_layered, + .block_width = std::min(regs.zeta.memory_layout.block_width.Value(), 5U), + .block_height = std::min(regs.zeta.memory_layout.block_height.Value(), 5U), + .block_depth = block_depth, + .tile_width_spacing = 1, + .width = regs.zeta_width, + .height = regs.zeta_height, + .depth = is_layered ? regs.zeta_layers.Value() : 1U, + .pitch = 0, + .num_levels = 1, + .emulated_levels = 1, + .pixel_format = pixel_format, + .type = GetFormatType(pixel_format), + .target = is_layered ? SurfaceTarget::Texture2DArray : SurfaceTarget::Texture2D, + }; } -SurfaceParams SurfaceParams::CreateForFramebuffer(Core::System& system, std::size_t index) { - const auto& config{system.GPU().Maxwell3D().regs.rt[index]}; +SurfaceParams SurfaceParams::CreateForFramebuffer(Tegra::Engines::Maxwell3D& maxwell3d, + std::size_t index) { + const auto& config{maxwell3d.regs.rt[index]}; SurfaceParams params; params.is_tiled = config.memory_layout.type == Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear; - params.srgb_conversion = config.format == Tegra::RenderTargetFormat::BGRA8_SRGB || - config.format == Tegra::RenderTargetFormat::RGBA8_SRGB; + params.srgb_conversion = config.format == Tegra::RenderTargetFormat::B8G8R8A8_SRGB || + config.format == Tegra::RenderTargetFormat::A8B8G8R8_SRGB; params.block_width = config.memory_layout.block_width; params.block_height = config.memory_layout.block_height; params.block_depth = config.memory_layout.block_depth; @@ -216,45 +215,60 @@ SurfaceParams SurfaceParams::CreateForFramebuffer(Core::System& system, std::siz params.num_levels = 1; params.emulated_levels = 1; - const bool is_layered = config.layers > 1 && params.block_depth == 0; - params.is_layered = is_layered; - params.depth = is_layered ? config.layers.Value() : 1; - params.target = is_layered ? SurfaceTarget::Texture2DArray : SurfaceTarget::Texture2D; + if (config.memory_layout.is_3d != 0) { + params.depth = config.layers.Value(); + params.is_layered = false; + params.target = SurfaceTarget::Texture3D; + } else if (config.layers > 1) { + params.depth = config.layers.Value(); + params.is_layered = true; + params.target = SurfaceTarget::Texture2DArray; + } else { + params.depth = 1; + params.is_layered = false; + params.target = SurfaceTarget::Texture2D; + } return params; } SurfaceParams SurfaceParams::CreateForFermiCopySurface( const Tegra::Engines::Fermi2D::Regs::Surface& config) { - SurfaceParams params{}; - params.is_tiled = !config.linear; - params.srgb_conversion = config.format == Tegra::RenderTargetFormat::BGRA8_SRGB || - config.format == Tegra::RenderTargetFormat::RGBA8_SRGB; - params.block_width = params.is_tiled ? std::min(config.BlockWidth(), 5U) : 0, - params.block_height = params.is_tiled ? std::min(config.BlockHeight(), 5U) : 0, - params.block_depth = params.is_tiled ? std::min(config.BlockDepth(), 5U) : 0, - params.tile_width_spacing = 1; - params.pixel_format = PixelFormatFromRenderTargetFormat(config.format); - params.type = GetFormatType(params.pixel_format); - params.width = config.width; - params.height = config.height; - params.pitch = config.pitch; - // TODO(Rodrigo): Try to guess the surface target from depth and layer parameters - params.target = SurfaceTarget::Texture2D; - params.depth = 1; - params.num_levels = 1; - params.emulated_levels = 1; + const bool is_tiled = !config.linear; + const auto pixel_format = PixelFormatFromRenderTargetFormat(config.format); + + SurfaceParams params{ + .is_tiled = is_tiled, + .srgb_conversion = config.format == Tegra::RenderTargetFormat::B8G8R8A8_SRGB || + config.format == Tegra::RenderTargetFormat::A8B8G8R8_SRGB, + .is_layered = false, + .block_width = is_tiled ? std::min(config.BlockWidth(), 5U) : 0U, + .block_height = is_tiled ? std::min(config.BlockHeight(), 5U) : 0U, + .block_depth = is_tiled ? std::min(config.BlockDepth(), 5U) : 0U, + .tile_width_spacing = 1, + .width = config.width, + .height = config.height, + .depth = 1, + .pitch = config.pitch, + .num_levels = 1, + .emulated_levels = 1, + .pixel_format = pixel_format, + .type = GetFormatType(pixel_format), + // TODO(Rodrigo): Try to guess texture arrays from parameters + .target = SurfaceTarget::Texture2D, + }; + params.is_layered = params.IsLayered(); return params; } VideoCore::Surface::SurfaceTarget SurfaceParams::ExpectedTarget( const VideoCommon::Shader::Sampler& entry) { - return TextureTypeToSurfaceTarget(entry.GetType(), entry.IsArray()); + return TextureTypeToSurfaceTarget(entry.type, entry.is_array); } VideoCore::Surface::SurfaceTarget SurfaceParams::ExpectedTarget( const VideoCommon::Shader::Image& entry) { - return ImageTypeToSurfaceTarget(entry.GetType()); + return ImageTypeToSurfaceTarget(entry.type); } bool SurfaceParams::IsLayered() const { @@ -335,8 +349,7 @@ std::size_t SurfaceParams::GetLayerSize(bool as_host_size, bool uncompressed) co size += GetInnerMipmapMemorySize(level, as_host_size, uncompressed); } if (is_tiled && is_layered) { - return Common::AlignBits(size, - Tegra::Texture::GetGOBSizeShift() + block_height + block_depth); + return Common::AlignBits(size, Tegra::Texture::GOB_SIZE_SHIFT + block_height + block_depth); } return size; } @@ -410,7 +423,7 @@ std::tuple<u32, u32, u32> SurfaceParams::GetBlockOffsetXYZ(u32 offset) const { const u32 block_size = GetBlockSize(); const u32 block_index = offset / block_size; const u32 gob_offset = offset % block_size; - const u32 gob_index = gob_offset / static_cast<u32>(Tegra::Texture::GetGOBSize()); + const u32 gob_index = gob_offset / static_cast<u32>(Tegra::Texture::GOB_SIZE); const u32 x_gob_pixels = 64U / GetBytesPerPixel(); const u32 x_block_pixels = x_gob_pixels << block_width; const u32 y_block_pixels = 8U << block_height; diff --git a/src/video_core/texture_cache/surface_params.h b/src/video_core/texture_cache/surface_params.h index 24957df8d..4466c3c34 100644 --- a/src/video_core/texture_cache/surface_params.h +++ b/src/video_core/texture_cache/surface_params.h @@ -33,10 +33,11 @@ public: const VideoCommon::Shader::Image& entry); /// Creates SurfaceCachedParams for a depth buffer configuration. - static SurfaceParams CreateForDepthBuffer(Core::System& system); + static SurfaceParams CreateForDepthBuffer(Tegra::Engines::Maxwell3D& maxwell3d); /// Creates SurfaceCachedParams from a framebuffer configuration. - static SurfaceParams CreateForFramebuffer(Core::System& system, std::size_t index); + static SurfaceParams CreateForFramebuffer(Tegra::Engines::Maxwell3D& maxwell3d, + std::size_t index); /// Creates SurfaceCachedParams from a Fermi2D surface configuration. static SurfaceParams CreateForFermiCopySurface( @@ -204,7 +205,7 @@ public: static std::size_t AlignLayered(const std::size_t out_size, const u32 block_height, const u32 block_depth) { return Common::AlignBits(out_size, - Tegra::Texture::GetGOBSizeShift() + block_height + block_depth); + Tegra::Texture::GOB_SIZE_SHIFT + block_height + block_depth); } /// Converts a width from a type of surface into another. This helps represent the diff --git a/src/video_core/texture_cache/surface_view.cpp b/src/video_core/texture_cache/surface_view.cpp index 57a1f5803..6b5f5984b 100644 --- a/src/video_core/texture_cache/surface_view.cpp +++ b/src/video_core/texture_cache/surface_view.cpp @@ -20,4 +20,8 @@ bool ViewParams::operator==(const ViewParams& rhs) const { std::tie(rhs.base_layer, rhs.num_layers, rhs.base_level, rhs.num_levels, rhs.target); } +bool ViewParams::operator!=(const ViewParams& rhs) const { + return !operator==(rhs); +} + } // namespace VideoCommon diff --git a/src/video_core/texture_cache/surface_view.h b/src/video_core/texture_cache/surface_view.h index b17fd11a9..90a8bb0ae 100644 --- a/src/video_core/texture_cache/surface_view.h +++ b/src/video_core/texture_cache/surface_view.h @@ -21,6 +21,7 @@ struct ViewParams { std::size_t Hash() const; bool operator==(const ViewParams& rhs) const; + bool operator!=(const ViewParams& rhs) const; bool IsLayered() const { switch (target) { diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index 4edd4313b..ea835c59f 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -6,6 +6,7 @@ #include <algorithm> #include <array> +#include <list> #include <memory> #include <mutex> #include <set> @@ -13,6 +14,7 @@ #include <unordered_map> #include <vector> +#include <boost/container/small_vector.hpp> #include <boost/icl/interval_map.hpp> #include <boost/range/iterator_range.hpp> @@ -22,6 +24,7 @@ #include "core/core.h" #include "core/memory.h" #include "core/settings.h" +#include "video_core/compatible_formats.h" #include "video_core/dirty_flags.h" #include "video_core/engines/fermi_2d.h" #include "video_core/engines/maxwell_3d.h" @@ -45,13 +48,14 @@ class RasterizerInterface; namespace VideoCommon { +using VideoCore::Surface::FormatCompatibility; using VideoCore::Surface::PixelFormat; - using VideoCore::Surface::SurfaceTarget; using RenderTargetConfig = Tegra::Engines::Maxwell3D::Regs::RenderTargetConfig; template <typename TSurface, typename TView> class TextureCache { + using VectorSurface = boost::container::small_vector<TSurface, 1>; public: void InvalidateRegion(VAddr addr, std::size_t size) { @@ -62,6 +66,30 @@ public: } } + void OnCPUWrite(VAddr addr, std::size_t size) { + std::lock_guard lock{mutex}; + + for (const auto& surface : GetSurfacesInRegion(addr, size)) { + if (surface->IsMemoryMarked()) { + UnmarkMemory(surface); + surface->SetSyncPending(true); + marked_for_unregister.emplace_back(surface); + } + } + } + + void SyncGuestHost() { + std::lock_guard lock{mutex}; + + for (const auto& surface : marked_for_unregister) { + if (surface->IsRegistered()) { + surface->SetSyncPending(false); + Unregister(surface); + } + } + marked_for_unregister.clear(); + } + /** * Guarantees that rendertargets don't unregister themselves if the * collide. Protection is currently only done on 3D slices. @@ -85,10 +113,20 @@ public: return a->GetModificationTick() < b->GetModificationTick(); }); for (const auto& surface : surfaces) { + mutex.unlock(); FlushSurface(surface); + mutex.lock(); } } + bool MustFlushRegion(VAddr addr, std::size_t size) { + std::lock_guard lock{mutex}; + + const auto surfaces = GetSurfacesInRegion(addr, size); + return std::any_of(surfaces.cbegin(), surfaces.cend(), + [](const TSurface& surface) { return surface->IsModified(); }); + } + TView GetTextureSurface(const Tegra::Texture::TICEntry& tic, const VideoCommon::Shader::Sampler& entry) { std::lock_guard lock{mutex}; @@ -97,8 +135,7 @@ public: return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); } - const std::optional<VAddr> cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); if (!cpu_addr) { return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); } @@ -108,7 +145,7 @@ public: } const auto params{SurfaceParams::CreateForTexture(format_lookup_table, tic, entry)}; - const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, false); + const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, true, false); if (guard_samplers) { sampled_textures.push_back(surface); } @@ -122,13 +159,12 @@ public: if (!gpu_addr) { return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); } - const std::optional<VAddr> cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); if (!cpu_addr) { return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); } const auto params{SurfaceParams::CreateForImage(format_lookup_table, tic, entry)}; - const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, false); + const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, true, false); if (guard_samplers) { sampled_textures.push_back(surface); } @@ -143,13 +179,13 @@ public: return any_rt; } - TView GetDepthBufferSurface() { + TView GetDepthBufferSurface(bool preserve_contents) { std::lock_guard lock{mutex}; - auto& maxwell3d = system.GPU().Maxwell3D(); - if (!maxwell3d.dirty.flags[VideoCommon::Dirty::ZetaBuffer]) { + auto& dirty = maxwell3d.dirty; + if (!dirty.flags[VideoCommon::Dirty::ZetaBuffer]) { return depth_buffer.view; } - maxwell3d.dirty.flags[VideoCommon::Dirty::ZetaBuffer] = false; + dirty.flags[VideoCommon::Dirty::ZetaBuffer] = false; const auto& regs{maxwell3d.regs}; const auto gpu_addr{regs.zeta.Address()}; @@ -157,14 +193,13 @@ public: SetEmptyDepthBuffer(); return {}; } - const std::optional<VAddr> cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); if (!cpu_addr) { SetEmptyDepthBuffer(); return {}; } - const auto depth_params{SurfaceParams::CreateForDepthBuffer(system)}; - auto surface_view = GetSurface(gpu_addr, *cpu_addr, depth_params, true); + const auto depth_params{SurfaceParams::CreateForDepthBuffer(maxwell3d)}; + auto surface_view = GetSurface(gpu_addr, *cpu_addr, depth_params, preserve_contents, true); if (depth_buffer.target) depth_buffer.target->MarkAsRenderTarget(false, NO_RT); depth_buffer.target = surface_view.first; @@ -174,10 +209,9 @@ public: return surface_view.second; } - TView GetColorBufferSurface(std::size_t index) { + TView GetColorBufferSurface(std::size_t index, bool preserve_contents) { std::lock_guard lock{mutex}; ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); - auto& maxwell3d = system.GPU().Maxwell3D(); if (!maxwell3d.dirty.flags[VideoCommon::Dirty::ColorBuffer0 + index]) { return render_targets[index].view; } @@ -197,17 +231,23 @@ public: return {}; } - const std::optional<VAddr> cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); if (!cpu_addr) { SetEmptyColorBuffer(index); return {}; } - auto surface_view = GetSurface(gpu_addr, *cpu_addr, - SurfaceParams::CreateForFramebuffer(system, index), true); - if (render_targets[index].target) - render_targets[index].target->MarkAsRenderTarget(false, NO_RT); + auto surface_view = + GetSurface(gpu_addr, *cpu_addr, SurfaceParams::CreateForFramebuffer(maxwell3d, index), + preserve_contents, true); + if (render_targets[index].target) { + auto& surface = render_targets[index].target; + surface->MarkAsRenderTarget(false, NO_RT); + const auto& cr_params = surface->GetSurfaceParams(); + if (!cr_params.is_tiled && Settings::values.use_asynchronous_gpu_emulation.GetValue()) { + AsyncFlushSurface(surface); + } + } render_targets[index].target = surface_view.first; render_targets[index].view = surface_view.second; if (render_targets[index].target) @@ -254,40 +294,69 @@ public: const GPUVAddr src_gpu_addr = src_config.Address(); const GPUVAddr dst_gpu_addr = dst_config.Address(); DeduceBestBlit(src_params, dst_params, src_gpu_addr, dst_gpu_addr); - const std::optional<VAddr> dst_cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(dst_gpu_addr); - const std::optional<VAddr> src_cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(src_gpu_addr); - std::pair<TSurface, TView> dst_surface = - GetSurface(dst_gpu_addr, *dst_cpu_addr, dst_params, false); - std::pair<TSurface, TView> src_surface = - GetSurface(src_gpu_addr, *src_cpu_addr, src_params, false); - ImageBlit(src_surface.second, dst_surface.second, copy_config); + + const std::optional<VAddr> dst_cpu_addr = gpu_memory.GpuToCpuAddress(dst_gpu_addr); + const std::optional<VAddr> src_cpu_addr = gpu_memory.GpuToCpuAddress(src_gpu_addr); + std::pair dst_surface = GetSurface(dst_gpu_addr, *dst_cpu_addr, dst_params, true, false); + TView src_surface = GetSurface(src_gpu_addr, *src_cpu_addr, src_params, true, false).second; + ImageBlit(src_surface, dst_surface.second, copy_config); dst_surface.first->MarkAsModified(true, Tick()); } - TSurface TryFindFramebufferSurface(VAddr addr) { + TSurface TryFindFramebufferSurface(VAddr addr) const { if (!addr) { return nullptr; } const VAddr page = addr >> registry_page_bits; - std::vector<TSurface>& list = registry[page]; - for (auto& surface : list) { - if (surface->GetCpuAddr() == addr) { - return surface; - } + const auto it = registry.find(page); + if (it == registry.end()) { + return nullptr; } - return nullptr; + const auto& list = it->second; + const auto found = std::find_if(list.begin(), list.end(), [addr](const auto& surface) { + return surface->GetCpuAddr() == addr; + }); + return found != list.end() ? *found : nullptr; } u64 Tick() { return ++ticks; } + void CommitAsyncFlushes() { + committed_flushes.push_back(uncommitted_flushes); + uncommitted_flushes.reset(); + } + + bool HasUncommittedFlushes() const { + return uncommitted_flushes != nullptr; + } + + bool ShouldWaitAsyncFlushes() const { + return !committed_flushes.empty() && committed_flushes.front() != nullptr; + } + + void PopAsyncFlushes() { + if (committed_flushes.empty()) { + return; + } + auto& flush_list = committed_flushes.front(); + if (!flush_list) { + committed_flushes.pop_front(); + return; + } + for (TSurface& surface : *flush_list) { + FlushSurface(surface); + } + committed_flushes.pop_front(); + } + protected: - explicit TextureCache(Core::System& system, VideoCore::RasterizerInterface& rasterizer, - bool is_astc_supported) - : system{system}, is_astc_supported{is_astc_supported}, rasterizer{rasterizer} { + explicit TextureCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, + bool is_astc_supported_) + : is_astc_supported{is_astc_supported_}, rasterizer{rasterizer_}, maxwell3d{maxwell3d_}, + gpu_memory{gpu_memory_} { for (std::size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { SetEmptyColorBuffer(i); } @@ -300,9 +369,9 @@ protected: siblings_table[static_cast<std::size_t>(b)] = a; }; std::fill(siblings_table.begin(), siblings_table.end(), PixelFormat::Invalid); - make_siblings(PixelFormat::Z16, PixelFormat::R16U); - make_siblings(PixelFormat::Z32F, PixelFormat::R32F); - make_siblings(PixelFormat::Z32FS8, PixelFormat::RG32F); + make_siblings(PixelFormat::D16_UNORM, PixelFormat::R16_UNORM); + make_siblings(PixelFormat::D32_FLOAT, PixelFormat::R32_FLOAT); + make_siblings(PixelFormat::D32_FLOAT_S8_UINT, PixelFormat::R32G32_FLOAT); sampled_textures.reserve(64); } @@ -322,7 +391,7 @@ protected: virtual void BufferCopy(TSurface& src_surface, TSurface& dst_surface) = 0; void ManageRenderTargetUnregister(TSurface& surface) { - auto& dirty = system.GPU().Maxwell3D().dirty; + auto& dirty = maxwell3d.dirty; const u32 index = surface->GetRenderTarget(); if (index == DEPTH_RT) { dirty.flags[VideoCommon::Dirty::ZetaBuffer] = true; @@ -335,8 +404,7 @@ protected: void Register(TSurface surface) { const GPUVAddr gpu_addr = surface->GetGpuAddr(); const std::size_t size = surface->GetSizeInBytes(); - const std::optional<VAddr> cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); if (!cpu_addr) { LOG_CRITICAL(HW_GPU, "Failed to register surface with unmapped gpu_address 0x{:016x}", gpu_addr); @@ -345,9 +413,20 @@ protected: surface->SetCpuAddr(*cpu_addr); RegisterInnerCache(surface); surface->MarkAsRegistered(true); + surface->SetMemoryMarked(true); rasterizer.UpdatePagesCachedCount(*cpu_addr, size, 1); } + void UnmarkMemory(TSurface surface) { + if (!surface->IsMemoryMarked()) { + return; + } + const std::size_t size = surface->GetSizeInBytes(); + const VAddr cpu_addr = surface->GetCpuAddr(); + rasterizer.UpdatePagesCachedCount(cpu_addr, size, -1); + surface->SetMemoryMarked(false); + } + void Unregister(TSurface surface) { if (guard_render_targets && surface->IsProtected()) { return; @@ -355,9 +434,11 @@ protected: if (!guard_render_targets && surface->IsRenderTarget()) { ManageRenderTargetUnregister(surface); } - const std::size_t size = surface->GetSizeInBytes(); - const VAddr cpu_addr = surface->GetCpuAddr(); - rasterizer.UpdatePagesCachedCount(cpu_addr, size, -1); + UnmarkMemory(surface); + if (surface->IsSyncPending()) { + marked_for_unregister.remove(surface); + surface->SetSyncPending(false); + } UnregisterInnerCache(surface); surface->MarkAsRegistered(false); ReserveSurface(surface->GetSurfaceParams(), surface); @@ -373,7 +454,6 @@ protected: return new_surface; } - Core::System& system; const bool is_astc_supported; private: @@ -415,18 +495,18 @@ private: * @param untopological Indicates to the recycler that the texture has no way * to match the overlaps due to topological reasons. **/ - RecycleStrategy PickStrategy(std::vector<TSurface>& overlaps, const SurfaceParams& params, + RecycleStrategy PickStrategy(VectorSurface& overlaps, const SurfaceParams& params, const GPUVAddr gpu_addr, const MatchTopologyResult untopological) { - if (Settings::values.use_accurate_gpu_emulation) { + if (Settings::IsGPULevelExtreme()) { return RecycleStrategy::Flush; } // 3D Textures decision - if (params.block_depth > 1 || params.target == SurfaceTarget::Texture3D) { + if (params.target == SurfaceTarget::Texture3D) { return RecycleStrategy::Flush; } for (const auto& s : overlaps) { const auto& s_params = s->GetSurfaceParams(); - if (s_params.block_depth > 1 || s_params.target == SurfaceTarget::Texture3D) { + if (s_params.target == SurfaceTarget::Texture3D) { return RecycleStrategy::Flush; } } @@ -450,18 +530,21 @@ private: * @param overlaps The overlapping surfaces registered in the cache. * @param params The parameters for the new surface. * @param gpu_addr The starting address of the new surface. + * @param preserve_contents Indicates that the new surface should be loaded from memory or left + * blank. * @param untopological Indicates to the recycler that the texture has no way to match the * overlaps due to topological reasons. **/ - std::pair<TSurface, TView> RecycleSurface(std::vector<TSurface>& overlaps, - const SurfaceParams& params, const GPUVAddr gpu_addr, + std::pair<TSurface, TView> RecycleSurface(VectorSurface& overlaps, const SurfaceParams& params, + const GPUVAddr gpu_addr, const bool preserve_contents, const MatchTopologyResult untopological) { + const bool do_load = preserve_contents && Settings::IsGPULevelExtreme(); for (auto& surface : overlaps) { Unregister(surface); } switch (PickStrategy(overlaps, params, gpu_addr, untopological)) { case RecycleStrategy::Ignore: { - return InitializeSurface(gpu_addr, params, Settings::values.use_accurate_gpu_emulation); + return InitializeSurface(gpu_addr, params, do_load); } case RecycleStrategy::Flush: { std::sort(overlaps.begin(), overlaps.end(), @@ -471,7 +554,7 @@ private: for (auto& surface : overlaps) { FlushSurface(surface); } - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } case RecycleStrategy::BufferCopy: { auto new_surface = GetUncachedSurface(gpu_addr, params); @@ -480,7 +563,7 @@ private: } default: { UNIMPLEMENTED_MSG("Unimplemented Texture Cache Recycling Strategy!"); - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, do_load); } } } @@ -507,15 +590,15 @@ private: } else { new_surface = GetUncachedSurface(gpu_addr, params); } - const auto& final_params = new_surface->GetSurfaceParams(); + const SurfaceParams& final_params = new_surface->GetSurfaceParams(); if (cr_params.type != final_params.type) { - if (Settings::values.use_accurate_gpu_emulation) { + if (Settings::IsGPULevelExtreme()) { BufferCopy(current_surface, new_surface); } } else { std::vector<CopyParams> bricks = current_surface->BreakDown(final_params); for (auto& brick : bricks) { - ImageCopy(current_surface, new_surface, brick); + TryCopyImage(current_surface, new_surface, brick); } } Unregister(current_surface); @@ -563,47 +646,65 @@ private: * @param params The parameters on the new surface. * @param gpu_addr The starting address of the new surface. **/ - std::optional<std::pair<TSurface, TView>> TryReconstructSurface(std::vector<TSurface>& overlaps, + std::optional<std::pair<TSurface, TView>> TryReconstructSurface(VectorSurface& overlaps, const SurfaceParams& params, - const GPUVAddr gpu_addr) { + GPUVAddr gpu_addr) { if (params.target == SurfaceTarget::Texture3D) { - return {}; + return std::nullopt; } - bool modified = false; + const auto test_modified = [](TSurface& surface) { return surface->IsModified(); }; TSurface new_surface = GetUncachedSurface(gpu_addr, params); - u32 passed_tests = 0; + + if (std::none_of(overlaps.begin(), overlaps.end(), test_modified)) { + LoadSurface(new_surface); + for (const auto& surface : overlaps) { + Unregister(surface); + } + Register(new_surface); + return {{new_surface, new_surface->GetMainView()}}; + } + + std::size_t passed_tests = 0; for (auto& surface : overlaps) { const SurfaceParams& src_params = surface->GetSurfaceParams(); - if (src_params.is_layered || src_params.num_levels > 1) { - // We send this cases to recycle as they are more complex to handle - return {}; - } - const std::size_t candidate_size = surface->GetSizeInBytes(); - auto mipmap_layer{new_surface->GetLayerMipmap(surface->GetGpuAddr())}; + const auto mipmap_layer{new_surface->GetLayerMipmap(surface->GetGpuAddr())}; if (!mipmap_layer) { continue; } - const auto [layer, mipmap] = *mipmap_layer; - if (new_surface->GetMipmapSize(mipmap) != candidate_size) { + const auto [base_layer, base_mipmap] = *mipmap_layer; + if (new_surface->GetMipmapSize(base_mipmap) != surface->GetMipmapSize(0)) { continue; } - modified |= surface->IsModified(); - // Now we got all the data set up - const u32 width = SurfaceParams::IntersectWidth(src_params, params, 0, mipmap); - const u32 height = SurfaceParams::IntersectHeight(src_params, params, 0, mipmap); - const CopyParams copy_params(0, 0, 0, 0, 0, layer, 0, mipmap, width, height, 1); - passed_tests++; - ImageCopy(surface, new_surface, copy_params); + ++passed_tests; + + // Copy all mipmaps and layers + const u32 block_width = params.GetDefaultBlockWidth(); + const u32 block_height = params.GetDefaultBlockHeight(); + for (u32 mipmap = base_mipmap; mipmap < base_mipmap + src_params.num_levels; ++mipmap) { + const u32 width = SurfaceParams::IntersectWidth(src_params, params, 0, mipmap); + const u32 height = SurfaceParams::IntersectHeight(src_params, params, 0, mipmap); + if (width < block_width || height < block_height) { + // Current APIs forbid copying small compressed textures, avoid errors + break; + } + const CopyParams copy_params(0, 0, 0, 0, 0, base_layer, 0, mipmap, width, height, + src_params.depth); + TryCopyImage(surface, new_surface, copy_params); + } } if (passed_tests == 0) { - return {}; + return std::nullopt; + } + if (Settings::IsGPULevelExtreme() && passed_tests != overlaps.size()) { // In Accurate GPU all tests should pass, else we recycle - } else if (Settings::values.use_accurate_gpu_emulation && passed_tests != overlaps.size()) { - return {}; + return std::nullopt; } + + const bool modified = std::any_of(overlaps.begin(), overlaps.end(), test_modified); for (const auto& surface : overlaps) { Unregister(surface); } + new_surface->MarkAsModified(modified, Tick()); Register(new_surface); return {{new_surface, new_surface->GetMainView()}}; @@ -614,64 +715,26 @@ private: * textures within the GPU if possible. Falls back to LLE when it isn't possible to use any of * the HLE methods. * - * @param overlaps The overlapping surfaces registered in the cache. - * @param params The parameters on the new surface. - * @param gpu_addr The starting address of the new surface. - * @param cache_addr The starting address of the new surface on physical memory. + * @param overlaps The overlapping surfaces registered in the cache. + * @param params The parameters on the new surface. + * @param gpu_addr The starting address of the new surface. + * @param cpu_addr The starting address of the new surface on physical memory. + * @param preserve_contents Indicates that the new surface should be loaded from memory or + * left blank. */ - std::optional<std::pair<TSurface, TView>> Manage3DSurfaces(std::vector<TSurface>& overlaps, + std::optional<std::pair<TSurface, TView>> Manage3DSurfaces(VectorSurface& overlaps, const SurfaceParams& params, - const GPUVAddr gpu_addr, - const VAddr cpu_addr) { - if (params.target == SurfaceTarget::Texture3D) { - bool failed = false; - if (params.num_levels > 1) { - // We can't handle mipmaps in 3D textures yet, better fallback to LLE approach - return std::nullopt; - } - TSurface new_surface = GetUncachedSurface(gpu_addr, params); - bool modified = false; - for (auto& surface : overlaps) { - const SurfaceParams& src_params = surface->GetSurfaceParams(); - if (src_params.target != SurfaceTarget::Texture2D) { - failed = true; - break; - } - if (src_params.height != params.height) { - failed = true; - break; - } - if (src_params.block_depth != params.block_depth || - src_params.block_height != params.block_height) { - failed = true; - break; - } - const u32 offset = static_cast<u32>(surface->GetCpuAddr() - cpu_addr); - const auto [x, y, z] = params.GetBlockOffsetXYZ(offset); - modified |= surface->IsModified(); - const CopyParams copy_params(0, 0, 0, 0, 0, z, 0, 0, params.width, params.height, - 1); - ImageCopy(surface, new_surface, copy_params); - } - if (failed) { - return std::nullopt; - } - for (const auto& surface : overlaps) { - Unregister(surface); - } - new_surface->MarkAsModified(modified, Tick()); - Register(new_surface); - auto view = new_surface->GetMainView(); - return {{std::move(new_surface), view}}; - } else { + GPUVAddr gpu_addr, VAddr cpu_addr, + bool preserve_contents) { + if (params.target != SurfaceTarget::Texture3D) { for (const auto& surface : overlaps) { if (!surface->MatchTarget(params.target)) { if (overlaps.size() == 1 && surface->GetCpuAddr() == cpu_addr) { - if (Settings::values.use_accurate_gpu_emulation) { + if (Settings::IsGPULevelExtreme()) { return std::nullopt; } Unregister(surface); - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } return std::nullopt; } @@ -679,11 +742,60 @@ private: continue; } if (surface->MatchesStructure(params) == MatchStructureResult::FullMatch) { - return {{surface, surface->GetMainView()}}; + return std::make_pair(surface, surface->GetMainView()); + } + } + return InitializeSurface(gpu_addr, params, preserve_contents); + } + + if (params.num_levels > 1) { + // We can't handle mipmaps in 3D textures yet, better fallback to LLE approach + return std::nullopt; + } + + if (overlaps.size() == 1) { + const auto& surface = overlaps[0]; + const SurfaceParams& overlap_params = surface->GetSurfaceParams(); + // Don't attempt to render to textures with more than one level for now + // The texture has to be to the right or the sample address if we want to render to it + if (overlap_params.num_levels == 1 && cpu_addr >= surface->GetCpuAddr()) { + const u32 offset = static_cast<u32>(cpu_addr - surface->GetCpuAddr()); + const u32 slice = std::get<2>(params.GetBlockOffsetXYZ(offset)); + if (slice < overlap_params.depth) { + auto view = surface->Emplace3DView(slice, params.depth, 0, 1); + return std::make_pair(std::move(surface), std::move(view)); } } - return InitializeSurface(gpu_addr, params); } + + TSurface new_surface = GetUncachedSurface(gpu_addr, params); + bool modified = false; + + for (auto& surface : overlaps) { + const SurfaceParams& src_params = surface->GetSurfaceParams(); + if (src_params.target != SurfaceTarget::Texture2D || + src_params.height != params.height || + src_params.block_depth != params.block_depth || + src_params.block_height != params.block_height) { + return std::nullopt; + } + modified |= surface->IsModified(); + + const u32 offset = static_cast<u32>(surface->GetCpuAddr() - cpu_addr); + const u32 slice = std::get<2>(params.GetBlockOffsetXYZ(offset)); + const u32 width = params.width; + const u32 height = params.height; + const CopyParams copy_params(0, 0, 0, 0, 0, slice, 0, 0, width, height, 1); + TryCopyImage(surface, new_surface, copy_params); + } + for (const auto& surface : overlaps) { + Unregister(surface); + } + new_surface->MarkAsModified(modified, Tick()); + Register(new_surface); + + TView view = new_surface->GetMainView(); + return std::make_pair(std::move(new_surface), std::move(view)); } /** @@ -705,10 +817,13 @@ private: * * @param gpu_addr The starting address of the candidate surface. * @param params The parameters on the candidate surface. + * @param preserve_contents Indicates that the new surface should be loaded from memory or + * left blank. * @param is_render Whether or not the surface is a render target. **/ std::pair<TSurface, TView> GetSurface(const GPUVAddr gpu_addr, const VAddr cpu_addr, - const SurfaceParams& params, bool is_render) { + const SurfaceParams& params, bool preserve_contents, + bool is_render) { // Step 1 // Check Level 1 Cache for a fast structural match. If candidate surface // matches at certain level we are pretty much done. @@ -716,8 +831,9 @@ private: TSurface& current_surface = iter->second; const auto topological_result = current_surface->MatchesTopology(params); if (topological_result != MatchTopologyResult::FullMatch) { - std::vector<TSurface> overlaps{current_surface}; - return RecycleSurface(overlaps, params, gpu_addr, topological_result); + VectorSurface overlaps{current_surface}; + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + topological_result); } const auto struct_result = current_surface->MatchesStructure(params); @@ -742,7 +858,7 @@ private: // If none are found, we are done. we just load the surface and create it. if (overlaps.empty()) { - return InitializeSurface(gpu_addr, params); + return InitializeSurface(gpu_addr, params, preserve_contents); } // Step 3 @@ -752,13 +868,15 @@ private: for (const auto& surface : overlaps) { const auto topological_result = surface->MatchesTopology(params); if (topological_result != MatchTopologyResult::FullMatch) { - return RecycleSurface(overlaps, params, gpu_addr, topological_result); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + topological_result); } } - // Check if it's a 3D texture + // Manage 3D textures if (params.block_depth > 0) { - auto surface = Manage3DSurfaces(overlaps, params, gpu_addr, cpu_addr); + auto surface = + Manage3DSurfaces(overlaps, params, gpu_addr, cpu_addr, preserve_contents); if (surface) { return *surface; } @@ -771,14 +889,12 @@ private: // two things either the candidate surface is a supertexture of the overlap // or they don't match in any known way. if (!current_surface->IsInside(gpu_addr, gpu_addr + candidate_size)) { - if (current_surface->GetGpuAddr() == gpu_addr) { - std::optional<std::pair<TSurface, TView>> view = - TryReconstructSurface(overlaps, params, gpu_addr); - if (view) { - return *view; - } + const std::optional view = TryReconstructSurface(overlaps, params, gpu_addr); + if (view) { + return *view; } - return RecycleSurface(overlaps, params, gpu_addr, MatchTopologyResult::FullMatch); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + MatchTopologyResult::FullMatch); } // Now we check if the candidate is a mipmap/layer of the overlap std::optional<TView> view = @@ -802,7 +918,7 @@ private: pair.first->EmplaceView(params, gpu_addr, candidate_size); if (mirage_view) return {pair.first, *mirage_view}; - return RecycleSurface(overlaps, params, gpu_addr, + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, MatchTopologyResult::FullMatch); } return {current_surface, *view}; @@ -818,7 +934,8 @@ private: } } // We failed all the tests, recycle the overlaps into a new texture. - return RecycleSurface(overlaps, params, gpu_addr, MatchTopologyResult::FullMatch); + return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, + MatchTopologyResult::FullMatch); } /** @@ -831,8 +948,7 @@ private: * @param params The parameters on the candidate surface. **/ Deduction DeduceSurface(const GPUVAddr gpu_addr, const SurfaceParams& params) { - const std::optional<VAddr> cpu_addr = - system.GPU().MemoryManager().GpuToCpuAddress(gpu_addr); + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); if (!cpu_addr) { Deduction result{}; @@ -892,7 +1008,9 @@ private: params.target = target; params.is_tiled = false; params.srgb_conversion = false; - params.is_layered = false; + params.is_layered = + target == SurfaceTarget::Texture1DArray || target == SurfaceTarget::Texture2DArray || + target == SurfaceTarget::TextureCubemap || target == SurfaceTarget::TextureCubeArray; params.block_width = 0; params.block_height = 0; params.block_depth = 0; @@ -906,7 +1024,7 @@ private: params.pitch = 4; params.num_levels = 1; params.emulated_levels = 1; - params.pixel_format = VideoCore::Surface::PixelFormat::R8U; + params.pixel_format = VideoCore::Surface::PixelFormat::R8_UNORM; params.type = VideoCore::Surface::SurfaceType::ColorTexture; auto surface = CreateSurface(0ULL, params); invalid_memory.resize(surface->GetHostSizeInBytes(), 0U); @@ -929,7 +1047,7 @@ private: void DeduceBestBlit(SurfaceParams& src_params, SurfaceParams& dst_params, const GPUVAddr src_gpu_addr, const GPUVAddr dst_gpu_addr) { auto deduced_src = DeduceSurface(src_gpu_addr, src_params); - auto deduced_dst = DeduceSurface(src_gpu_addr, src_params); + auto deduced_dst = DeduceSurface(dst_gpu_addr, dst_params); if (deduced_src.Failed() || deduced_dst.Failed()) { return; } @@ -976,10 +1094,10 @@ private: } std::pair<TSurface, TView> InitializeSurface(GPUVAddr gpu_addr, const SurfaceParams& params, - bool do_load = true) { + bool preserve_contents) { auto new_surface{GetUncachedSurface(gpu_addr, params)}; Register(new_surface); - if (do_load) { + if (preserve_contents) { LoadSurface(new_surface); } return {new_surface, new_surface->GetMainView()}; @@ -987,7 +1105,7 @@ private: void LoadSurface(const TSurface& surface) { staging_cache.GetBuffer(0).resize(surface->GetHostSizeInBytes()); - surface->LoadBuffer(system.GPU().MemoryManager(), staging_cache); + surface->LoadBuffer(gpu_memory, staging_cache); surface->UploadTexture(staging_cache.GetBuffer(0)); surface->MarkAsModified(false, Tick()); } @@ -998,7 +1116,7 @@ private: } staging_cache.GetBuffer(0).resize(surface->GetHostSizeInBytes()); surface->DownloadTexture(staging_cache.GetBuffer(0)); - surface->FlushBuffer(system.GPU().MemoryManager(), staging_cache); + surface->FlushBuffer(gpu_memory, staging_cache); surface->MarkAsModified(false, Tick()); } @@ -1025,23 +1143,25 @@ private: } } - std::vector<TSurface> GetSurfacesInRegion(const VAddr cpu_addr, const std::size_t size) { + VectorSurface GetSurfacesInRegion(const VAddr cpu_addr, const std::size_t size) { if (size == 0) { return {}; } const VAddr cpu_addr_end = cpu_addr + size; - VAddr start = cpu_addr >> registry_page_bits; const VAddr end = (cpu_addr_end - 1) >> registry_page_bits; - std::vector<TSurface> surfaces; - while (start <= end) { - std::vector<TSurface>& list = registry[start]; - for (auto& surface : list) { - if (!surface->IsPicked() && surface->Overlaps(cpu_addr, cpu_addr_end)) { - surface->MarkAsPicked(true); - surfaces.push_back(surface); + VectorSurface surfaces; + for (VAddr start = cpu_addr >> registry_page_bits; start <= end; ++start) { + const auto it = registry.find(start); + if (it == registry.end()) { + continue; + } + for (auto& surface : it->second) { + if (surface->IsPicked() || !surface->Overlaps(cpu_addr, cpu_addr_end)) { + continue; } + surface->MarkAsPicked(true); + surfaces.push_back(surface); } - start++; } for (auto& surface : surfaces) { surface->MarkAsPicked(false); @@ -1066,6 +1186,19 @@ private: return {}; } + /// Try to do an image copy logging when formats are incompatible. + void TryCopyImage(TSurface& src, TSurface& dst, const CopyParams& copy) { + const SurfaceParams& src_params = src->GetSurfaceParams(); + const SurfaceParams& dst_params = dst->GetSurfaceParams(); + if (!format_compatibility.TestCopy(src_params.pixel_format, dst_params.pixel_format)) { + LOG_ERROR(HW_GPU, "Illegal copy between formats={{{}, {}}}", + static_cast<int>(dst_params.pixel_format), + static_cast<int>(src_params.pixel_format)); + return; + } + ImageCopy(src, dst, copy); + } + constexpr PixelFormat GetSiblingFormat(PixelFormat format) const { return siblings_table[static_cast<std::size_t>(format)]; } @@ -1073,7 +1206,7 @@ private: /// Returns true the shader sampler entry is compatible with the TIC texture type. static bool IsTypeCompatible(Tegra::Texture::TextureType tic_type, const VideoCommon::Shader::Sampler& entry) { - const auto shader_type = entry.GetType(); + const auto shader_type = entry.type; switch (tic_type) { case Tegra::Texture::TextureType::Texture1D: case Tegra::Texture::TextureType::Texture1DArray: @@ -1094,7 +1227,7 @@ private: if (shader_type == Tegra::Shader::TextureType::TextureCube) { return true; } - return shader_type == Tegra::Shader::TextureType::Texture2D && entry.IsArray(); + return shader_type == Tegra::Shader::TextureType::Texture2D && entry.is_array; } UNREACHABLE(); return true; @@ -1105,9 +1238,19 @@ private: TView view; }; + void AsyncFlushSurface(TSurface& surface) { + if (!uncommitted_flushes) { + uncommitted_flushes = std::make_shared<std::list<TSurface>>(); + } + uncommitted_flushes->push_back(surface); + } + VideoCore::RasterizerInterface& rasterizer; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::MemoryManager& gpu_memory; FormatLookupTable format_lookup_table; + FormatCompatibility format_compatibility; u64 ticks{}; @@ -1149,6 +1292,11 @@ private: std::unordered_map<u32, TSurface> invalid_cache; std::vector<u8> invalid_memory; + std::list<TSurface> marked_for_unregister; + + std::shared_ptr<std::list<TSurface>> uncommitted_flushes{}; + std::list<std::shared_ptr<std::list<TSurface>>> committed_flushes; + StagingCache staging_cache; std::recursive_mutex mutex; }; diff --git a/src/video_core/textures/convert.cpp b/src/video_core/textures/convert.cpp index f3efa7eb0..962921483 100644 --- a/src/video_core/textures/convert.cpp +++ b/src/video_core/textures/convert.cpp @@ -35,7 +35,7 @@ void SwapS8Z24ToZ24S8(u8* data, u32 width, u32 height) { S8Z24 s8z24_pixel{}; Z24S8 z24s8_pixel{}; constexpr auto bpp{ - VideoCore::Surface::GetBytesPerPixel(VideoCore::Surface::PixelFormat::S8Z24)}; + VideoCore::Surface::GetBytesPerPixel(VideoCore::Surface::PixelFormat::S8_UINT_D24_UNORM)}; for (std::size_t y = 0; y < height; ++y) { for (std::size_t x = 0; x < width; ++x) { const std::size_t offset{bpp * (y * width + x)}; @@ -73,7 +73,7 @@ void ConvertFromGuestToHost(u8* in_data, u8* out_data, PixelFormat pixel_format, in_data, width, height, depth, block_width, block_height); std::copy(rgba8_data.begin(), rgba8_data.end(), out_data); - } else if (convert_s8z24 && pixel_format == PixelFormat::S8Z24) { + } else if (convert_s8z24 && pixel_format == PixelFormat::S8_UINT_D24_UNORM) { Tegra::Texture::ConvertS8Z24ToZ24S8(in_data, width, height); } } @@ -85,7 +85,7 @@ void ConvertFromHostToGuest(u8* data, PixelFormat pixel_format, u32 width, u32 h static_cast<u32>(pixel_format)); UNREACHABLE(); - } else if (convert_s8z24 && pixel_format == PixelFormat::S8Z24) { + } else if (convert_s8z24 && pixel_format == PixelFormat::S8_UINT_D24_UNORM) { Tegra::Texture::ConvertZ24S8ToS8Z24(data, width, height); } } diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index 7df5f1452..16d46a018 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -6,11 +6,13 @@ #include <cstring> #include "common/alignment.h" #include "common/assert.h" +#include "common/bit_util.h" #include "video_core/gpu.h" #include "video_core/textures/decoders.h" #include "video_core/textures/texture.h" namespace Tegra::Texture { +namespace { /** * This table represents the internal swizzle of a gob, @@ -36,20 +38,10 @@ struct alignas(64) SwizzleTable { std::array<std::array<u16, M>, N> values{}; }; -constexpr u32 gob_size_x_shift = 6; -constexpr u32 gob_size_y_shift = 3; -constexpr u32 gob_size_z_shift = 0; -constexpr u32 gob_size_shift = gob_size_x_shift + gob_size_y_shift + gob_size_z_shift; +constexpr u32 FAST_SWIZZLE_ALIGN = 16; -constexpr u32 gob_size_x = 1U << gob_size_x_shift; -constexpr u32 gob_size_y = 1U << gob_size_y_shift; -constexpr u32 gob_size_z = 1U << gob_size_z_shift; -constexpr u32 gob_size = 1U << gob_size_shift; - -constexpr u32 fast_swizzle_align = 16; - -constexpr auto legacy_swizzle_table = SwizzleTable<gob_size_y, gob_size_x, gob_size_z>(); -constexpr auto fast_swizzle_table = SwizzleTable<gob_size_y, 4, fast_swizzle_align>(); +constexpr auto LEGACY_SWIZZLE_TABLE = SwizzleTable<GOB_SIZE_X, GOB_SIZE_X, GOB_SIZE_Z>(); +constexpr auto FAST_SWIZZLE_TABLE = SwizzleTable<GOB_SIZE_Y, 4, FAST_SWIZZLE_ALIGN>(); /** * This function manages ALL the GOBs(Group of Bytes) Inside a single block. @@ -68,17 +60,17 @@ void PreciseProcessBlock(u8* const swizzled_data, u8* const unswizzled_data, con u32 y_address = z_address; u32 pixel_base = layer_z * z + y_start * stride_x; for (u32 y = y_start; y < y_end; y++) { - const auto& table = legacy_swizzle_table[y % gob_size_y]; + const auto& table = LEGACY_SWIZZLE_TABLE[y % GOB_SIZE_Y]; for (u32 x = x_start; x < x_end; x++) { - const u32 swizzle_offset{y_address + table[x * bytes_per_pixel % gob_size_x]}; + const u32 swizzle_offset{y_address + table[x * bytes_per_pixel % GOB_SIZE_X]}; const u32 pixel_index{x * out_bytes_per_pixel + pixel_base}; data_ptrs[unswizzle] = swizzled_data + swizzle_offset; data_ptrs[!unswizzle] = unswizzled_data + pixel_index; std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel); } pixel_base += stride_x; - if ((y + 1) % gob_size_y == 0) - y_address += gob_size; + if ((y + 1) % GOB_SIZE_Y == 0) + y_address += GOB_SIZE; } z_address += xy_block_size; } @@ -103,18 +95,18 @@ void FastProcessBlock(u8* const swizzled_data, u8* const unswizzled_data, const u32 y_address = z_address; u32 pixel_base = layer_z * z + y_start * stride_x; for (u32 y = y_start; y < y_end; y++) { - const auto& table = fast_swizzle_table[y % gob_size_y]; - for (u32 xb = x_startb; xb < x_endb; xb += fast_swizzle_align) { - const u32 swizzle_offset{y_address + table[(xb / fast_swizzle_align) % 4]}; + const auto& table = FAST_SWIZZLE_TABLE[y % GOB_SIZE_Y]; + for (u32 xb = x_startb; xb < x_endb; xb += FAST_SWIZZLE_ALIGN) { + const u32 swizzle_offset{y_address + table[(xb / FAST_SWIZZLE_ALIGN) % 4]}; const u32 out_x = xb * out_bytes_per_pixel / bytes_per_pixel; const u32 pixel_index{out_x + pixel_base}; data_ptrs[unswizzle ? 1 : 0] = swizzled_data + swizzle_offset; data_ptrs[unswizzle ? 0 : 1] = unswizzled_data + pixel_index; - std::memcpy(data_ptrs[0], data_ptrs[1], fast_swizzle_align); + std::memcpy(data_ptrs[0], data_ptrs[1], FAST_SWIZZLE_ALIGN); } pixel_base += stride_x; - if ((y + 1) % gob_size_y == 0) - y_address += gob_size; + if ((y + 1) % GOB_SIZE_Y == 0) + y_address += GOB_SIZE; } z_address += xy_block_size; } @@ -137,9 +129,9 @@ void SwizzledData(u8* const swizzled_data, u8* const unswizzled_data, const bool auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); }; const u32 stride_x = width * out_bytes_per_pixel; const u32 layer_z = height * stride_x; - const u32 gob_elements_x = gob_size_x / bytes_per_pixel; - constexpr u32 gob_elements_y = gob_size_y; - constexpr u32 gob_elements_z = gob_size_z; + const u32 gob_elements_x = GOB_SIZE_X / bytes_per_pixel; + constexpr u32 gob_elements_y = GOB_SIZE_Y; + constexpr u32 gob_elements_z = GOB_SIZE_Z; const u32 block_x_elements = gob_elements_x; const u32 block_y_elements = gob_elements_y * block_height; const u32 block_z_elements = gob_elements_z * block_depth; @@ -147,7 +139,7 @@ void SwizzledData(u8* const swizzled_data, u8* const unswizzled_data, const bool const u32 blocks_on_x = div_ceil(aligned_width, block_x_elements); const u32 blocks_on_y = div_ceil(height, block_y_elements); const u32 blocks_on_z = div_ceil(depth, block_z_elements); - const u32 xy_block_size = gob_size * block_height; + const u32 xy_block_size = GOB_SIZE * block_height; const u32 block_size = xy_block_size * block_depth; u32 tile_offset = 0; for (u32 zb = 0; zb < blocks_on_z; zb++) { @@ -174,12 +166,14 @@ void SwizzledData(u8* const swizzled_data, u8* const unswizzled_data, const bool } } +} // Anonymous namespace + void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel, u32 out_bytes_per_pixel, u8* const swizzled_data, u8* const unswizzled_data, bool unswizzle, u32 block_height, u32 block_depth, u32 width_spacing) { const u32 block_height_size{1U << block_height}; const u32 block_depth_size{1U << block_depth}; - if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % fast_swizzle_align == 0) { + if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % FAST_SWIZZLE_ALIGN == 0) { SwizzledData<true>(swizzled_data, unswizzled_data, unswizzle, width, height, depth, bytes_per_pixel, out_bytes_per_pixel, block_height_size, block_depth_size, width_spacing); @@ -190,53 +184,6 @@ void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel, } } -u32 BytesPerPixel(TextureFormat format) { - switch (format) { - case TextureFormat::DXT1: - case TextureFormat::DXN1: - // In this case a 'pixel' actually refers to a 4x4 tile. - return 8; - case TextureFormat::DXT23: - case TextureFormat::DXT45: - case TextureFormat::DXN2: - case TextureFormat::BC7U: - case TextureFormat::BC6H_UF16: - case TextureFormat::BC6H_SF16: - // In this case a 'pixel' actually refers to a 4x4 tile. - return 16; - case TextureFormat::R32_G32_B32: - return 12; - case TextureFormat::ASTC_2D_4X4: - case TextureFormat::ASTC_2D_5X4: - case TextureFormat::ASTC_2D_8X8: - case TextureFormat::ASTC_2D_8X5: - case TextureFormat::ASTC_2D_10X8: - case TextureFormat::ASTC_2D_5X5: - case TextureFormat::A8R8G8B8: - case TextureFormat::A2B10G10R10: - case TextureFormat::BF10GF11RF11: - case TextureFormat::R32: - case TextureFormat::R16_G16: - return 4; - case TextureFormat::A1B5G5R5: - case TextureFormat::B5G6R5: - case TextureFormat::G8R8: - case TextureFormat::R16: - return 2; - case TextureFormat::R8: - return 1; - case TextureFormat::R16_G16_B16_A16: - return 8; - case TextureFormat::R32_G32_B32_A32: - return 16; - case TextureFormat::R32_G32: - return 8; - default: - UNIMPLEMENTED_MSG("Format not implemented"); - return 1; - } -} - void UnswizzleTexture(u8* const unswizzled_data, u8* address, u32 tile_size_x, u32 tile_size_y, u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth, u32 width_spacing) { @@ -256,47 +203,82 @@ std::vector<u8> UnswizzleTexture(u8* address, u32 tile_size_x, u32 tile_size_y, } void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width, - u32 bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data, + u32 bytes_per_pixel, u8* swizzled_data, const u8* unswizzled_data, u32 block_height_bit, u32 offset_x, u32 offset_y) { const u32 block_height = 1U << block_height_bit; - const u32 image_width_in_gobs{(swizzled_width * bytes_per_pixel + (gob_size_x - 1)) / - gob_size_x}; + const u32 image_width_in_gobs = + (swizzled_width * bytes_per_pixel + (GOB_SIZE_X - 1)) / GOB_SIZE_X; for (u32 line = 0; line < subrect_height; ++line) { const u32 dst_y = line + offset_y; const u32 gob_address_y = - (dst_y / (gob_size_y * block_height)) * gob_size * block_height * image_width_in_gobs + - ((dst_y % (gob_size_y * block_height)) / gob_size_y) * gob_size; - const auto& table = legacy_swizzle_table[dst_y % gob_size_y]; + (dst_y / (GOB_SIZE_Y * block_height)) * GOB_SIZE * block_height * image_width_in_gobs + + ((dst_y % (GOB_SIZE_Y * block_height)) / GOB_SIZE_Y) * GOB_SIZE; + const auto& table = LEGACY_SWIZZLE_TABLE[dst_y % GOB_SIZE_Y]; for (u32 x = 0; x < subrect_width; ++x) { const u32 dst_x = x + offset_x; const u32 gob_address = - gob_address_y + (dst_x * bytes_per_pixel / gob_size_x) * gob_size * block_height; - const u32 swizzled_offset = gob_address + table[(dst_x * bytes_per_pixel) % gob_size_x]; - u8* source_line = unswizzled_data + line * source_pitch + x * bytes_per_pixel; - u8* dest_addr = swizzled_data + swizzled_offset; + gob_address_y + (dst_x * bytes_per_pixel / GOB_SIZE_X) * GOB_SIZE * block_height; + const u32 swizzled_offset = gob_address + table[(dst_x * bytes_per_pixel) % GOB_SIZE_X]; + const u32 unswizzled_offset = line * source_pitch + x * bytes_per_pixel; + const u8* const source_line = unswizzled_data + unswizzled_offset; + u8* const dest_addr = swizzled_data + swizzled_offset; std::memcpy(dest_addr, source_line, bytes_per_pixel); } } } -void UnswizzleSubrect(u32 subrect_width, u32 subrect_height, u32 dest_pitch, u32 swizzled_width, - u32 bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data, - u32 block_height_bit, u32 offset_x, u32 offset_y) { - const u32 block_height = 1U << block_height_bit; - for (u32 line = 0; line < subrect_height; ++line) { - const u32 y2 = line + offset_y; - const u32 gob_address_y = (y2 / (gob_size_y * block_height)) * gob_size * block_height + - ((y2 % (gob_size_y * block_height)) / gob_size_y) * gob_size; - const auto& table = legacy_swizzle_table[y2 % gob_size_y]; - for (u32 x = 0; x < subrect_width; ++x) { - const u32 x2 = (x + offset_x) * bytes_per_pixel; - const u32 gob_address = gob_address_y + (x2 / gob_size_x) * gob_size * block_height; - const u32 swizzled_offset = gob_address + table[x2 % gob_size_x]; - u8* dest_line = unswizzled_data + line * dest_pitch + x * bytes_per_pixel; - u8* source_addr = swizzled_data + swizzled_offset; +void UnswizzleSubrect(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 bytes_per_pixel, + u32 block_height, u32 origin_x, u32 origin_y, u8* output, const u8* input) { + const u32 stride = width * bytes_per_pixel; + const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) / GOB_SIZE_X; + const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height); + + const u32 block_height_mask = (1U << block_height) - 1; + const u32 x_shift = static_cast<u32>(GOB_SIZE_SHIFT) + block_height; + + for (u32 line = 0; line < line_count; ++line) { + const u32 src_y = line + origin_y; + const auto& table = LEGACY_SWIZZLE_TABLE[src_y % GOB_SIZE_Y]; + + const u32 block_y = src_y >> GOB_SIZE_Y_SHIFT; + const u32 src_offset_y = (block_y >> block_height) * block_size + + ((block_y & block_height_mask) << GOB_SIZE_SHIFT); + for (u32 column = 0; column < line_length_in; ++column) { + const u32 src_x = (column + origin_x) * bytes_per_pixel; + const u32 src_offset_x = (src_x >> GOB_SIZE_X_SHIFT) << x_shift; + + const u32 swizzled_offset = src_offset_y + src_offset_x + table[src_x % GOB_SIZE_X]; + const u32 unswizzled_offset = line * pitch + column * bytes_per_pixel; + + std::memcpy(output + unswizzled_offset, input + swizzled_offset, bytes_per_pixel); + } + } +} + +void SwizzleSliceToVoxel(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 height, + u32 bytes_per_pixel, u32 block_height, u32 block_depth, u32 origin_x, + u32 origin_y, u8* output, const u8* input) { + UNIMPLEMENTED_IF(origin_x > 0); + UNIMPLEMENTED_IF(origin_y > 0); - std::memcpy(dest_line, source_addr, bytes_per_pixel); + const u32 stride = width * bytes_per_pixel; + const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) / GOB_SIZE_X; + const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height + block_depth); + + const u32 block_height_mask = (1U << block_height) - 1; + const u32 x_shift = static_cast<u32>(GOB_SIZE_SHIFT) + block_height + block_depth; + + for (u32 line = 0; line < line_count; ++line) { + const auto& table = LEGACY_SWIZZLE_TABLE[line % GOB_SIZE_Y]; + const u32 block_y = line / GOB_SIZE_Y; + const u32 dst_offset_y = + (block_y >> block_height) * block_size + (block_y & block_height_mask) * GOB_SIZE; + for (u32 x = 0; x < line_length_in; ++x) { + const u32 dst_offset = + ((x / GOB_SIZE_X) << x_shift) + dst_offset_y + table[x % GOB_SIZE_X]; + const u32 src_offset = x * bytes_per_pixel + line * pitch; + std::memcpy(output + dst_offset, input + src_offset, bytes_per_pixel); } } } @@ -305,17 +287,17 @@ void SwizzleKepler(const u32 width, const u32 height, const u32 dst_x, const u32 const u32 block_height_bit, const std::size_t copy_size, const u8* source_data, u8* swizzle_data) { const u32 block_height = 1U << block_height_bit; - const u32 image_width_in_gobs{(width + gob_size_x - 1) / gob_size_x}; + const u32 image_width_in_gobs{(width + GOB_SIZE_X - 1) / GOB_SIZE_X}; std::size_t count = 0; for (std::size_t y = dst_y; y < height && count < copy_size; ++y) { const std::size_t gob_address_y = - (y / (gob_size_y * block_height)) * gob_size * block_height * image_width_in_gobs + - ((y % (gob_size_y * block_height)) / gob_size_y) * gob_size; - const auto& table = legacy_swizzle_table[y % gob_size_y]; + (y / (GOB_SIZE_Y * block_height)) * GOB_SIZE * block_height * image_width_in_gobs + + ((y % (GOB_SIZE_Y * block_height)) / GOB_SIZE_Y) * GOB_SIZE; + const auto& table = LEGACY_SWIZZLE_TABLE[y % GOB_SIZE_Y]; for (std::size_t x = dst_x; x < width && count < copy_size; ++x) { const std::size_t gob_address = - gob_address_y + (x / gob_size_x) * gob_size * block_height; - const std::size_t swizzled_offset = gob_address + table[x % gob_size_x]; + gob_address_y + (x / GOB_SIZE_X) * GOB_SIZE * block_height; + const std::size_t swizzled_offset = gob_address + table[x % GOB_SIZE_X]; const u8* source_line = source_data + count; u8* dest_addr = swizzle_data + swizzled_offset; count++; @@ -325,58 +307,30 @@ void SwizzleKepler(const u32 width, const u32 height, const u32 dst_x, const u32 } } -std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat format, u32 width, - u32 height) { - std::vector<u8> rgba_data; - - // TODO(Subv): Implement. - switch (format) { - case TextureFormat::DXT1: - case TextureFormat::DXT23: - case TextureFormat::DXT45: - case TextureFormat::DXN1: - case TextureFormat::DXN2: - case TextureFormat::BC7U: - case TextureFormat::BC6H_UF16: - case TextureFormat::BC6H_SF16: - case TextureFormat::ASTC_2D_4X4: - case TextureFormat::ASTC_2D_8X8: - case TextureFormat::ASTC_2D_5X5: - case TextureFormat::ASTC_2D_10X8: - case TextureFormat::A8R8G8B8: - case TextureFormat::A2B10G10R10: - case TextureFormat::A1B5G5R5: - case TextureFormat::B5G6R5: - case TextureFormat::R8: - case TextureFormat::G8R8: - case TextureFormat::BF10GF11RF11: - case TextureFormat::R32_G32_B32_A32: - case TextureFormat::R32_G32: - case TextureFormat::R32: - case TextureFormat::R16: - case TextureFormat::R16_G16: - case TextureFormat::R32_G32_B32: - // TODO(Subv): For the time being just forward the same data without any decoding. - rgba_data = texture_data; - break; - default: - UNIMPLEMENTED_MSG("Format not implemented"); - break; - } - - return rgba_data; -} - std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth) { if (tiled) { - const u32 aligned_width = Common::AlignBits(width * bytes_per_pixel, gob_size_x_shift); - const u32 aligned_height = Common::AlignBits(height, gob_size_y_shift + block_height); - const u32 aligned_depth = Common::AlignBits(depth, gob_size_z_shift + block_depth); + const u32 aligned_width = Common::AlignBits(width * bytes_per_pixel, GOB_SIZE_X_SHIFT); + const u32 aligned_height = Common::AlignBits(height, GOB_SIZE_Y_SHIFT + block_height); + const u32 aligned_depth = Common::AlignBits(depth, GOB_SIZE_Z_SHIFT + block_depth); return aligned_width * aligned_height * aligned_depth; } else { return width * height * depth * bytes_per_pixel; } } +u64 GetGOBOffset(u32 width, u32 height, u32 dst_x, u32 dst_y, u32 block_height, + u32 bytes_per_pixel) { + auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); }; + const u32 gobs_in_block = 1 << block_height; + const u32 y_blocks = GOB_SIZE_Y << block_height; + const u32 x_per_gob = GOB_SIZE_X / bytes_per_pixel; + const u32 x_blocks = div_ceil(width, x_per_gob); + const u32 block_size = GOB_SIZE * gobs_in_block; + const u32 stride = block_size * x_blocks; + const u32 base = (dst_y / y_blocks) * stride + (dst_x / x_per_gob) * block_size; + const u32 relative_y = dst_y % y_blocks; + return base + (relative_y / GOB_SIZE_Y) * GOB_SIZE; +} + } // namespace Tegra::Texture diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index e5eac3f3b..01e156bc8 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -10,15 +10,15 @@ namespace Tegra::Texture { -// GOBSize constant. Calculated by 64 bytes in x multiplied by 8 y coords, represents -// an small rect of (64/bytes_per_pixel)X8. -inline std::size_t GetGOBSize() { - return 512; -} +constexpr u32 GOB_SIZE_X = 64; +constexpr u32 GOB_SIZE_Y = 8; +constexpr u32 GOB_SIZE_Z = 1; +constexpr u32 GOB_SIZE = GOB_SIZE_X * GOB_SIZE_Y * GOB_SIZE_Z; -inline std::size_t GetGOBSizeShift() { - return 9; -} +constexpr std::size_t GOB_SIZE_X_SHIFT = 6; +constexpr std::size_t GOB_SIZE_Y_SHIFT = 3; +constexpr std::size_t GOB_SIZE_Z_SHIFT = 0; +constexpr std::size_t GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT; /// Unswizzles a swizzled texture without changing its format. void UnswizzleTexture(u8* unswizzled_data, u8* address, u32 tile_size_x, u32 tile_size_y, @@ -38,26 +38,42 @@ void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel, u32 out_bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data, bool unswizzle, u32 block_height, u32 block_depth, u32 width_spacing); -/// Decodes an unswizzled texture into a A8R8G8B8 texture. -std::vector<u8> DecodeTexture(const std::vector<u8>& texture_data, TextureFormat format, u32 width, - u32 height); - /// This function calculates the correct size of a texture depending if it's tiled or not. std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth); /// Copies an untiled subrectangle into a tiled surface. void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width, - u32 bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data, u32 block_height, - u32 offset_x, u32 offset_y); + u32 bytes_per_pixel, u8* swizzled_data, const u8* unswizzled_data, + u32 block_height_bit, u32 offset_x, u32 offset_y); /// Copies a tiled subrectangle into a linear surface. -void UnswizzleSubrect(u32 subrect_width, u32 subrect_height, u32 dest_pitch, u32 swizzled_width, - u32 bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data, u32 block_height, - u32 offset_x, u32 offset_y); +void UnswizzleSubrect(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 bytes_per_pixel, + u32 block_height, u32 origin_x, u32 origin_y, u8* output, const u8* input); + +/// @brief Swizzles a 2D array of pixels into a 3D texture +/// @param line_length_in Number of pixels per line +/// @param line_count Number of lines +/// @param pitch Number of bytes per line +/// @param width Width of the swizzled texture +/// @param height Height of the swizzled texture +/// @param bytes_per_pixel Number of bytes used per pixel +/// @param block_height Block height shift +/// @param block_depth Block depth shift +/// @param origin_x Column offset in pixels of the swizzled texture +/// @param origin_y Row offset in pixels of the swizzled texture +/// @param output Pointer to the pixels of the swizzled texture +/// @param input Pointer to the 2D array of pixels used as input +/// @pre input and output points to an array large enough to hold the number of bytes used +void SwizzleSliceToVoxel(u32 line_length_in, u32 line_count, u32 pitch, u32 width, u32 height, + u32 bytes_per_pixel, u32 block_height, u32 block_depth, u32 origin_x, + u32 origin_y, u8* output, const u8* input); + +void SwizzleKepler(u32 width, u32 height, u32 dst_x, u32 dst_y, u32 block_height, + std::size_t copy_size, const u8* source_data, u8* swizzle_data); -void SwizzleKepler(const u32 width, const u32 height, const u32 dst_x, const u32 dst_y, - const u32 block_height, const std::size_t copy_size, const u8* source_data, - u8* swizzle_data); +/// Obtains the offset of the gob for positions 'dst_x' & 'dst_y' +u64 GetGOBOffset(u32 width, u32 height, u32 dst_x, u32 dst_y, u32 block_height, + u32 bytes_per_pixel); } // namespace Tegra::Texture diff --git a/src/video_core/textures/texture.cpp b/src/video_core/textures/texture.cpp index d1939d744..4171e3ef2 100644 --- a/src/video_core/textures/texture.cpp +++ b/src/video_core/textures/texture.cpp @@ -48,7 +48,7 @@ constexpr std::array<float, 256> SRGB_CONVERSION_LUT = { }; unsigned SettingsMinimumAnisotropy() noexcept { - switch (static_cast<Anisotropy>(Settings::values.max_anisotropy)) { + switch (static_cast<Anisotropy>(Settings::values.max_anisotropy.GetValue())) { default: case Anisotropy::Default: return 1U; diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h index eba05aced..0574fef12 100644 --- a/src/video_core/textures/texture.h +++ b/src/video_core/textures/texture.h @@ -12,10 +12,10 @@ namespace Tegra::Texture { enum class TextureFormat : u32 { - R32_G32_B32_A32 = 0x01, - R32_G32_B32 = 0x02, - R16_G16_B16_A16 = 0x03, - R32_G32 = 0x04, + R32G32B32A32 = 0x01, + R32G32B32 = 0x02, + R16G16B16A16 = 0x03, + R32G32 = 0x04, R32_B24G8 = 0x05, ETC2_RGB = 0x06, X8B8G8R8 = 0x07, @@ -23,19 +23,19 @@ enum class TextureFormat : u32 { A2B10G10R10 = 0x09, ETC2_RGB_PTA = 0x0a, ETC2_RGBA = 0x0b, - R16_G16 = 0x0c, - G8R24 = 0x0d, - G24R8 = 0x0e, + R16G16 = 0x0c, + R24G8 = 0x0d, + R8G24 = 0x0e, R32 = 0x0f, - BC6H_SF16 = 0x10, - BC6H_UF16 = 0x11, + BC6H_SFLOAT = 0x10, + BC6H_UFLOAT = 0x11, A4B4G4R4 = 0x12, A5B5G5R1 = 0x13, A1B5G5R5 = 0x14, B5G6R5 = 0x15, B6G5R5 = 0x16, - BC7U = 0x17, - G8R8 = 0x18, + BC7 = 0x17, + R8G8 = 0x18, EAC = 0x19, EACX2 = 0x1a, R16 = 0x1b, @@ -43,23 +43,23 @@ enum class TextureFormat : u32 { R8 = 0x1d, G4R4 = 0x1e, R1 = 0x1f, - E5B9G9R9_SHAREDEXP = 0x20, - BF10GF11RF11 = 0x21, + E5B9G9R9 = 0x20, + B10G11R11 = 0x21, G8B8G8R8 = 0x22, B8G8R8G8 = 0x23, - DXT1 = 0x24, - DXT23 = 0x25, - DXT45 = 0x26, - DXN1 = 0x27, - DXN2 = 0x28, - S8Z24 = 0x29, + BC1_RGBA = 0x24, + BC2 = 0x25, + BC3 = 0x26, + BC4 = 0x27, + BC5 = 0x28, + S8D24 = 0x29, X8Z24 = 0x2a, - Z24S8 = 0x2b, + D24S8 = 0x2b, X4V4Z24__COV4R4V = 0x2c, X4V4Z24__COV8R8V = 0x2d, V8Z24__COV4R12V = 0x2e, - ZF32 = 0x2f, - ZF32_X24S8 = 0x30, + D32 = 0x2f, + D32S8 = 0x30, X8Z24_X20V4S8__COV4R4V = 0x31, X8Z24_X20V4S8__COV8R8V = 0x32, ZF32_X20V4X8__COV4R4V = 0x33, @@ -69,7 +69,7 @@ enum class TextureFormat : u32 { X8Z24_X16V8S8__COV4R12V = 0x37, ZF32_X16V8X8__COV4R12V = 0x38, ZF32_X16V8S8__COV4R12V = 0x39, - Z16 = 0x3a, + D16 = 0x3a, V8Z24__COV8R24V = 0x3b, X8Z24_X16V8S8__COV8R24V = 0x3c, ZF32_X16V8X8__COV8R24V = 0x3d, @@ -375,7 +375,4 @@ struct FullTextureInfo { TSCEntry tsc; }; -/// Returns the number of bytes per pixel of the input texture format. -u32 BytesPerPixel(TextureFormat format); - } // namespace Tegra::Texture diff --git a/src/video_core/video_core.cpp b/src/video_core/video_core.cpp index f60bdc60a..dd5cee4a1 100644 --- a/src/video_core/video_core.cpp +++ b/src/video_core/video_core.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <memory> + #include "common/logging/log.h" #include "core/core.h" #include "core/settings.h" @@ -16,43 +17,56 @@ #include "video_core/video_core.h" namespace { -std::unique_ptr<VideoCore::RendererBase> CreateRenderer(Core::Frontend::EmuWindow& emu_window, - Core::System& system, - Core::Frontend::GraphicsContext& context) { - switch (Settings::values.renderer_backend) { + +std::unique_ptr<VideoCore::RendererBase> CreateRenderer( + Core::System& system, Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu, + std::unique_ptr<Core::Frontend::GraphicsContext> context) { + auto& telemetry_session = system.TelemetrySession(); + auto& cpu_memory = system.Memory(); + + switch (Settings::values.renderer_backend.GetValue()) { case Settings::RendererBackend::OpenGL: - return std::make_unique<OpenGL::RendererOpenGL>(emu_window, system, context); + return std::make_unique<OpenGL::RendererOpenGL>(telemetry_session, emu_window, cpu_memory, + gpu, std::move(context)); #ifdef HAS_VULKAN case Settings::RendererBackend::Vulkan: - return std::make_unique<Vulkan::RendererVulkan>(emu_window, system); + return std::make_unique<Vulkan::RendererVulkan>(telemetry_session, emu_window, cpu_memory, + gpu, std::move(context)); #endif default: return nullptr; } } + } // Anonymous namespace namespace VideoCore { std::unique_ptr<Tegra::GPU> CreateGPU(Core::Frontend::EmuWindow& emu_window, Core::System& system) { + std::unique_ptr<Tegra::GPU> gpu; + const bool use_nvdec = Settings::values.use_nvdec_emulation.GetValue(); + if (Settings::values.use_asynchronous_gpu_emulation.GetValue()) { + gpu = std::make_unique<VideoCommon::GPUAsynch>(system, use_nvdec); + } else { + gpu = std::make_unique<VideoCommon::GPUSynch>(system, use_nvdec); + } + auto context = emu_window.CreateSharedContext(); const auto scope = context->Acquire(); - auto renderer = CreateRenderer(emu_window, system, *context); + + auto renderer = CreateRenderer(system, emu_window, *gpu, std::move(context)); if (!renderer->Init()) { return nullptr; } - if (Settings::values.use_asynchronous_gpu_emulation) { - return std::make_unique<VideoCommon::GPUAsynch>(system, std::move(renderer), - std::move(context)); - } - return std::make_unique<VideoCommon::GPUSynch>(system, std::move(renderer), std::move(context)); + gpu->BindRenderer(std::move(renderer)); + return gpu; } u16 GetResolutionScaleFactor(const RendererBase& renderer) { return static_cast<u16>( - Settings::values.resolution_factor != 0 - ? Settings::values.resolution_factor + Settings::values.resolution_factor.GetValue() != 0 + ? Settings::values.resolution_factor.GetValue() : renderer.GetRenderWindow().GetFramebufferLayout().GetScalingRatio()); } diff --git a/src/web_service/CMakeLists.txt b/src/web_service/CMakeLists.txt index 01f2d129d..ae85a72ea 100644 --- a/src/web_service/CMakeLists.txt +++ b/src/web_service/CMakeLists.txt @@ -5,12 +5,8 @@ add_library(web_service STATIC verify_login.h web_backend.cpp web_backend.h + web_result.h ) create_target_directory_groups(web_service) - -get_directory_property(OPENSSL_LIBS - DIRECTORY ${PROJECT_SOURCE_DIR}/externals/libressl - DEFINITION OPENSSL_LIBS) -target_compile_definitions(web_service PRIVATE -DCPPHTTPLIB_OPENSSL_SUPPORT) -target_link_libraries(web_service PRIVATE common json-headers ${OPENSSL_LIBS} httplib lurlparser) +target_link_libraries(web_service PRIVATE common nlohmann_json::nlohmann_json httplib) diff --git a/src/web_service/telemetry_json.cpp b/src/web_service/telemetry_json.cpp index 7538389bf..6215c914f 100644 --- a/src/web_service/telemetry_json.cpp +++ b/src/web_service/telemetry_json.cpp @@ -2,14 +2,16 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <json.hpp> +#include <nlohmann/json.hpp> #include "common/detached_tasks.h" -#include "common/web_result.h" #include "web_service/telemetry_json.h" #include "web_service/web_backend.h" +#include "web_service/web_result.h" namespace WebService { +namespace Telemetry = Common::Telemetry; + struct TelemetryJson::Impl { Impl(std::string host, std::string username, std::string token) : host{std::move(host)}, username{std::move(username)}, token{std::move(token)} {} @@ -123,7 +125,7 @@ bool TelemetryJson::SubmitTestcase() { Client client(impl->host, impl->username, impl->token); auto value = client.PostJson("/gamedb/testcase", content, false); - return value.result_code == Common::WebResult::Code::Success; + return value.result_code == WebResult::Code::Success; } } // namespace WebService diff --git a/src/web_service/telemetry_json.h b/src/web_service/telemetry_json.h index dfd202829..df51e00f8 100644 --- a/src/web_service/telemetry_json.h +++ b/src/web_service/telemetry_json.h @@ -14,25 +14,25 @@ namespace WebService { * Implementation of VisitorInterface that serialized telemetry into JSON, and submits it to the * yuzu web service */ -class TelemetryJson : public Telemetry::VisitorInterface { +class TelemetryJson : public Common::Telemetry::VisitorInterface { public: TelemetryJson(std::string host, std::string username, std::string token); ~TelemetryJson() override; - void Visit(const Telemetry::Field<bool>& field) override; - void Visit(const Telemetry::Field<double>& field) override; - void Visit(const Telemetry::Field<float>& field) override; - void Visit(const Telemetry::Field<u8>& field) override; - void Visit(const Telemetry::Field<u16>& field) override; - void Visit(const Telemetry::Field<u32>& field) override; - void Visit(const Telemetry::Field<u64>& field) override; - void Visit(const Telemetry::Field<s8>& field) override; - void Visit(const Telemetry::Field<s16>& field) override; - void Visit(const Telemetry::Field<s32>& field) override; - void Visit(const Telemetry::Field<s64>& field) override; - void Visit(const Telemetry::Field<std::string>& field) override; - void Visit(const Telemetry::Field<const char*>& field) override; - void Visit(const Telemetry::Field<std::chrono::microseconds>& field) override; + void Visit(const Common::Telemetry::Field<bool>& field) override; + void Visit(const Common::Telemetry::Field<double>& field) override; + void Visit(const Common::Telemetry::Field<float>& field) override; + void Visit(const Common::Telemetry::Field<u8>& field) override; + void Visit(const Common::Telemetry::Field<u16>& field) override; + void Visit(const Common::Telemetry::Field<u32>& field) override; + void Visit(const Common::Telemetry::Field<u64>& field) override; + void Visit(const Common::Telemetry::Field<s8>& field) override; + void Visit(const Common::Telemetry::Field<s16>& field) override; + void Visit(const Common::Telemetry::Field<s32>& field) override; + void Visit(const Common::Telemetry::Field<s64>& field) override; + void Visit(const Common::Telemetry::Field<std::string>& field) override; + void Visit(const Common::Telemetry::Field<const char*>& field) override; + void Visit(const Common::Telemetry::Field<std::chrono::microseconds>& field) override; void Complete() override; bool SubmitTestcase() override; diff --git a/src/web_service/verify_login.cpp b/src/web_service/verify_login.cpp index ca4b43b93..ceb55ca6b 100644 --- a/src/web_service/verify_login.cpp +++ b/src/web_service/verify_login.cpp @@ -2,10 +2,10 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <json.hpp> -#include "common/web_result.h" +#include <nlohmann/json.hpp> #include "web_service/verify_login.h" #include "web_service/web_backend.h" +#include "web_service/web_result.h" namespace WebService { diff --git a/src/web_service/web_backend.cpp b/src/web_service/web_backend.cpp index 737ffe409..67183e64c 100644 --- a/src/web_service/web_backend.cpp +++ b/src/web_service/web_backend.cpp @@ -6,21 +6,18 @@ #include <cstdlib> #include <mutex> #include <string> -#include <LUrlParser.h> + #include <fmt/format.h> #include <httplib.h> -#include "common/common_types.h" + #include "common/logging/log.h" -#include "common/web_result.h" #include "web_service/web_backend.h" +#include "web_service/web_result.h" namespace WebService { constexpr std::array<const char, 1> API_VERSION{'1'}; -constexpr int HTTP_PORT = 80; -constexpr int HTTPS_PORT = 443; - constexpr std::size_t TIMEOUT_SECONDS = 30; struct Client::Impl { @@ -33,17 +30,16 @@ struct Client::Impl { } /// A generic function handles POST, GET and DELETE request together - Common::WebResult GenericRequest(const std::string& method, const std::string& path, - const std::string& data, bool allow_anonymous, - const std::string& accept) { + WebResult GenericRequest(const std::string& method, const std::string& path, + const std::string& data, bool allow_anonymous, + const std::string& accept) { if (jwt.empty()) { UpdateJWT(); } if (jwt.empty() && !allow_anonymous) { LOG_ERROR(WebService, "Credentials must be provided for authenticated requests"); - return Common::WebResult{Common::WebResult::Code::CredentialsMissing, - "Credentials needed"}; + return WebResult{WebResult::Code::CredentialsMissing, "Credentials needed", ""}; } auto result = GenericRequest(method, path, data, accept, jwt); @@ -62,33 +58,22 @@ struct Client::Impl { * username + token is used if jwt is empty but username and token are * not empty anonymous if all of jwt, username and token are empty */ - Common::WebResult GenericRequest(const std::string& method, const std::string& path, - const std::string& data, const std::string& accept, - const std::string& jwt = "", const std::string& username = "", - const std::string& token = "") { + WebResult GenericRequest(const std::string& method, const std::string& path, + const std::string& data, const std::string& accept, + const std::string& jwt = "", const std::string& username = "", + const std::string& token = "") { if (cli == nullptr) { - auto parsedUrl = LUrlParser::clParseURL::ParseURL(host); - int port; - if (parsedUrl.m_Scheme == "http") { - if (!parsedUrl.GetPort(&port)) { - port = HTTP_PORT; - } - cli = std::make_unique<httplib::Client>(parsedUrl.m_Host.c_str(), port); - } else if (parsedUrl.m_Scheme == "https") { - if (!parsedUrl.GetPort(&port)) { - port = HTTPS_PORT; - } - cli = std::make_unique<httplib::SSLClient>(parsedUrl.m_Host.c_str(), port); - } else { - LOG_ERROR(WebService, "Bad URL scheme {}", parsedUrl.m_Scheme); - return Common::WebResult{Common::WebResult::Code::InvalidURL, "Bad URL scheme"}; - } + cli = std::make_unique<httplib::Client>(host.c_str()); } - if (cli == nullptr) { - LOG_ERROR(WebService, "Invalid URL {}", host + path); - return Common::WebResult{Common::WebResult::Code::InvalidURL, "Invalid URL"}; + + if (!cli->is_valid()) { + LOG_ERROR(WebService, "Client is invalid, skipping request!"); + return {}; } - cli->set_timeout_sec(TIMEOUT_SECONDS); + + cli->set_connection_timeout(TIMEOUT_SECONDS); + cli->set_read_timeout(TIMEOUT_SECONDS); + cli->set_write_timeout(TIMEOUT_SECONDS); httplib::Headers params; if (!jwt.empty()) { @@ -106,7 +91,7 @@ struct Client::Impl { std::string(API_VERSION.begin(), API_VERSION.end())); if (method != "GET") { params.emplace(std::string("Content-Type"), std::string("application/json")); - }; + } httplib::Request request; request.method = method; @@ -118,29 +103,28 @@ struct Client::Impl { if (!cli->send(request, response)) { LOG_ERROR(WebService, "{} to {} returned null", method, host + path); - return Common::WebResult{Common::WebResult::Code::LibError, "Null response"}; + return WebResult{WebResult::Code::LibError, "Null response", ""}; } if (response.status >= 400) { LOG_ERROR(WebService, "{} to {} returned error status code: {}", method, host + path, response.status); - return Common::WebResult{Common::WebResult::Code::HttpError, - std::to_string(response.status)}; + return WebResult{WebResult::Code::HttpError, std::to_string(response.status), ""}; } auto content_type = response.headers.find("content-type"); if (content_type == response.headers.end()) { LOG_ERROR(WebService, "{} to {} returned no content", method, host + path); - return Common::WebResult{Common::WebResult::Code::WrongContent, ""}; + return WebResult{WebResult::Code::WrongContent, "", ""}; } if (content_type->second.find(accept) == std::string::npos) { LOG_ERROR(WebService, "{} to {} returned wrong content: {}", method, host + path, content_type->second); - return Common::WebResult{Common::WebResult::Code::WrongContent, "Wrong content"}; + return WebResult{WebResult::Code::WrongContent, "Wrong content", ""}; } - return Common::WebResult{Common::WebResult::Code::Success, "", response.body}; + return WebResult{WebResult::Code::Success, "", response.body}; } // Retrieve a new JWT from given username and token @@ -150,7 +134,7 @@ struct Client::Impl { } auto result = GenericRequest("POST", "/jwt/internal", "", "text/html", "", username, token); - if (result.result_code != Common::WebResult::Code::Success) { + if (result.result_code != WebResult::Code::Success) { LOG_ERROR(WebService, "UpdateJWT failed"); } else { std::lock_guard lock{jwt_cache.mutex}; @@ -180,29 +164,28 @@ Client::Client(std::string host, std::string username, std::string token) Client::~Client() = default; -Common::WebResult Client::PostJson(const std::string& path, const std::string& data, - bool allow_anonymous) { +WebResult Client::PostJson(const std::string& path, const std::string& data, bool allow_anonymous) { return impl->GenericRequest("POST", path, data, allow_anonymous, "application/json"); } -Common::WebResult Client::GetJson(const std::string& path, bool allow_anonymous) { +WebResult Client::GetJson(const std::string& path, bool allow_anonymous) { return impl->GenericRequest("GET", path, "", allow_anonymous, "application/json"); } -Common::WebResult Client::DeleteJson(const std::string& path, const std::string& data, - bool allow_anonymous) { +WebResult Client::DeleteJson(const std::string& path, const std::string& data, + bool allow_anonymous) { return impl->GenericRequest("DELETE", path, data, allow_anonymous, "application/json"); } -Common::WebResult Client::GetPlain(const std::string& path, bool allow_anonymous) { +WebResult Client::GetPlain(const std::string& path, bool allow_anonymous) { return impl->GenericRequest("GET", path, "", allow_anonymous, "text/plain"); } -Common::WebResult Client::GetImage(const std::string& path, bool allow_anonymous) { +WebResult Client::GetImage(const std::string& path, bool allow_anonymous) { return impl->GenericRequest("GET", path, "", allow_anonymous, "image/png"); } -Common::WebResult Client::GetExternalJWT(const std::string& audience) { +WebResult Client::GetExternalJWT(const std::string& audience) { return impl->GenericRequest("POST", fmt::format("/jwt/external/{}", audience), "", false, "text/html"); } diff --git a/src/web_service/web_backend.h b/src/web_service/web_backend.h index 04121f17e..81f58583c 100644 --- a/src/web_service/web_backend.h +++ b/src/web_service/web_backend.h @@ -7,12 +7,10 @@ #include <memory> #include <string> -namespace Common { -struct WebResult; -} - namespace WebService { +struct WebResult; + class Client { public: Client(std::string host, std::string username, std::string token); @@ -25,8 +23,7 @@ public: * @param allow_anonymous If true, allow anonymous unauthenticated requests. * @return the result of the request. */ - Common::WebResult PostJson(const std::string& path, const std::string& data, - bool allow_anonymous); + WebResult PostJson(const std::string& path, const std::string& data, bool allow_anonymous); /** * Gets JSON from the specified path. @@ -34,7 +31,7 @@ public: * @param allow_anonymous If true, allow anonymous unauthenticated requests. * @return the result of the request. */ - Common::WebResult GetJson(const std::string& path, bool allow_anonymous); + WebResult GetJson(const std::string& path, bool allow_anonymous); /** * Deletes JSON to the specified path. @@ -43,8 +40,7 @@ public: * @param allow_anonymous If true, allow anonymous unauthenticated requests. * @return the result of the request. */ - Common::WebResult DeleteJson(const std::string& path, const std::string& data, - bool allow_anonymous); + WebResult DeleteJson(const std::string& path, const std::string& data, bool allow_anonymous); /** * Gets a plain string from the specified path. @@ -52,7 +48,7 @@ public: * @param allow_anonymous If true, allow anonymous unauthenticated requests. * @return the result of the request. */ - Common::WebResult GetPlain(const std::string& path, bool allow_anonymous); + WebResult GetPlain(const std::string& path, bool allow_anonymous); /** * Gets an PNG image from the specified path. @@ -60,14 +56,14 @@ public: * @param allow_anonymous If true, allow anonymous unauthenticated requests. * @return the result of the request. */ - Common::WebResult GetImage(const std::string& path, bool allow_anonymous); + WebResult GetImage(const std::string& path, bool allow_anonymous); /** * Requests an external JWT for the specific audience provided. * @param audience the audience of the JWT requested. * @return the result of the request. */ - Common::WebResult GetExternalJWT(const std::string& audience); + WebResult GetExternalJWT(const std::string& audience); private: struct Impl; diff --git a/src/common/web_result.h b/src/web_service/web_result.h index 8bfa2141d..3aeeb5288 100644 --- a/src/common/web_result.h +++ b/src/web_service/web_result.h @@ -7,7 +7,7 @@ #include <string> #include "common/common_types.h" -namespace Common { +namespace WebService { struct WebResult { enum class Code : u32 { Success, @@ -22,4 +22,4 @@ struct WebResult { std::string result_string; std::string returned_data; }; -} // namespace Common +} // namespace WebService diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt index 8b9404718..b16b54032 100644 --- a/src/yuzu/CMakeLists.txt +++ b/src/yuzu/CMakeLists.txt @@ -9,6 +9,9 @@ add_executable(yuzu about_dialog.cpp about_dialog.h aboutdialog.ui + applets/controller.cpp + applets/controller.h + applets/controller.ui applets/error.cpp applets/error.h applets/profile_select.cpp @@ -24,13 +27,24 @@ add_executable(yuzu compatibility_list.h configuration/config.cpp configuration/config.h + configuration/configuration_shared.cpp + configuration/configuration_shared.h configuration/configure.ui configuration/configure_audio.cpp configuration/configure_audio.h configuration/configure_audio.ui + configuration/configure_cpu.cpp + configuration/configure_cpu.h + configuration/configure_cpu.ui + configuration/configure_cpu_debug.cpp + configuration/configure_cpu_debug.h + configuration/configure_cpu_debug.ui configuration/configure_debug.cpp configuration/configure_debug.h configuration/configure_debug.ui + configuration/configure_debug_controller.cpp + configuration/configure_debug_controller.h + configuration/configure_debug_controller.ui configuration/configure_dialog.cpp configuration/configure_dialog.h configuration/configure_filesystem.cpp @@ -51,18 +65,27 @@ add_executable(yuzu configuration/configure_input.cpp configuration/configure_input.h configuration/configure_input.ui + configuration/configure_input_advanced.cpp + configuration/configure_input_advanced.h + configuration/configure_input_advanced.ui configuration/configure_input_player.cpp configuration/configure_input_player.h configuration/configure_input_player.ui - configuration/configure_input_simple.cpp - configuration/configure_input_simple.h - configuration/configure_input_simple.ui + configuration/configure_input_profile_dialog.cpp + configuration/configure_input_profile_dialog.h + configuration/configure_input_profile_dialog.ui + configuration/configure_motion_touch.cpp + configuration/configure_motion_touch.h + configuration/configure_motion_touch.ui configuration/configure_mouse_advanced.cpp configuration/configure_mouse_advanced.h configuration/configure_mouse_advanced.ui - configuration/configure_per_general.cpp - configuration/configure_per_general.h - configuration/configure_per_general.ui + configuration/configure_per_game.cpp + configuration/configure_per_game.h + configuration/configure_per_game.ui + configuration/configure_per_game_addons.cpp + configuration/configure_per_game_addons.h + configuration/configure_per_game_addons.ui configuration/configure_profile_manager.cpp configuration/configure_profile_manager.h configuration/configure_profile_manager.ui @@ -72,15 +95,24 @@ add_executable(yuzu configuration/configure_system.cpp configuration/configure_system.h configuration/configure_system.ui + configuration/configure_touch_from_button.cpp + configuration/configure_touch_from_button.h + configuration/configure_touch_from_button.ui configuration/configure_touchscreen_advanced.cpp configuration/configure_touchscreen_advanced.h configuration/configure_touchscreen_advanced.ui + configuration/configure_touch_widget.h configuration/configure_ui.cpp configuration/configure_ui.h configuration/configure_ui.ui + configuration/configure_vibration.cpp + configuration/configure_vibration.h + configuration/configure_vibration.ui configuration/configure_web.cpp configuration/configure_web.h configuration/configure_web.ui + configuration/input_profiles.cpp + configuration/input_profiles.h debugger/console.cpp debugger/console.h debugger/profiler.cpp @@ -93,11 +125,13 @@ add_executable(yuzu game_list_p.h game_list_worker.cpp game_list_worker.h + hotkeys.cpp + hotkeys.h + install_dialog.cpp + install_dialog.h loading_screen.cpp loading_screen.h loading_screen.ui - hotkeys.cpp - hotkeys.h main.cpp main.h main.ui @@ -120,11 +154,44 @@ file(GLOB COMPAT_LIST file(GLOB_RECURSE ICONS ${PROJECT_SOURCE_DIR}/dist/icons/*) file(GLOB_RECURSE THEMES ${PROJECT_SOURCE_DIR}/dist/qt_themes/*) +if (ENABLE_QT_TRANSLATION) + set(YUZU_QT_LANGUAGES "${PROJECT_SOURCE_DIR}/dist/languages" CACHE PATH "Path to the translation bundle for the Qt frontend") + option(GENERATE_QT_TRANSLATION "Generate en.ts as the translation source file" OFF) + + # Update source TS file if enabled + if (GENERATE_QT_TRANSLATION) + get_target_property(SRCS yuzu SOURCES) + qt5_create_translation(QM_FILES ${SRCS} ${UIS} ${YUZU_QT_LANGUAGES}/en.ts) + add_custom_target(translation ALL DEPENDS ${YUZU_QT_LANGUAGES}/en.ts) + endif() + + # Find all TS files except en.ts + file(GLOB_RECURSE LANGUAGES_TS ${YUZU_QT_LANGUAGES}/*.ts) + list(REMOVE_ITEM LANGUAGES_TS ${YUZU_QT_LANGUAGES}/en.ts) + + # Compile TS files to QM files + qt5_add_translation(LANGUAGES_QM ${LANGUAGES_TS}) + + # Build a QRC file from the QM file list + set(LANGUAGES_QRC ${CMAKE_CURRENT_BINARY_DIR}/languages.qrc) + file(WRITE ${LANGUAGES_QRC} "<RCC><qresource prefix=\"languages\">\n") + foreach (QM ${LANGUAGES_QM}) + get_filename_component(QM_FILE ${QM} NAME) + file(APPEND ${LANGUAGES_QRC} "<file>${QM_FILE}</file>\n") + endforeach (QM) + file(APPEND ${LANGUAGES_QRC} "</qresource></RCC>") + + # Add the QRC file to package in all QM files + qt5_add_resources(LANGUAGES ${LANGUAGES_QRC}) +else() + set(LANGUAGES) +endif() target_sources(yuzu PRIVATE ${COMPAT_LIST} ${ICONS} + ${LANGUAGES} ${THEMES} ) @@ -147,7 +214,7 @@ endif() create_target_directory_groups(yuzu) target_link_libraries(yuzu PRIVATE common core input_common video_core) -target_link_libraries(yuzu PRIVATE Boost::boost glad Qt5::OpenGL Qt5::Widgets) +target_link_libraries(yuzu PRIVATE Boost::boost glad Qt5::Widgets) target_link_libraries(yuzu PRIVATE ${PLATFORM_LIBRARIES} Threads::Threads) if (ENABLE_VULKAN AND NOT WIN32) @@ -202,10 +269,14 @@ endif() if (MSVC) include(CopyYuzuQt5Deps) include(CopyYuzuSDLDeps) - include(CopyYuzuUnicornDeps) + include(CopyYuzuFFmpegDeps) copy_yuzu_Qt5_deps(yuzu) copy_yuzu_SDL_deps(yuzu) - copy_yuzu_unicorn_deps(yuzu) + copy_yuzu_FFmpeg_deps(yuzu) +endif() + +if (NOT APPLE) + target_compile_definitions(yuzu PRIVATE HAS_OPENGL) endif() if (ENABLE_VULKAN) diff --git a/src/yuzu/aboutdialog.ui b/src/yuzu/aboutdialog.ui index f122ba39d..1b320630c 100644 --- a/src/yuzu/aboutdialog.ui +++ b/src/yuzu/aboutdialog.ui @@ -160,32 +160,12 @@ p, li { white-space: pre-wrap; } <signal>accepted()</signal> <receiver>AboutDialog</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>248</x> - <y>254</y> - </hint> - <hint type="destinationlabel"> - <x>157</x> - <y>274</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>AboutDialog</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>316</x> - <y>260</y> - </hint> - <hint type="destinationlabel"> - <x>286</x> - <y>274</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/applets/controller.cpp b/src/yuzu/applets/controller.cpp new file mode 100644 index 000000000..8ecfec770 --- /dev/null +++ b/src/yuzu/applets/controller.cpp @@ -0,0 +1,637 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <thread> + +#include "common/assert.h" +#include "common/string_util.h" +#include "core/core.h" +#include "core/hle/lock.h" +#include "core/hle/service/hid/controllers/npad.h" +#include "core/hle/service/hid/hid.h" +#include "core/hle/service/sm/sm.h" +#include "ui_controller.h" +#include "yuzu/applets/controller.h" +#include "yuzu/configuration/configure_input.h" +#include "yuzu/configuration/configure_input_profile_dialog.h" +#include "yuzu/configuration/configure_vibration.h" +#include "yuzu/configuration/input_profiles.h" +#include "yuzu/main.h" + +namespace { + +constexpr std::size_t HANDHELD_INDEX = 8; + +constexpr std::array<std::array<bool, 4>, 8> led_patterns{{ + {true, false, false, false}, + {true, true, false, false}, + {true, true, true, false}, + {true, true, true, true}, + {true, false, false, true}, + {true, false, true, false}, + {true, false, true, true}, + {false, true, true, false}, +}}; + +void UpdateController(Settings::ControllerType controller_type, std::size_t npad_index, + bool connected) { + Core::System& system{Core::System::GetInstance()}; + + if (!system.IsPoweredOn()) { + return; + } + + Service::SM::ServiceManager& sm = system.ServiceManager(); + + auto& npad = + sm.GetService<Service::HID::Hid>("hid") + ->GetAppletResource() + ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad); + + npad.UpdateControllerAt(npad.MapSettingsTypeToNPad(controller_type), npad_index, connected); +} + +// Returns true if the given controller type is compatible with the given parameters. +bool IsControllerCompatible(Settings::ControllerType controller_type, + Core::Frontend::ControllerParameters parameters) { + switch (controller_type) { + case Settings::ControllerType::ProController: + return parameters.allow_pro_controller; + case Settings::ControllerType::DualJoyconDetached: + return parameters.allow_dual_joycons; + case Settings::ControllerType::LeftJoycon: + return parameters.allow_left_joycon; + case Settings::ControllerType::RightJoycon: + return parameters.allow_right_joycon; + case Settings::ControllerType::Handheld: + return parameters.enable_single_mode && parameters.allow_handheld; + default: + return false; + } +} + +/// Maps the controller type combobox index to Controller Type enum +constexpr Settings::ControllerType GetControllerTypeFromIndex(int index) { + switch (index) { + case 0: + default: + return Settings::ControllerType::ProController; + case 1: + return Settings::ControllerType::DualJoyconDetached; + case 2: + return Settings::ControllerType::LeftJoycon; + case 3: + return Settings::ControllerType::RightJoycon; + case 4: + return Settings::ControllerType::Handheld; + } +} + +/// Maps the Controller Type enum to controller type combobox index +constexpr int GetIndexFromControllerType(Settings::ControllerType type) { + switch (type) { + case Settings::ControllerType::ProController: + default: + return 0; + case Settings::ControllerType::DualJoyconDetached: + return 1; + case Settings::ControllerType::LeftJoycon: + return 2; + case Settings::ControllerType::RightJoycon: + return 3; + case Settings::ControllerType::Handheld: + return 4; + } +} + +} // namespace + +QtControllerSelectorDialog::QtControllerSelectorDialog( + QWidget* parent, Core::Frontend::ControllerParameters parameters_, + InputCommon::InputSubsystem* input_subsystem_) + : QDialog(parent), ui(std::make_unique<Ui::QtControllerSelectorDialog>()), + parameters(std::move(parameters_)), input_subsystem{input_subsystem_}, + input_profiles(std::make_unique<InputProfiles>()) { + ui->setupUi(this); + + player_widgets = { + ui->widgetPlayer1, ui->widgetPlayer2, ui->widgetPlayer3, ui->widgetPlayer4, + ui->widgetPlayer5, ui->widgetPlayer6, ui->widgetPlayer7, ui->widgetPlayer8, + }; + + player_groupboxes = { + ui->groupPlayer1Connected, ui->groupPlayer2Connected, ui->groupPlayer3Connected, + ui->groupPlayer4Connected, ui->groupPlayer5Connected, ui->groupPlayer6Connected, + ui->groupPlayer7Connected, ui->groupPlayer8Connected, + }; + + connected_controller_icons = { + ui->controllerPlayer1, ui->controllerPlayer2, ui->controllerPlayer3, ui->controllerPlayer4, + ui->controllerPlayer5, ui->controllerPlayer6, ui->controllerPlayer7, ui->controllerPlayer8, + }; + + led_patterns_boxes = {{ + {ui->checkboxPlayer1LED1, ui->checkboxPlayer1LED2, ui->checkboxPlayer1LED3, + ui->checkboxPlayer1LED4}, + {ui->checkboxPlayer2LED1, ui->checkboxPlayer2LED2, ui->checkboxPlayer2LED3, + ui->checkboxPlayer2LED4}, + {ui->checkboxPlayer3LED1, ui->checkboxPlayer3LED2, ui->checkboxPlayer3LED3, + ui->checkboxPlayer3LED4}, + {ui->checkboxPlayer4LED1, ui->checkboxPlayer4LED2, ui->checkboxPlayer4LED3, + ui->checkboxPlayer4LED4}, + {ui->checkboxPlayer5LED1, ui->checkboxPlayer5LED2, ui->checkboxPlayer5LED3, + ui->checkboxPlayer5LED4}, + {ui->checkboxPlayer6LED1, ui->checkboxPlayer6LED2, ui->checkboxPlayer6LED3, + ui->checkboxPlayer6LED4}, + {ui->checkboxPlayer7LED1, ui->checkboxPlayer7LED2, ui->checkboxPlayer7LED3, + ui->checkboxPlayer7LED4}, + {ui->checkboxPlayer8LED1, ui->checkboxPlayer8LED2, ui->checkboxPlayer8LED3, + ui->checkboxPlayer8LED4}, + }}; + + explain_text_labels = { + ui->labelPlayer1Explain, ui->labelPlayer2Explain, ui->labelPlayer3Explain, + ui->labelPlayer4Explain, ui->labelPlayer5Explain, ui->labelPlayer6Explain, + ui->labelPlayer7Explain, ui->labelPlayer8Explain, + }; + + emulated_controllers = { + ui->comboPlayer1Emulated, ui->comboPlayer2Emulated, ui->comboPlayer3Emulated, + ui->comboPlayer4Emulated, ui->comboPlayer5Emulated, ui->comboPlayer6Emulated, + ui->comboPlayer7Emulated, ui->comboPlayer8Emulated, + }; + + player_labels = { + ui->labelPlayer1, ui->labelPlayer2, ui->labelPlayer3, ui->labelPlayer4, + ui->labelPlayer5, ui->labelPlayer6, ui->labelPlayer7, ui->labelPlayer8, + }; + + connected_controller_labels = { + ui->labelConnectedPlayer1, ui->labelConnectedPlayer2, ui->labelConnectedPlayer3, + ui->labelConnectedPlayer4, ui->labelConnectedPlayer5, ui->labelConnectedPlayer6, + ui->labelConnectedPlayer7, ui->labelConnectedPlayer8, + }; + + connected_controller_checkboxes = { + ui->checkboxPlayer1Connected, ui->checkboxPlayer2Connected, ui->checkboxPlayer3Connected, + ui->checkboxPlayer4Connected, ui->checkboxPlayer5Connected, ui->checkboxPlayer6Connected, + ui->checkboxPlayer7Connected, ui->checkboxPlayer8Connected, + }; + + // Setup/load everything prior to setting up connections. + // This avoids unintentionally changing the states of elements while loading them in. + SetSupportedControllers(); + DisableUnsupportedPlayers(); + LoadConfiguration(); + + for (std::size_t i = 0; i < NUM_PLAYERS; ++i) { + SetExplainText(i); + UpdateControllerIcon(i); + UpdateLEDPattern(i); + UpdateBorderColor(i); + + connect(player_groupboxes[i], &QGroupBox::toggled, [this, i](bool checked) { + if (checked) { + for (std::size_t index = 0; index <= i; ++index) { + connected_controller_checkboxes[index]->setChecked(checked); + } + } else { + for (std::size_t index = i; index < NUM_PLAYERS; ++index) { + connected_controller_checkboxes[index]->setChecked(checked); + } + } + }); + + connect(emulated_controllers[i], qOverload<int>(&QComboBox::currentIndexChanged), + [this, i](int) { + UpdateControllerIcon(i); + UpdateControllerState(i); + UpdateLEDPattern(i); + CheckIfParametersMet(); + }); + + connect(connected_controller_checkboxes[i], &QCheckBox::stateChanged, [this, i](int state) { + player_groupboxes[i]->setChecked(state == Qt::Checked); + UpdateControllerIcon(i); + UpdateControllerState(i); + UpdateLEDPattern(i); + UpdateBorderColor(i); + CheckIfParametersMet(); + }); + + if (i == 0) { + connect(emulated_controllers[i], qOverload<int>(&QComboBox::currentIndexChanged), + [this](int index) { + UpdateDockedState(GetControllerTypeFromIndex(index) == + Settings::ControllerType::Handheld); + }); + } + } + + connect(ui->vibrationButton, &QPushButton::clicked, this, + &QtControllerSelectorDialog::CallConfigureVibrationDialog); + + connect(ui->inputConfigButton, &QPushButton::clicked, this, + &QtControllerSelectorDialog::CallConfigureInputProfileDialog); + + connect(ui->buttonBox, &QDialogButtonBox::accepted, this, + &QtControllerSelectorDialog::ApplyConfiguration); + + // Enhancement: Check if the parameters have already been met before disconnecting controllers. + // If all the parameters are met AND only allows a single player, + // stop the constructor here as we do not need to continue. + if (CheckIfParametersMet() && parameters.enable_single_mode) { + return; + } + + // If keep_controllers_connected is false, forcefully disconnect all controllers + if (!parameters.keep_controllers_connected) { + for (auto player : player_groupboxes) { + player->setChecked(false); + } + } + + resize(0, 0); +} + +QtControllerSelectorDialog::~QtControllerSelectorDialog() = default; + +int QtControllerSelectorDialog::exec() { + if (parameters_met && parameters.enable_single_mode) { + return QDialog::Accepted; + } + return QDialog::exec(); +} + +void QtControllerSelectorDialog::ApplyConfiguration() { + const bool pre_docked_mode = Settings::values.use_docked_mode.GetValue(); + Settings::values.use_docked_mode.SetValue(ui->radioDocked->isChecked()); + OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode.GetValue()); + + Settings::values.vibration_enabled.SetValue(ui->vibrationGroup->isChecked()); + Settings::values.motion_enabled.SetValue(ui->motionGroup->isChecked()); +} + +void QtControllerSelectorDialog::LoadConfiguration() { + for (std::size_t index = 0; index < NUM_PLAYERS; ++index) { + const auto connected = + Settings::values.players.GetValue()[index].connected || + (index == 0 && Settings::values.players.GetValue()[HANDHELD_INDEX].connected); + player_groupboxes[index]->setChecked(connected); + connected_controller_checkboxes[index]->setChecked(connected); + emulated_controllers[index]->setCurrentIndex( + GetIndexFromControllerType(Settings::values.players.GetValue()[index].controller_type)); + } + + UpdateDockedState(Settings::values.players.GetValue()[HANDHELD_INDEX].connected); + + ui->vibrationGroup->setChecked(Settings::values.vibration_enabled.GetValue()); + ui->motionGroup->setChecked(Settings::values.motion_enabled.GetValue()); +} + +void QtControllerSelectorDialog::CallConfigureVibrationDialog() { + ConfigureVibration dialog(this); + + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | + Qt::WindowSystemMenuHint); + dialog.setWindowModality(Qt::WindowModal); + + if (dialog.exec() == QDialog::Accepted) { + dialog.ApplyConfiguration(); + } +} + +void QtControllerSelectorDialog::CallConfigureInputProfileDialog() { + ConfigureInputProfileDialog dialog(this, input_subsystem, input_profiles.get()); + + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | + Qt::WindowSystemMenuHint); + dialog.setWindowModality(Qt::WindowModal); + dialog.exec(); +} + +bool QtControllerSelectorDialog::CheckIfParametersMet() { + // Here, we check and validate the current configuration against all applicable parameters. + const auto num_connected_players = static_cast<int>( + std::count_if(player_groupboxes.begin(), player_groupboxes.end(), + [this](const QGroupBox* player) { return player->isChecked(); })); + + const auto min_supported_players = parameters.enable_single_mode ? 1 : parameters.min_players; + const auto max_supported_players = parameters.enable_single_mode ? 1 : parameters.max_players; + + // First, check against the number of connected players. + if (num_connected_players < min_supported_players || + num_connected_players > max_supported_players) { + parameters_met = false; + ui->buttonBox->setEnabled(parameters_met); + return parameters_met; + } + + // Next, check against all connected controllers. + const auto all_controllers_compatible = [this] { + for (std::size_t index = 0; index < NUM_PLAYERS; ++index) { + // Skip controllers that are not used, we only care about the currently connected ones. + if (!player_groupboxes[index]->isChecked() || !player_groupboxes[index]->isEnabled()) { + continue; + } + + const auto compatible = IsControllerCompatible( + GetControllerTypeFromIndex(emulated_controllers[index]->currentIndex()), + parameters); + + // If any controller is found to be incompatible, return false early. + if (!compatible) { + return false; + } + } + + // Reaching here means all currently connected controllers are compatible. + return true; + }(); + + parameters_met = all_controllers_compatible; + ui->buttonBox->setEnabled(parameters_met); + return parameters_met; +} + +void QtControllerSelectorDialog::SetSupportedControllers() { + const QString theme = [] { + if (QIcon::themeName().contains(QStringLiteral("dark"))) { + return QStringLiteral("_dark"); + } else if (QIcon::themeName().contains(QStringLiteral("midnight"))) { + return QStringLiteral("_midnight"); + } else { + return QString{}; + } + }(); + + if (parameters.enable_single_mode && parameters.allow_handheld) { + ui->controllerSupported1->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_handheld%0); ").arg(theme)); + } else { + ui->controllerSupported1->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_handheld%0_disabled); ").arg(theme)); + } + + if (parameters.allow_dual_joycons) { + ui->controllerSupported2->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_dual_joycon%0); ").arg(theme)); + } else { + ui->controllerSupported2->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_dual_joycon%0_disabled); ").arg(theme)); + } + + if (parameters.allow_left_joycon) { + ui->controllerSupported3->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_joycon_left%0); ").arg(theme)); + } else { + ui->controllerSupported3->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_joycon_left%0_disabled); ").arg(theme)); + } + + if (parameters.allow_right_joycon) { + ui->controllerSupported4->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_joycon_right%0); ").arg(theme)); + } else { + ui->controllerSupported4->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_joycon_right%0_disabled); ").arg(theme)); + } + + if (parameters.allow_pro_controller) { + ui->controllerSupported5->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_pro_controller%0); ").arg(theme)); + } else { + ui->controllerSupported5->setStyleSheet( + QStringLiteral("image: url(:/controller/applet_pro_controller%0_disabled); ") + .arg(theme)); + } + + // enable_single_mode overrides min_players and max_players. + if (parameters.enable_single_mode) { + ui->numberSupportedLabel->setText(QStringLiteral("1")); + return; + } + + if (parameters.min_players == parameters.max_players) { + ui->numberSupportedLabel->setText(QStringLiteral("%1").arg(parameters.max_players)); + } else { + ui->numberSupportedLabel->setText( + QStringLiteral("%1 - %2").arg(parameters.min_players).arg(parameters.max_players)); + } +} + +void QtControllerSelectorDialog::UpdateControllerIcon(std::size_t player_index) { + if (!player_groupboxes[player_index]->isChecked()) { + connected_controller_icons[player_index]->setStyleSheet(QString{}); + player_labels[player_index]->show(); + return; + } + + const QString stylesheet = [this, player_index] { + switch (GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex())) { + case Settings::ControllerType::ProController: + return QStringLiteral("image: url(:/controller/applet_pro_controller%0); "); + case Settings::ControllerType::DualJoyconDetached: + return QStringLiteral("image: url(:/controller/applet_dual_joycon%0); "); + case Settings::ControllerType::LeftJoycon: + return QStringLiteral("image: url(:/controller/applet_joycon_left%0); "); + case Settings::ControllerType::RightJoycon: + return QStringLiteral("image: url(:/controller/applet_joycon_right%0); "); + case Settings::ControllerType::Handheld: + return QStringLiteral("image: url(:/controller/applet_handheld%0); "); + default: + return QString{}; + } + }(); + + const QString theme = [] { + if (QIcon::themeName().contains(QStringLiteral("dark"))) { + return QStringLiteral("_dark"); + } else if (QIcon::themeName().contains(QStringLiteral("midnight"))) { + return QStringLiteral("_midnight"); + } else { + return QString{}; + } + }(); + + connected_controller_icons[player_index]->setStyleSheet(stylesheet.arg(theme)); + player_labels[player_index]->hide(); +} + +void QtControllerSelectorDialog::UpdateControllerState(std::size_t player_index) { + auto& player = Settings::values.players.GetValue()[player_index]; + + const auto controller_type = + GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex()); + const auto player_connected = player_groupboxes[player_index]->isChecked() && + controller_type != Settings::ControllerType::Handheld; + + if (player.controller_type == controller_type && player.connected == player_connected) { + // Set vibration devices in the event that the input device has changed. + ConfigureVibration::SetVibrationDevices(player_index); + return; + } + + // Disconnect the controller first. + UpdateController(controller_type, player_index, false); + + player.controller_type = controller_type; + player.connected = player_connected; + + ConfigureVibration::SetVibrationDevices(player_index); + + // Handheld + if (player_index == 0) { + auto& handheld = Settings::values.players.GetValue()[HANDHELD_INDEX]; + if (controller_type == Settings::ControllerType::Handheld) { + handheld = player; + } + handheld.connected = player_groupboxes[player_index]->isChecked() && + controller_type == Settings::ControllerType::Handheld; + UpdateController(Settings::ControllerType::Handheld, 8, handheld.connected); + } + + if (!player.connected) { + return; + } + + // This emulates a delay between disconnecting and reconnecting controllers as some games + // do not respond to a change in controller type if it was instantaneous. + using namespace std::chrono_literals; + std::this_thread::sleep_for(20ms); + + UpdateController(controller_type, player_index, player_connected); +} + +void QtControllerSelectorDialog::UpdateLEDPattern(std::size_t player_index) { + if (!player_groupboxes[player_index]->isChecked() || + GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex()) == + Settings::ControllerType::Handheld) { + led_patterns_boxes[player_index][0]->setChecked(false); + led_patterns_boxes[player_index][1]->setChecked(false); + led_patterns_boxes[player_index][2]->setChecked(false); + led_patterns_boxes[player_index][3]->setChecked(false); + return; + } + + led_patterns_boxes[player_index][0]->setChecked(led_patterns[player_index][0]); + led_patterns_boxes[player_index][1]->setChecked(led_patterns[player_index][1]); + led_patterns_boxes[player_index][2]->setChecked(led_patterns[player_index][2]); + led_patterns_boxes[player_index][3]->setChecked(led_patterns[player_index][3]); +} + +void QtControllerSelectorDialog::UpdateBorderColor(std::size_t player_index) { + if (!parameters.enable_border_color || + player_index >= static_cast<std::size_t>(parameters.max_players) || + player_groupboxes[player_index]->styleSheet().contains(QStringLiteral("QGroupBox"))) { + return; + } + + player_groupboxes[player_index]->setStyleSheet( + player_groupboxes[player_index]->styleSheet().append( + QStringLiteral("QGroupBox#groupPlayer%1Connected:checked " + "{ border: 1px solid rgba(%2, %3, %4, %5); }") + .arg(player_index + 1) + .arg(parameters.border_colors[player_index][0]) + .arg(parameters.border_colors[player_index][1]) + .arg(parameters.border_colors[player_index][2]) + .arg(parameters.border_colors[player_index][3]))); +} + +void QtControllerSelectorDialog::SetExplainText(std::size_t player_index) { + if (!parameters.enable_explain_text || + player_index >= static_cast<std::size_t>(parameters.max_players)) { + return; + } + + explain_text_labels[player_index]->setText(QString::fromStdString( + Common::StringFromFixedZeroTerminatedBuffer(parameters.explain_text[player_index].data(), + parameters.explain_text[player_index].size()))); +} + +void QtControllerSelectorDialog::UpdateDockedState(bool is_handheld) { + // Disallow changing the console mode if the controller type is handheld. + ui->radioDocked->setEnabled(!is_handheld); + ui->radioUndocked->setEnabled(!is_handheld); + + ui->radioDocked->setChecked(Settings::values.use_docked_mode.GetValue()); + ui->radioUndocked->setChecked(!Settings::values.use_docked_mode.GetValue()); + + // Also force into undocked mode if the controller type is handheld. + if (is_handheld) { + ui->radioUndocked->setChecked(true); + } +} + +void QtControllerSelectorDialog::DisableUnsupportedPlayers() { + const auto max_supported_players = parameters.enable_single_mode ? 1 : parameters.max_players; + + switch (max_supported_players) { + case 0: + default: + UNREACHABLE(); + return; + case 1: + ui->widgetSpacer->hide(); + ui->widgetSpacer2->hide(); + ui->widgetSpacer3->hide(); + ui->widgetSpacer4->hide(); + break; + case 2: + ui->widgetSpacer->hide(); + ui->widgetSpacer2->hide(); + ui->widgetSpacer3->hide(); + break; + case 3: + ui->widgetSpacer->hide(); + ui->widgetSpacer2->hide(); + break; + case 4: + ui->widgetSpacer->hide(); + break; + case 5: + case 6: + case 7: + case 8: + break; + } + + for (std::size_t index = max_supported_players; index < NUM_PLAYERS; ++index) { + // Disconnect any unsupported players here and disable or hide them if applicable. + Settings::values.players.GetValue()[index].connected = false; + UpdateController(Settings::values.players.GetValue()[index].controller_type, index, false); + // Hide the player widgets when max_supported_controllers is less than or equal to 4. + if (max_supported_players <= 4) { + player_widgets[index]->hide(); + } + + // Disable and hide the following to prevent these from interaction. + player_widgets[index]->setDisabled(true); + connected_controller_checkboxes[index]->setDisabled(true); + connected_controller_labels[index]->hide(); + connected_controller_checkboxes[index]->hide(); + } +} + +QtControllerSelector::QtControllerSelector(GMainWindow& parent) { + connect(this, &QtControllerSelector::MainWindowReconfigureControllers, &parent, + &GMainWindow::ControllerSelectorReconfigureControllers, Qt::QueuedConnection); + connect(&parent, &GMainWindow::ControllerSelectorReconfigureFinished, this, + &QtControllerSelector::MainWindowReconfigureFinished, Qt::QueuedConnection); +} + +QtControllerSelector::~QtControllerSelector() = default; + +void QtControllerSelector::ReconfigureControllers( + std::function<void()> callback, const Core::Frontend::ControllerParameters& parameters) const { + this->callback = std::move(callback); + emit MainWindowReconfigureControllers(parameters); +} + +void QtControllerSelector::MainWindowReconfigureFinished() { + // Acquire the HLE mutex + std::lock_guard lock(HLE::g_hle_lock); + callback(); +} diff --git a/src/yuzu/applets/controller.h b/src/yuzu/applets/controller.h new file mode 100644 index 000000000..4344e1dd0 --- /dev/null +++ b/src/yuzu/applets/controller.h @@ -0,0 +1,144 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <memory> +#include <QDialog> +#include "core/frontend/applets/controller.h" + +class GMainWindow; +class QCheckBox; +class QComboBox; +class QDialogButtonBox; +class QGroupBox; +class QLabel; + +class InputProfiles; + +namespace InputCommon { +class InputSubsystem; +} + +namespace Ui { +class QtControllerSelectorDialog; +} + +class QtControllerSelectorDialog final : public QDialog { + Q_OBJECT + +public: + explicit QtControllerSelectorDialog(QWidget* parent, + Core::Frontend::ControllerParameters parameters_, + InputCommon::InputSubsystem* input_subsystem_); + ~QtControllerSelectorDialog() override; + + int exec() override; + +private: + // Applies the current configuration. + void ApplyConfiguration(); + + // Loads the current input configuration into the frontend applet. + void LoadConfiguration(); + + // Initializes the "Configure Vibration" Dialog. + void CallConfigureVibrationDialog(); + + // Initializes the "Create Input Profile" Dialog. + void CallConfigureInputProfileDialog(); + + // Checks the current configuration against the given parameters. + // This sets and returns the value of parameters_met. + bool CheckIfParametersMet(); + + // Sets the controller icons for "Supported Controller Types". + void SetSupportedControllers(); + + // Updates the controller icons per player. + void UpdateControllerIcon(std::size_t player_index); + + // Updates the controller state (type and connection status) per player. + void UpdateControllerState(std::size_t player_index); + + // Updates the LED pattern per player. + void UpdateLEDPattern(std::size_t player_index); + + // Updates the border color per player. + void UpdateBorderColor(std::size_t player_index); + + // Sets the "Explain Text" per player. + void SetExplainText(std::size_t player_index); + + // Updates the console mode. + void UpdateDockedState(bool is_handheld); + + // Disables and disconnects unsupported players based on the given parameters. + void DisableUnsupportedPlayers(); + + std::unique_ptr<Ui::QtControllerSelectorDialog> ui; + + // Parameters sent in from the backend HLE applet. + Core::Frontend::ControllerParameters parameters; + + InputCommon::InputSubsystem* input_subsystem; + + std::unique_ptr<InputProfiles> input_profiles; + + // This is true if and only if all parameters are met. Otherwise, this is false. + // This determines whether the "OK" button can be clicked to exit the applet. + bool parameters_met{false}; + + static constexpr std::size_t NUM_PLAYERS = 8; + + // Widgets encapsulating the groupboxes and comboboxes per player. + std::array<QWidget*, NUM_PLAYERS> player_widgets; + + // Groupboxes encapsulating the controller icons and LED patterns per player. + std::array<QGroupBox*, NUM_PLAYERS> player_groupboxes; + + // Icons for currently connected controllers/players. + std::array<QWidget*, NUM_PLAYERS> connected_controller_icons; + + // Labels that represent the player numbers in place of the controller icons. + std::array<QLabel*, NUM_PLAYERS> player_labels; + + // LED patterns for currently connected controllers/players. + std::array<std::array<QCheckBox*, 4>, NUM_PLAYERS> led_patterns_boxes; + + // Labels representing additional information known as "Explain Text" per player. + std::array<QLabel*, NUM_PLAYERS> explain_text_labels; + + // Comboboxes with a list of emulated controllers per player. + std::array<QComboBox*, NUM_PLAYERS> emulated_controllers; + + // Labels representing the number of connected controllers + // above the "Connected Controllers" checkboxes. + std::array<QLabel*, NUM_PLAYERS> connected_controller_labels; + + // Checkboxes representing the "Connected Controllers". + std::array<QCheckBox*, NUM_PLAYERS> connected_controller_checkboxes; +}; + +class QtControllerSelector final : public QObject, public Core::Frontend::ControllerApplet { + Q_OBJECT + +public: + explicit QtControllerSelector(GMainWindow& parent); + ~QtControllerSelector() override; + + void ReconfigureControllers( + std::function<void()> callback, + const Core::Frontend::ControllerParameters& parameters) const override; + +signals: + void MainWindowReconfigureControllers( + const Core::Frontend::ControllerParameters& parameters) const; + +private: + void MainWindowReconfigureFinished(); + + mutable std::function<void()> callback; +}; diff --git a/src/yuzu/applets/controller.ui b/src/yuzu/applets/controller.ui new file mode 100644 index 000000000..c8cb6bcf3 --- /dev/null +++ b/src/yuzu/applets/controller.ui @@ -0,0 +1,2653 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>QtControllerSelectorDialog</class> + <widget class="QDialog" name="QtControllerSelectorDialog"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>839</width> + <height>630</height> + </rect> + </property> + <property name="windowTitle"> + <string>Controller Applet</string> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout" stretch="0"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="mainControllerApplet" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_2" stretch="0,3,0"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="topControllerApplet" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout"> + <property name="spacing"> + <number>10</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>10</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>10</number> + </property> + <item> + <spacer name="controllerAppletHorizontalSpacer2"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QWidget" name="controllersSupported" native="true"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_21"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="controllersSupportedLabel"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="font"> + <font> + <weight>75</weight> + <bold>true</bold> + </font> + </property> + <property name="text"> + <string>Supported Controller Types:</string> + </property> + <property name="alignment"> + <set>Qt::AlignRight|Qt::AlignTrailing|Qt::AlignVCenter</set> + </property> + <property name="wordWrap"> + <bool>true</bool> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="controllerSupported1" native="true"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Preferred" vsizetype="Preferred"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="controllerSupported2" native="true"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="controllerSupported3" native="true"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="controllerSupported4" native="true"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="controllerSupported5" native="true"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="playersSupported" native="true"> + <property name="minimumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>70</width> + <height>70</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_20"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>16</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>16</number> + </property> + <item> + <widget class="QLabel" name="maxSupportedLabel"> + <property name="font"> + <font> + <weight>75</weight> + <bold>true</bold> + </font> + </property> + <property name="text"> + <string>Players:</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <property name="wordWrap"> + <bool>false</bool> + </property> + </widget> + </item> + <item> + <widget class="QLabel" name="numberSupportedLabel"> + <property name="font"> + <font> + <pointsize>14</pointsize> + </font> + </property> + <property name="text"> + <string>1 - 8</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="controllerAppletHorizontalSpacer3"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="middleControllerApplet" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_3"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <layout class="QGridLayout" name="gridLayout"> + <property name="spacing"> + <number>5</number> + </property> + <item row="1" column="7"> + <widget class="QWidget" name="widgetPlayer4" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_27"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer4Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_7" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer4" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_15"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer4"> + <property name="text"> + <string>P4</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player4LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_10"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer4LED1"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer4LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer4LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer4LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player4Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_39"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer4Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer4Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer4Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="1" column="3"> + <widget class="QWidget" name="widgetPlayer2" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_29"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer2Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_5" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer2" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_13"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer2"> + <property name="text"> + <string>P2</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player2LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_8"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer2LED1"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer2LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer2LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer2LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player2Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_37"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer2Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer2Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer2Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="1" column="1"> + <widget class="QWidget" name="widgetPlayer1" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_30"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer1Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_4" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer1" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_12"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer1"> + <property name="text"> + <string>P1</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player1LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer1LED1"> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer1LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer1LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer1LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player1Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_36"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer1Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer1Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Handheld</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer1Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="1" column="8"> + <widget class="QWidget" name="widgetSpacer2" native="true"> + <property name="minimumSize"> + <size> + <width>25</width> + <height>0</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_31"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletHorizontalSpacer8"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>25</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="1" column="4"> + <widget class="QWidget" name="widgetSpacer4" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_33"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletHorizontalSpacer6"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="1" column="6"> + <widget class="QWidget" name="widgetSpacer3" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_32"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletHorizontalSpacer7"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="1" column="5"> + <widget class="QWidget" name="widgetPlayer3" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_28"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer3Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_6" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer3" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_14"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer3"> + <property name="text"> + <string>P3</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player3LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_9"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer3LED1"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer3LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer3LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer3LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player3Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_38"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer3Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer3Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer3Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="0" column="1"> + <widget class="QWidget" name="widgetSpacer5" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>25</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_34"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletVerticalSpacer3"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>25</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="6" column="5"> + <widget class="QWidget" name="widgetPlayer7" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_25"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer7Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_10" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer7" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_18"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer7"> + <property name="text"> + <string>P7</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player7LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_13"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer7LED1"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer7LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer7LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer7LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player7Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_42"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer7Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer7Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer7Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="6" column="7"> + <widget class="QWidget" name="widgetPlayer8" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_26"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer8Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_11" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer8" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_19"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer8"> + <property name="text"> + <string>P8</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player8LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_14"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer8LED1"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer8LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer8LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer8LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player8Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_35"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer8Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer8Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer8Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="6" column="1"> + <widget class="QWidget" name="widgetPlayer5" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_23"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer5Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_8" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer5" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_16"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer5"> + <property name="text"> + <string>P5</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player5LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_11"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer5LED1"> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer5LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer5LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer5LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player5Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_40"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer5Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer5Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer5Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="6" column="3"> + <widget class="QWidget" name="widgetPlayer6" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_24"> + <property name="spacing"> + <number>5</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupPlayer6Connected"> + <property name="minimumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>100</width> + <height>100</height> + </size> + </property> + <property name="title"> + <string/> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <property name="checked"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_9" stretch="1,0"> + <property name="spacing"> + <number>7</number> + </property> + <property name="leftMargin"> + <number>14</number> + </property> + <property name="topMargin"> + <number>7</number> + </property> + <property name="rightMargin"> + <number>14</number> + </property> + <property name="bottomMargin"> + <number>4</number> + </property> + <item> + <widget class="QWidget" name="controllerPlayer6" native="true"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout" name="verticalLayout_17"> + <property name="topMargin"> + <number>16</number> + </property> + <item alignment="Qt::AlignHCenter|Qt::AlignVCenter"> + <widget class="QLabel" name="labelPlayer6"> + <property name="text"> + <string>P6</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QWidget" name="Player6LEDs" native="true"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_12"> + <property name="spacing"> + <number>4</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QCheckBox" name="checkboxPlayer6LED1"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer6LED2"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer6LED3"/> + </item> + <item> + <widget class="QCheckBox" name="checkboxPlayer6LED4"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="Player6Explain" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>10</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>150</width> + <height>16777215</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_41"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="labelPlayer6Explain"> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer6Emulated"> + <item> + <property name="text"> + <string>Pro Controller</string> + </property> + </item> + <item> + <property name="text"> + <string>Dual Joycons</string> + </property> + </item> + <item> + <property name="text"> + <string>Left Joycon</string> + </property> + </item> + <item> + <property name="text"> + <string>Right Joycon</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QComboBox" name="comboPlayer6Profile"> + <item> + <property name="text"> + <string>Use Current Config</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item row="10" column="1"> + <widget class="QWidget" name="widgetSpacer" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>25</height> + </size> + </property> + <layout class="QVBoxLayout" name="verticalLayout_22"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletVerticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>25</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="1" column="2"> + <widget class="QWidget" name="widgetSpacer6" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_15"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletHorizontalSpacer5"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="1" column="0"> + <widget class="QWidget" name="widgetSpacer7" native="true"> + <property name="minimumSize"> + <size> + <width>25</width> + <height>0</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_16"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletHorizontalSpacer4"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>25</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item row="2" column="1"> + <widget class="QWidget" name="widgetSpacer9" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>25</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_17"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="controllerAppletVerticalSpacer2"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>25</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="bottomControllerApplet" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_6"> + <property name="spacing"> + <number>15</number> + </property> + <property name="leftMargin"> + <number>15</number> + </property> + <property name="topMargin"> + <number>8</number> + </property> + <property name="rightMargin"> + <number>15</number> + </property> + <property name="bottomMargin"> + <number>15</number> + </property> + <item> + <widget class="QGroupBox" name="handheldGroup"> + <property name="maximumSize"> + <size> + <width>16777215</width> + <height>16777215</height> + </size> + </property> + <property name="title"> + <string>Console Mode</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>8</number> + </property> + <property name="topMargin"> + <number>6</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QRadioButton" name="radioDocked"> + <property name="text"> + <string>Docked</string> + </property> + <property name="checked"> + <bool>true</bool> + </property> + </widget> + </item> + <item> + <widget class="QRadioButton" name="radioUndocked"> + <property name="text"> + <string>Undocked</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroup"> + <property name="title"> + <string>Vibration</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_5"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="vibrationButton"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="motionGroup"> + <property name="title"> + <string>Motion</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_4"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="motionButton"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="inputConfigGroup"> + <property name="title"> + <string>Profiles</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_7"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="inputConfigButton"> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Create</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="connectedControllers" native="true"> + <layout class="QGridLayout" name="gridLayout_2"> + <property name="leftMargin"> + <number>5</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <property name="spacing"> + <number>3</number> + </property> + <item row="1" column="4"> + <widget class="QCheckBox" name="checkboxPlayer4Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="0"> + <widget class="QLabel" name="labelControllers"> + <property name="text"> + <string>Controllers</string> + </property> + </widget> + </item> + <item row="1" column="2"> + <widget class="QCheckBox" name="checkboxPlayer2Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="0" column="1"> + <widget class="QLabel" name="labelConnectedPlayer1"> + <property name="text"> + <string>1</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="1" column="3"> + <widget class="QCheckBox" name="checkboxPlayer3Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="1"> + <widget class="QCheckBox" name="checkboxPlayer1Connected"> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + <property name="checked"> + <bool>false</bool> + </property> + </widget> + </item> + <item row="0" column="2"> + <widget class="QLabel" name="labelConnectedPlayer2"> + <property name="text"> + <string>2</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="4"> + <widget class="QLabel" name="labelConnectedPlayer4"> + <property name="text"> + <string>4</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="3"> + <widget class="QLabel" name="labelConnectedPlayer3"> + <property name="text"> + <string>3</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="0"> + <widget class="QLabel" name="labelConnected"> + <property name="text"> + <string>Connected</string> + </property> + </widget> + </item> + <item row="1" column="7"> + <widget class="QCheckBox" name="checkboxPlayer7Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="0" column="5"> + <widget class="QLabel" name="labelConnectedPlayer5"> + <property name="text"> + <string>5</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="1" column="6"> + <widget class="QCheckBox" name="checkboxPlayer6Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="0" column="7"> + <widget class="QLabel" name="labelConnectedPlayer7"> + <property name="text"> + <string>7</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="1" column="5"> + <widget class="QCheckBox" name="checkboxPlayer5Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="0" column="6"> + <widget class="QLabel" name="labelConnectedPlayer6"> + <property name="text"> + <string>6</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="8"> + <widget class="QLabel" name="labelConnectedPlayer8"> + <property name="text"> + <string>8</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="1" column="8"> + <widget class="QCheckBox" name="checkboxPlayer8Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="controllerAppletHorizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignBottom"> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="standardButtons"> + <set>QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections> + <connection> + <sender>buttonBox</sender> + <signal>accepted()</signal> + <receiver>QtControllerSelectorDialog</receiver> + <slot>accept()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/applets/profile_select.cpp b/src/yuzu/applets/profile_select.cpp index 6aff38735..c9a2f8601 100644 --- a/src/yuzu/applets/profile_select.cpp +++ b/src/yuzu/applets/profile_select.cpp @@ -17,6 +17,7 @@ #include "yuzu/applets/profile_select.h" #include "yuzu/main.h" +namespace { QString FormatUserEntryText(const QString& username, Common::UUID uuid) { return QtProfileSelectionDialog::tr( "%1\n%2", "%1 is the profile username, %2 is the formatted UUID (e.g. " @@ -25,7 +26,7 @@ QString FormatUserEntryText(const QString& username, Common::UUID uuid) { } QString GetImagePath(Common::UUID uuid) { - const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + + const auto path = Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg"; return QString::fromStdString(path); } @@ -41,6 +42,7 @@ QPixmap GetIcon(Common::UUID uuid) { return icon.scaled(64, 64, Qt::IgnoreAspectRatio, Qt::SmoothTransformation); } +} // Anonymous namespace QtProfileSelectionDialog::QtProfileSelectionDialog(QWidget* parent) : QDialog(parent), profile_manager(std::make_unique<Service::Account::ProfileManager>()) { @@ -112,6 +114,15 @@ QtProfileSelectionDialog::QtProfileSelectionDialog(QWidget* parent) QtProfileSelectionDialog::~QtProfileSelectionDialog() = default; +int QtProfileSelectionDialog::exec() { + // Skip profile selection when there's only one. + if (profile_manager->GetUserCount() == 1) { + user_index = 0; + return QDialog::Accepted; + } + return QDialog::exec(); +} + void QtProfileSelectionDialog::accept() { QDialog::accept(); } diff --git a/src/yuzu/applets/profile_select.h b/src/yuzu/applets/profile_select.h index cee886a77..29c33cca0 100644 --- a/src/yuzu/applets/profile_select.h +++ b/src/yuzu/applets/profile_select.h @@ -27,6 +27,7 @@ public: explicit QtProfileSelectionDialog(QWidget* parent); ~QtProfileSelectionDialog() override; + int exec() override; void accept() override; void reject() override; diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp index 1cac2f942..d62b0efc2 100644 --- a/src/yuzu/bootmanager.cpp +++ b/src/yuzu/bootmanager.cpp @@ -8,13 +8,17 @@ #include <QHBoxLayout> #include <QKeyEvent> #include <QMessageBox> -#include <QOffscreenSurface> -#include <QOpenGLContext> #include <QPainter> #include <QScreen> +#include <QString> #include <QStringList> #include <QWindow> +#ifdef HAS_OPENGL +#include <QOffscreenSurface> +#include <QOpenGLContext> +#endif + #if !defined(WIN32) && HAS_VULKAN #include <qpa/qplatformnativeinterface.h> #endif @@ -27,6 +31,7 @@ #include "common/scope_exit.h" #include "core/core.h" #include "core/frontend/framebuffer_layout.h" +#include "core/hle/kernel/process.h" #include "core/settings.h" #include "input_common/keyboard.h" #include "input_common/main.h" @@ -41,49 +46,66 @@ EmuThread::EmuThread() = default; EmuThread::~EmuThread() = default; void EmuThread::run() { - MicroProfileOnThreadCreate("EmuThread"); + std::string name = "yuzu:EmuControlThread"; + MicroProfileOnThreadCreate(name.c_str()); + Common::SetCurrentThreadName(name.c_str()); + + auto& system = Core::System::GetInstance(); + + system.RegisterHostThread(); + + auto& gpu = system.GPU(); // Main process has been loaded. Make the context current to this thread and begin GPU and CPU // execution. - Core::System::GetInstance().GPU().Start(); + gpu.Start(); + + gpu.ObtainContext(); emit LoadProgress(VideoCore::LoadCallbackStage::Prepare, 0, 0); - Core::System::GetInstance().Renderer().Rasterizer().LoadDiskResources( - stop_run, [this](VideoCore::LoadCallbackStage stage, std::size_t value, std::size_t total) { + system.Renderer().Rasterizer().LoadDiskResources( + system.CurrentProcess()->GetTitleID(), stop_run, + [this](VideoCore::LoadCallbackStage stage, std::size_t value, std::size_t total) { emit LoadProgress(stage, value, total); }); emit LoadProgress(VideoCore::LoadCallbackStage::Complete, 0, 0); + gpu.ReleaseContext(); + // Holds whether the cpu was running during the last iteration, // so that the DebugModeLeft signal can be emitted before the // next execution step bool was_active = false; while (!stop_run) { if (running) { - if (!was_active) + if (was_active) { emit DebugModeLeft(); + } - Core::System::ResultStatus result = Core::System::GetInstance().RunLoop(); + running_guard = true; + Core::System::ResultStatus result = system.Run(); if (result != Core::System::ResultStatus::Success) { + running_guard = false; this->SetRunning(false); - emit ErrorThrown(result, Core::System::GetInstance().GetStatusDetails()); + emit ErrorThrown(result, system.GetStatusDetails()); } + running_wait.Wait(); + result = system.Pause(); + if (result != Core::System::ResultStatus::Success) { + running_guard = false; + this->SetRunning(false); + emit ErrorThrown(result, system.GetStatusDetails()); + } + running_guard = false; - was_active = running || exec_step; - if (!was_active && !stop_run) + if (!stop_run) { + was_active = true; emit DebugModeEntered(); + } } else if (exec_step) { - if (!was_active) - emit DebugModeLeft(); - - exec_step = false; - Core::System::GetInstance().SingleStep(); - emit DebugModeEntered(); - yieldCurrentThread(); - - was_active = false; + UNIMPLEMENTED(); } else { std::unique_lock lock{running_mutex}; running_cv.wait(lock, [this] { return IsRunning() || exec_step || stop_run; }); @@ -91,13 +113,14 @@ void EmuThread::run() { } // Shutdown the core emulation - Core::System::GetInstance().Shutdown(); + system.Shutdown(); #if MICROPROFILE_ENABLED MicroProfileOnThreadExit(); #endif } +#ifdef HAS_OPENGL class OpenGLSharedContext : public Core::Frontend::GraphicsContext { public: /// Create the original context that should be shared from @@ -106,6 +129,9 @@ public: format.setVersion(4, 3); format.setProfile(QSurfaceFormat::CompatibilityProfile); format.setOption(QSurfaceFormat::FormatOption::DeprecatedFunctions); + if (Settings::values.renderer_debug) { + format.setOption(QSurfaceFormat::FormatOption::DebugContext); + } // TODO: expose a setting for buffer value (ie default/single/double/triple) format.setSwapBehavior(QSurfaceFormat::DefaultSwapBehavior); format.setSwapInterval(0); @@ -122,7 +148,7 @@ public: // disable vsync for any shared contexts auto format = share_context->format(); - format.setSwapInterval(main_surface ? Settings::values.use_vsync : 0); + format.setSwapInterval(main_surface ? Settings::values.use_vsync.GetValue() : 0); context = std::make_unique<QOpenGLContext>(); context->setShareContext(share_context); @@ -150,18 +176,19 @@ public: } void MakeCurrent() override { - if (is_current) { - return; + // We can't track the current state of the underlying context in this wrapper class because + // Qt may make the underlying context not current for one reason or another. In particular, + // the WebBrowser uses GL, so it seems to conflict if we aren't careful. + // Instead of always just making the context current (which does not have any caching to + // check if the underlying context is already current) we can check for the current context + // in the thread local data by calling `currentContext()` and checking if its ours. + if (QOpenGLContext::currentContext() != context.get()) { + context->makeCurrent(surface); } - is_current = context->makeCurrent(surface); } void DoneCurrent() override { - if (!is_current) { - return; - } context->doneCurrent(); - is_current = false; } QOpenGLContext* GetShareContext() { @@ -178,8 +205,8 @@ private: std::unique_ptr<QOpenGLContext> context; std::unique_ptr<QOffscreenSurface> offscreen_surface{}; QSurface* surface; - bool is_current = false; }; +#endif class DummyContext : public Core::Frontend::GraphicsContext {}; @@ -192,15 +219,6 @@ public: virtual ~RenderWidget() = default; - /// Called on the UI thread when this Widget is ready to draw - /// Dervied classes can override this to draw the latest frame. - virtual void Present() {} - - void paintEvent(QPaintEvent* event) override { - Present(); - update(); - } - QPaintEngine* paintEngine() const override { return nullptr; } @@ -219,20 +237,8 @@ public: context = std::move(context_); } - void Present() override { - if (!isVisible()) { - return; - } - - context->MakeCurrent(); - if (Core::System::GetInstance().Renderer().TryPresent(100)) { - context->SwapBuffers(); - glFinish(); - } - } - private: - std::unique_ptr<Core::Frontend::GraphicsContext> context{}; + std::unique_ptr<Core::Frontend::GraphicsContext> context; }; #ifdef HAS_VULKAN @@ -280,8 +286,9 @@ static Core::Frontend::EmuWindow::WindowSystemInfo GetWindowSystemInfo(QWindow* return wsi; } -GRenderWindow::GRenderWindow(GMainWindow* parent_, EmuThread* emu_thread_) - : QWidget(parent_), emu_thread(emu_thread_) { +GRenderWindow::GRenderWindow(GMainWindow* parent, EmuThread* emu_thread_, + std::shared_ptr<InputCommon::InputSubsystem> input_subsystem_) + : QWidget(parent), emu_thread(emu_thread_), input_subsystem{std::move(input_subsystem_)} { setWindowTitle(QStringLiteral("yuzu %1 | %2-%3") .arg(QString::fromUtf8(Common::g_build_name), QString::fromUtf8(Common::g_scm_branch), @@ -290,13 +297,15 @@ GRenderWindow::GRenderWindow(GMainWindow* parent_, EmuThread* emu_thread_) auto layout = new QHBoxLayout(this); layout->setMargin(0); setLayout(layout); - InputCommon::Init(); + input_subsystem->Initialize(); - connect(this, &GRenderWindow::FirstFrameDisplayed, parent_, &GMainWindow::OnLoadComplete); + this->setMouseTracking(true); + + connect(this, &GRenderWindow::FirstFrameDisplayed, parent, &GMainWindow::OnLoadComplete); } GRenderWindow::~GRenderWindow() { - InputCommon::Shutdown(); + input_subsystem->Shutdown(); } void GRenderWindow::PollEvents() { @@ -350,7 +359,7 @@ QByteArray GRenderWindow::saveGeometry() { } qreal GRenderWindow::windowPixelRatio() const { - return devicePixelRatio(); + return devicePixelRatioF(); } std::pair<u32, u32> GRenderWindow::ScaleTouch(const QPointF& pos) const { @@ -365,15 +374,20 @@ void GRenderWindow::closeEvent(QCloseEvent* event) { } void GRenderWindow::keyPressEvent(QKeyEvent* event) { - InputCommon::GetKeyboard()->PressKey(event->key()); + input_subsystem->GetKeyboard()->PressKey(event->key()); } void GRenderWindow::keyReleaseEvent(QKeyEvent* event) { - InputCommon::GetKeyboard()->ReleaseKey(event->key()); + input_subsystem->GetKeyboard()->ReleaseKey(event->key()); } void GRenderWindow::mousePressEvent(QMouseEvent* event) { - // touch input is handled in TouchBeginEvent + if (!Settings::values.touchscreen.enabled) { + input_subsystem->GetKeyboard()->PressKey(event->button()); + return; + } + + // Touch input is handled in TouchBeginEvent if (event->source() == Qt::MouseEventSynthesizedBySystem) { return; } @@ -383,12 +397,13 @@ void GRenderWindow::mousePressEvent(QMouseEvent* event) { const auto [x, y] = ScaleTouch(pos); this->TouchPressed(x, y); } else if (event->button() == Qt::RightButton) { - InputCommon::GetMotionEmu()->BeginTilt(pos.x(), pos.y()); + input_subsystem->GetMotionEmu()->BeginTilt(pos.x(), pos.y()); } + QWidget::mousePressEvent(event); } void GRenderWindow::mouseMoveEvent(QMouseEvent* event) { - // touch input is handled in TouchUpdateEvent + // Touch input is handled in TouchUpdateEvent if (event->source() == Qt::MouseEventSynthesizedBySystem) { return; } @@ -396,11 +411,17 @@ void GRenderWindow::mouseMoveEvent(QMouseEvent* event) { auto pos = event->pos(); const auto [x, y] = ScaleTouch(pos); this->TouchMoved(x, y); - InputCommon::GetMotionEmu()->Tilt(pos.x(), pos.y()); + input_subsystem->GetMotionEmu()->Tilt(pos.x(), pos.y()); + QWidget::mouseMoveEvent(event); } void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) { - // touch input is handled in TouchEndEvent + if (!Settings::values.touchscreen.enabled) { + input_subsystem->GetKeyboard()->ReleaseKey(event->button()); + return; + } + + // Touch input is handled in TouchEndEvent if (event->source() == Qt::MouseEventSynthesizedBySystem) { return; } @@ -408,7 +429,7 @@ void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) { if (event->button() == Qt::LeftButton) { this->TouchReleased(); } else if (event->button() == Qt::RightButton) { - InputCommon::GetMotionEmu()->EndTilt(); + input_subsystem->GetMotionEmu()->EndTilt(); } } @@ -423,7 +444,7 @@ void GRenderWindow::TouchUpdateEvent(const QTouchEvent* event) { int active_points = 0; // average all active touch points - for (const auto tp : event->touchPoints()) { + for (const auto& tp : event->touchPoints()) { if (tp.state() & (Qt::TouchPointPressed | Qt::TouchPointMoved | Qt::TouchPointStationary)) { active_points++; pos += tp.pos(); @@ -457,7 +478,7 @@ bool GRenderWindow::event(QEvent* event) { void GRenderWindow::focusOutEvent(QFocusEvent* event) { QWidget::focusOutEvent(event); - InputCommon::GetKeyboard()->ReleaseAllKeys(); + input_subsystem->GetKeyboard()->ReleaseAllKeys(); } void GRenderWindow::resizeEvent(QResizeEvent* event) { @@ -466,13 +487,15 @@ void GRenderWindow::resizeEvent(QResizeEvent* event) { } std::unique_ptr<Core::Frontend::GraphicsContext> GRenderWindow::CreateSharedContext() const { - if (Settings::values.renderer_backend == Settings::RendererBackend::OpenGL) { +#ifdef HAS_OPENGL + if (Settings::values.renderer_backend.GetValue() == Settings::RendererBackend::OpenGL) { auto c = static_cast<OpenGLSharedContext*>(main_context.get()); // Bind the shared contexts to the main surface in case the backend wants to take over // presentation return std::make_unique<OpenGLSharedContext>(c->GetShareContext(), child_widget->windowHandle()); } +#endif return std::make_unique<DummyContext>(); } @@ -481,7 +504,7 @@ bool GRenderWindow::InitRenderTarget() { first_frame = false; - switch (Settings::values.renderer_backend) { + switch (Settings::values.renderer_backend.GetValue()) { case Settings::RendererBackend::OpenGL: if (!InitializeOpenGL()) { return false; @@ -508,7 +531,7 @@ bool GRenderWindow::InitRenderTarget() { OnFramebufferSizeChanged(); BackupGeometry(); - if (Settings::values.renderer_backend == Settings::RendererBackend::OpenGL) { + if (Settings::values.renderer_backend.GetValue() == Settings::RendererBackend::OpenGL) { if (!LoadOpenGL()) { return false; } @@ -537,7 +560,7 @@ void GRenderWindow::CaptureScreenshot(u32 res_scale, const QString& screenshot_p screenshot_image = QImage(QSize(layout.width, layout.height), QImage::Format_RGB32); renderer.RequestScreenshot( screenshot_image.bits(), - [=] { + [=, this] { const std::string std_screenshot_path = screenshot_path.toStdString(); if (screenshot_image.mirrored(false, true).save(screenshot_path)) { LOG_INFO(Frontend, "Screenshot saved to \"{}\"", std_screenshot_path); @@ -553,6 +576,7 @@ void GRenderWindow::OnMinimalClientAreaChangeRequest(std::pair<u32, u32> minimal } bool GRenderWindow::InitializeOpenGL() { +#ifdef HAS_OPENGL // TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground, // WA_DontShowOnScreen, WA_DeleteOnClose auto child = new OpenGLRenderWidget(this); @@ -564,6 +588,11 @@ bool GRenderWindow::InitializeOpenGL() { std::make_unique<OpenGLSharedContext>(context->GetShareContext(), child->windowHandle())); return true; +#else + QMessageBox::warning(this, tr("OpenGL not available!"), + tr("yuzu has not been compiled with OpenGL support.")); + return false; +#endif } bool GRenderWindow::InitializeVulkan() { @@ -585,19 +614,33 @@ bool GRenderWindow::LoadOpenGL() { auto context = CreateSharedContext(); auto scope = context->Acquire(); if (!gladLoadGL()) { - QMessageBox::critical(this, tr("Error while initializing OpenGL 4.3!"), - tr("Your GPU may not support OpenGL 4.3, or you do not have the " - "latest graphics driver.")); + QMessageBox::warning( + this, tr("Error while initializing OpenGL!"), + tr("Your GPU may not support OpenGL, or you do not have the latest graphics driver.")); + return false; + } + + const QString renderer = + QString::fromUtf8(reinterpret_cast<const char*>(glGetString(GL_RENDERER))); + + if (!GLAD_GL_VERSION_4_3) { + LOG_ERROR(Frontend, "GPU does not support OpenGL 4.3: {}", renderer.toStdString()); + QMessageBox::warning(this, tr("Error while initializing OpenGL 4.3!"), + tr("Your GPU may not support OpenGL 4.3, or you do not have the " + "latest graphics driver.<br><br>GL Renderer:<br>%1") + .arg(renderer)); return false; } QStringList unsupported_gl_extensions = GetUnsupportedGLExtensions(); if (!unsupported_gl_extensions.empty()) { - QMessageBox::critical( + QMessageBox::warning( this, tr("Error while initializing OpenGL!"), tr("Your GPU may not support one or more required OpenGL extensions. Please ensure you " - "have the latest graphics driver.<br><br>Unsupported extensions:<br>") + - unsupported_gl_extensions.join(QStringLiteral("<br>"))); + "have the latest graphics driver.<br><br>GL Renderer:<br>%1<br><br>Unsupported " + "extensions:<br>%2") + .arg(renderer) + .arg(unsupported_gl_extensions.join(QStringLiteral("<br>")))); return false; } return true; @@ -627,8 +670,13 @@ QStringList GRenderWindow::GetUnsupportedGLExtensions() const { if (!GLAD_GL_ARB_depth_buffer_float) unsupported_ext.append(QStringLiteral("ARB_depth_buffer_float")); - for (const QString& ext : unsupported_ext) - LOG_CRITICAL(Frontend, "Unsupported GL extension: {}", ext.toStdString()); + if (!unsupported_ext.empty()) { + LOG_ERROR(Frontend, "GPU does not support all required extensions: {}", + glGetString(GL_RENDERER)); + } + for (const QString& ext : unsupported_ext) { + LOG_ERROR(Frontend, "Unsupported GL extension: {}", ext.toStdString()); + } return unsupported_ext; } diff --git a/src/yuzu/bootmanager.h b/src/yuzu/bootmanager.h index 3626604ca..ca35cf831 100644 --- a/src/yuzu/bootmanager.h +++ b/src/yuzu/bootmanager.h @@ -6,6 +6,7 @@ #include <atomic> #include <condition_variable> +#include <memory> #include <mutex> #include <QImage> @@ -23,6 +24,10 @@ class QKeyEvent; class QTouchEvent; class QStringList; +namespace InputCommon { +class InputSubsystem; +} + namespace VideoCore { enum class LoadCallbackStage; } @@ -59,6 +64,12 @@ public: this->running = running; lock.unlock(); running_cv.notify_all(); + if (!running) { + running_wait.Set(); + /// Wait until effectively paused + while (running_guard) + ; + } } /** @@ -84,6 +95,8 @@ private: std::atomic_bool stop_run{false}; std::mutex running_mutex; std::condition_variable running_cv; + Common::Event running_wait{}; + std::atomic_bool running_guard{false}; signals: /** @@ -113,7 +126,8 @@ class GRenderWindow : public QWidget, public Core::Frontend::EmuWindow { Q_OBJECT public: - GRenderWindow(GMainWindow* parent, EmuThread* emu_thread); + explicit GRenderWindow(GMainWindow* parent, EmuThread* emu_thread_, + std::shared_ptr<InputCommon::InputSubsystem> input_subsystem_); ~GRenderWindow() override; // EmuWindow implementation. @@ -175,6 +189,7 @@ private: QStringList GetUnsupportedGLExtensions() const; EmuThread* emu_thread; + std::shared_ptr<InputCommon::InputSubsystem> input_subsystem; // Main context that will be shared with all other contexts that are requested. // If this is used in a shared context setting, then this should not be used directly, but diff --git a/src/yuzu/compatdb.cpp b/src/yuzu/compatdb.cpp index 5477f050c..649912557 100644 --- a/src/yuzu/compatdb.cpp +++ b/src/yuzu/compatdb.cpp @@ -54,7 +54,8 @@ void CompatDB::Submit() { back(); LOG_DEBUG(Frontend, "Compatibility Rating: {}", compatibility->checkedId()); Core::System::GetInstance().TelemetrySession().AddField( - Telemetry::FieldType::UserFeedback, "Compatibility", compatibility->checkedId()); + Common::Telemetry::FieldType::UserFeedback, "Compatibility", + compatibility->checkedId()); button(NextButton)->setEnabled(false); button(NextButton)->setText(tr("Submitting")); diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index 3b9ab38dd..3c423a271 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -5,52 +5,59 @@ #include <array> #include <QKeySequence> #include <QSettings> +#include "common/common_paths.h" #include "common/file_util.h" -#include "configure_input_simple.h" #include "core/hle/service/acc/profile_manager.h" #include "core/hle/service/hid/controllers/npad.h" #include "input_common/main.h" #include "input_common/udp/client.h" #include "yuzu/configuration/config.h" -#include "yuzu/uisettings.h" -Config::Config() { - // TODO: Don't hardcode the path; let the frontend decide where to put the config files. - qt_config_loc = FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + "qt-config.ini"; - FileUtil::CreateFullPath(qt_config_loc); - qt_config = - std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), QSettings::IniFormat); - Reload(); +namespace FS = Common::FS; + +Config::Config(const std::string& config_name, ConfigType config_type) : type(config_type) { + global = config_type == ConfigType::GlobalConfig; + + Initialize(config_name); } Config::~Config() { - Save(); + if (global) { + Save(); + } } const std::array<int, Settings::NativeButton::NumButtons> Config::default_buttons = { - Qt::Key_A, Qt::Key_S, Qt::Key_Z, Qt::Key_X, Qt::Key_3, Qt::Key_4, Qt::Key_Q, - Qt::Key_W, Qt::Key_1, Qt::Key_2, Qt::Key_N, Qt::Key_M, Qt::Key_F, Qt::Key_T, - Qt::Key_H, Qt::Key_G, Qt::Key_Left, Qt::Key_Up, Qt::Key_Right, Qt::Key_Down, Qt::Key_J, - Qt::Key_I, Qt::Key_L, Qt::Key_K, Qt::Key_D, Qt::Key_C, Qt::Key_B, Qt::Key_V, + Qt::Key_A, Qt::Key_S, Qt::Key_Z, Qt::Key_X, Qt::Key_3, Qt::Key_4, Qt::Key_Q, + Qt::Key_W, Qt::Key_1, Qt::Key_2, Qt::Key_N, Qt::Key_M, Qt::Key_F, Qt::Key_T, + Qt::Key_H, Qt::Key_G, Qt::Key_D, Qt::Key_C, Qt::Key_B, Qt::Key_V, +}; + +const std::array<int, Settings::NativeMotion::NumMotions> Config::default_motions = { + Qt::Key_7, + Qt::Key_8, }; -const std::array<std::array<int, 5>, Settings::NativeAnalog::NumAnalogs> Config::default_analogs{{ +const std::array<std::array<int, 4>, Settings::NativeAnalog::NumAnalogs> Config::default_analogs{{ { Qt::Key_Up, Qt::Key_Down, Qt::Key_Left, Qt::Key_Right, - Qt::Key_E, }, { Qt::Key_I, Qt::Key_K, Qt::Key_J, Qt::Key_L, - Qt::Key_R, }, }}; +const std::array<int, 2> Config::default_stick_mod = { + Qt::Key_E, + Qt::Key_R, +}; + const std::array<int, Settings::NativeMouseButton::NumMouseButtons> Config::default_mouse_buttons = { Qt::Key_BracketLeft, Qt::Key_BracketRight, Qt::Key_Apostrophe, Qt::Key_Minus, Qt::Key_Equal, @@ -212,95 +219,174 @@ const std::array<int, Settings::NativeKeyboard::NumKeyboardMods> Config::default // This must be in alphabetical order according to action name as it must have the same order as // UISetting::values.shortcuts, which is alphabetically ordered. // clang-format off -const std::array<UISettings::Shortcut, 15> default_hotkeys{{ - {QStringLiteral("Capture Screenshot"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+P"), Qt::ApplicationShortcut}}, +const std::array<UISettings::Shortcut, 16> Config::default_hotkeys{{ + {QStringLiteral("Capture Screenshot"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+P"), Qt::WidgetWithChildrenShortcut}}, + {QStringLiteral("Change Docked Mode"), QStringLiteral("Main Window"), {QStringLiteral("F10"), Qt::ApplicationShortcut}}, {QStringLiteral("Continue/Pause Emulation"), QStringLiteral("Main Window"), {QStringLiteral("F4"), Qt::WindowShortcut}}, {QStringLiteral("Decrease Speed Limit"), QStringLiteral("Main Window"), {QStringLiteral("-"), Qt::ApplicationShortcut}}, - {QStringLiteral("Exit yuzu"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+Q"), Qt::WindowShortcut}}, {QStringLiteral("Exit Fullscreen"), QStringLiteral("Main Window"), {QStringLiteral("Esc"), Qt::WindowShortcut}}, + {QStringLiteral("Exit yuzu"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+Q"), Qt::WindowShortcut}}, {QStringLiteral("Fullscreen"), QStringLiteral("Main Window"), {QStringLiteral("F11"), Qt::WindowShortcut}}, {QStringLiteral("Increase Speed Limit"), QStringLiteral("Main Window"), {QStringLiteral("+"), Qt::ApplicationShortcut}}, - {QStringLiteral("Load Amiibo"), QStringLiteral("Main Window"), {QStringLiteral("F2"), Qt::ApplicationShortcut}}, - {QStringLiteral("Load File"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+O"), Qt::WindowShortcut}}, + {QStringLiteral("Load Amiibo"), QStringLiteral("Main Window"), {QStringLiteral("F2"), Qt::WidgetWithChildrenShortcut}}, + {QStringLiteral("Load File"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+O"), Qt::WidgetWithChildrenShortcut}}, + {QStringLiteral("Mute Audio"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+M"), Qt::WindowShortcut}}, {QStringLiteral("Restart Emulation"), QStringLiteral("Main Window"), {QStringLiteral("F6"), Qt::WindowShortcut}}, {QStringLiteral("Stop Emulation"), QStringLiteral("Main Window"), {QStringLiteral("F5"), Qt::WindowShortcut}}, {QStringLiteral("Toggle Filter Bar"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+F"), Qt::WindowShortcut}}, {QStringLiteral("Toggle Speed Limit"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+Z"), Qt::ApplicationShortcut}}, {QStringLiteral("Toggle Status Bar"), QStringLiteral("Main Window"), {QStringLiteral("Ctrl+S"), Qt::WindowShortcut}}, - {QStringLiteral("Change Docked Mode"), QStringLiteral("Main Window"), {QStringLiteral("F10"), Qt::ApplicationShortcut}}, }}; // clang-format on -void Config::ReadPlayerValues() { - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { - auto& player = Settings::values.players[p]; +void Config::Initialize(const std::string& config_name) { + switch (type) { + case ConfigType::GlobalConfig: + qt_config_loc = fmt::format("{}" DIR_SEP "{}.ini", FS::GetUserPath(FS::UserPath::ConfigDir), + config_name); + FS::CreateFullPath(qt_config_loc); + qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), + QSettings::IniFormat); + Reload(); + break; + case ConfigType::PerGameConfig: + qt_config_loc = fmt::format("{}custom" DIR_SEP "{}.ini", + FS::GetUserPath(FS::UserPath::ConfigDir), config_name); + FS::CreateFullPath(qt_config_loc); + qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), + QSettings::IniFormat); + Reload(); + break; + case ConfigType::InputProfile: + qt_config_loc = fmt::format("{}input" DIR_SEP "{}.ini", + FS::GetUserPath(FS::UserPath::ConfigDir), config_name); + FS::CreateFullPath(qt_config_loc); + qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), + QSettings::IniFormat); + break; + } +} + +void Config::ReadPlayerValue(std::size_t player_index) { + const QString player_prefix = [this, player_index] { + if (type == ConfigType::InputProfile) { + return QString{}; + } else { + return QStringLiteral("player_%1_").arg(player_index); + } + }(); + + auto& player = Settings::values.players.GetValue()[player_index]; + + if (player_prefix.isEmpty()) { + const auto controller = static_cast<Settings::ControllerType>( + qt_config + ->value(QStringLiteral("%1type").arg(player_prefix), + static_cast<u8>(Settings::ControllerType::ProController)) + .toUInt()); + if (controller == Settings::ControllerType::LeftJoycon || + controller == Settings::ControllerType::RightJoycon) { + player.controller_type = controller; + } + } else { player.connected = - ReadSetting(QStringLiteral("player_%1_connected").arg(p), false).toBool(); + ReadSetting(QStringLiteral("%1connected").arg(player_prefix), player_index == 0) + .toBool(); - player.type = static_cast<Settings::ControllerType>( + player.controller_type = static_cast<Settings::ControllerType>( qt_config - ->value(QStringLiteral("player_%1_type").arg(p), - static_cast<u8>(Settings::ControllerType::DualJoycon)) + ->value(QStringLiteral("%1type").arg(player_prefix), + static_cast<u8>(Settings::ControllerType::ProController)) .toUInt()); + player.vibration_enabled = + qt_config->value(QStringLiteral("%1vibration_enabled").arg(player_prefix), true) + .toBool(); + + player.vibration_strength = + qt_config->value(QStringLiteral("%1vibration_strength").arg(player_prefix), 100) + .toInt(); + player.body_color_left = qt_config - ->value(QStringLiteral("player_%1_body_color_left").arg(p), + ->value(QStringLiteral("%1body_color_left").arg(player_prefix), Settings::JOYCON_BODY_NEON_BLUE) .toUInt(); - player.body_color_right = qt_config - ->value(QStringLiteral("player_%1_body_color_right").arg(p), - Settings::JOYCON_BODY_NEON_RED) - .toUInt(); - player.button_color_left = qt_config - ->value(QStringLiteral("player_%1_button_color_left").arg(p), - Settings::JOYCON_BUTTONS_NEON_BLUE) - .toUInt(); + player.body_color_right = + qt_config + ->value(QStringLiteral("%1body_color_right").arg(player_prefix), + Settings::JOYCON_BODY_NEON_RED) + .toUInt(); + player.button_color_left = + qt_config + ->value(QStringLiteral("%1button_color_left").arg(player_prefix), + Settings::JOYCON_BUTTONS_NEON_BLUE) + .toUInt(); player.button_color_right = qt_config - ->value(QStringLiteral("player_%1_button_color_right").arg(p), + ->value(QStringLiteral("%1button_color_right").arg(player_prefix), Settings::JOYCON_BUTTONS_NEON_RED) .toUInt(); + } - for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { - const std::string default_param = - InputCommon::GenerateKeyboardParam(default_buttons[i]); - auto& player_buttons = player.buttons[i]; - - player_buttons = qt_config - ->value(QStringLiteral("player_%1_").arg(p) + - QString::fromUtf8(Settings::NativeButton::mapping[i]), - QString::fromStdString(default_param)) - .toString() - .toStdString(); - if (player_buttons.empty()) { - player_buttons = default_param; - } + for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_buttons[i]); + auto& player_buttons = player.buttons[i]; + + player_buttons = qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeButton::mapping[i]), + QString::fromStdString(default_param)) + .toString() + .toStdString(); + if (player_buttons.empty()) { + player_buttons = default_param; } + } - for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { - const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( - default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], - default_analogs[i][3], default_analogs[i][4], 0.5f); - auto& player_analogs = player.analogs[i]; - - player_analogs = qt_config - ->value(QStringLiteral("player_%1_").arg(p) + - QString::fromUtf8(Settings::NativeAnalog::mapping[i]), - QString::fromStdString(default_param)) - .toString() - .toStdString(); - if (player_analogs.empty()) { - player_analogs = default_param; - } + for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { + const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( + default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], + default_analogs[i][3], default_stick_mod[i], 0.5f); + auto& player_analogs = player.analogs[i]; + + player_analogs = qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeAnalog::mapping[i]), + QString::fromStdString(default_param)) + .toString() + .toStdString(); + if (player_analogs.empty()) { + player_analogs = default_param; } } - std::stable_partition( - Settings::values.players.begin(), - Settings::values.players.begin() + - Service::HID::Controller_NPad::NPadIdToIndex(Service::HID::NPAD_HANDHELD), - [](const auto& player) { return player.connected; }); + for (int i = 0; i < Settings::NativeVibration::NumVibrations; ++i) { + auto& player_vibrations = player.vibrations[i]; + + player_vibrations = + qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeVibration::mapping[i]), + QString{}) + .toString() + .toStdString(); + } + + for (int i = 0; i < Settings::NativeMotion::NumMotions; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_motions[i]); + auto& player_motions = player.motions[i]; + + player_motions = qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeMotion::mapping[i]), + QString::fromStdString(default_param)) + .toString() + .toStdString(); + if (player_motions.empty()) { + player_motions = default_param; + } + } } void Config::ReadDebugValues() { @@ -325,7 +411,7 @@ void Config::ReadDebugValues() { for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], - default_analogs[i][3], default_analogs[i][4], 0.5f); + default_analogs[i][3], default_stick_mod[i], 0.5f); auto& debug_pad_analogs = Settings::values.debug_pad_analogs[i]; debug_pad_analogs = qt_config @@ -392,26 +478,22 @@ void Config::ReadTouchscreenValues() { ReadSetting(QStringLiteral("touchscreen_diameter_y"), 15).toUInt(); } -void Config::ApplyDefaultProfileIfInputInvalid() { - if (!std::any_of(Settings::values.players.begin(), Settings::values.players.end(), - [](const Settings::PlayerInput& in) { return in.connected; })) { - ApplyInputProfileConfiguration(UISettings::values.profile_index); - } -} - void Config::ReadAudioValues() { qt_config->beginGroup(QStringLiteral("Audio")); - Settings::values.sink_id = ReadSetting(QStringLiteral("output_engine"), QStringLiteral("auto")) - .toString() - .toStdString(); - Settings::values.enable_audio_stretching = - ReadSetting(QStringLiteral("enable_audio_stretching"), true).toBool(); - Settings::values.audio_device_id = - ReadSetting(QStringLiteral("output_device"), QStringLiteral("auto")) - .toString() - .toStdString(); - Settings::values.volume = ReadSetting(QStringLiteral("volume"), 1).toFloat(); + if (global) { + Settings::values.sink_id = + ReadSetting(QStringLiteral("output_engine"), QStringLiteral("auto")) + .toString() + .toStdString(); + Settings::values.audio_device_id = + ReadSetting(QStringLiteral("output_device"), QStringLiteral("auto")) + .toString() + .toStdString(); + } + ReadSettingGlobal(Settings::values.enable_audio_stretching, + QStringLiteral("enable_audio_stretching"), true); + ReadSettingGlobal(Settings::values.volume, QStringLiteral("volume"), 1); qt_config->endGroup(); } @@ -419,17 +501,73 @@ void Config::ReadAudioValues() { void Config::ReadControlValues() { qt_config->beginGroup(QStringLiteral("Controls")); - ReadPlayerValues(); + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { + ReadPlayerValue(p); + } ReadDebugValues(); ReadKeyboardValues(); ReadMouseValues(); ReadTouchscreenValues(); + ReadMotionTouchValues(); + + ReadSettingGlobal(Settings::values.use_docked_mode, QStringLiteral("use_docked_mode"), false); + ReadSettingGlobal(Settings::values.vibration_enabled, QStringLiteral("vibration_enabled"), + true); + ReadSettingGlobal(Settings::values.enable_accurate_vibrations, + QStringLiteral("enable_accurate_vibrations"), false); + ReadSettingGlobal(Settings::values.motion_enabled, QStringLiteral("motion_enabled"), true); + + qt_config->endGroup(); +} + +void Config::ReadMotionTouchValues() { + int num_touch_from_button_maps = + qt_config->beginReadArray(QStringLiteral("touch_from_button_maps")); + + if (num_touch_from_button_maps > 0) { + const auto append_touch_from_button_map = [this] { + Settings::TouchFromButtonMap map; + map.name = ReadSetting(QStringLiteral("name"), QStringLiteral("default")) + .toString() + .toStdString(); + const int num_touch_maps = qt_config->beginReadArray(QStringLiteral("entries")); + map.buttons.reserve(num_touch_maps); + for (int i = 0; i < num_touch_maps; i++) { + qt_config->setArrayIndex(i); + std::string touch_mapping = + ReadSetting(QStringLiteral("bind")).toString().toStdString(); + map.buttons.emplace_back(std::move(touch_mapping)); + } + qt_config->endArray(); // entries + Settings::values.touch_from_button_maps.emplace_back(std::move(map)); + }; + + for (int i = 0; i < num_touch_from_button_maps; ++i) { + qt_config->setArrayIndex(i); + append_touch_from_button_map(); + } + } else { + Settings::values.touch_from_button_maps.emplace_back( + Settings::TouchFromButtonMap{"default", {}}); + num_touch_from_button_maps = 1; + } + qt_config->endArray(); Settings::values.motion_device = ReadSetting(QStringLiteral("motion_device"), QStringLiteral("engine:motion_emu,update_period:100,sensitivity:0.01")) .toString() .toStdString(); + Settings::values.touch_device = + ReadSetting(QStringLiteral("touch_device"), QStringLiteral("engine:emu_window")) + .toString() + .toStdString(); + Settings::values.use_touch_from_button = + ReadSetting(QStringLiteral("use_touch_from_button"), false).toBool(); + Settings::values.touch_from_button_map_index = + ReadSetting(QStringLiteral("touch_from_button_map"), 0).toInt(); + Settings::values.touch_from_button_map_index = + std::clamp(Settings::values.touch_from_button_map_index, 0, num_touch_from_button_maps - 1); Settings::values.udp_input_address = ReadSetting(QStringLiteral("udp_input_address"), QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_ADDR)) @@ -440,14 +578,12 @@ void Config::ReadControlValues() { .toInt()); Settings::values.udp_pad_index = static_cast<u8>(ReadSetting(QStringLiteral("udp_pad_index"), 0).toUInt()); - - qt_config->endGroup(); } void Config::ReadCoreValues() { qt_config->beginGroup(QStringLiteral("Core")); - Settings::values.use_multi_core = ReadSetting(QStringLiteral("use_multi_core"), false).toBool(); + ReadSettingGlobal(Settings::values.use_multi_core, QStringLiteral("use_multi_core"), true); qt_config->endGroup(); } @@ -456,63 +592,42 @@ void Config::ReadDataStorageValues() { qt_config->beginGroup(QStringLiteral("Data Storage")); Settings::values.use_virtual_sd = ReadSetting(QStringLiteral("use_virtual_sd"), true).toBool(); - FileUtil::GetUserPath( - FileUtil::UserPath::NANDDir, - qt_config - ->value(QStringLiteral("nand_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir))) - .toString() - .toStdString()); - FileUtil::GetUserPath( - FileUtil::UserPath::SDMCDir, - qt_config - ->value(QStringLiteral("sdmc_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir))) - .toString() - .toStdString()); - FileUtil::GetUserPath( - FileUtil::UserPath::LoadDir, - qt_config - ->value(QStringLiteral("load_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir))) - .toString() - .toStdString()); - FileUtil::GetUserPath( - FileUtil::UserPath::DumpDir, - qt_config - ->value(QStringLiteral("dump_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::DumpDir))) - .toString() - .toStdString()); - FileUtil::GetUserPath( - FileUtil::UserPath::CacheDir, - qt_config - ->value(QStringLiteral("cache_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir))) - .toString() - .toStdString()); + FS::GetUserPath(FS::UserPath::NANDDir, + qt_config + ->value(QStringLiteral("nand_directory"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::NANDDir))) + .toString() + .toStdString()); + FS::GetUserPath(FS::UserPath::SDMCDir, + qt_config + ->value(QStringLiteral("sdmc_directory"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::SDMCDir))) + .toString() + .toStdString()); + FS::GetUserPath(FS::UserPath::LoadDir, + qt_config + ->value(QStringLiteral("load_directory"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::LoadDir))) + .toString() + .toStdString()); + FS::GetUserPath(FS::UserPath::DumpDir, + qt_config + ->value(QStringLiteral("dump_directory"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::DumpDir))) + .toString() + .toStdString()); + FS::GetUserPath(FS::UserPath::CacheDir, + qt_config + ->value(QStringLiteral("cache_directory"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::CacheDir))) + .toString() + .toStdString()); Settings::values.gamecard_inserted = ReadSetting(QStringLiteral("gamecard_inserted"), false).toBool(); Settings::values.gamecard_current_game = ReadSetting(QStringLiteral("gamecard_current_game"), false).toBool(); Settings::values.gamecard_path = - ReadSetting(QStringLiteral("gamecard_path"), QStringLiteral("")).toString().toStdString(); - Settings::values.nand_total_size = static_cast<Settings::NANDTotalSize>( - ReadSetting(QStringLiteral("nand_total_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::NANDTotalSize::S29_1GB))) - .toULongLong()); - Settings::values.nand_user_size = static_cast<Settings::NANDUserSize>( - ReadSetting(QStringLiteral("nand_user_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::NANDUserSize::S26GB))) - .toULongLong()); - Settings::values.nand_system_size = static_cast<Settings::NANDSystemSize>( - ReadSetting(QStringLiteral("nand_system_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::NANDSystemSize::S2_5GB))) - .toULongLong()); - Settings::values.sdmc_size = static_cast<Settings::SDMCSize>( - ReadSetting(QStringLiteral("sdmc_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::SDMCSize::S16GB))) - .toULongLong()); + ReadSetting(QStringLiteral("gamecard_path"), QString{}).toString().toStdString(); qt_config->endGroup(); } @@ -526,12 +641,16 @@ void Config::ReadDebuggingValues() { Settings::values.use_gdbstub = ReadSetting(QStringLiteral("use_gdbstub"), false).toBool(); Settings::values.gdbstub_port = ReadSetting(QStringLiteral("gdbstub_port"), 24689).toInt(); Settings::values.program_args = - ReadSetting(QStringLiteral("program_args"), QStringLiteral("")).toString().toStdString(); + ReadSetting(QStringLiteral("program_args"), QString{}).toString().toStdString(); Settings::values.dump_exefs = ReadSetting(QStringLiteral("dump_exefs"), false).toBool(); Settings::values.dump_nso = ReadSetting(QStringLiteral("dump_nso"), false).toBool(); Settings::values.reporting_services = ReadSetting(QStringLiteral("reporting_services"), false).toBool(); Settings::values.quest_flag = ReadSetting(QStringLiteral("quest_flag"), false).toBool(); + Settings::values.disable_macro_jit = + ReadSetting(QStringLiteral("disable_macro_jit"), false).toBool(); + Settings::values.extended_logging = + ReadSetting(QStringLiteral("extended_logging"), false).toBool(); qt_config->endGroup(); } @@ -557,8 +676,7 @@ void Config::ReadDisabledAddOnValues() { const auto d_size = qt_config->beginReadArray(QStringLiteral("disabled")); for (int j = 0; j < d_size; ++j) { qt_config->setArrayIndex(j); - out.push_back( - ReadSetting(QStringLiteral("d"), QStringLiteral("")).toString().toStdString()); + out.push_back(ReadSetting(QStringLiteral("d"), QString{}).toString().toStdString()); } qt_config->endArray(); Settings::values.disabled_addons.insert_or_assign(title_id, out); @@ -584,7 +702,6 @@ void Config::ReadPathValues() { UISettings::values.roms_path = ReadSetting(QStringLiteral("romsPath")).toString(); UISettings::values.symbols_path = ReadSetting(QStringLiteral("symbolsPath")).toString(); - UISettings::values.screenshot_path = ReadSetting(QStringLiteral("screenshotPath")).toString(); UISettings::values.game_dir_deprecated = ReadSetting(QStringLiteral("gameListRootDir"), QStringLiteral(".")).toString(); UISettings::values.game_dir_deprecated_deepscan = @@ -617,6 +734,40 @@ void Config::ReadPathValues() { } } UISettings::values.recent_files = ReadSetting(QStringLiteral("recentFiles")).toStringList(); + UISettings::values.language = ReadSetting(QStringLiteral("language"), QString{}).toString(); + + qt_config->endGroup(); +} + +void Config::ReadCpuValues() { + qt_config->beginGroup(QStringLiteral("Cpu")); + + if (global) { + Settings::values.cpu_accuracy = static_cast<Settings::CPUAccuracy>( + ReadSetting(QStringLiteral("cpu_accuracy"), 0).toInt()); + + Settings::values.cpuopt_page_tables = + ReadSetting(QStringLiteral("cpuopt_page_tables"), true).toBool(); + Settings::values.cpuopt_block_linking = + ReadSetting(QStringLiteral("cpuopt_block_linking"), true).toBool(); + Settings::values.cpuopt_return_stack_buffer = + ReadSetting(QStringLiteral("cpuopt_return_stack_buffer"), true).toBool(); + Settings::values.cpuopt_fast_dispatcher = + ReadSetting(QStringLiteral("cpuopt_fast_dispatcher"), true).toBool(); + Settings::values.cpuopt_context_elimination = + ReadSetting(QStringLiteral("cpuopt_context_elimination"), true).toBool(); + Settings::values.cpuopt_const_prop = + ReadSetting(QStringLiteral("cpuopt_const_prop"), true).toBool(); + Settings::values.cpuopt_misc_ir = + ReadSetting(QStringLiteral("cpuopt_misc_ir"), true).toBool(); + Settings::values.cpuopt_reduce_misalign_checks = + ReadSetting(QStringLiteral("cpuopt_reduce_misalign_checks"), true).toBool(); + + Settings::values.cpuopt_unsafe_unfuse_fma = + ReadSetting(QStringLiteral("cpuopt_unsafe_unfuse_fma"), true).toBool(); + Settings::values.cpuopt_unsafe_reduce_fp_error = + ReadSetting(QStringLiteral("cpuopt_unsafe_reduce_fp_error"), true).toBool(); + } qt_config->endGroup(); } @@ -624,30 +775,46 @@ void Config::ReadPathValues() { void Config::ReadRendererValues() { qt_config->beginGroup(QStringLiteral("Renderer")); - Settings::values.renderer_backend = - static_cast<Settings::RendererBackend>(ReadSetting(QStringLiteral("backend"), 0).toInt()); - Settings::values.renderer_debug = ReadSetting(QStringLiteral("debug"), false).toBool(); - Settings::values.vulkan_device = ReadSetting(QStringLiteral("vulkan_device"), 0).toInt(); - Settings::values.resolution_factor = - ReadSetting(QStringLiteral("resolution_factor"), 1.0).toFloat(); - Settings::values.aspect_ratio = ReadSetting(QStringLiteral("aspect_ratio"), 0).toInt(); - Settings::values.max_anisotropy = ReadSetting(QStringLiteral("max_anisotropy"), 0).toInt(); - Settings::values.use_frame_limit = - ReadSetting(QStringLiteral("use_frame_limit"), true).toBool(); - Settings::values.frame_limit = ReadSetting(QStringLiteral("frame_limit"), 100).toInt(); - Settings::values.use_disk_shader_cache = - ReadSetting(QStringLiteral("use_disk_shader_cache"), true).toBool(); - Settings::values.use_accurate_gpu_emulation = - ReadSetting(QStringLiteral("use_accurate_gpu_emulation"), false).toBool(); - Settings::values.use_asynchronous_gpu_emulation = - ReadSetting(QStringLiteral("use_asynchronous_gpu_emulation"), false).toBool(); - Settings::values.use_vsync = ReadSetting(QStringLiteral("use_vsync"), true).toBool(); - Settings::values.force_30fps_mode = - ReadSetting(QStringLiteral("force_30fps_mode"), false).toBool(); - - Settings::values.bg_red = ReadSetting(QStringLiteral("bg_red"), 0.0).toFloat(); - Settings::values.bg_green = ReadSetting(QStringLiteral("bg_green"), 0.0).toFloat(); - Settings::values.bg_blue = ReadSetting(QStringLiteral("bg_blue"), 0.0).toFloat(); + ReadSettingGlobal(Settings::values.renderer_backend, QStringLiteral("backend"), 0); + ReadSettingGlobal(Settings::values.renderer_debug, QStringLiteral("debug"), false); + ReadSettingGlobal(Settings::values.vulkan_device, QStringLiteral("vulkan_device"), 0); + ReadSettingGlobal(Settings::values.aspect_ratio, QStringLiteral("aspect_ratio"), 0); + ReadSettingGlobal(Settings::values.max_anisotropy, QStringLiteral("max_anisotropy"), 0); + ReadSettingGlobal(Settings::values.use_frame_limit, QStringLiteral("use_frame_limit"), true); + ReadSettingGlobal(Settings::values.frame_limit, QStringLiteral("frame_limit"), 100); + ReadSettingGlobal(Settings::values.use_disk_shader_cache, + QStringLiteral("use_disk_shader_cache"), true); + ReadSettingGlobal(Settings::values.gpu_accuracy, QStringLiteral("gpu_accuracy"), 0); + ReadSettingGlobal(Settings::values.use_asynchronous_gpu_emulation, + QStringLiteral("use_asynchronous_gpu_emulation"), true); + ReadSettingGlobal(Settings::values.use_nvdec_emulation, QStringLiteral("use_nvdec_emulation"), + true); + ReadSettingGlobal(Settings::values.use_vsync, QStringLiteral("use_vsync"), true); + ReadSettingGlobal(Settings::values.use_assembly_shaders, QStringLiteral("use_assembly_shaders"), + true); + ReadSettingGlobal(Settings::values.use_asynchronous_shaders, + QStringLiteral("use_asynchronous_shaders"), false); + ReadSettingGlobal(Settings::values.use_fast_gpu_time, QStringLiteral("use_fast_gpu_time"), + true); + ReadSettingGlobal(Settings::values.bg_red, QStringLiteral("bg_red"), 0.0); + ReadSettingGlobal(Settings::values.bg_green, QStringLiteral("bg_green"), 0.0); + ReadSettingGlobal(Settings::values.bg_blue, QStringLiteral("bg_blue"), 0.0); + + qt_config->endGroup(); +} + +void Config::ReadScreenshotValues() { + qt_config->beginGroup(QStringLiteral("Screenshots")); + + UISettings::values.enable_screenshot_save_as = + ReadSetting(QStringLiteral("enable_screenshot_save_as"), true).toBool(); + FS::GetUserPath( + FS::UserPath::ScreenshotsDir, + qt_config + ->value(QStringLiteral("screenshot_path"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::ScreenshotsDir))) + .toString() + .toStdString()); qt_config->endGroup(); } @@ -659,11 +826,13 @@ void Config::ReadShortcutValues() { const auto& [keyseq, context] = shortcut; qt_config->beginGroup(group); qt_config->beginGroup(name); + // No longer using ReadSetting for shortcut.second as it innacurately returns a value of 1 + // for WidgetWithChildrenShortcut which is a value of 3. Needed to fix shortcuts the open + // a file dialog in windowed mode UISettings::values.shortcuts.push_back( {name, group, - {ReadSetting(QStringLiteral("KeySeq"), keyseq).toString(), - ReadSetting(QStringLiteral("Context"), context).toInt()}}); + {ReadSetting(QStringLiteral("KeySeq"), keyseq).toString(), shortcut.second}}); qt_config->endGroup(); qt_config->endGroup(); } @@ -674,33 +843,45 @@ void Config::ReadShortcutValues() { void Config::ReadSystemValues() { qt_config->beginGroup(QStringLiteral("System")); - Settings::values.use_docked_mode = - ReadSetting(QStringLiteral("use_docked_mode"), false).toBool(); + ReadSettingGlobal(Settings::values.current_user, QStringLiteral("current_user"), 0); + Settings::values.current_user = + std::clamp<int>(Settings::values.current_user, 0, Service::Account::MAX_USERS - 1); - Settings::values.current_user = std::clamp<int>( - ReadSetting(QStringLiteral("current_user"), 0).toInt(), 0, Service::Account::MAX_USERS - 1); + ReadSettingGlobal(Settings::values.language_index, QStringLiteral("language_index"), 1); - Settings::values.language_index = ReadSetting(QStringLiteral("language_index"), 1).toInt(); + ReadSettingGlobal(Settings::values.region_index, QStringLiteral("region_index"), 1); - Settings::values.region_index = ReadSetting(QStringLiteral("region_index"), 1).toInt(); + ReadSettingGlobal(Settings::values.time_zone_index, QStringLiteral("time_zone_index"), 0); - const auto rng_seed_enabled = ReadSetting(QStringLiteral("rng_seed_enabled"), false).toBool(); - if (rng_seed_enabled) { - Settings::values.rng_seed = ReadSetting(QStringLiteral("rng_seed"), 0).toULongLong(); - } else { - Settings::values.rng_seed = std::nullopt; + bool rng_seed_enabled; + ReadSettingGlobal(rng_seed_enabled, QStringLiteral("rng_seed_enabled"), false); + bool rng_seed_global = + global || qt_config->value(QStringLiteral("rng_seed/use_global"), true).toBool(); + Settings::values.rng_seed.SetGlobal(rng_seed_global); + if (global || !rng_seed_global) { + if (rng_seed_enabled) { + Settings::values.rng_seed.SetValue( + ReadSetting(QStringLiteral("rng_seed"), 0).toULongLong()); + } else { + Settings::values.rng_seed.SetValue(std::nullopt); + } } - const auto custom_rtc_enabled = - ReadSetting(QStringLiteral("custom_rtc_enabled"), false).toBool(); - if (custom_rtc_enabled) { - Settings::values.custom_rtc = - std::chrono::seconds(ReadSetting(QStringLiteral("custom_rtc"), 0).toULongLong()); - } else { - Settings::values.custom_rtc = std::nullopt; + bool custom_rtc_enabled; + ReadSettingGlobal(custom_rtc_enabled, QStringLiteral("custom_rtc_enabled"), false); + bool custom_rtc_global = + global || qt_config->value(QStringLiteral("custom_rtc/use_global"), true).toBool(); + Settings::values.custom_rtc.SetGlobal(custom_rtc_global); + if (global || !custom_rtc_global) { + if (custom_rtc_enabled) { + Settings::values.custom_rtc.SetValue( + std::chrono::seconds(ReadSetting(QStringLiteral("custom_rtc"), 0).toULongLong())); + } else { + Settings::values.custom_rtc.SetValue(std::nullopt); + } } - Settings::values.sound_index = ReadSetting(QStringLiteral("sound_index"), 1).toInt(); + ReadSettingGlobal(Settings::values.sound_index, QStringLiteral("sound_index"), 1); qt_config->endGroup(); } @@ -713,14 +894,13 @@ void Config::ReadUIValues() { .toString(); UISettings::values.enable_discord_presence = ReadSetting(QStringLiteral("enable_discord_presence"), true).toBool(); - UISettings::values.screenshot_resolution_factor = - static_cast<u16>(ReadSetting(QStringLiteral("screenshot_resolution_factor"), 0).toUInt()); UISettings::values.select_user_on_boot = ReadSetting(QStringLiteral("select_user_on_boot"), false).toBool(); ReadUIGamelistValues(); ReadUILayoutValues(); ReadPathValues(); + ReadScreenshotValues(); ReadShortcutValues(); UISettings::values.single_window_mode = @@ -737,11 +917,10 @@ void Config::ReadUIValues() { UISettings::values.first_start = ReadSetting(QStringLiteral("firstStart"), true).toBool(); UISettings::values.callout_flags = ReadSetting(QStringLiteral("calloutFlags"), 0).toUInt(); UISettings::values.show_console = ReadSetting(QStringLiteral("showConsole"), false).toBool(); - UISettings::values.profile_index = ReadSetting(QStringLiteral("profileIndex"), 0).toUInt(); UISettings::values.pause_when_in_background = ReadSetting(QStringLiteral("pauseWhenInBackground"), false).toBool(); - - ApplyDefaultProfileIfInputInvalid(); + UISettings::values.hide_mouse = + ReadSetting(QStringLiteral("hideInactiveMouse"), false).toBool(); qt_config->endGroup(); } @@ -794,54 +973,81 @@ void Config::ReadWebServiceValues() { } void Config::ReadValues() { - ReadControlValues(); + if (global) { + ReadControlValues(); + ReadDataStorageValues(); + ReadDebuggingValues(); + ReadDisabledAddOnValues(); + ReadServiceValues(); + ReadUIValues(); + ReadWebServiceValues(); + ReadMiscellaneousValues(); + } ReadCoreValues(); + ReadCpuValues(); ReadRendererValues(); ReadAudioValues(); - ReadDataStorageValues(); ReadSystemValues(); - ReadMiscellaneousValues(); - ReadDebuggingValues(); - ReadWebServiceValues(); - ReadServiceValues(); - ReadDisabledAddOnValues(); - ReadUIValues(); } -void Config::SavePlayerValues() { - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { - const auto& player = Settings::values.players[p]; +void Config::SavePlayerValue(std::size_t player_index) { + const QString player_prefix = [this, player_index] { + if (type == ConfigType::InputProfile) { + return QString{}; + } else { + return QStringLiteral("player_%1_").arg(player_index); + } + }(); + + const auto& player = Settings::values.players.GetValue()[player_index]; - WriteSetting(QStringLiteral("player_%1_connected").arg(p), player.connected, false); - WriteSetting(QStringLiteral("player_%1_type").arg(p), static_cast<u8>(player.type), - static_cast<u8>(Settings::ControllerType::DualJoycon)); + WriteSetting(QStringLiteral("%1type").arg(player_prefix), + static_cast<u8>(player.controller_type), + static_cast<u8>(Settings::ControllerType::ProController)); - WriteSetting(QStringLiteral("player_%1_body_color_left").arg(p), player.body_color_left, + if (!player_prefix.isEmpty()) { + WriteSetting(QStringLiteral("%1connected").arg(player_prefix), player.connected, false); + WriteSetting(QStringLiteral("%1vibration_enabled").arg(player_prefix), + player.vibration_enabled, true); + WriteSetting(QStringLiteral("%1vibration_strength").arg(player_prefix), + player.vibration_strength, 100); + WriteSetting(QStringLiteral("%1body_color_left").arg(player_prefix), player.body_color_left, Settings::JOYCON_BODY_NEON_BLUE); - WriteSetting(QStringLiteral("player_%1_body_color_right").arg(p), player.body_color_right, - Settings::JOYCON_BODY_NEON_RED); - WriteSetting(QStringLiteral("player_%1_button_color_left").arg(p), player.button_color_left, - Settings::JOYCON_BUTTONS_NEON_BLUE); - WriteSetting(QStringLiteral("player_%1_button_color_right").arg(p), + WriteSetting(QStringLiteral("%1body_color_right").arg(player_prefix), + player.body_color_right, Settings::JOYCON_BODY_NEON_RED); + WriteSetting(QStringLiteral("%1button_color_left").arg(player_prefix), + player.button_color_left, Settings::JOYCON_BUTTONS_NEON_BLUE); + WriteSetting(QStringLiteral("%1button_color_right").arg(player_prefix), player.button_color_right, Settings::JOYCON_BUTTONS_NEON_RED); + } - for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { - const std::string default_param = - InputCommon::GenerateKeyboardParam(default_buttons[i]); - WriteSetting(QStringLiteral("player_%1_").arg(p) + - QString::fromStdString(Settings::NativeButton::mapping[i]), - QString::fromStdString(player.buttons[i]), - QString::fromStdString(default_param)); - } - for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { - const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( - default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], - default_analogs[i][3], default_analogs[i][4], 0.5f); - WriteSetting(QStringLiteral("player_%1_").arg(p) + - QString::fromStdString(Settings::NativeAnalog::mapping[i]), - QString::fromStdString(player.analogs[i]), - QString::fromStdString(default_param)); - } + for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_buttons[i]); + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeButton::mapping[i]), + QString::fromStdString(player.buttons[i]), + QString::fromStdString(default_param)); + } + for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { + const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( + default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], + default_analogs[i][3], default_stick_mod[i], 0.5f); + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeAnalog::mapping[i]), + QString::fromStdString(player.analogs[i]), + QString::fromStdString(default_param)); + } + for (int i = 0; i < Settings::NativeVibration::NumVibrations; ++i) { + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeVibration::mapping[i]), + QString::fromStdString(player.vibrations[i]), QString{}); + } + for (int i = 0; i < Settings::NativeMotion::NumMotions; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_motions[i]); + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeMotion::mapping[i]), + QString::fromStdString(player.motions[i]), + QString::fromStdString(default_param)); } } @@ -857,7 +1063,7 @@ void Config::SaveDebugValues() { for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], - default_analogs[i][3], default_analogs[i][4], 0.5f); + default_analogs[i][3], default_stick_mod[i], 0.5f); WriteSetting(QStringLiteral("debug_pad_") + QString::fromStdString(Settings::NativeAnalog::mapping[i]), QString::fromStdString(Settings::values.debug_pad_analogs[i]), @@ -891,31 +1097,74 @@ void Config::SaveTouchscreenValues() { WriteSetting(QStringLiteral("touchscreen_diameter_y"), touchscreen.diameter_y, 15); } +void Config::SaveMotionTouchValues() { + WriteSetting(QStringLiteral("motion_device"), + QString::fromStdString(Settings::values.motion_device), + QStringLiteral("engine:motion_emu,update_period:100,sensitivity:0.01")); + WriteSetting(QStringLiteral("touch_device"), + QString::fromStdString(Settings::values.touch_device), + QStringLiteral("engine:emu_window")); + WriteSetting(QStringLiteral("use_touch_from_button"), Settings::values.use_touch_from_button, + false); + WriteSetting(QStringLiteral("touch_from_button_map"), + Settings::values.touch_from_button_map_index, 0); + WriteSetting(QStringLiteral("udp_input_address"), + QString::fromStdString(Settings::values.udp_input_address), + QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_ADDR)); + WriteSetting(QStringLiteral("udp_input_port"), Settings::values.udp_input_port, + InputCommon::CemuhookUDP::DEFAULT_PORT); + WriteSetting(QStringLiteral("udp_pad_index"), Settings::values.udp_pad_index, 0); + + qt_config->beginWriteArray(QStringLiteral("touch_from_button_maps")); + for (std::size_t p = 0; p < Settings::values.touch_from_button_maps.size(); ++p) { + qt_config->setArrayIndex(static_cast<int>(p)); + WriteSetting(QStringLiteral("name"), + QString::fromStdString(Settings::values.touch_from_button_maps[p].name), + QStringLiteral("default")); + qt_config->beginWriteArray(QStringLiteral("entries")); + for (std::size_t q = 0; q < Settings::values.touch_from_button_maps[p].buttons.size(); + ++q) { + qt_config->setArrayIndex(static_cast<int>(q)); + WriteSetting( + QStringLiteral("bind"), + QString::fromStdString(Settings::values.touch_from_button_maps[p].buttons[q])); + } + qt_config->endArray(); + } + qt_config->endArray(); +} + void Config::SaveValues() { - SaveControlValues(); + if (global) { + SaveControlValues(); + SaveDataStorageValues(); + SaveDebuggingValues(); + SaveDisabledAddOnValues(); + SaveServiceValues(); + SaveUIValues(); + SaveWebServiceValues(); + SaveMiscellaneousValues(); + } SaveCoreValues(); + SaveCpuValues(); SaveRendererValues(); SaveAudioValues(); - SaveDataStorageValues(); SaveSystemValues(); - SaveMiscellaneousValues(); - SaveDebuggingValues(); - SaveWebServiceValues(); - SaveServiceValues(); - SaveDisabledAddOnValues(); - SaveUIValues(); } void Config::SaveAudioValues() { qt_config->beginGroup(QStringLiteral("Audio")); - WriteSetting(QStringLiteral("output_engine"), QString::fromStdString(Settings::values.sink_id), - QStringLiteral("auto")); - WriteSetting(QStringLiteral("enable_audio_stretching"), - Settings::values.enable_audio_stretching, true); - WriteSetting(QStringLiteral("output_device"), - QString::fromStdString(Settings::values.audio_device_id), QStringLiteral("auto")); - WriteSetting(QStringLiteral("volume"), Settings::values.volume, 1.0f); + if (global) { + WriteSetting(QStringLiteral("output_engine"), + QString::fromStdString(Settings::values.sink_id), QStringLiteral("auto")); + WriteSetting(QStringLiteral("output_device"), + QString::fromStdString(Settings::values.audio_device_id), + QStringLiteral("auto")); + } + WriteSettingGlobal(QStringLiteral("enable_audio_stretching"), + Settings::values.enable_audio_stretching, true); + WriteSettingGlobal(QStringLiteral("volume"), Settings::values.volume, 1.0f); qt_config->endGroup(); } @@ -923,21 +1172,27 @@ void Config::SaveAudioValues() { void Config::SaveControlValues() { qt_config->beginGroup(QStringLiteral("Controls")); - SavePlayerValues(); + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { + SavePlayerValue(p); + } SaveDebugValues(); SaveMouseValues(); SaveTouchscreenValues(); - + SaveMotionTouchValues(); + + WriteSettingGlobal(QStringLiteral("use_docked_mode"), Settings::values.use_docked_mode, false); + WriteSettingGlobal(QStringLiteral("vibration_enabled"), Settings::values.vibration_enabled, + true); + WriteSettingGlobal(QStringLiteral("enable_accurate_vibrations"), + Settings::values.enable_accurate_vibrations, false); + WriteSettingGlobal(QStringLiteral("motion_enabled"), Settings::values.motion_enabled, true); WriteSetting(QStringLiteral("motion_device"), QString::fromStdString(Settings::values.motion_device), QStringLiteral("engine:motion_emu,update_period:100,sensitivity:0.01")); + WriteSetting(QStringLiteral("touch_device"), + QString::fromStdString(Settings::values.touch_device), + QStringLiteral("engine:emu_window")); WriteSetting(QStringLiteral("keyboard_enabled"), Settings::values.keyboard_enabled, false); - WriteSetting(QStringLiteral("udp_input_address"), - QString::fromStdString(Settings::values.udp_input_address), - QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_ADDR)); - WriteSetting(QStringLiteral("udp_input_port"), Settings::values.udp_input_port, - InputCommon::CemuhookUDP::DEFAULT_PORT); - WriteSetting(QStringLiteral("udp_pad_index"), Settings::values.udp_pad_index, 0); qt_config->endGroup(); } @@ -945,7 +1200,7 @@ void Config::SaveControlValues() { void Config::SaveCoreValues() { qt_config->beginGroup(QStringLiteral("Core")); - WriteSetting(QStringLiteral("use_multi_core"), Settings::values.use_multi_core, false); + WriteSettingGlobal(QStringLiteral("use_multi_core"), Settings::values.use_multi_core, true); qt_config->endGroup(); } @@ -955,37 +1210,26 @@ void Config::SaveDataStorageValues() { WriteSetting(QStringLiteral("use_virtual_sd"), Settings::values.use_virtual_sd, true); WriteSetting(QStringLiteral("nand_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir)), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir))); + QString::fromStdString(FS::GetUserPath(FS::UserPath::NANDDir)), + QString::fromStdString(FS::GetUserPath(FS::UserPath::NANDDir))); WriteSetting(QStringLiteral("sdmc_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir)), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir))); + QString::fromStdString(FS::GetUserPath(FS::UserPath::SDMCDir)), + QString::fromStdString(FS::GetUserPath(FS::UserPath::SDMCDir))); WriteSetting(QStringLiteral("load_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir)), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir))); + QString::fromStdString(FS::GetUserPath(FS::UserPath::LoadDir)), + QString::fromStdString(FS::GetUserPath(FS::UserPath::LoadDir))); WriteSetting(QStringLiteral("dump_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::DumpDir)), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::DumpDir))); + QString::fromStdString(FS::GetUserPath(FS::UserPath::DumpDir)), + QString::fromStdString(FS::GetUserPath(FS::UserPath::DumpDir))); WriteSetting(QStringLiteral("cache_directory"), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir)), - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir))); + QString::fromStdString(FS::GetUserPath(FS::UserPath::CacheDir)), + QString::fromStdString(FS::GetUserPath(FS::UserPath::CacheDir))); WriteSetting(QStringLiteral("gamecard_inserted"), Settings::values.gamecard_inserted, false); WriteSetting(QStringLiteral("gamecard_current_game"), Settings::values.gamecard_current_game, false); WriteSetting(QStringLiteral("gamecard_path"), - QString::fromStdString(Settings::values.gamecard_path), QStringLiteral("")); - WriteSetting(QStringLiteral("nand_total_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::values.nand_total_size)), - QVariant::fromValue<u64>(static_cast<u64>(Settings::NANDTotalSize::S29_1GB))); - WriteSetting(QStringLiteral("nand_user_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::values.nand_user_size)), - QVariant::fromValue<u64>(static_cast<u64>(Settings::NANDUserSize::S26GB))); - WriteSetting(QStringLiteral("nand_system_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::values.nand_system_size)), - QVariant::fromValue<u64>(static_cast<u64>(Settings::NANDSystemSize::S2_5GB))); - WriteSetting(QStringLiteral("sdmc_size"), - QVariant::fromValue<u64>(static_cast<u64>(Settings::values.sdmc_size)), - QVariant::fromValue<u64>(static_cast<u64>(Settings::SDMCSize::S16GB))); + QString::fromStdString(Settings::values.gamecard_path), QString{}); + qt_config->endGroup(); } @@ -997,10 +1241,11 @@ void Config::SaveDebuggingValues() { WriteSetting(QStringLiteral("use_gdbstub"), Settings::values.use_gdbstub, false); WriteSetting(QStringLiteral("gdbstub_port"), Settings::values.gdbstub_port, 24689); WriteSetting(QStringLiteral("program_args"), - QString::fromStdString(Settings::values.program_args), QStringLiteral("")); + QString::fromStdString(Settings::values.program_args), QString{}); WriteSetting(QStringLiteral("dump_exefs"), Settings::values.dump_exefs, false); WriteSetting(QStringLiteral("dump_nso"), Settings::values.dump_nso, false); WriteSetting(QStringLiteral("quest_flag"), Settings::values.quest_flag, false); + WriteSetting(QStringLiteral("disable_macro_jit"), Settings::values.disable_macro_jit, false); qt_config->endGroup(); } @@ -1023,8 +1268,7 @@ void Config::SaveDisabledAddOnValues() { qt_config->beginWriteArray(QStringLiteral("disabled")); for (std::size_t j = 0; j < elem.second.size(); ++j) { qt_config->setArrayIndex(static_cast<int>(j)); - WriteSetting(QStringLiteral("d"), QString::fromStdString(elem.second[j]), - QStringLiteral("")); + WriteSetting(QStringLiteral("d"), QString::fromStdString(elem.second[j]), QString{}); } qt_config->endArray(); ++i; @@ -1048,7 +1292,6 @@ void Config::SavePathValues() { WriteSetting(QStringLiteral("romsPath"), UISettings::values.roms_path); WriteSetting(QStringLiteral("symbolsPath"), UISettings::values.symbols_path); - WriteSetting(QStringLiteral("screenshotPath"), UISettings::values.screenshot_path); qt_config->beginWriteArray(QStringLiteral("gamedirs")); for (int i = 0; i < UISettings::values.game_dirs.size(); ++i) { qt_config->setArrayIndex(i); @@ -1059,6 +1302,38 @@ void Config::SavePathValues() { } qt_config->endArray(); WriteSetting(QStringLiteral("recentFiles"), UISettings::values.recent_files); + WriteSetting(QStringLiteral("language"), UISettings::values.language, QString{}); + + qt_config->endGroup(); +} + +void Config::SaveCpuValues() { + qt_config->beginGroup(QStringLiteral("Cpu")); + + if (global) { + WriteSetting(QStringLiteral("cpu_accuracy"), + static_cast<int>(Settings::values.cpu_accuracy), 0); + + WriteSetting(QStringLiteral("cpuopt_page_tables"), Settings::values.cpuopt_page_tables, + true); + WriteSetting(QStringLiteral("cpuopt_block_linking"), Settings::values.cpuopt_block_linking, + true); + WriteSetting(QStringLiteral("cpuopt_return_stack_buffer"), + Settings::values.cpuopt_return_stack_buffer, true); + WriteSetting(QStringLiteral("cpuopt_fast_dispatcher"), + Settings::values.cpuopt_fast_dispatcher, true); + WriteSetting(QStringLiteral("cpuopt_context_elimination"), + Settings::values.cpuopt_context_elimination, true); + WriteSetting(QStringLiteral("cpuopt_const_prop"), Settings::values.cpuopt_const_prop, true); + WriteSetting(QStringLiteral("cpuopt_misc_ir"), Settings::values.cpuopt_misc_ir, true); + WriteSetting(QStringLiteral("cpuopt_reduce_misalign_checks"), + Settings::values.cpuopt_reduce_misalign_checks, true); + + WriteSetting(QStringLiteral("cpuopt_unsafe_unfuse_fma"), + Settings::values.cpuopt_unsafe_unfuse_fma, true); + WriteSetting(QStringLiteral("cpuopt_unsafe_reduce_fp_error"), + Settings::values.cpuopt_unsafe_reduce_fp_error, true); + } qt_config->endGroup(); } @@ -1066,28 +1341,46 @@ void Config::SavePathValues() { void Config::SaveRendererValues() { qt_config->beginGroup(QStringLiteral("Renderer")); - WriteSetting(QStringLiteral("backend"), static_cast<int>(Settings::values.renderer_backend), 0); + WriteSettingGlobal(QStringLiteral("backend"), + static_cast<int>(Settings::values.renderer_backend.GetValue(global)), + Settings::values.renderer_backend.UsingGlobal(), 0); WriteSetting(QStringLiteral("debug"), Settings::values.renderer_debug, false); - WriteSetting(QStringLiteral("vulkan_device"), Settings::values.vulkan_device, 0); - WriteSetting(QStringLiteral("resolution_factor"), - static_cast<double>(Settings::values.resolution_factor), 1.0); - WriteSetting(QStringLiteral("aspect_ratio"), Settings::values.aspect_ratio, 0); - WriteSetting(QStringLiteral("max_anisotropy"), Settings::values.max_anisotropy, 0); - WriteSetting(QStringLiteral("use_frame_limit"), Settings::values.use_frame_limit, true); - WriteSetting(QStringLiteral("frame_limit"), Settings::values.frame_limit, 100); - WriteSetting(QStringLiteral("use_disk_shader_cache"), Settings::values.use_disk_shader_cache, - true); - WriteSetting(QStringLiteral("use_accurate_gpu_emulation"), - Settings::values.use_accurate_gpu_emulation, false); - WriteSetting(QStringLiteral("use_asynchronous_gpu_emulation"), - Settings::values.use_asynchronous_gpu_emulation, false); - WriteSetting(QStringLiteral("use_vsync"), Settings::values.use_vsync, true); - WriteSetting(QStringLiteral("force_30fps_mode"), Settings::values.force_30fps_mode, false); - + WriteSettingGlobal(QStringLiteral("vulkan_device"), Settings::values.vulkan_device, 0); + WriteSettingGlobal(QStringLiteral("aspect_ratio"), Settings::values.aspect_ratio, 0); + WriteSettingGlobal(QStringLiteral("max_anisotropy"), Settings::values.max_anisotropy, 0); + WriteSettingGlobal(QStringLiteral("use_frame_limit"), Settings::values.use_frame_limit, true); + WriteSettingGlobal(QStringLiteral("frame_limit"), Settings::values.frame_limit, 100); + WriteSettingGlobal(QStringLiteral("use_disk_shader_cache"), + Settings::values.use_disk_shader_cache, true); + WriteSettingGlobal(QStringLiteral("gpu_accuracy"), + static_cast<int>(Settings::values.gpu_accuracy.GetValue(global)), + Settings::values.gpu_accuracy.UsingGlobal(), 0); + WriteSettingGlobal(QStringLiteral("use_asynchronous_gpu_emulation"), + Settings::values.use_asynchronous_gpu_emulation, true); + WriteSettingGlobal(QStringLiteral("use_nvdec_emulation"), Settings::values.use_nvdec_emulation, + true); + WriteSettingGlobal(QStringLiteral("use_vsync"), Settings::values.use_vsync, true); + WriteSettingGlobal(QStringLiteral("use_assembly_shaders"), + Settings::values.use_assembly_shaders, true); + WriteSettingGlobal(QStringLiteral("use_asynchronous_shaders"), + Settings::values.use_asynchronous_shaders, false); + WriteSettingGlobal(QStringLiteral("use_fast_gpu_time"), Settings::values.use_fast_gpu_time, + true); // Cast to double because Qt's written float values are not human-readable - WriteSetting(QStringLiteral("bg_red"), static_cast<double>(Settings::values.bg_red), 0.0); - WriteSetting(QStringLiteral("bg_green"), static_cast<double>(Settings::values.bg_green), 0.0); - WriteSetting(QStringLiteral("bg_blue"), static_cast<double>(Settings::values.bg_blue), 0.0); + WriteSettingGlobal(QStringLiteral("bg_red"), Settings::values.bg_red, 0.0); + WriteSettingGlobal(QStringLiteral("bg_green"), Settings::values.bg_green, 0.0); + WriteSettingGlobal(QStringLiteral("bg_blue"), Settings::values.bg_blue, 0.0); + + qt_config->endGroup(); +} + +void Config::SaveScreenshotValues() { + qt_config->beginGroup(QStringLiteral("Screenshots")); + + WriteSetting(QStringLiteral("enable_screenshot_save_as"), + UISettings::values.enable_screenshot_save_as); + WriteSetting(QStringLiteral("screenshot_path"), + QString::fromStdString(FS::GetUserPath(FS::UserPath::ScreenshotsDir))); qt_config->endGroup(); } @@ -1115,22 +1408,28 @@ void Config::SaveShortcutValues() { void Config::SaveSystemValues() { qt_config->beginGroup(QStringLiteral("System")); - WriteSetting(QStringLiteral("use_docked_mode"), Settings::values.use_docked_mode, false); WriteSetting(QStringLiteral("current_user"), Settings::values.current_user, 0); - WriteSetting(QStringLiteral("language_index"), Settings::values.language_index, 1); - WriteSetting(QStringLiteral("region_index"), Settings::values.region_index, 1); - - WriteSetting(QStringLiteral("rng_seed_enabled"), Settings::values.rng_seed.has_value(), false); - WriteSetting(QStringLiteral("rng_seed"), Settings::values.rng_seed.value_or(0), 0); - - WriteSetting(QStringLiteral("custom_rtc_enabled"), Settings::values.custom_rtc.has_value(), - false); - WriteSetting(QStringLiteral("custom_rtc"), - QVariant::fromValue<long long>( - Settings::values.custom_rtc.value_or(std::chrono::seconds{}).count()), - 0); - - WriteSetting(QStringLiteral("sound_index"), Settings::values.sound_index, 1); + WriteSettingGlobal(QStringLiteral("language_index"), Settings::values.language_index, 1); + WriteSettingGlobal(QStringLiteral("region_index"), Settings::values.region_index, 1); + WriteSettingGlobal(QStringLiteral("time_zone_index"), Settings::values.time_zone_index, 0); + + WriteSettingGlobal(QStringLiteral("rng_seed_enabled"), + Settings::values.rng_seed.GetValue(global).has_value(), + Settings::values.rng_seed.UsingGlobal(), false); + WriteSettingGlobal(QStringLiteral("rng_seed"), + Settings::values.rng_seed.GetValue(global).value_or(0), + Settings::values.rng_seed.UsingGlobal(), 0); + + WriteSettingGlobal(QStringLiteral("custom_rtc_enabled"), + Settings::values.custom_rtc.GetValue(global).has_value(), + Settings::values.custom_rtc.UsingGlobal(), false); + WriteSettingGlobal( + QStringLiteral("custom_rtc"), + QVariant::fromValue<long long>( + Settings::values.custom_rtc.GetValue(global).value_or(std::chrono::seconds{}).count()), + Settings::values.custom_rtc.UsingGlobal(), 0); + + WriteSettingGlobal(QStringLiteral("sound_index"), Settings::values.sound_index, 1); qt_config->endGroup(); } @@ -1142,14 +1441,13 @@ void Config::SaveUIValues() { QString::fromUtf8(UISettings::themes[0].second)); WriteSetting(QStringLiteral("enable_discord_presence"), UISettings::values.enable_discord_presence, true); - WriteSetting(QStringLiteral("screenshot_resolution_factor"), - UISettings::values.screenshot_resolution_factor, 0); WriteSetting(QStringLiteral("select_user_on_boot"), UISettings::values.select_user_on_boot, false); SaveUIGamelistValues(); SaveUILayoutValues(); SavePathValues(); + SaveScreenshotValues(); SaveShortcutValues(); WriteSetting(QStringLiteral("singleWindowMode"), UISettings::values.single_window_mode, true); @@ -1161,9 +1459,9 @@ void Config::SaveUIValues() { WriteSetting(QStringLiteral("firstStart"), UISettings::values.first_start, true); WriteSetting(QStringLiteral("calloutFlags"), UISettings::values.callout_flags, 0); WriteSetting(QStringLiteral("showConsole"), UISettings::values.show_console, false); - WriteSetting(QStringLiteral("profileIndex"), UISettings::values.profile_index, 0); WriteSetting(QStringLiteral("pauseWhenInBackground"), UISettings::values.pause_when_in_background, false); + WriteSetting(QStringLiteral("hideInactiveMouse"), UISettings::values.hide_mouse, false); qt_config->endGroup(); } @@ -1223,6 +1521,34 @@ QVariant Config::ReadSetting(const QString& name, const QVariant& default_value) return result; } +template <typename Type> +void Config::ReadSettingGlobal(Settings::Setting<Type>& setting, const QString& name) { + const bool use_global = qt_config->value(name + QStringLiteral("/use_global"), true).toBool(); + setting.SetGlobal(use_global); + if (global || !use_global) { + setting.SetValue(ReadSetting(name).value<Type>()); + } +} + +template <typename Type> +void Config::ReadSettingGlobal(Settings::Setting<Type>& setting, const QString& name, + const QVariant& default_value) { + const bool use_global = qt_config->value(name + QStringLiteral("/use_global"), true).toBool(); + setting.SetGlobal(use_global); + if (global || !use_global) { + setting.SetValue(ReadSetting(name, default_value).value<Type>()); + } +} + +template <typename Type> +void Config::ReadSettingGlobal(Type& setting, const QString& name, + const QVariant& default_value) const { + const bool use_global = qt_config->value(name + QStringLiteral("/use_global"), true).toBool(); + if (global || !use_global) { + setting = ReadSetting(name, default_value).value<Type>(); + } +} + void Config::WriteSetting(const QString& name, const QVariant& value) { qt_config->setValue(name, value); } @@ -1233,13 +1559,65 @@ void Config::WriteSetting(const QString& name, const QVariant& value, qt_config->setValue(name, value); } +template <typename Type> +void Config::WriteSettingGlobal(const QString& name, const Settings::Setting<Type>& setting) { + if (!global) { + qt_config->setValue(name + QStringLiteral("/use_global"), setting.UsingGlobal()); + } + if (global || !setting.UsingGlobal()) { + qt_config->setValue(name, setting.GetValue(global)); + } +} + +template <typename Type> +void Config::WriteSettingGlobal(const QString& name, const Settings::Setting<Type>& setting, + const QVariant& default_value) { + if (!global) { + qt_config->setValue(name + QStringLiteral("/use_global"), setting.UsingGlobal()); + } + if (global || !setting.UsingGlobal()) { + qt_config->setValue(name + QStringLiteral("/default"), + setting.GetValue(global) == default_value.value<Type>()); + qt_config->setValue(name, setting.GetValue(global)); + } +} + +void Config::WriteSettingGlobal(const QString& name, const QVariant& value, bool use_global, + const QVariant& default_value) { + if (!global) { + qt_config->setValue(name + QStringLiteral("/use_global"), use_global); + } + if (global || !use_global) { + qt_config->setValue(name + QStringLiteral("/default"), value == default_value); + qt_config->setValue(name, value); + } +} + void Config::Reload() { ReadValues(); + Settings::Sanitize(); // To apply default value changes SaveValues(); Settings::Apply(); } void Config::Save() { + Settings::Sanitize(); SaveValues(); } + +void Config::ReadControlPlayerValue(std::size_t player_index) { + qt_config->beginGroup(QStringLiteral("Controls")); + ReadPlayerValue(player_index); + qt_config->endGroup(); +} + +void Config::SaveControlPlayerValue(std::size_t player_index) { + qt_config->beginGroup(QStringLiteral("Controls")); + SavePlayerValue(player_index); + qt_config->endGroup(); +} + +const std::string& Config::GetConfigFilePath() const { + return qt_config_loc; +} diff --git a/src/yuzu/configuration/config.h b/src/yuzu/configuration/config.h index ba6888004..8a600e19d 100644 --- a/src/yuzu/configuration/config.h +++ b/src/yuzu/configuration/config.h @@ -7,34 +7,53 @@ #include <array> #include <memory> #include <string> +#include <QMetaType> #include <QVariant> #include "core/settings.h" +#include "yuzu/uisettings.h" class QSettings; class Config { public: - Config(); + enum class ConfigType { + GlobalConfig, + PerGameConfig, + InputProfile, + }; + + explicit Config(const std::string& config_name = "qt-config", + ConfigType config_type = ConfigType::GlobalConfig); ~Config(); void Reload(); void Save(); + void ReadControlPlayerValue(std::size_t player_index); + void SaveControlPlayerValue(std::size_t player_index); + + const std::string& GetConfigFilePath() const; + static const std::array<int, Settings::NativeButton::NumButtons> default_buttons; - static const std::array<std::array<int, 5>, Settings::NativeAnalog::NumAnalogs> default_analogs; + static const std::array<int, Settings::NativeMotion::NumMotions> default_motions; + static const std::array<std::array<int, 4>, Settings::NativeAnalog::NumAnalogs> default_analogs; + static const std::array<int, 2> default_stick_mod; static const std::array<int, Settings::NativeMouseButton::NumMouseButtons> default_mouse_buttons; static const std::array<int, Settings::NativeKeyboard::NumKeyboardKeys> default_keyboard_keys; static const std::array<int, Settings::NativeKeyboard::NumKeyboardMods> default_keyboard_mods; + static const std::array<UISettings::Shortcut, 16> default_hotkeys; private: + void Initialize(const std::string& config_name); + void ReadValues(); - void ReadPlayerValues(); + void ReadPlayerValue(std::size_t player_index); void ReadDebugValues(); void ReadKeyboardValues(); void ReadMouseValues(); void ReadTouchscreenValues(); - void ApplyDefaultProfileIfInputInvalid(); + void ReadMotionTouchValues(); // Read functions bases off the respective config section names. void ReadAudioValues(); @@ -46,7 +65,9 @@ private: void ReadDisabledAddOnValues(); void ReadMiscellaneousValues(); void ReadPathValues(); + void ReadCpuValues(); void ReadRendererValues(); + void ReadScreenshotValues(); void ReadShortcutValues(); void ReadSystemValues(); void ReadUIValues(); @@ -55,10 +76,11 @@ private: void ReadWebServiceValues(); void SaveValues(); - void SavePlayerValues(); + void SavePlayerValue(std::size_t player_index); void SaveDebugValues(); void SaveMouseValues(); void SaveTouchscreenValues(); + void SaveMotionTouchValues(); // Save functions based off the respective config section names. void SaveAudioValues(); @@ -70,7 +92,9 @@ private: void SaveDisabledAddOnValues(); void SaveMiscellaneousValues(); void SavePathValues(); + void SaveCpuValues(); void SaveRendererValues(); + void SaveScreenshotValues(); void SaveShortcutValues(); void SaveSystemValues(); void SaveUIValues(); @@ -80,9 +104,33 @@ private: QVariant ReadSetting(const QString& name) const; QVariant ReadSetting(const QString& name, const QVariant& default_value) const; + // Templated ReadSettingGlobal functions will also look for the use_global setting and set + // both the value and the global state properly + template <typename Type> + void ReadSettingGlobal(Settings::Setting<Type>& setting, const QString& name); + template <typename Type> + void ReadSettingGlobal(Settings::Setting<Type>& setting, const QString& name, + const QVariant& default_value); + template <typename Type> + void ReadSettingGlobal(Type& setting, const QString& name, const QVariant& default_value) const; + // Templated WriteSettingGlobal functions will also write the global state if needed and will + // skip writing the actual setting if it defers to the global value void WriteSetting(const QString& name, const QVariant& value); void WriteSetting(const QString& name, const QVariant& value, const QVariant& default_value); + template <typename Type> + void WriteSettingGlobal(const QString& name, const Settings::Setting<Type>& setting); + template <typename Type> + void WriteSettingGlobal(const QString& name, const Settings::Setting<Type>& setting, + const QVariant& default_value); + void WriteSettingGlobal(const QString& name, const QVariant& value, bool use_global, + const QVariant& default_value); + ConfigType type; std::unique_ptr<QSettings> qt_config; std::string qt_config_loc; + bool global; }; + +// These metatype declarations cannot be in core/settings.h because core is devoid of QT +Q_DECLARE_METATYPE(Settings::RendererBackend); +Q_DECLARE_METATYPE(Settings::GPUAccuracy); diff --git a/src/yuzu/configuration/configuration_shared.cpp b/src/yuzu/configuration/configuration_shared.cpp new file mode 100644 index 000000000..18482795c --- /dev/null +++ b/src/yuzu/configuration/configuration_shared.cpp @@ -0,0 +1,134 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QCheckBox> +#include <QComboBox> +#include <QObject> +#include <QString> +#include "core/settings.h" +#include "yuzu/configuration/configuration_shared.h" +#include "yuzu/configuration/configure_per_game.h" + +void ConfigurationShared::ApplyPerGameSetting(Settings::Setting<bool>* setting, + const QCheckBox* checkbox, + const CheckState& tracker) { + if (tracker == CheckState::Global) { + setting->SetGlobal(true); + } else { + setting->SetGlobal(false); + setting->SetValue(checkbox->checkState()); + } +} + +void ConfigurationShared::ApplyPerGameSetting(Settings::Setting<int>* setting, + const QComboBox* combobox) { + if (combobox->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + setting->SetGlobal(true); + } else { + setting->SetGlobal(false); + setting->SetValue(combobox->currentIndex() - ConfigurationShared::USE_GLOBAL_OFFSET); + } +} + +void ConfigurationShared::ApplyPerGameSetting(Settings::Setting<Settings::RendererBackend>* setting, + const QComboBox* combobox) { + if (combobox->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + setting->SetGlobal(true); + } else { + setting->SetGlobal(false); + setting->SetValue(static_cast<Settings::RendererBackend>( + combobox->currentIndex() - ConfigurationShared::USE_GLOBAL_OFFSET)); + } +} + +void ConfigurationShared::SetPerGameSetting(QCheckBox* checkbox, + const Settings::Setting<bool>* setting) { + if (setting->UsingGlobal()) { + checkbox->setCheckState(Qt::PartiallyChecked); + } else { + checkbox->setCheckState(setting->GetValue() ? Qt::Checked : Qt::Unchecked); + } +} + +void ConfigurationShared::SetPerGameSetting(QComboBox* combobox, + const Settings::Setting<int>* setting) { + combobox->setCurrentIndex(setting->UsingGlobal() + ? ConfigurationShared::USE_GLOBAL_INDEX + : setting->GetValue() + ConfigurationShared::USE_GLOBAL_OFFSET); +} + +void ConfigurationShared::SetPerGameSetting( + QComboBox* combobox, const Settings::Setting<Settings::RendererBackend>* setting) { + combobox->setCurrentIndex(setting->UsingGlobal() ? ConfigurationShared::USE_GLOBAL_INDEX + : static_cast<int>(setting->GetValue()) + + ConfigurationShared::USE_GLOBAL_OFFSET); +} + +void ConfigurationShared::SetPerGameSetting( + QComboBox* combobox, const Settings::Setting<Settings::GPUAccuracy>* setting) { + combobox->setCurrentIndex(setting->UsingGlobal() ? ConfigurationShared::USE_GLOBAL_INDEX + : static_cast<int>(setting->GetValue()) + + ConfigurationShared::USE_GLOBAL_OFFSET); +} + +void ConfigurationShared::SetHighlight(QWidget* widget, bool highlighted) { + if (highlighted) { + widget->setStyleSheet(QStringLiteral("QWidget#%1 { background-color:rgba(0,203,255,0.5) }") + .arg(widget->objectName())); + } else { + widget->setStyleSheet(QStringLiteral("QWidget#%1 { background-color:rgba(0,0,0,0) }") + .arg(widget->objectName())); + } + widget->show(); +} + +void ConfigurationShared::SetColoredTristate(QCheckBox* checkbox, + const Settings::Setting<bool>& setting, + CheckState& tracker) { + if (setting.UsingGlobal()) { + tracker = CheckState::Global; + } else { + tracker = (setting.GetValue() == setting.GetValue(true)) ? CheckState::On : CheckState::Off; + } + SetHighlight(checkbox, tracker != CheckState::Global); + QObject::connect(checkbox, &QCheckBox::clicked, checkbox, [checkbox, setting, &tracker] { + tracker = static_cast<CheckState>((static_cast<int>(tracker) + 1) % + static_cast<int>(CheckState::Count)); + if (tracker == CheckState::Global) { + checkbox->setChecked(setting.GetValue(true)); + } + SetHighlight(checkbox, tracker != CheckState::Global); + }); +} + +void ConfigurationShared::SetColoredTristate(QCheckBox* checkbox, bool global, bool state, + bool global_state, CheckState& tracker) { + if (global) { + tracker = CheckState::Global; + } else { + tracker = (state == global_state) ? CheckState::On : CheckState::Off; + } + SetHighlight(checkbox, tracker != CheckState::Global); + QObject::connect(checkbox, &QCheckBox::clicked, checkbox, [checkbox, global_state, &tracker] { + tracker = static_cast<CheckState>((static_cast<int>(tracker) + 1) % + static_cast<int>(CheckState::Count)); + if (tracker == CheckState::Global) { + checkbox->setChecked(global_state); + } + SetHighlight(checkbox, tracker != CheckState::Global); + }); +} + +void ConfigurationShared::SetColoredComboBox(QComboBox* combobox, QWidget* target, int global) { + InsertGlobalItem(combobox, global); + QObject::connect(combobox, qOverload<int>(&QComboBox::activated), target, + [target](int index) { SetHighlight(target, index != 0); }); +} + +void ConfigurationShared::InsertGlobalItem(QComboBox* combobox, int global_index) { + const QString use_global_text = + ConfigurePerGame::tr("Use global configuration (%1)").arg(combobox->itemText(global_index)); + combobox->insertItem(ConfigurationShared::USE_GLOBAL_INDEX, use_global_text); + combobox->insertSeparator(ConfigurationShared::USE_GLOBAL_SEPARATOR_INDEX); +} diff --git a/src/yuzu/configuration/configuration_shared.h b/src/yuzu/configuration/configuration_shared.h new file mode 100644 index 000000000..312b9e549 --- /dev/null +++ b/src/yuzu/configuration/configuration_shared.h @@ -0,0 +1,51 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <QCheckBox> +#include <QComboBox> +#include <QString> +#include "core/settings.h" + +namespace ConfigurationShared { + +constexpr int USE_GLOBAL_INDEX = 0; +constexpr int USE_GLOBAL_SEPARATOR_INDEX = 1; +constexpr int USE_GLOBAL_OFFSET = 2; + +enum class CheckState { + Off, + On, + Global, + Count, +}; + +// Global-aware apply and set functions + +void ApplyPerGameSetting(Settings::Setting<bool>* setting, const QCheckBox* checkbox, + const CheckState& tracker); +void ApplyPerGameSetting(Settings::Setting<int>* setting, const QComboBox* combobox); +void ApplyPerGameSetting(Settings::Setting<Settings::RendererBackend>* setting, + const QComboBox* combobox); +void ApplyPerGameSetting(Settings::Setting<Settings::GPUAccuracy>* setting, + const QComboBox* combobox); + +void SetPerGameSetting(QCheckBox* checkbox, const Settings::Setting<bool>* setting); +void SetPerGameSetting(QComboBox* combobox, const Settings::Setting<int>* setting); +void SetPerGameSetting(QComboBox* combobox, + const Settings::Setting<Settings::RendererBackend>* setting); +void SetPerGameSetting(QComboBox* combobox, + const Settings::Setting<Settings::GPUAccuracy>* setting); + +void SetHighlight(QWidget* widget, bool highlighted); +void SetColoredTristate(QCheckBox* checkbox, const Settings::Setting<bool>& setting, + CheckState& tracker); +void SetColoredTristate(QCheckBox* checkbox, bool global, bool state, bool global_state, + CheckState& tracker); +void SetColoredComboBox(QComboBox* combobox, QWidget* target, int global); + +void InsertGlobalItem(QComboBox* combobox, int global_index); + +} // namespace ConfigurationShared diff --git a/src/yuzu/configuration/configure.ui b/src/yuzu/configuration/configure.ui index 9aec1bd09..f92c3aff3 100644 --- a/src/yuzu/configuration/configure.ui +++ b/src/yuzu/configuration/configure.ui @@ -6,7 +6,7 @@ <rect> <x>0</x> <y>0</y> - <width>382</width> + <width>650</width> <height>650</height> </rect> </property> @@ -26,13 +26,13 @@ <widget class="QListWidget" name="selectorList"> <property name="minimumSize"> <size> - <width>150</width> + <width>120</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>150</width> + <width>120</width> <height>16777215</height> </size> </property> @@ -44,66 +44,121 @@ <number>0</number> </property> <widget class="ConfigureGeneral" name="generalTab"> + <property name="accessibleName"> + <string>General</string> + </property> <attribute name="title"> <string>General</string> </attribute> </widget> <widget class="ConfigureUi" name="uiTab"> + <property name="accessibleName"> + <string>UI</string> + </property> <attribute name="title"> <string>Game List</string> </attribute> </widget> <widget class="ConfigureSystem" name="systemTab"> + <property name="accessibleName"> + <string>System</string> + </property> <attribute name="title"> <string>System</string> </attribute> </widget> <widget class="ConfigureProfileManager" name="profileManagerTab"> + <property name="accessibleName"> + <string>Profiles</string> + </property> <attribute name="title"> <string>Profiles</string> </attribute> </widget> <widget class="ConfigureFilesystem" name="filesystemTab"> + <property name="accessibleName"> + <string>Filesystem</string> + </property> <attribute name="title"> <string>Filesystem</string> </attribute> </widget> - <widget class="ConfigureInputSimple" name="inputTab"> + <widget class="ConfigureInput" name="inputTab"> + <property name="accessibleName"> + <string>Controls</string> + </property> <attribute name="title"> - <string>Input</string> + <string>Controls</string> </attribute> </widget> <widget class="ConfigureHotkeys" name="hotkeysTab"> + <property name="accessibleName"> + <string>Hotkeys</string> + </property> <attribute name="title"> <string>Hotkeys</string> </attribute> </widget> + <widget class="ConfigureCpu" name="cpuTab"> + <property name="accessibleName"> + <string>CPU</string> + </property> + <attribute name="title"> + <string>CPU</string> + </attribute> + </widget> + <widget class="ConfigureCpuDebug" name="cpuDebugTab"> + <property name="accessibleName"> + <string>Debug</string> + </property> + <attribute name="title"> + <string>Debug</string> + </attribute> + </widget> <widget class="ConfigureGraphics" name="graphicsTab"> + <property name="accessibleName"> + <string>Graphics</string> + </property> <attribute name="title"> <string>Graphics</string> </attribute> </widget> <widget class="ConfigureGraphicsAdvanced" name="graphicsAdvancedTab"> + <property name="accessibleName"> + <string>Advanced</string> + </property> <attribute name="title"> <string>GraphicsAdvanced</string> </attribute> </widget> <widget class="ConfigureAudio" name="audioTab"> + <property name="accessibleName"> + <string>Audio</string> + </property> <attribute name="title"> <string>Audio</string> </attribute> </widget> <widget class="ConfigureDebug" name="debugTab"> + <property name="accessibleName"> + <string>Debug</string> + </property> <attribute name="title"> <string>Debug</string> </attribute> </widget> <widget class="ConfigureWeb" name="webTab"> + <property name="accessibleName"> + <string>Web</string> + </property> <attribute name="title"> <string>Web</string> </attribute> </widget> <widget class="ConfigureService" name="serviceTab"> + <property name="accessibleName"> + <string>Services</string> + </property> <attribute name="title"> <string>Services</string> </attribute> @@ -159,6 +214,18 @@ <container>1</container> </customwidget> <customwidget> + <class>ConfigureCpu</class> + <extends>QWidget</extends> + <header>configuration/configure_cpu.h</header> + <container>1</container> + </customwidget> + <customwidget> + <class>ConfigureCpuDebug</class> + <extends>QWidget</extends> + <header>configuration/configure_cpu_debug.h</header> + <container>1</container> + </customwidget> + <customwidget> <class>ConfigureGraphics</class> <extends>QWidget</extends> <header>configuration/configure_graphics.h</header> @@ -183,9 +250,9 @@ <container>1</container> </customwidget> <customwidget> - <class>ConfigureInputSimple</class> + <class>ConfigureInput</class> <extends>QWidget</extends> - <header>configuration/configure_input_simple.h</header> + <header>configuration/configure_input.h</header> <container>1</container> </customwidget> <customwidget> @@ -208,32 +275,12 @@ <signal>accepted()</signal> <receiver>ConfigureDialog</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>220</x> - <y>380</y> - </hint> - <hint type="destinationlabel"> - <x>220</x> - <y>200</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureDialog</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>220</x> - <y>380</y> - </hint> - <hint type="destinationlabel"> - <x>220</x> - <y>200</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_audio.cpp b/src/yuzu/configuration/configure_audio.cpp index f370c690f..db9518798 100644 --- a/src/yuzu/configuration/configure_audio.cpp +++ b/src/yuzu/configuration/configure_audio.cpp @@ -11,6 +11,7 @@ #include "core/core.h" #include "core/settings.h" #include "ui_configure_audio.h" +#include "yuzu/configuration/configuration_shared.h" #include "yuzu/configuration/configure_audio.h" ConfigureAudio::ConfigureAudio(QWidget* parent) @@ -24,6 +25,11 @@ ConfigureAudio::ConfigureAudio(QWidget* parent) connect(ui->output_sink_combo_box, qOverload<int>(&QComboBox::currentIndexChanged), this, &ConfigureAudio::UpdateAudioDevices); + ui->volume_label->setVisible(Settings::IsConfiguringGlobal()); + ui->volume_combo_box->setVisible(!Settings::IsConfiguringGlobal()); + + SetupPerGameUI(); + SetConfiguration(); const bool is_powered_on = Core::System::GetInstance().IsPoweredOn(); @@ -41,8 +47,21 @@ void ConfigureAudio::SetConfiguration() { SetAudioDeviceFromDeviceID(); - ui->toggle_audio_stretching->setChecked(Settings::values.enable_audio_stretching); - ui->volume_slider->setValue(Settings::values.volume * ui->volume_slider->maximum()); + ui->volume_slider->setValue(Settings::values.volume.GetValue() * ui->volume_slider->maximum()); + + ui->toggle_audio_stretching->setChecked(Settings::values.enable_audio_stretching.GetValue()); + + if (!Settings::IsConfiguringGlobal()) { + if (Settings::values.volume.UsingGlobal()) { + ui->volume_combo_box->setCurrentIndex(0); + ui->volume_slider->setEnabled(false); + } else { + ui->volume_combo_box->setCurrentIndex(1); + ui->volume_slider->setEnabled(true); + } + ConfigurationShared::SetHighlight(ui->volume_layout, + !Settings::values.volume.UsingGlobal()); + } SetVolumeIndicatorText(ui->volume_slider->sliderPosition()); } @@ -80,15 +99,37 @@ void ConfigureAudio::SetVolumeIndicatorText(int percentage) { } void ConfigureAudio::ApplyConfiguration() { - Settings::values.sink_id = - ui->output_sink_combo_box->itemText(ui->output_sink_combo_box->currentIndex()) - .toStdString(); - Settings::values.enable_audio_stretching = ui->toggle_audio_stretching->isChecked(); - Settings::values.audio_device_id = - ui->audio_device_combo_box->itemText(ui->audio_device_combo_box->currentIndex()) - .toStdString(); - Settings::values.volume = - static_cast<float>(ui->volume_slider->sliderPosition()) / ui->volume_slider->maximum(); + if (Settings::IsConfiguringGlobal()) { + Settings::values.sink_id = + ui->output_sink_combo_box->itemText(ui->output_sink_combo_box->currentIndex()) + .toStdString(); + Settings::values.audio_device_id = + ui->audio_device_combo_box->itemText(ui->audio_device_combo_box->currentIndex()) + .toStdString(); + + // Guard if during game and set to game-specific value + if (Settings::values.enable_audio_stretching.UsingGlobal()) { + Settings::values.enable_audio_stretching.SetValue( + ui->toggle_audio_stretching->isChecked()); + } + if (Settings::values.volume.UsingGlobal()) { + Settings::values.volume.SetValue( + static_cast<float>(ui->volume_slider->sliderPosition()) / + ui->volume_slider->maximum()); + } + } else { + ConfigurationShared::ApplyPerGameSetting(&Settings::values.enable_audio_stretching, + ui->toggle_audio_stretching, + enable_audio_stretching); + if (ui->volume_combo_box->currentIndex() == 0) { + Settings::values.volume.SetGlobal(true); + } else { + Settings::values.volume.SetGlobal(false); + Settings::values.volume.SetValue( + static_cast<float>(ui->volume_slider->sliderPosition()) / + ui->volume_slider->maximum()); + } + } } void ConfigureAudio::changeEvent(QEvent* event) { @@ -122,3 +163,26 @@ void ConfigureAudio::RetranslateUI() { ui->retranslateUi(this); SetVolumeIndicatorText(ui->volume_slider->sliderPosition()); } + +void ConfigureAudio::SetupPerGameUI() { + if (Settings::IsConfiguringGlobal()) { + ui->volume_slider->setEnabled(Settings::values.volume.UsingGlobal()); + ui->toggle_audio_stretching->setEnabled( + Settings::values.enable_audio_stretching.UsingGlobal()); + + return; + } + + ConfigurationShared::SetColoredTristate(ui->toggle_audio_stretching, + Settings::values.enable_audio_stretching, + enable_audio_stretching); + connect(ui->volume_combo_box, qOverload<int>(&QComboBox::activated), this, [this](int index) { + ui->volume_slider->setEnabled(index == 1); + ConfigurationShared::SetHighlight(ui->volume_layout, index == 1); + }); + + ui->output_sink_combo_box->setVisible(false); + ui->output_sink_label->setVisible(false); + ui->audio_device_combo_box->setVisible(false); + ui->audio_device_label->setVisible(false); +} diff --git a/src/yuzu/configuration/configure_audio.h b/src/yuzu/configuration/configure_audio.h index ea83bd72d..9dbd3d93e 100644 --- a/src/yuzu/configuration/configure_audio.h +++ b/src/yuzu/configuration/configure_audio.h @@ -7,6 +7,10 @@ #include <memory> #include <QWidget> +namespace ConfigurationShared { +enum class CheckState; +} + namespace Ui { class ConfigureAudio; } @@ -34,5 +38,9 @@ private: void SetAudioDeviceFromDeviceID(); void SetVolumeIndicatorText(int percentage); + void SetupPerGameUI(); + std::unique_ptr<Ui::ConfigureAudio> ui; + + ConfigurationShared::CheckState enable_audio_stretching; }; diff --git a/src/yuzu/configuration/configure_audio.ui b/src/yuzu/configuration/configure_audio.ui index a098b9acc..9bd0cca96 100644 --- a/src/yuzu/configuration/configure_audio.ui +++ b/src/yuzu/configuration/configure_audio.ui @@ -6,8 +6,8 @@ <rect> <x>0</x> <y>0</y> - <width>188</width> - <height>246</height> + <width>367</width> + <height>368</height> </rect> </property> <layout class="QVBoxLayout"> @@ -18,9 +18,9 @@ </property> <layout class="QVBoxLayout"> <item> - <layout class="QHBoxLayout"> + <layout class="QHBoxLayout" name="_3"> <item> - <widget class="QLabel" name="label_1"> + <widget class="QLabel" name="output_sink_label"> <property name="text"> <string>Output Engine:</string> </property> @@ -31,20 +31,20 @@ </item> </layout> </item> - <item> - <widget class="QCheckBox" name="toggle_audio_stretching"> - <property name="toolTip"> - <string>This post-processing effect adjusts audio speed to match emulation speed and helps prevent audio stutter. This however increases audio latency.</string> - </property> - <property name="text"> - <string>Enable audio stretching</string> - </property> - </widget> - </item> <item> - <layout class="QHBoxLayout"> + <widget class="QCheckBox" name="toggle_audio_stretching"> + <property name="toolTip"> + <string>This post-processing effect adjusts audio speed to match emulation speed and helps prevent audio stutter. This however increases audio latency.</string> + </property> + <property name="text"> + <string>Enable audio stretching</string> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="_2"> <item> - <widget class="QLabel" name="label_2"> + <widget class="QLabel" name="audio_device_label"> <property name="text"> <string>Audio Device:</string> </property> @@ -56,66 +56,91 @@ </layout> </item> <item> - <layout class="QHBoxLayout" name="horizontalLayout_2"> - <property name="topMargin"> - <number>0</number> - </property> - <item> - <widget class="QLabel" name="label_3"> - <property name="text"> - <string>Volume:</string> - </property> - </widget> - </item> - <item> - <spacer name="horizontalSpacer"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item> - <widget class="QSlider" name="volume_slider"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Expanding" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="maximum"> - <number>100</number> - </property> - <property name="pageStep"> - <number>10</number> - </property> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - </widget> - </item> - <item> - <widget class="QLabel" name="volume_indicator"> - <property name="minimumSize"> - <size> - <width>32</width> - <height>0</height> - </size> - </property> - <property name="text"> - <string>0 %</string> - </property> - <property name="alignment"> - <set>Qt::AlignCenter</set> - </property> - </widget> - </item> - </layout> + <widget class="QWidget" name="volume_layout" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QComboBox" name="volume_combo_box"> + <item> + <property name="text"> + <string>Use global volume</string> + </property> + </item> + <item> + <property name="text"> + <string>Set volume:</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QLabel" name="volume_label"> + <property name="text"> + <string>Volume:</string> + </property> + </widget> + </item> + <item> + <spacer name="horizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>30</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QSlider" name="volume_slider"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Expanding" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="maximum"> + <number>100</number> + </property> + <property name="pageStep"> + <number>10</number> + </property> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + </widget> + </item> + <item> + <widget class="QLabel" name="volume_indicator"> + <property name="minimumSize"> + <size> + <width>32</width> + <height>0</height> + </size> + </property> + <property name="text"> + <string>0 %</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + </layout> + </widget> </item> </layout> </widget> diff --git a/src/yuzu/configuration/configure_cpu.cpp b/src/yuzu/configuration/configure_cpu.cpp new file mode 100644 index 000000000..37fcd6adc --- /dev/null +++ b/src/yuzu/configuration/configure_cpu.cpp @@ -0,0 +1,76 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QComboBox> +#include <QMessageBox> + +#include "common/common_types.h" +#include "common/logging/log.h" +#include "core/core.h" +#include "core/settings.h" +#include "ui_configure_cpu.h" +#include "yuzu/configuration/configure_cpu.h" + +ConfigureCpu::ConfigureCpu(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureCpu) { + ui->setupUi(this); + + SetConfiguration(); + + connect(ui->accuracy, qOverload<int>(&QComboBox::activated), this, + &ConfigureCpu::AccuracyUpdated); + connect(ui->accuracy, qOverload<int>(&QComboBox::currentIndexChanged), this, + &ConfigureCpu::UpdateGroup); +} + +ConfigureCpu::~ConfigureCpu() = default; + +void ConfigureCpu::SetConfiguration() { + const bool runtime_lock = !Core::System::GetInstance().IsPoweredOn(); + + ui->accuracy->setEnabled(runtime_lock); + ui->accuracy->setCurrentIndex(static_cast<int>(Settings::values.cpu_accuracy)); + UpdateGroup(static_cast<int>(Settings::values.cpu_accuracy)); + + ui->cpuopt_unsafe_unfuse_fma->setEnabled(runtime_lock); + ui->cpuopt_unsafe_unfuse_fma->setChecked(Settings::values.cpuopt_unsafe_unfuse_fma); + ui->cpuopt_unsafe_reduce_fp_error->setEnabled(runtime_lock); + ui->cpuopt_unsafe_reduce_fp_error->setChecked(Settings::values.cpuopt_unsafe_reduce_fp_error); +} + +void ConfigureCpu::AccuracyUpdated(int index) { + if (static_cast<Settings::CPUAccuracy>(index) == Settings::CPUAccuracy::DebugMode) { + const auto result = QMessageBox::warning(this, tr("Setting CPU to Debug Mode"), + tr("CPU Debug Mode is only intended for developer " + "use. Are you sure you want to enable this?"), + QMessageBox::Yes | QMessageBox::No); + if (result == QMessageBox::No) { + ui->accuracy->setCurrentIndex(static_cast<int>(Settings::CPUAccuracy::Accurate)); + UpdateGroup(static_cast<int>(Settings::CPUAccuracy::Accurate)); + } + } +} + +void ConfigureCpu::UpdateGroup(int index) { + ui->unsafe_group->setVisible(static_cast<Settings::CPUAccuracy>(index) == + Settings::CPUAccuracy::Unsafe); +} + +void ConfigureCpu::ApplyConfiguration() { + Settings::values.cpu_accuracy = + static_cast<Settings::CPUAccuracy>(ui->accuracy->currentIndex()); + Settings::values.cpuopt_unsafe_unfuse_fma = ui->cpuopt_unsafe_unfuse_fma->isChecked(); + Settings::values.cpuopt_unsafe_reduce_fp_error = ui->cpuopt_unsafe_reduce_fp_error->isChecked(); +} + +void ConfigureCpu::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QWidget::changeEvent(event); +} + +void ConfigureCpu::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_cpu.h b/src/yuzu/configuration/configure_cpu.h new file mode 100644 index 000000000..3c5683d81 --- /dev/null +++ b/src/yuzu/configuration/configure_cpu.h @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QWidget> +#include "core/settings.h" + +namespace Ui { +class ConfigureCpu; +} + +class ConfigureCpu : public QWidget { + Q_OBJECT + +public: + explicit ConfigureCpu(QWidget* parent = nullptr); + ~ConfigureCpu() override; + + void ApplyConfiguration(); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + void AccuracyUpdated(int index); + void UpdateGroup(int index); + + void SetConfiguration(); + + std::unique_ptr<Ui::ConfigureCpu> ui; +}; diff --git a/src/yuzu/configuration/configure_cpu.ui b/src/yuzu/configuration/configure_cpu.ui new file mode 100644 index 000000000..ebdd2e6e9 --- /dev/null +++ b/src/yuzu/configuration/configure_cpu.ui @@ -0,0 +1,144 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureCpu</class> + <widget class="QWidget" name="ConfigureCpu"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>400</width> + <height>321</height> + </rect> + </property> + <property name="windowTitle"> + <string>Form</string> + </property> + <layout class="QVBoxLayout"> + <item> + <layout class="QVBoxLayout"> + <item> + <widget class="QGroupBox"> + <property name="title"> + <string>General</string> + </property> + <layout class="QVBoxLayout"> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel"> + <property name="text"> + <string>Accuracy:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="accuracy"> + <item> + <property name="text"> + <string>Accurate</string> + </property> + </item> + <item> + <property name="text"> + <string>Unsafe</string> + </property> + </item> + <item> + <property name="text"> + <string>Enable Debug Mode</string> + </property> + </item> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QLabel"> + <property name="wordWrap"> + <bool>1</bool> + </property> + <property name="text"> + <string>We recommend setting accuracy to "Accurate".</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QVBoxLayout"> + <item> + <widget class="QGroupBox" name="unsafe_group"> + <property name="title"> + <string>Unsafe CPU Optimization Settings</string> + </property> + <layout class="QVBoxLayout"> + <item> + <widget class="QLabel"> + <property name="wordWrap"> + <bool>1</bool> + </property> + <property name="text"> + <string>These settings reduce accuracy for speed.</string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_unsafe_unfuse_fma"> + <property name="text"> + <string>Unfuse FMA (improve performance on CPUs without FMA)</string> + </property> + <property name="toolTip"> + <string> + <div>This option improves speed by reducing accuracy of fused-multiply-add instructions on CPUs without native FMA support.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_unsafe_reduce_fp_error"> + <property name="text"> + <string>Faster FRSQRTE and FRECPE</string> + </property> + <property name="toolTip"> + <string> + <div>This option improves the speed of some approximate floating-point functions by using less accurate native approximations.</div> + </string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QLabel" name="label_disable_info"> + <property name="text"> + <string>CPU settings are available only when game is not running.</string> + </property> + <property name="wordWrap"> + <bool>true</bool> + </property> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui> diff --git a/src/yuzu/configuration/configure_cpu_debug.cpp b/src/yuzu/configuration/configure_cpu_debug.cpp new file mode 100644 index 000000000..3385b2cf6 --- /dev/null +++ b/src/yuzu/configuration/configure_cpu_debug.cpp @@ -0,0 +1,65 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QComboBox> + +#include "common/common_types.h" +#include "common/logging/log.h" +#include "core/core.h" +#include "core/settings.h" +#include "ui_configure_cpu_debug.h" +#include "yuzu/configuration/configure_cpu_debug.h" + +ConfigureCpuDebug::ConfigureCpuDebug(QWidget* parent) + : QWidget(parent), ui(new Ui::ConfigureCpuDebug) { + ui->setupUi(this); + + SetConfiguration(); +} + +ConfigureCpuDebug::~ConfigureCpuDebug() = default; + +void ConfigureCpuDebug::SetConfiguration() { + const bool runtime_lock = !Core::System::GetInstance().IsPoweredOn(); + + ui->cpuopt_page_tables->setEnabled(runtime_lock); + ui->cpuopt_page_tables->setChecked(Settings::values.cpuopt_page_tables); + ui->cpuopt_block_linking->setEnabled(runtime_lock); + ui->cpuopt_block_linking->setChecked(Settings::values.cpuopt_block_linking); + ui->cpuopt_return_stack_buffer->setEnabled(runtime_lock); + ui->cpuopt_return_stack_buffer->setChecked(Settings::values.cpuopt_return_stack_buffer); + ui->cpuopt_fast_dispatcher->setEnabled(runtime_lock); + ui->cpuopt_fast_dispatcher->setChecked(Settings::values.cpuopt_fast_dispatcher); + ui->cpuopt_context_elimination->setEnabled(runtime_lock); + ui->cpuopt_context_elimination->setChecked(Settings::values.cpuopt_context_elimination); + ui->cpuopt_const_prop->setEnabled(runtime_lock); + ui->cpuopt_const_prop->setChecked(Settings::values.cpuopt_const_prop); + ui->cpuopt_misc_ir->setEnabled(runtime_lock); + ui->cpuopt_misc_ir->setChecked(Settings::values.cpuopt_misc_ir); + ui->cpuopt_reduce_misalign_checks->setEnabled(runtime_lock); + ui->cpuopt_reduce_misalign_checks->setChecked(Settings::values.cpuopt_reduce_misalign_checks); +} + +void ConfigureCpuDebug::ApplyConfiguration() { + Settings::values.cpuopt_page_tables = ui->cpuopt_page_tables->isChecked(); + Settings::values.cpuopt_block_linking = ui->cpuopt_block_linking->isChecked(); + Settings::values.cpuopt_return_stack_buffer = ui->cpuopt_return_stack_buffer->isChecked(); + Settings::values.cpuopt_fast_dispatcher = ui->cpuopt_fast_dispatcher->isChecked(); + Settings::values.cpuopt_context_elimination = ui->cpuopt_context_elimination->isChecked(); + Settings::values.cpuopt_const_prop = ui->cpuopt_const_prop->isChecked(); + Settings::values.cpuopt_misc_ir = ui->cpuopt_misc_ir->isChecked(); + Settings::values.cpuopt_reduce_misalign_checks = ui->cpuopt_reduce_misalign_checks->isChecked(); +} + +void ConfigureCpuDebug::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QWidget::changeEvent(event); +} + +void ConfigureCpuDebug::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_cpu_debug.h b/src/yuzu/configuration/configure_cpu_debug.h new file mode 100644 index 000000000..c9941ef3b --- /dev/null +++ b/src/yuzu/configuration/configure_cpu_debug.h @@ -0,0 +1,31 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QWidget> +#include "core/settings.h" + +namespace Ui { +class ConfigureCpuDebug; +} + +class ConfigureCpuDebug : public QWidget { + Q_OBJECT + +public: + explicit ConfigureCpuDebug(QWidget* parent = nullptr); + ~ConfigureCpuDebug() override; + + void ApplyConfiguration(); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + void SetConfiguration(); + + std::unique_ptr<Ui::ConfigureCpuDebug> ui; +}; diff --git a/src/yuzu/configuration/configure_cpu_debug.ui b/src/yuzu/configuration/configure_cpu_debug.ui new file mode 100644 index 000000000..a90dc64fe --- /dev/null +++ b/src/yuzu/configuration/configure_cpu_debug.ui @@ -0,0 +1,174 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureCpuDebug</class> + <widget class="QWidget" name="ConfigureCpuDebug"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>400</width> + <height>321</height> + </rect> + </property> + <property name="windowTitle"> + <string>Form</string> + </property> + <layout class="QVBoxLayout"> + <item> + <layout class="QVBoxLayout"> + <item> + <widget class="QGroupBox"> + <property name="title"> + <string>Toggle CPU Optimizations</string> + </property> + <layout class="QVBoxLayout"> + <item> + <widget class="QLabel"> + <property name="wordWrap"> + <bool>1</bool> + </property> + <property name="text"> + <string> + <div> + <b>For debugging only.</b> + <br> + If you're not sure what these do, keep all of these enabled. + <br> + These settings only take effect when CPU Accuracy is "Debug Mode". + </div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_page_tables"> + <property name="text"> + <string>Enable inline page tables</string> + </property> + <property name="toolTip"> + <string> + <div style="white-space: nowrap">This optimization speeds up memory accesses by the guest program.</div> + <div style="white-space: nowrap">Enabling it inlines accesses to PageTable::pointers into emitted code.</div> + <div style="white-space: nowrap">Disabling this forces all memory accesses to go through the Memory::Read/Memory::Write functions.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_block_linking"> + <property name="text"> + <string>Enable block linking</string> + </property> + <property name="toolTip"> + <string> + <div>This optimization avoids dispatcher lookups by allowing emitted basic blocks to jump directly to other basic blocks if the destination PC is static.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_return_stack_buffer"> + <property name="text"> + <string>Enable return stack buffer</string> + </property> + <property name="toolTip"> + <string> + <div>This optimization avoids dispatcher lookups by keeping track potential return addresses of BL instructions. This approximates what happens with a return stack buffer on a real CPU.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_fast_dispatcher"> + <property name="text"> + <string>Enable fast dispatcher</string> + </property> + <property name="toolTip"> + <string> + <div>Enable a two-tiered dispatch system. A faster dispatcher written in assembly has a small MRU cache of jump destinations is used first. If that fails, dispatch falls back to the slower C++ dispatcher.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_context_elimination"> + <property name="text"> + <string>Enable context elimination</string> + </property> + <property name="toolTip"> + <string> + <div>Enables an IR optimization that reduces unnecessary accesses to the CPU context structure.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_const_prop"> + <property name="text"> + <string>Enable constant propagation</string> + </property> + <property name="toolTip"> + <string> + <div>Enables IR optimizations that involve constant propagation.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_misc_ir"> + <property name="text"> + <string>Enable miscellaneous optimizations</string> + </property> + <property name="toolTip"> + <string> + <div>Enables miscellaneous IR optimizations.</div> + </string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="cpuopt_reduce_misalign_checks"> + <property name="text"> + <string>Enable misalignment check reduction</string> + </property> + <property name="toolTip"> + <string> + <div style="white-space: nowrap">When enabled, a misalignment is only triggered when an access crosses a page boundary.</div> + <div style="white-space: nowrap">When disabled, a misalignment is triggered on all misaligned accesses.</div> + </string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QLabel" name="label_disable_info"> + <property name="text"> + <string>CPU settings are available only when game is not running.</string> + </property> + <property name="wordWrap"> + <bool>true</bool> + </property> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui> diff --git a/src/yuzu/configuration/configure_debug.cpp b/src/yuzu/configuration/configure_debug.cpp index 9631059c7..027099ab7 100644 --- a/src/yuzu/configuration/configure_debug.cpp +++ b/src/yuzu/configuration/configure_debug.cpp @@ -19,7 +19,8 @@ ConfigureDebug::ConfigureDebug(QWidget* parent) : QWidget(parent), ui(new Ui::Co SetConfiguration(); connect(ui->open_log_button, &QPushButton::clicked, []() { - QString path = QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::LogDir)); + const auto path = + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::LogDir)); QDesktopServices::openUrl(QUrl::fromLocalFile(path)); }); } @@ -38,6 +39,9 @@ void ConfigureDebug::SetConfiguration() { ui->quest_flag->setChecked(Settings::values.quest_flag); ui->enable_graphics_debugging->setEnabled(!Core::System::GetInstance().IsPoweredOn()); ui->enable_graphics_debugging->setChecked(Settings::values.renderer_debug); + ui->disable_macro_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn()); + ui->disable_macro_jit->setChecked(Settings::values.disable_macro_jit); + ui->extended_logging->setChecked(Settings::values.extended_logging); } void ConfigureDebug::ApplyConfiguration() { @@ -49,6 +53,8 @@ void ConfigureDebug::ApplyConfiguration() { Settings::values.reporting_services = ui->reporting_services->isChecked(); Settings::values.quest_flag = ui->quest_flag->isChecked(); Settings::values.renderer_debug = ui->enable_graphics_debugging->isChecked(); + Settings::values.disable_macro_jit = ui->disable_macro_jit->isChecked(); + Settings::values.extended_logging = ui->extended_logging->isChecked(); Debugger::ToggleConsole(); Log::Filter filter; filter.ParseFilterString(Settings::values.log_filter); diff --git a/src/yuzu/configuration/configure_debug.ui b/src/yuzu/configuration/configure_debug.ui index e028c4c80..6f94fe304 100644 --- a/src/yuzu/configuration/configure_debug.ui +++ b/src/yuzu/configuration/configure_debug.ui @@ -90,7 +90,7 @@ <item> <widget class="QCheckBox" name="toggle_console"> <property name="text"> - <string>Show Log Console (Windows Only)</string> + <string>Show Log in Console</string> </property> </widget> </item> @@ -103,6 +103,34 @@ </item> </layout> </item> + <item> + <widget class="QCheckBox" name="extended_logging"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="toolTip"> + <string>When checked, the max size of the log increases from 100 MB to 1 GB</string> + </property> + <property name="text"> + <string>Enable Extended Logging</string> + </property> + </widget> + </item> + <item> + <widget class="QLabel" name="label_3"> + <property name="font"> + <font> + <italic>true</italic> + </font> + </property> + <property name="text"> + <string>This will be reset automatically when yuzu closes.</string> + </property> + <property name="indent"> + <number>20</number> + </property> + </widget> + </item> </layout> </widget> </item> @@ -115,7 +143,7 @@ <item> <layout class="QHBoxLayout" name="horizontalLayout_4"> <item> - <widget class="QLabel" name="label_3"> + <widget class="QLabel" name="label_4"> <property name="text"> <string>Arguments String</string> </property> @@ -140,14 +168,27 @@ <property name="enabled"> <bool>true</bool> </property> - <property name="whatsThis"> - <string>When checked, the graphics API enters in a slower debugging mode</string> + <property name="toolTip"> + <string>When checked, the graphics API enters a slower debugging mode</string> </property> <property name="text"> <string>Enable Graphics Debugging</string> </property> </widget> </item> + <item> + <widget class="QCheckBox" name="disable_macro_jit"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="toolTip"> + <string>When checked, it disables the macro Just In Time compiler. Enabling this makes games run slower</string> + </property> + <property name="text"> + <string>Disable Macro JIT</string> + </property> + </widget> + </item> </layout> </widget> </item> @@ -156,27 +197,7 @@ <property name="title"> <string>Dump</string> </property> - <layout class="QVBoxLayout" name="verticalLayout_6"> - <item> - <widget class="QCheckBox" name="dump_decompressed_nso"> - <property name="whatsThis"> - <string>When checked, any NSO yuzu tries to load or patch will be copied decompressed to the yuzu/dump directory.</string> - </property> - <property name="text"> - <string>Dump Decompressed NSOs</string> - </property> - </widget> - </item> - <item> - <widget class="QCheckBox" name="dump_exefs"> - <property name="whatsThis"> - <string>When checked, any game that yuzu loads will have its ExeFS dumped to the yuzu/dump directory.</string> - </property> - <property name="text"> - <string>Dump ExeFS</string> - </property> - </widget> - </item> + <layout class="QVBoxLayout" name="verticalLayout_7"> <item> <widget class="QCheckBox" name="reporting_services"> <property name="text"> @@ -185,7 +206,7 @@ </widget> </item> <item> - <widget class="QLabel" name="label"> + <widget class="QLabel" name="label_5"> <property name="font"> <font> <italic>true</italic> @@ -207,7 +228,7 @@ <property name="title"> <string>Advanced</string> </property> - <layout class="QVBoxLayout" name="verticalLayout_7"> + <layout class="QVBoxLayout" name="verticalLayout_8"> <item> <widget class="QCheckBox" name="quest_flag"> <property name="text"> @@ -244,8 +265,6 @@ <tabstop>open_log_button</tabstop> <tabstop>homebrew_args_edit</tabstop> <tabstop>enable_graphics_debugging</tabstop> - <tabstop>dump_decompressed_nso</tabstop> - <tabstop>dump_exefs</tabstop> <tabstop>reporting_services</tabstop> <tabstop>quest_flag</tabstop> </tabstops> diff --git a/src/yuzu/configuration/configure_debug_controller.cpp b/src/yuzu/configuration/configure_debug_controller.cpp new file mode 100644 index 000000000..a878ef9c6 --- /dev/null +++ b/src/yuzu/configuration/configure_debug_controller.cpp @@ -0,0 +1,43 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "ui_configure_debug_controller.h" +#include "yuzu/configuration/configure_debug_controller.h" +#include "yuzu/configuration/configure_input_player.h" + +ConfigureDebugController::ConfigureDebugController(QWidget* parent, + InputCommon::InputSubsystem* input_subsystem, + InputProfiles* profiles) + : QDialog(parent), ui(std::make_unique<Ui::ConfigureDebugController>()), + debug_controller( + new ConfigureInputPlayer(this, 9, nullptr, input_subsystem, profiles, true)) { + ui->setupUi(this); + + ui->controllerLayout->addWidget(debug_controller); + + connect(ui->clear_all_button, &QPushButton::clicked, this, + [this] { debug_controller->ClearAll(); }); + connect(ui->restore_defaults_button, &QPushButton::clicked, this, + [this] { debug_controller->RestoreDefaults(); }); + + RetranslateUI(); +} + +ConfigureDebugController::~ConfigureDebugController() = default; + +void ConfigureDebugController::ApplyConfiguration() { + debug_controller->ApplyConfiguration(); +} + +void ConfigureDebugController::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QDialog::changeEvent(event); +} + +void ConfigureDebugController::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_debug_controller.h b/src/yuzu/configuration/configure_debug_controller.h new file mode 100644 index 000000000..b4f53fad5 --- /dev/null +++ b/src/yuzu/configuration/configure_debug_controller.h @@ -0,0 +1,41 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QDialog> + +class QPushButton; + +class ConfigureInputPlayer; + +class InputProfiles; + +namespace InputCommon { +class InputSubsystem; +} + +namespace Ui { +class ConfigureDebugController; +} + +class ConfigureDebugController : public QDialog { + Q_OBJECT + +public: + explicit ConfigureDebugController(QWidget* parent, InputCommon::InputSubsystem* input_subsystem, + InputProfiles* profiles); + ~ConfigureDebugController() override; + + void ApplyConfiguration(); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + std::unique_ptr<Ui::ConfigureDebugController> ui; + + ConfigureInputPlayer* debug_controller; +}; diff --git a/src/yuzu/configuration/configure_debug_controller.ui b/src/yuzu/configuration/configure_debug_controller.ui new file mode 100644 index 000000000..7b7e6582c --- /dev/null +++ b/src/yuzu/configuration/configure_debug_controller.ui @@ -0,0 +1,77 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureDebugController</class> + <widget class="QDialog" name="ConfigureDebugController"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>780</width> + <height>500</height> + </rect> + </property> + <property name="windowTitle"> + <string>Configure Debug Controller</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <property name="spacing"> + <number>2</number> + </property> + <property name="leftMargin"> + <number>9</number> + </property> + <property name="topMargin"> + <number>9</number> + </property> + <property name="rightMargin"> + <number>9</number> + </property> + <property name="bottomMargin"> + <number>9</number> + </property> + <item> + <layout class="QHBoxLayout" name="controllerLayout"/> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout"> + <item> + <widget class="QPushButton" name="clear_all_button"> + <property name="text"> + <string>Clear</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="restore_defaults_button"> + <property name="text"> + <string>Defaults</string> + </property> + </widget> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + <resources/> + <connections> + <connection> + <sender>buttonBox</sender> + <signal>accepted()</signal> + <receiver>ConfigureDebugController</receiver> + <slot>accept()</slot> + </connection> + <connection> + <sender>buttonBox</sender> + <signal>rejected()</signal> + <receiver>ConfigureDebugController</receiver> + <slot>reject()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/configure_dialog.cpp b/src/yuzu/configuration/configure_dialog.cpp index df4473b46..5041e0bf8 100644 --- a/src/yuzu/configuration/configure_dialog.cpp +++ b/src/yuzu/configuration/configure_dialog.cpp @@ -12,15 +12,21 @@ #include "yuzu/configuration/configure_input_player.h" #include "yuzu/hotkeys.h" -ConfigureDialog::ConfigureDialog(QWidget* parent, HotkeyRegistry& registry) +ConfigureDialog::ConfigureDialog(QWidget* parent, HotkeyRegistry& registry, + InputCommon::InputSubsystem* input_subsystem) : QDialog(parent), ui(new Ui::ConfigureDialog), registry(registry) { + Settings::SetConfiguringGlobal(true); + ui->setupUi(this); ui->hotkeysTab->Populate(registry); setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint); + ui->inputTab->Initialize(input_subsystem); + SetConfiguration(); PopulateSelectionList(); + connect(ui->uiTab, &ConfigureUi::LanguageChanged, this, &ConfigureDialog::OnLanguageChanged); connect(ui->selectorList, &QListWidget::itemSelectionChanged, this, &ConfigureDialog::UpdateVisibleTabs); @@ -40,6 +46,8 @@ void ConfigureDialog::ApplyConfiguration() { ui->filesystemTab->applyConfiguration(); ui->inputTab->ApplyConfiguration(); ui->hotkeysTab->ApplyConfiguration(registry); + ui->cpuTab->ApplyConfiguration(); + ui->cpuDebugTab->ApplyConfiguration(); ui->graphicsTab->ApplyConfiguration(); ui->graphicsAdvancedTab->ApplyConfiguration(); ui->audioTab->ApplyConfiguration(); @@ -74,12 +82,13 @@ void ConfigureDialog::RetranslateUI() { Q_DECLARE_METATYPE(QList<QWidget*>); void ConfigureDialog::PopulateSelectionList() { - const std::array<std::pair<QString, QList<QWidget*>>, 5> items{ - {{tr("General"), {ui->generalTab, ui->webTab, ui->debugTab, ui->uiTab}}, + const std::array<std::pair<QString, QList<QWidget*>>, 6> items{ + {{tr("General"), {ui->generalTab, ui->hotkeysTab, ui->uiTab, ui->webTab, ui->debugTab}}, {tr("System"), {ui->systemTab, ui->profileManagerTab, ui->serviceTab, ui->filesystemTab}}, + {tr("CPU"), {ui->cpuTab, ui->cpuDebugTab}}, {tr("Graphics"), {ui->graphicsTab, ui->graphicsAdvancedTab}}, {tr("Audio"), {ui->audioTab}}, - {tr("Controls"), {ui->inputTab, ui->hotkeysTab}}}, + {tr("Controls"), ui->inputTab->GetSubTabs()}}, }; [[maybe_unused]] const QSignalBlocker blocker(ui->selectorList); @@ -93,6 +102,14 @@ void ConfigureDialog::PopulateSelectionList() { } } +void ConfigureDialog::OnLanguageChanged(const QString& locale) { + emit LanguageChanged(locale); + // first apply the configuration, and then restore the display + ApplyConfiguration(); + RetranslateUI(); + SetConfiguration(); +} + void ConfigureDialog::UpdateVisibleTabs() { const auto items = ui->selectorList->selectedItems(); if (items.isEmpty()) { @@ -103,8 +120,10 @@ void ConfigureDialog::UpdateVisibleTabs() { {ui->generalTab, tr("General")}, {ui->systemTab, tr("System")}, {ui->profileManagerTab, tr("Profiles")}, - {ui->inputTab, tr("Input")}, + {ui->inputTab, tr("Controls")}, {ui->hotkeysTab, tr("Hotkeys")}, + {ui->cpuTab, tr("CPU")}, + {ui->cpuDebugTab, tr("Debug")}, {ui->graphicsTab, tr("Graphics")}, {ui->graphicsAdvancedTab, tr("Advanced")}, {ui->audioTab, tr("Audio")}, @@ -122,6 +141,6 @@ void ConfigureDialog::UpdateVisibleTabs() { const QList<QWidget*> tabs = qvariant_cast<QList<QWidget*>>(items[0]->data(Qt::UserRole)); for (const auto tab : tabs) { - ui->tabWidget->addTab(tab, widgets.at(tab)); + ui->tabWidget->addTab(tab, tab->accessibleName()); } } diff --git a/src/yuzu/configuration/configure_dialog.h b/src/yuzu/configuration/configure_dialog.h index 2d3bfc2da..570c3b941 100644 --- a/src/yuzu/configuration/configure_dialog.h +++ b/src/yuzu/configuration/configure_dialog.h @@ -9,6 +9,10 @@ class HotkeyRegistry; +namespace InputCommon { +class InputSubsystem; +} + namespace Ui { class ConfigureDialog; } @@ -17,11 +21,18 @@ class ConfigureDialog : public QDialog { Q_OBJECT public: - explicit ConfigureDialog(QWidget* parent, HotkeyRegistry& registry); + explicit ConfigureDialog(QWidget* parent, HotkeyRegistry& registry, + InputCommon::InputSubsystem* input_subsystem); ~ConfigureDialog() override; void ApplyConfiguration(); +private slots: + void OnLanguageChanged(const QString& locale); + +signals: + void LanguageChanged(const QString& locale); + private: void changeEvent(QEvent* event) override; diff --git a/src/yuzu/configuration/configure_filesystem.cpp b/src/yuzu/configuration/configure_filesystem.cpp index 29f540eb7..7ab4a80f7 100644 --- a/src/yuzu/configuration/configure_filesystem.cpp +++ b/src/yuzu/configuration/configure_filesystem.cpp @@ -11,19 +11,6 @@ #include "yuzu/configuration/configure_filesystem.h" #include "yuzu/uisettings.h" -namespace { - -template <typename T> -void SetComboBoxFromData(QComboBox* combo_box, T data) { - const auto index = combo_box->findData(QVariant::fromValue(static_cast<u64>(data))); - if (index >= combo_box->count() || index < 0) - return; - - combo_box->setCurrentIndex(index); -} - -} // Anonymous namespace - ConfigureFilesystem::ConfigureFilesystem(QWidget* parent) : QWidget(parent), ui(std::make_unique<Ui::ConfigureFilesystem>()) { ui->setupUi(this); @@ -55,16 +42,16 @@ ConfigureFilesystem::~ConfigureFilesystem() = default; void ConfigureFilesystem::setConfiguration() { ui->nand_directory_edit->setText( - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir))); + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir))); ui->sdmc_directory_edit->setText( - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir))); + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir))); ui->gamecard_path_edit->setText(QString::fromStdString(Settings::values.gamecard_path)); ui->dump_path_edit->setText( - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::DumpDir))); + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::DumpDir))); ui->load_path_edit->setText( - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::LoadDir))); + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::LoadDir))); ui->cache_directory_edit->setText( - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir))); + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir))); ui->gamecard_inserted->setChecked(Settings::values.gamecard_inserted); ui->gamecard_current_game->setChecked(Settings::values.gamecard_current_game); @@ -73,23 +60,20 @@ void ConfigureFilesystem::setConfiguration() { ui->cache_game_list->setChecked(UISettings::values.cache_game_list); - SetComboBoxFromData(ui->nand_size, Settings::values.nand_total_size); - SetComboBoxFromData(ui->usrnand_size, Settings::values.nand_user_size); - SetComboBoxFromData(ui->sysnand_size, Settings::values.nand_system_size); - SetComboBoxFromData(ui->sdmc_size, Settings::values.sdmc_size); - UpdateEnabledControls(); } void ConfigureFilesystem::applyConfiguration() { - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir, - ui->nand_directory_edit->text().toStdString()); - FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir, - ui->sdmc_directory_edit->text().toStdString()); - FileUtil::GetUserPath(FileUtil::UserPath::DumpDir, ui->dump_path_edit->text().toStdString()); - FileUtil::GetUserPath(FileUtil::UserPath::LoadDir, ui->load_path_edit->text().toStdString()); - FileUtil::GetUserPath(FileUtil::UserPath::CacheDir, - ui->cache_directory_edit->text().toStdString()); + Common::FS::GetUserPath(Common::FS::UserPath::NANDDir, + ui->nand_directory_edit->text().toStdString()); + Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir, + ui->sdmc_directory_edit->text().toStdString()); + Common::FS::GetUserPath(Common::FS::UserPath::DumpDir, + ui->dump_path_edit->text().toStdString()); + Common::FS::GetUserPath(Common::FS::UserPath::LoadDir, + ui->load_path_edit->text().toStdString()); + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir, + ui->cache_directory_edit->text().toStdString()); Settings::values.gamecard_path = ui->gamecard_path_edit->text().toStdString(); Settings::values.gamecard_inserted = ui->gamecard_inserted->isChecked(); @@ -98,15 +82,6 @@ void ConfigureFilesystem::applyConfiguration() { Settings::values.dump_nso = ui->dump_nso->isChecked(); UISettings::values.cache_game_list = ui->cache_game_list->isChecked(); - - Settings::values.nand_total_size = static_cast<Settings::NANDTotalSize>( - ui->nand_size->itemData(ui->nand_size->currentIndex()).toULongLong()); - Settings::values.nand_system_size = static_cast<Settings::NANDSystemSize>( - ui->nand_size->itemData(ui->sysnand_size->currentIndex()).toULongLong()); - Settings::values.nand_user_size = static_cast<Settings::NANDUserSize>( - ui->nand_size->itemData(ui->usrnand_size->currentIndex()).toULongLong()); - Settings::values.sdmc_size = static_cast<Settings::SDMCSize>( - ui->nand_size->itemData(ui->sdmc_size->currentIndex()).toULongLong()); } void ConfigureFilesystem::SetDirectory(DirectoryTarget target, QLineEdit* edit) { @@ -138,7 +113,7 @@ void ConfigureFilesystem::SetDirectory(DirectoryTarget target, QLineEdit* edit) str = QFileDialog::getOpenFileName(this, caption, QFileInfo(edit->text()).dir().path(), QStringLiteral("NX Gamecard;*.xci")); } else { - str = QFileDialog::getExistingDirectory(this, caption, edit->text()); + str = QFileDialog::getExistingDirectory(this, caption, edit->text()) + QDir::separator(); } if (str.isEmpty()) @@ -148,12 +123,13 @@ void ConfigureFilesystem::SetDirectory(DirectoryTarget target, QLineEdit* edit) } void ConfigureFilesystem::ResetMetadata() { - if (!FileUtil::Exists(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + DIR_SEP + - "game_list")) { + if (!Common::FS::Exists(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + DIR_SEP + + "game_list")) { QMessageBox::information(this, tr("Reset Metadata Cache"), tr("The metadata cache is already empty.")); - } else if (FileUtil::DeleteDirRecursively(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + - DIR_SEP + "game_list")) { + } else if (Common::FS::DeleteDirRecursively( + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + DIR_SEP + + "game_list")) { QMessageBox::information(this, tr("Reset Metadata Cache"), tr("The operation completed successfully.")); UISettings::values.is_game_list_reload_pending.exchange(true); diff --git a/src/yuzu/configuration/configure_filesystem.ui b/src/yuzu/configuration/configure_filesystem.ui index 58cd07f52..84bea0600 100644 --- a/src/yuzu/configuration/configure_filesystem.ui +++ b/src/yuzu/configuration/configure_filesystem.ui @@ -116,127 +116,6 @@ </widget> </item> <item> - <widget class="QGroupBox" name="groupBox_3"> - <property name="title"> - <string>Storage Sizes</string> - </property> - <layout class="QGridLayout" name="gridLayout_3"> - <item row="3" column="0"> - <widget class="QLabel" name="label_5"> - <property name="text"> - <string>SD Card</string> - </property> - </widget> - </item> - <item row="1" column="0"> - <widget class="QLabel" name="label_4"> - <property name="text"> - <string>System NAND</string> - </property> - </widget> - </item> - <item row="1" column="1"> - <widget class="QComboBox" name="sysnand_size"> - <item> - <property name="text"> - <string>2.5 GB</string> - </property> - </item> - </widget> - </item> - <item row="3" column="1"> - <widget class="QComboBox" name="sdmc_size"> - <property name="currentText"> - <string>32 GB</string> - </property> - <item> - <property name="text"> - <string>1 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>2 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>4 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>8 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>16 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>32 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>64 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>128 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>256 GB</string> - </property> - </item> - <item> - <property name="text"> - <string>1 TB</string> - </property> - </item> - </widget> - </item> - <item row="2" column="1"> - <widget class="QComboBox" name="usrnand_size"> - <item> - <property name="text"> - <string>26 GB</string> - </property> - </item> - </widget> - </item> - <item row="2" column="0"> - <widget class="QLabel" name="label_6"> - <property name="text"> - <string>User NAND</string> - </property> - </widget> - </item> - <item row="0" column="0"> - <widget class="QLabel" name="label_7"> - <property name="text"> - <string>NAND</string> - </property> - </widget> - </item> - <item row="0" column="1"> - <widget class="QComboBox" name="nand_size"> - <item> - <property name="text"> - <string>29.1 GB</string> - </property> - </item> - </widget> - </item> - </layout> - </widget> - </item> - <item> <widget class="QGroupBox" name="groupBox_4"> <property name="title"> <string>Patch Manager</string> diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp index 5ef927114..d4d29d422 100644 --- a/src/yuzu/configuration/configure_general.cpp +++ b/src/yuzu/configuration/configure_general.cpp @@ -7,38 +7,77 @@ #include "core/core.h" #include "core/settings.h" #include "ui_configure_general.h" +#include "yuzu/configuration/configuration_shared.h" #include "yuzu/configuration/configure_general.h" #include "yuzu/uisettings.h" ConfigureGeneral::ConfigureGeneral(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureGeneral) { - ui->setupUi(this); + SetupPerGameUI(); + SetConfiguration(); - connect(ui->toggle_frame_limit, &QCheckBox::toggled, ui->frame_limit, &QSpinBox::setEnabled); + if (Settings::IsConfiguringGlobal()) { + connect(ui->toggle_frame_limit, &QCheckBox::clicked, ui->frame_limit, + [this]() { ui->frame_limit->setEnabled(ui->toggle_frame_limit->isChecked()); }); + } } ConfigureGeneral::~ConfigureGeneral() = default; void ConfigureGeneral::SetConfiguration() { + const bool runtime_lock = !Core::System::GetInstance().IsPoweredOn(); + + ui->use_multi_core->setEnabled(runtime_lock); + ui->use_multi_core->setChecked(Settings::values.use_multi_core.GetValue()); + ui->toggle_check_exit->setChecked(UISettings::values.confirm_before_closing); ui->toggle_user_on_boot->setChecked(UISettings::values.select_user_on_boot); ui->toggle_background_pause->setChecked(UISettings::values.pause_when_in_background); + ui->toggle_hide_mouse->setChecked(UISettings::values.hide_mouse); + + ui->toggle_frame_limit->setChecked(Settings::values.use_frame_limit.GetValue()); + ui->frame_limit->setValue(Settings::values.frame_limit.GetValue()); - ui->toggle_frame_limit->setChecked(Settings::values.use_frame_limit); - ui->frame_limit->setEnabled(ui->toggle_frame_limit->isChecked()); - ui->frame_limit->setValue(Settings::values.frame_limit); + if (Settings::IsConfiguringGlobal()) { + ui->frame_limit->setEnabled(Settings::values.use_frame_limit.GetValue()); + } else { + ui->frame_limit->setEnabled(Settings::values.use_frame_limit.GetValue() && + use_frame_limit != ConfigurationShared::CheckState::Global); + } } void ConfigureGeneral::ApplyConfiguration() { - UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); - UISettings::values.select_user_on_boot = ui->toggle_user_on_boot->isChecked(); - UISettings::values.pause_when_in_background = ui->toggle_background_pause->isChecked(); + if (Settings::IsConfiguringGlobal()) { + UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); + UISettings::values.select_user_on_boot = ui->toggle_user_on_boot->isChecked(); + UISettings::values.pause_when_in_background = ui->toggle_background_pause->isChecked(); + UISettings::values.hide_mouse = ui->toggle_hide_mouse->isChecked(); + + // Guard if during game and set to game-specific value + if (Settings::values.use_frame_limit.UsingGlobal()) { + Settings::values.use_frame_limit.SetValue(ui->toggle_frame_limit->checkState() == + Qt::Checked); + Settings::values.frame_limit.SetValue(ui->frame_limit->value()); + } + if (Settings::values.use_multi_core.UsingGlobal()) { + Settings::values.use_multi_core.SetValue(ui->use_multi_core->isChecked()); + } + } else { + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_multi_core, + ui->use_multi_core, use_multi_core); - Settings::values.use_frame_limit = ui->toggle_frame_limit->isChecked(); - Settings::values.frame_limit = ui->frame_limit->value(); + bool global_frame_limit = use_frame_limit == ConfigurationShared::CheckState::Global; + Settings::values.use_frame_limit.SetGlobal(global_frame_limit); + Settings::values.frame_limit.SetGlobal(global_frame_limit); + if (!global_frame_limit) { + Settings::values.use_frame_limit.SetValue(ui->toggle_frame_limit->checkState() == + Qt::Checked); + Settings::values.frame_limit.SetValue(ui->frame_limit->value()); + } + } } void ConfigureGeneral::changeEvent(QEvent* event) { @@ -52,3 +91,27 @@ void ConfigureGeneral::changeEvent(QEvent* event) { void ConfigureGeneral::RetranslateUI() { ui->retranslateUi(this); } + +void ConfigureGeneral::SetupPerGameUI() { + if (Settings::IsConfiguringGlobal()) { + ui->toggle_frame_limit->setEnabled(Settings::values.use_frame_limit.UsingGlobal()); + ui->frame_limit->setEnabled(Settings::values.frame_limit.UsingGlobal()); + + return; + } + + ui->toggle_check_exit->setVisible(false); + ui->toggle_user_on_boot->setVisible(false); + ui->toggle_background_pause->setVisible(false); + ui->toggle_hide_mouse->setVisible(false); + + ConfigurationShared::SetColoredTristate(ui->toggle_frame_limit, + Settings::values.use_frame_limit, use_frame_limit); + ConfigurationShared::SetColoredTristate(ui->use_multi_core, Settings::values.use_multi_core, + use_multi_core); + + connect(ui->toggle_frame_limit, &QCheckBox::clicked, ui->frame_limit, [this]() { + ui->frame_limit->setEnabled(ui->toggle_frame_limit->isChecked() && + (use_frame_limit != ConfigurationShared::CheckState::Global)); + }); +} diff --git a/src/yuzu/configuration/configure_general.h b/src/yuzu/configuration/configure_general.h index ef05ce065..323ffbd8f 100644 --- a/src/yuzu/configuration/configure_general.h +++ b/src/yuzu/configuration/configure_general.h @@ -7,6 +7,10 @@ #include <memory> #include <QWidget> +namespace ConfigurationShared { +enum class CheckState; +} + class HotkeyRegistry; namespace Ui { @@ -28,5 +32,10 @@ private: void SetConfiguration(); + void SetupPerGameUI(); + std::unique_ptr<Ui::ConfigureGeneral> ui; + + ConfigurationShared::CheckState use_frame_limit; + ConfigurationShared::CheckState use_multi_core; }; diff --git a/src/yuzu/configuration/configure_general.ui b/src/yuzu/configuration/configure_general.ui index 857119bb3..2711116a2 100644 --- a/src/yuzu/configuration/configure_general.ui +++ b/src/yuzu/configuration/configure_general.ui @@ -52,6 +52,13 @@ </layout> </item> <item> + <widget class="QCheckBox" name="use_multi_core"> + <property name="text"> + <string>Multicore CPU Emulation</string> + </property> + </widget> + </item> + <item> <widget class="QCheckBox" name="toggle_check_exit"> <property name="text"> <string>Confirm exit while emulation is running</string> @@ -72,6 +79,13 @@ </property> </widget> </item> + <item> + <widget class="QCheckBox" name="toggle_hide_mouse"> + <property name="text"> + <string>Hide mouse on inactivity</string> + </property> + </widget> + </item> </layout> </item> </layout> diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp index ea667caef..6fda0ce35 100644 --- a/src/yuzu/configuration/configure_graphics.cpp +++ b/src/yuzu/configuration/configure_graphics.cpp @@ -13,65 +13,32 @@ #include "core/core.h" #include "core/settings.h" #include "ui_configure_graphics.h" +#include "yuzu/configuration/configuration_shared.h" #include "yuzu/configuration/configure_graphics.h" #ifdef HAS_VULKAN #include "video_core/renderer_vulkan/renderer_vulkan.h" #endif -namespace { -enum class Resolution : int { - Auto, - Scale1x, - Scale2x, - Scale3x, - Scale4x, -}; - -float ToResolutionFactor(Resolution option) { - switch (option) { - case Resolution::Auto: - return 0.f; - case Resolution::Scale1x: - return 1.f; - case Resolution::Scale2x: - return 2.f; - case Resolution::Scale3x: - return 3.f; - case Resolution::Scale4x: - return 4.f; - } - return 0.f; -} - -Resolution FromResolutionFactor(float factor) { - if (factor == 0.f) { - return Resolution::Auto; - } else if (factor == 1.f) { - return Resolution::Scale1x; - } else if (factor == 2.f) { - return Resolution::Scale2x; - } else if (factor == 3.f) { - return Resolution::Scale3x; - } else if (factor == 4.f) { - return Resolution::Scale4x; - } - return Resolution::Auto; -} -} // Anonymous namespace - ConfigureGraphics::ConfigureGraphics(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureGraphics) { - vulkan_device = Settings::values.vulkan_device; + vulkan_device = Settings::values.vulkan_device.GetValue(); RetrieveVulkanDevices(); ui->setupUi(this); + SetupPerGameUI(); + SetConfiguration(); - connect(ui->api, static_cast<void (QComboBox::*)(int)>(&QComboBox::currentIndexChanged), this, - [this] { UpdateDeviceComboBox(); }); - connect(ui->device, static_cast<void (QComboBox::*)(int)>(&QComboBox::activated), this, + connect(ui->api, qOverload<int>(&QComboBox::currentIndexChanged), this, [this] { + UpdateDeviceComboBox(); + if (!Settings::IsConfiguringGlobal()) { + ConfigurationShared::SetHighlight( + ui->api_layout, ui->api->currentIndex() != ConfigurationShared::USE_GLOBAL_INDEX); + } + }); + connect(ui->device, qOverload<int>(&QComboBox::activated), this, [this](int device) { UpdateDeviceSelection(device); }); connect(ui->bg_button, &QPushButton::clicked, this, [this] { @@ -81,6 +48,9 @@ ConfigureGraphics::ConfigureGraphics(QWidget* parent) } UpdateBackgroundColorButton(new_bg_color); }); + + ui->bg_label->setVisible(Settings::IsConfiguringGlobal()); + ui->bg_combobox->setVisible(!Settings::IsConfiguringGlobal()); } void ConfigureGraphics::UpdateDeviceSelection(int device) { @@ -98,31 +68,103 @@ void ConfigureGraphics::SetConfiguration() { const bool runtime_lock = !Core::System::GetInstance().IsPoweredOn(); ui->api->setEnabled(runtime_lock); - ui->api->setCurrentIndex(static_cast<int>(Settings::values.renderer_backend)); - ui->resolution_factor_combobox->setCurrentIndex( - static_cast<int>(FromResolutionFactor(Settings::values.resolution_factor))); - ui->aspect_ratio_combobox->setCurrentIndex(Settings::values.aspect_ratio); - ui->use_disk_shader_cache->setEnabled(runtime_lock); - ui->use_disk_shader_cache->setChecked(Settings::values.use_disk_shader_cache); ui->use_asynchronous_gpu_emulation->setEnabled(runtime_lock); - ui->use_asynchronous_gpu_emulation->setChecked(Settings::values.use_asynchronous_gpu_emulation); - UpdateBackgroundColorButton(QColor::fromRgbF(Settings::values.bg_red, Settings::values.bg_green, - Settings::values.bg_blue)); + ui->use_disk_shader_cache->setEnabled(runtime_lock); + ui->use_nvdec_emulation->setEnabled(runtime_lock); + ui->use_disk_shader_cache->setChecked(Settings::values.use_disk_shader_cache.GetValue()); + ui->use_asynchronous_gpu_emulation->setChecked( + Settings::values.use_asynchronous_gpu_emulation.GetValue()); + ui->use_nvdec_emulation->setChecked(Settings::values.use_nvdec_emulation.GetValue()); + + if (Settings::IsConfiguringGlobal()) { + ui->api->setCurrentIndex(static_cast<int>(Settings::values.renderer_backend.GetValue())); + ui->aspect_ratio_combobox->setCurrentIndex(Settings::values.aspect_ratio.GetValue()); + } else { + ConfigurationShared::SetPerGameSetting(ui->api, &Settings::values.renderer_backend); + ConfigurationShared::SetHighlight(ui->api_layout, + !Settings::values.renderer_backend.UsingGlobal()); + ConfigurationShared::SetPerGameSetting(ui->aspect_ratio_combobox, + &Settings::values.aspect_ratio); + + ui->bg_combobox->setCurrentIndex(Settings::values.bg_red.UsingGlobal() ? 0 : 1); + ui->bg_button->setEnabled(!Settings::values.bg_red.UsingGlobal()); + ConfigurationShared::SetHighlight(ui->ar_label, + !Settings::values.aspect_ratio.UsingGlobal()); + ConfigurationShared::SetHighlight(ui->bg_layout, !Settings::values.bg_red.UsingGlobal()); + } + + UpdateBackgroundColorButton(QColor::fromRgbF(Settings::values.bg_red.GetValue(), + Settings::values.bg_green.GetValue(), + Settings::values.bg_blue.GetValue())); UpdateDeviceComboBox(); } void ConfigureGraphics::ApplyConfiguration() { - Settings::values.renderer_backend = GetCurrentGraphicsBackend(); - Settings::values.vulkan_device = vulkan_device; - Settings::values.resolution_factor = - ToResolutionFactor(static_cast<Resolution>(ui->resolution_factor_combobox->currentIndex())); - Settings::values.aspect_ratio = ui->aspect_ratio_combobox->currentIndex(); - Settings::values.use_disk_shader_cache = ui->use_disk_shader_cache->isChecked(); - Settings::values.use_asynchronous_gpu_emulation = - ui->use_asynchronous_gpu_emulation->isChecked(); - Settings::values.bg_red = static_cast<float>(bg_color.redF()); - Settings::values.bg_green = static_cast<float>(bg_color.greenF()); - Settings::values.bg_blue = static_cast<float>(bg_color.blueF()); + if (Settings::IsConfiguringGlobal()) { + // Guard if during game and set to game-specific value + if (Settings::values.renderer_backend.UsingGlobal()) { + Settings::values.renderer_backend.SetValue(GetCurrentGraphicsBackend()); + } + if (Settings::values.vulkan_device.UsingGlobal()) { + Settings::values.vulkan_device.SetValue(vulkan_device); + } + if (Settings::values.aspect_ratio.UsingGlobal()) { + Settings::values.aspect_ratio.SetValue(ui->aspect_ratio_combobox->currentIndex()); + } + if (Settings::values.use_disk_shader_cache.UsingGlobal()) { + Settings::values.use_disk_shader_cache.SetValue(ui->use_disk_shader_cache->isChecked()); + } + if (Settings::values.use_asynchronous_gpu_emulation.UsingGlobal()) { + Settings::values.use_asynchronous_gpu_emulation.SetValue( + ui->use_asynchronous_gpu_emulation->isChecked()); + } + if (Settings::values.use_nvdec_emulation.UsingGlobal()) { + Settings::values.use_nvdec_emulation.SetValue(ui->use_nvdec_emulation->isChecked()); + } + if (Settings::values.bg_red.UsingGlobal()) { + Settings::values.bg_red.SetValue(static_cast<float>(bg_color.redF())); + Settings::values.bg_green.SetValue(static_cast<float>(bg_color.greenF())); + Settings::values.bg_blue.SetValue(static_cast<float>(bg_color.blueF())); + } + } else { + if (ui->api->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + Settings::values.renderer_backend.SetGlobal(true); + Settings::values.vulkan_device.SetGlobal(true); + } else { + Settings::values.renderer_backend.SetGlobal(false); + Settings::values.renderer_backend.SetValue(GetCurrentGraphicsBackend()); + if (GetCurrentGraphicsBackend() == Settings::RendererBackend::Vulkan) { + Settings::values.vulkan_device.SetGlobal(false); + Settings::values.vulkan_device.SetValue(vulkan_device); + } else { + Settings::values.vulkan_device.SetGlobal(true); + } + } + + ConfigurationShared::ApplyPerGameSetting(&Settings::values.aspect_ratio, + ui->aspect_ratio_combobox); + + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_disk_shader_cache, + ui->use_disk_shader_cache, use_disk_shader_cache); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_asynchronous_gpu_emulation, + ui->use_asynchronous_gpu_emulation, + use_asynchronous_gpu_emulation); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_nvdec_emulation, + ui->use_nvdec_emulation, use_nvdec_emulation); + + if (ui->bg_combobox->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + Settings::values.bg_red.SetGlobal(true); + Settings::values.bg_green.SetGlobal(true); + Settings::values.bg_blue.SetGlobal(true); + } else { + Settings::values.bg_red.SetGlobal(false); + Settings::values.bg_green.SetGlobal(false); + Settings::values.bg_blue.SetGlobal(false); + Settings::values.bg_red.SetValue(static_cast<float>(bg_color.redF())); + Settings::values.bg_green.SetValue(static_cast<float>(bg_color.greenF())); + Settings::values.bg_blue.SetValue(static_cast<float>(bg_color.blueF())); + } + } } void ConfigureGraphics::changeEvent(QEvent* event) { @@ -151,19 +193,27 @@ void ConfigureGraphics::UpdateDeviceComboBox() { ui->device->clear(); bool enabled = false; + + if (!Settings::IsConfiguringGlobal() && + ui->api->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + vulkan_device = Settings::values.vulkan_device.GetValue(); + } switch (GetCurrentGraphicsBackend()) { case Settings::RendererBackend::OpenGL: ui->device->addItem(tr("OpenGL Graphics Device")); enabled = false; break; case Settings::RendererBackend::Vulkan: - for (const auto device : vulkan_devices) { + for (const auto& device : vulkan_devices) { ui->device->addItem(device); } ui->device->setCurrentIndex(vulkan_device); enabled = !vulkan_devices.empty(); break; } + // If in per-game config and use global is selected, don't enable. + enabled &= !(!Settings::IsConfiguringGlobal() && + ui->api->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX); ui->device->setEnabled(enabled && !Core::System::GetInstance().IsPoweredOn()); } @@ -177,5 +227,48 @@ void ConfigureGraphics::RetrieveVulkanDevices() { } Settings::RendererBackend ConfigureGraphics::GetCurrentGraphicsBackend() const { - return static_cast<Settings::RendererBackend>(ui->api->currentIndex()); + if (Settings::IsConfiguringGlobal()) { + return static_cast<Settings::RendererBackend>(ui->api->currentIndex()); + } + + if (ui->api->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + Settings::values.renderer_backend.SetGlobal(true); + return Settings::values.renderer_backend.GetValue(); + } + Settings::values.renderer_backend.SetGlobal(false); + return static_cast<Settings::RendererBackend>(ui->api->currentIndex() - + ConfigurationShared::USE_GLOBAL_OFFSET); +} + +void ConfigureGraphics::SetupPerGameUI() { + if (Settings::IsConfiguringGlobal()) { + ui->api->setEnabled(Settings::values.renderer_backend.UsingGlobal()); + ui->device->setEnabled(Settings::values.renderer_backend.UsingGlobal()); + ui->aspect_ratio_combobox->setEnabled(Settings::values.aspect_ratio.UsingGlobal()); + ui->use_asynchronous_gpu_emulation->setEnabled( + Settings::values.use_asynchronous_gpu_emulation.UsingGlobal()); + ui->use_nvdec_emulation->setEnabled(Settings::values.use_nvdec_emulation.UsingGlobal()); + ui->use_disk_shader_cache->setEnabled(Settings::values.use_disk_shader_cache.UsingGlobal()); + ui->bg_button->setEnabled(Settings::values.bg_red.UsingGlobal()); + + return; + } + + connect(ui->bg_combobox, qOverload<int>(&QComboBox::activated), this, [this](int index) { + ui->bg_button->setEnabled(index == 1); + ConfigurationShared::SetHighlight(ui->bg_layout, index == 1); + }); + + ConfigurationShared::SetColoredTristate( + ui->use_disk_shader_cache, Settings::values.use_disk_shader_cache, use_disk_shader_cache); + ConfigurationShared::SetColoredTristate( + ui->use_nvdec_emulation, Settings::values.use_nvdec_emulation, use_nvdec_emulation); + ConfigurationShared::SetColoredTristate(ui->use_asynchronous_gpu_emulation, + Settings::values.use_asynchronous_gpu_emulation, + use_asynchronous_gpu_emulation); + + ConfigurationShared::SetColoredComboBox(ui->aspect_ratio_combobox, ui->ar_label, + Settings::values.aspect_ratio.GetValue(true)); + ConfigurationShared::InsertGlobalItem( + ui->api, static_cast<int>(Settings::values.renderer_backend.GetValue(true))); } diff --git a/src/yuzu/configuration/configure_graphics.h b/src/yuzu/configuration/configure_graphics.h index 7e0596d9c..1fefc88eb 100644 --- a/src/yuzu/configuration/configure_graphics.h +++ b/src/yuzu/configuration/configure_graphics.h @@ -10,6 +10,10 @@ #include <QWidget> #include "core/settings.h" +namespace ConfigurationShared { +enum class CheckState; +} + namespace Ui { class ConfigureGraphics; } @@ -35,11 +39,17 @@ private: void RetrieveVulkanDevices(); + void SetupPerGameUI(); + Settings::RendererBackend GetCurrentGraphicsBackend() const; std::unique_ptr<Ui::ConfigureGraphics> ui; QColor bg_color; + ConfigurationShared::CheckState use_nvdec_emulation; + ConfigurationShared::CheckState use_disk_shader_cache; + ConfigurationShared::CheckState use_asynchronous_gpu_emulation; + std::vector<QString> vulkan_devices; u32 vulkan_device{}; }; diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui index c816d6108..58486eb1e 100644 --- a/src/yuzu/configuration/configure_graphics.ui +++ b/src/yuzu/configuration/configure_graphics.ui @@ -6,7 +6,7 @@ <rect> <x>0</x> <y>0</y> - <width>400</width> + <width>437</width> <height>321</height> </rect> </property> @@ -23,43 +23,56 @@ </property> <layout class="QVBoxLayout" name="verticalLayout_3"> <item> - <layout class="QHBoxLayout" name="horizontalLayout_4"> - <item> - <widget class="QLabel" name="label_2"> - <property name="text"> - <string>API:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="api"> - <item> + <widget class="QWidget" name="api_layout" native="true"> + <layout class="QGridLayout" name="gridLayout"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <property name="horizontalSpacing"> + <number>6</number> + </property> + <item row="0" column="0"> + <widget class="QLabel" name="api_label"> <property name="text"> - <string notr="true">OpenGL</string> + <string>API:</string> </property> - </item> - <item> + </widget> + </item> + <item row="0" column="1"> + <widget class="QComboBox" name="api"> + <item> + <property name="text"> + <string notr="true">OpenGL</string> + </property> + </item> + <item> + <property name="text"> + <string notr="true">Vulkan</string> + </property> + </item> + </widget> + </item> + <item row="1" column="0"> + <widget class="QLabel" name="device_label"> <property name="text"> - <string notr="true">Vulkan</string> + <string>Device:</string> </property> - </item> - </widget> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_5"> - <item> - <widget class="QLabel" name="label_3"> - <property name="text"> - <string>Device:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="device"/> - </item> - </layout> + </widget> + </item> + <item row="1" column="1"> + <widget class="QComboBox" name="device"/> + </item> + </layout> + </widget> </item> </layout> </widget> @@ -85,100 +98,140 @@ </widget> </item> <item> - <layout class="QHBoxLayout" name="horizontalLayout_2"> - <item> - <widget class="QLabel" name="label"> - <property name="text"> - <string>Internal Resolution:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="resolution_factor_combobox"> - <item> - <property name="text"> - <string>Auto (Window Size)</string> - </property> - </item> - <item> + <widget class="QCheckBox" name="use_nvdec_emulation"> + <property name="text"> + <string>Use NVDEC emulation</string> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="aspect_ratio_layout" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_6"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="ar_label"> <property name="text"> - <string>Native (1280x720)</string> + <string>Aspect Ratio:</string> </property> - </item> - <item> - <property name="text"> - <string>2x Native (2560x1440)</string> + </widget> + </item> + <item> + <widget class="QComboBox" name="aspect_ratio_combobox"> + <item> + <property name="text"> + <string>Default (16:9)</string> + </property> + </item> + <item> + <property name="text"> + <string>Force 4:3</string> + </property> + </item> + <item> + <property name="text"> + <string>Force 21:9</string> + </property> + </item> + <item> + <property name="text"> + <string>Stretch to Window</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="bg_layout" native="true"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Preferred" vsizetype="Preferred"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QComboBox" name="bg_combobox"> + <property name="currentText"> + <string>Use global background color</string> </property> - </item> - <item> - <property name="text"> - <string>3x Native (3840x2160)</string> + <property name="currentIndex"> + <number>0</number> </property> - </item> - <item> - <property name="text"> - <string>4x Native (5120x2880)</string> + <property name="maxVisibleItems"> + <number>10</number> </property> - </item> - </widget> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_6"> - <item> - <widget class="QLabel" name="ar_label"> - <property name="text"> - <string>Aspect Ratio:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="aspect_ratio_combobox"> - <item> + <item> + <property name="text"> + <string>Use global background color</string> + </property> + </item> + <item> + <property name="text"> + <string>Set background color:</string> + </property> + </item> + </widget> + </item> + <item> + <widget class="QLabel" name="bg_label"> <property name="text"> - <string>Default (16:9)</string> + <string>Background Color:</string> </property> - </item> - <item> - <property name="text"> - <string>Force 4:3</string> + </widget> + </item> + <item> + <spacer name="horizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> </property> - </item> - <item> - <property name="text"> - <string>Force 21:9</string> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> </property> - </item> - <item> - <property name="text"> - <string>Stretch to Window</string> + </spacer> + </item> + <item> + <widget class="QPushButton" name="bg_button"> + <property name="maximumSize"> + <size> + <width>40</width> + <height>16777215</height> + </size> </property> - </item> - </widget> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_3"> - <item> - <widget class="QLabel" name="bg_label"> - <property name="text"> - <string>Background Color:</string> - </property> - </widget> - </item> - <item> - <widget class="QPushButton" name="bg_button"> - <property name="maximumSize"> - <size> - <width>40</width> - <height>16777215</height> - </size> - </property> - </widget> - </item> - </layout> + </widget> + </item> + </layout> + </widget> </item> </layout> </widget> diff --git a/src/yuzu/configuration/configure_graphics_advanced.cpp b/src/yuzu/configuration/configure_graphics_advanced.cpp index b9f429f84..383c7bac8 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.cpp +++ b/src/yuzu/configuration/configure_graphics_advanced.cpp @@ -5,6 +5,7 @@ #include "core/core.h" #include "core/settings.h" #include "ui_configure_graphics_advanced.h" +#include "yuzu/configuration/configuration_shared.h" #include "yuzu/configuration/configure_graphics_advanced.h" ConfigureGraphicsAdvanced::ConfigureGraphicsAdvanced(QWidget* parent) @@ -12,6 +13,8 @@ ConfigureGraphicsAdvanced::ConfigureGraphicsAdvanced(QWidget* parent) ui->setupUi(this); + SetupPerGameUI(); + SetConfiguration(); } @@ -19,20 +22,86 @@ ConfigureGraphicsAdvanced::~ConfigureGraphicsAdvanced() = default; void ConfigureGraphicsAdvanced::SetConfiguration() { const bool runtime_lock = !Core::System::GetInstance().IsPoweredOn(); - ui->use_accurate_gpu_emulation->setChecked(Settings::values.use_accurate_gpu_emulation); ui->use_vsync->setEnabled(runtime_lock); - ui->use_vsync->setChecked(Settings::values.use_vsync); - ui->force_30fps_mode->setEnabled(runtime_lock); - ui->force_30fps_mode->setChecked(Settings::values.force_30fps_mode); + ui->use_assembly_shaders->setEnabled(runtime_lock); + ui->use_asynchronous_shaders->setEnabled(runtime_lock); ui->anisotropic_filtering_combobox->setEnabled(runtime_lock); - ui->anisotropic_filtering_combobox->setCurrentIndex(Settings::values.max_anisotropy); + + ui->use_vsync->setChecked(Settings::values.use_vsync.GetValue()); + ui->use_assembly_shaders->setChecked(Settings::values.use_assembly_shaders.GetValue()); + ui->use_asynchronous_shaders->setChecked(Settings::values.use_asynchronous_shaders.GetValue()); + ui->use_fast_gpu_time->setChecked(Settings::values.use_fast_gpu_time.GetValue()); + + if (Settings::IsConfiguringGlobal()) { + ui->gpu_accuracy->setCurrentIndex( + static_cast<int>(Settings::values.gpu_accuracy.GetValue())); + ui->anisotropic_filtering_combobox->setCurrentIndex( + Settings::values.max_anisotropy.GetValue()); + } else { + ConfigurationShared::SetPerGameSetting(ui->gpu_accuracy, &Settings::values.gpu_accuracy); + ConfigurationShared::SetPerGameSetting(ui->anisotropic_filtering_combobox, + &Settings::values.max_anisotropy); + ConfigurationShared::SetHighlight(ui->label_gpu_accuracy, + !Settings::values.gpu_accuracy.UsingGlobal()); + ConfigurationShared::SetHighlight(ui->af_label, + !Settings::values.max_anisotropy.UsingGlobal()); + } } void ConfigureGraphicsAdvanced::ApplyConfiguration() { - Settings::values.use_accurate_gpu_emulation = ui->use_accurate_gpu_emulation->isChecked(); - Settings::values.use_vsync = ui->use_vsync->isChecked(); - Settings::values.force_30fps_mode = ui->force_30fps_mode->isChecked(); - Settings::values.max_anisotropy = ui->anisotropic_filtering_combobox->currentIndex(); + // Subtract 2 if configuring per-game (separator and "use global configuration" take 2 slots) + const auto gpu_accuracy = static_cast<Settings::GPUAccuracy>( + ui->gpu_accuracy->currentIndex() - + ((Settings::IsConfiguringGlobal()) ? 0 : ConfigurationShared::USE_GLOBAL_OFFSET)); + + if (Settings::IsConfiguringGlobal()) { + // Must guard in case of a during-game configuration when set to be game-specific. + if (Settings::values.gpu_accuracy.UsingGlobal()) { + Settings::values.gpu_accuracy.SetValue(gpu_accuracy); + } + if (Settings::values.use_vsync.UsingGlobal()) { + Settings::values.use_vsync.SetValue(ui->use_vsync->isChecked()); + } + if (Settings::values.use_assembly_shaders.UsingGlobal()) { + Settings::values.use_assembly_shaders.SetValue(ui->use_assembly_shaders->isChecked()); + } + if (Settings::values.use_asynchronous_shaders.UsingGlobal()) { + Settings::values.use_asynchronous_shaders.SetValue( + ui->use_asynchronous_shaders->isChecked()); + } + if (Settings::values.use_asynchronous_shaders.UsingGlobal()) { + Settings::values.use_asynchronous_shaders.SetValue( + ui->use_asynchronous_shaders->isChecked()); + } + if (Settings::values.use_fast_gpu_time.UsingGlobal()) { + Settings::values.use_fast_gpu_time.SetValue(ui->use_fast_gpu_time->isChecked()); + } + if (Settings::values.max_anisotropy.UsingGlobal()) { + Settings::values.max_anisotropy.SetValue( + ui->anisotropic_filtering_combobox->currentIndex()); + } + } else { + ConfigurationShared::ApplyPerGameSetting(&Settings::values.max_anisotropy, + ui->anisotropic_filtering_combobox); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_vsync, ui->use_vsync, + use_vsync); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_assembly_shaders, + ui->use_assembly_shaders, use_assembly_shaders); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_asynchronous_shaders, + ui->use_asynchronous_shaders, + use_asynchronous_shaders); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_fast_gpu_time, + ui->use_fast_gpu_time, use_fast_gpu_time); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.max_anisotropy, + ui->anisotropic_filtering_combobox); + + if (ui->gpu_accuracy->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { + Settings::values.gpu_accuracy.SetGlobal(true); + } else { + Settings::values.gpu_accuracy.SetGlobal(false); + Settings::values.gpu_accuracy.SetValue(gpu_accuracy); + } + } } void ConfigureGraphicsAdvanced::changeEvent(QEvent* event) { @@ -46,3 +115,34 @@ void ConfigureGraphicsAdvanced::changeEvent(QEvent* event) { void ConfigureGraphicsAdvanced::RetranslateUI() { ui->retranslateUi(this); } + +void ConfigureGraphicsAdvanced::SetupPerGameUI() { + // Disable if not global (only happens during game) + if (Settings::IsConfiguringGlobal()) { + ui->gpu_accuracy->setEnabled(Settings::values.gpu_accuracy.UsingGlobal()); + ui->use_vsync->setEnabled(Settings::values.use_vsync.UsingGlobal()); + ui->use_assembly_shaders->setEnabled(Settings::values.use_assembly_shaders.UsingGlobal()); + ui->use_asynchronous_shaders->setEnabled( + Settings::values.use_asynchronous_shaders.UsingGlobal()); + ui->use_fast_gpu_time->setEnabled(Settings::values.use_fast_gpu_time.UsingGlobal()); + ui->anisotropic_filtering_combobox->setEnabled( + Settings::values.max_anisotropy.UsingGlobal()); + + return; + } + + ConfigurationShared::SetColoredTristate(ui->use_vsync, Settings::values.use_vsync, use_vsync); + ConfigurationShared::SetColoredTristate( + ui->use_assembly_shaders, Settings::values.use_assembly_shaders, use_assembly_shaders); + ConfigurationShared::SetColoredTristate(ui->use_asynchronous_shaders, + Settings::values.use_asynchronous_shaders, + use_asynchronous_shaders); + ConfigurationShared::SetColoredTristate(ui->use_fast_gpu_time, + Settings::values.use_fast_gpu_time, use_fast_gpu_time); + ConfigurationShared::SetColoredComboBox( + ui->gpu_accuracy, ui->label_gpu_accuracy, + static_cast<int>(Settings::values.gpu_accuracy.GetValue(true))); + ConfigurationShared::SetColoredComboBox( + ui->anisotropic_filtering_combobox, ui->af_label, + static_cast<int>(Settings::values.max_anisotropy.GetValue(true))); +} diff --git a/src/yuzu/configuration/configure_graphics_advanced.h b/src/yuzu/configuration/configure_graphics_advanced.h index bbc9d4355..e61b571c7 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.h +++ b/src/yuzu/configuration/configure_graphics_advanced.h @@ -7,6 +7,10 @@ #include <memory> #include <QWidget> +namespace ConfigurationShared { +enum class CheckState; +} + namespace Ui { class ConfigureGraphicsAdvanced; } @@ -26,5 +30,12 @@ private: void SetConfiguration(); + void SetupPerGameUI(); + std::unique_ptr<Ui::ConfigureGraphicsAdvanced> ui; + + ConfigurationShared::CheckState use_vsync; + ConfigurationShared::CheckState use_assembly_shaders; + ConfigurationShared::CheckState use_asynchronous_shaders; + ConfigurationShared::CheckState use_fast_gpu_time; }; diff --git a/src/yuzu/configuration/configure_graphics_advanced.ui b/src/yuzu/configuration/configure_graphics_advanced.ui index 42eec278e..846a30586 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.ui +++ b/src/yuzu/configuration/configure_graphics_advanced.ui @@ -6,7 +6,7 @@ <rect> <x>0</x> <y>0</y> - <width>400</width> + <width>404</width> <height>321</height> </rect> </property> @@ -23,10 +23,47 @@ </property> <layout class="QVBoxLayout" name="verticalLayout_3"> <item> - <widget class="QCheckBox" name="use_accurate_gpu_emulation"> - <property name="text"> - <string>Use accurate GPU emulation (slow)</string> - </property> + <widget class="QWidget" name="gpu_accuracy_layout" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="label_gpu_accuracy"> + <property name="text"> + <string>Accuracy Level:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="gpu_accuracy"> + <item> + <property name="text"> + <string notr="true">Normal</string> + </property> + </item> + <item> + <property name="text"> + <string notr="true">High</string> + </property> + </item> + <item> + <property name="text"> + <string notr="true">Extreme(very slow)</string> + </property> + </item> + </widget> + </item> + </layout> </widget> </item> <item> @@ -40,51 +77,85 @@ </widget> </item> <item> - <widget class="QCheckBox" name="force_30fps_mode"> + <widget class="QCheckBox" name="use_assembly_shaders"> + <property name="toolTip"> + <string>Enabling this reduces shader stutter. Enables OpenGL assembly shaders on supported Nvidia devices (NV_gpu_program5 is required). This feature is experimental.</string> + </property> <property name="text"> - <string>Force 30 FPS mode</string> + <string>Use assembly shaders (experimental, Nvidia OpenGL only)</string> </property> </widget> </item> <item> - <layout class="QHBoxLayout" name="horizontalLayout_1"> - <item> - <widget class="QLabel" name="af_label"> - <property name="text"> - <string>Anisotropic Filtering:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="anisotropic_filtering_combobox"> - <item> - <property name="text"> - <string>Default</string> - </property> - </item> - <item> - <property name="text"> - <string>2x</string> - </property> - </item> - <item> - <property name="text"> - <string>4x</string> - </property> - </item> - <item> - <property name="text"> - <string>8x</string> - </property> - </item> - <item> + <widget class="QCheckBox" name="use_asynchronous_shaders"> + <property name="toolTip"> + <string>Enables asynchronous shader compilation, which may reduce shader stutter. This feature is experimental.</string> + </property> + <property name="text"> + <string>Use asynchronous shader building (experimental)</string> + </property> + </widget> + </item> + <item> + <widget class="QCheckBox" name="use_fast_gpu_time"> + <property name="text"> + <string>Use Fast GPU Time</string> + </property> + </widget> + </item> + <item> + <widget class="QWidget" name="af_layout" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_1"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="af_label"> <property name="text"> - <string>16x</string> + <string>Anisotropic Filtering:</string> </property> - </item> - </widget> - </item> - </layout> + </widget> + </item> + <item> + <widget class="QComboBox" name="anisotropic_filtering_combobox"> + <item> + <property name="text"> + <string>Default</string> + </property> + </item> + <item> + <property name="text"> + <string>2x</string> + </property> + </item> + <item> + <property name="text"> + <string>4x</string> + </property> + </item> + <item> + <property name="text"> + <string>8x</string> + </property> + </item> + <item> + <property name="text"> + <string>16x</string> + </property> + </item> + </widget> + </item> + </layout> + </widget> </item> </layout> </widget> diff --git a/src/yuzu/configuration/configure_hotkeys.cpp b/src/yuzu/configuration/configure_hotkeys.cpp index fa9052136..cbee51a5e 100644 --- a/src/yuzu/configuration/configure_hotkeys.cpp +++ b/src/yuzu/configuration/configure_hotkeys.cpp @@ -2,10 +2,12 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <QMenu> #include <QMessageBox> #include <QStandardItemModel> #include "core/settings.h" #include "ui_configure_hotkeys.h" +#include "yuzu/configuration/config.h" #include "yuzu/configuration/configure_hotkeys.h" #include "yuzu/hotkeys.h" #include "yuzu/util/sequence_dialog/sequence_dialog.h" @@ -19,6 +21,9 @@ ConfigureHotkeys::ConfigureHotkeys(QWidget* parent) model->setColumnCount(3); connect(ui->hotkey_list, &QTreeView::doubleClicked, this, &ConfigureHotkeys::Configure); + connect(ui->hotkey_list, &QTreeView::customContextMenuRequested, this, + &ConfigureHotkeys::PopupContextMenu); + ui->hotkey_list->setContextMenuPolicy(Qt::CustomContextMenu); ui->hotkey_list->setModel(model); // TODO(Kloen): Make context configurable as well (hiding the column for now) @@ -27,6 +32,10 @@ ConfigureHotkeys::ConfigureHotkeys(QWidget* parent) ui->hotkey_list->setColumnWidth(0, 200); ui->hotkey_list->resizeColumnToContents(1); + connect(ui->button_restore_defaults, &QPushButton::clicked, this, + &ConfigureHotkeys::RestoreDefaults); + connect(ui->button_clear_all, &QPushButton::clicked, this, &ConfigureHotkeys::ClearAll); + RetranslateUI(); } @@ -71,7 +80,6 @@ void ConfigureHotkeys::Configure(QModelIndex index) { } index = index.sibling(index.row(), 1); - auto* const model = ui->hotkey_list->model(); const auto previous_key = model->data(index); SequenceDialog hotkey_dialog{this}; @@ -81,31 +89,33 @@ void ConfigureHotkeys::Configure(QModelIndex index) { if (return_code == QDialog::Rejected || key_sequence.isEmpty()) { return; } + const auto [key_sequence_used, used_action] = IsUsedKey(key_sequence); - if (IsUsedKey(key_sequence) && key_sequence != QKeySequence(previous_key.toString())) { - QMessageBox::warning(this, tr("Conflicting Key Sequence"), - tr("The entered key sequence is already assigned to another hotkey.")); + if (key_sequence_used && key_sequence != QKeySequence(previous_key.toString())) { + QMessageBox::warning( + this, tr("Conflicting Key Sequence"), + tr("The entered key sequence is already assigned to: %1").arg(used_action)); } else { model->setData(index, key_sequence.toString(QKeySequence::NativeText)); } } -bool ConfigureHotkeys::IsUsedKey(QKeySequence key_sequence) const { - for (int r = 0; r < model->rowCount(); r++) { +std::pair<bool, QString> ConfigureHotkeys::IsUsedKey(QKeySequence key_sequence) const { + for (int r = 0; r < model->rowCount(); ++r) { const QStandardItem* const parent = model->item(r, 0); - for (int r2 = 0; r2 < parent->rowCount(); r2++) { + for (int r2 = 0; r2 < parent->rowCount(); ++r2) { const QStandardItem* const key_seq_item = parent->child(r2, 1); const auto key_seq_str = key_seq_item->text(); const auto key_seq = QKeySequence::fromString(key_seq_str, QKeySequence::NativeText); if (key_sequence == key_seq) { - return true; + return std::make_pair(true, parent->child(r2, 0)->text()); } } } - return false; + return std::make_pair(false, QString()); } void ConfigureHotkeys::ApplyConfiguration(HotkeyRegistry& registry) { @@ -128,3 +138,55 @@ void ConfigureHotkeys::ApplyConfiguration(HotkeyRegistry& registry) { registry.SaveHotkeys(); } + +void ConfigureHotkeys::RestoreDefaults() { + for (int r = 0; r < model->rowCount(); ++r) { + const QStandardItem* parent = model->item(r, 0); + + for (int r2 = 0; r2 < parent->rowCount(); ++r2) { + model->item(r, 0)->child(r2, 1)->setText(Config::default_hotkeys[r2].shortcut.first); + } + } +} + +void ConfigureHotkeys::ClearAll() { + for (int r = 0; r < model->rowCount(); ++r) { + const QStandardItem* parent = model->item(r, 0); + + for (int r2 = 0; r2 < parent->rowCount(); ++r2) { + model->item(r, 0)->child(r2, 1)->setText(QString{}); + } + } +} + +void ConfigureHotkeys::PopupContextMenu(const QPoint& menu_location) { + QModelIndex index = ui->hotkey_list->indexAt(menu_location); + if (!index.parent().isValid()) { + return; + } + + const auto selected = index.sibling(index.row(), 1); + QMenu context_menu; + + QAction* restore_default = context_menu.addAction(tr("Restore Default")); + QAction* clear = context_menu.addAction(tr("Clear")); + + connect(restore_default, &QAction::triggered, [this, selected] { + const QKeySequence& default_key_sequence = QKeySequence::fromString( + Config::default_hotkeys[selected.row()].shortcut.first, QKeySequence::NativeText); + const auto [key_sequence_used, used_action] = IsUsedKey(default_key_sequence); + + if (key_sequence_used && + default_key_sequence != QKeySequence(model->data(selected).toString())) { + + QMessageBox::warning( + this, tr("Conflicting Key Sequence"), + tr("The default key sequence is already assigned to: %1").arg(used_action)); + } else { + model->setData(selected, default_key_sequence.toString(QKeySequence::NativeText)); + } + }); + connect(clear, &QAction::triggered, [this, selected] { model->setData(selected, QString{}); }); + + context_menu.exec(ui->hotkey_list->viewport()->mapToGlobal(menu_location)); +} diff --git a/src/yuzu/configuration/configure_hotkeys.h b/src/yuzu/configuration/configure_hotkeys.h index 8f8c6173b..a2ec3323e 100644 --- a/src/yuzu/configuration/configure_hotkeys.h +++ b/src/yuzu/configuration/configure_hotkeys.h @@ -35,7 +35,11 @@ private: void RetranslateUI(); void Configure(QModelIndex index); - bool IsUsedKey(QKeySequence key_sequence) const; + std::pair<bool, QString> IsUsedKey(QKeySequence key_sequence) const; + + void RestoreDefaults(); + void ClearAll(); + void PopupContextMenu(const QPoint& menu_location); std::unique_ptr<Ui::ConfigureHotkeys> ui; diff --git a/src/yuzu/configuration/configure_hotkeys.ui b/src/yuzu/configuration/configure_hotkeys.ui index 0d0b70f38..6d9f861e3 100644 --- a/src/yuzu/configuration/configure_hotkeys.ui +++ b/src/yuzu/configuration/configure_hotkeys.ui @@ -6,8 +6,8 @@ <rect> <x>0</x> <y>0</y> - <width>363</width> - <height>388</height> + <width>439</width> + <height>510</height> </rect> </property> <property name="windowTitle"> @@ -15,7 +15,7 @@ </property> <layout class="QVBoxLayout" name="verticalLayout"> <item> - <layout class="QVBoxLayout" name="verticalLayout_2"> + <layout class="QHBoxLayout" name="horizontalLayout"> <item> <widget class="QLabel" name="label_2"> <property name="text"> @@ -24,6 +24,37 @@ </widget> </item> <item> + <spacer name="horizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QPushButton" name="button_clear_all"> + <property name="text"> + <string>Clear All</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="button_restore_defaults"> + <property name="text"> + <string>Restore Defaults</string> + </property> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QVBoxLayout" name="verticalLayout_2"> + <item> <widget class="QTreeView" name="hotkey_list"> <property name="editTriggers"> <set>QAbstractItemView::NoEditTriggers</set> @@ -39,4 +70,4 @@ </widget> <resources/> <connections/> -</ui>
\ No newline at end of file +</ui> diff --git a/src/yuzu/configuration/configure_input.cpp b/src/yuzu/configuration/configure_input.cpp index f2977719c..d9009091b 100644 --- a/src/yuzu/configuration/configure_input.cpp +++ b/src/yuzu/configuration/configure_input.cpp @@ -8,18 +8,35 @@ #include <QSignalBlocker> #include <QTimer> -#include "configuration/configure_touchscreen_advanced.h" #include "core/core.h" #include "core/hle/service/am/am.h" #include "core/hle/service/am/applet_ae.h" #include "core/hle/service/am/applet_oe.h" -#include "core/hle/service/hid/controllers/npad.h" #include "core/hle/service/sm/sm.h" #include "ui_configure_input.h" +#include "ui_configure_input_advanced.h" #include "ui_configure_input_player.h" +#include "yuzu/configuration/configure_debug_controller.h" #include "yuzu/configuration/configure_input.h" +#include "yuzu/configuration/configure_input_advanced.h" #include "yuzu/configuration/configure_input_player.h" +#include "yuzu/configuration/configure_motion_touch.h" #include "yuzu/configuration/configure_mouse_advanced.h" +#include "yuzu/configuration/configure_touchscreen_advanced.h" +#include "yuzu/configuration/configure_vibration.h" +#include "yuzu/configuration/input_profiles.h" + +namespace { +template <typename Dialog, typename... Args> +void CallConfigureDialog(ConfigureInput& parent, Args&&... args) { + Dialog dialog(&parent, std::forward<Args>(args)...); + + const auto res = dialog.exec(); + if (res == QDialog::Accepted) { + dialog.ApplyConfiguration(); + } +} +} // Anonymous namespace void OnDockedModeChanged(bool last_state, bool new_state) { if (last_state == new_state) { @@ -48,97 +65,134 @@ void OnDockedModeChanged(bool last_state, bool new_state) { } } -namespace { -template <typename Dialog, typename... Args> -void CallConfigureDialog(ConfigureInput& parent, Args&&... args) { - parent.ApplyConfiguration(); - Dialog dialog(&parent, std::forward<Args>(args)...); - - const auto res = dialog.exec(); - if (res == QDialog::Accepted) { - dialog.ApplyConfiguration(); - } -} -} // Anonymous namespace - ConfigureInput::ConfigureInput(QWidget* parent) - : QDialog(parent), ui(std::make_unique<Ui::ConfigureInput>()) { + : QWidget(parent), ui(std::make_unique<Ui::ConfigureInput>()), + profiles(std::make_unique<InputProfiles>()) { ui->setupUi(this); +} + +ConfigureInput::~ConfigureInput() = default; - players_controller = { - ui->player1_combobox, ui->player2_combobox, ui->player3_combobox, ui->player4_combobox, - ui->player5_combobox, ui->player6_combobox, ui->player7_combobox, ui->player8_combobox, +void ConfigureInput::Initialize(InputCommon::InputSubsystem* input_subsystem, + std::size_t max_players) { + player_controllers = { + new ConfigureInputPlayer(this, 0, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 1, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 2, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 3, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 4, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 5, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 6, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 7, ui->consoleInputSettings, input_subsystem, + profiles.get()), }; - players_configure = { - ui->player1_configure, ui->player2_configure, ui->player3_configure, ui->player4_configure, - ui->player5_configure, ui->player6_configure, ui->player7_configure, ui->player8_configure, + player_tabs = { + ui->tabPlayer1, ui->tabPlayer2, ui->tabPlayer3, ui->tabPlayer4, + ui->tabPlayer5, ui->tabPlayer6, ui->tabPlayer7, ui->tabPlayer8, }; - RetranslateUI(); - LoadConfiguration(); - UpdateUIEnabled(); + player_connected = { + ui->checkboxPlayer1Connected, ui->checkboxPlayer2Connected, ui->checkboxPlayer3Connected, + ui->checkboxPlayer4Connected, ui->checkboxPlayer5Connected, ui->checkboxPlayer6Connected, + ui->checkboxPlayer7Connected, ui->checkboxPlayer8Connected, + }; - connect(ui->restore_defaults_button, &QPushButton::clicked, this, - &ConfigureInput::RestoreDefaults); + std::array<QLabel*, 8> player_connected_labels = { + ui->label, ui->label_3, ui->label_4, ui->label_5, + ui->label_6, ui->label_7, ui->label_8, ui->label_9, + }; - for (auto* enabled : players_controller) { - connect(enabled, QOverload<int>::of(&QComboBox::currentIndexChanged), this, - &ConfigureInput::UpdateUIEnabled); - } - connect(ui->use_docked_mode, &QCheckBox::stateChanged, this, &ConfigureInput::UpdateUIEnabled); - connect(ui->handheld_connected, &QCheckBox::stateChanged, this, - &ConfigureInput::UpdateUIEnabled); - connect(ui->mouse_enabled, &QCheckBox::stateChanged, this, &ConfigureInput::UpdateUIEnabled); - connect(ui->keyboard_enabled, &QCheckBox::stateChanged, this, &ConfigureInput::UpdateUIEnabled); - connect(ui->debug_enabled, &QCheckBox::stateChanged, this, &ConfigureInput::UpdateUIEnabled); - connect(ui->touchscreen_enabled, &QCheckBox::stateChanged, this, - &ConfigureInput::UpdateUIEnabled); - - for (std::size_t i = 0; i < players_configure.size(); ++i) { - connect(players_configure[i], &QPushButton::clicked, this, - [this, i] { CallConfigureDialog<ConfigureInputPlayer>(*this, i, false); }); + for (std::size_t i = 0; i < player_tabs.size(); ++i) { + player_tabs[i]->setLayout(new QHBoxLayout(player_tabs[i])); + player_tabs[i]->layout()->addWidget(player_controllers[i]); + connect(player_controllers[i], &ConfigureInputPlayer::Connected, [&, i](bool is_connected) { + if (is_connected) { + for (std::size_t index = 0; index <= i; ++index) { + player_connected[index]->setChecked(is_connected); + } + } else { + for (std::size_t index = i; index < player_tabs.size(); ++index) { + player_connected[index]->setChecked(is_connected); + } + } + }); + connect(player_controllers[i], &ConfigureInputPlayer::RefreshInputDevices, this, + &ConfigureInput::UpdateAllInputDevices); + connect(player_controllers[i], &ConfigureInputPlayer::RefreshInputProfiles, this, + &ConfigureInput::UpdateAllInputProfiles, Qt::QueuedConnection); + connect(player_connected[i], &QCheckBox::stateChanged, [this, i](int state) { + player_controllers[i]->ConnectPlayer(state == Qt::Checked); + }); + + // Remove/hide all the elements that exceed max_players, if applicable. + if (i >= max_players) { + ui->tabWidget->removeTab(static_cast<int>(max_players)); + player_connected[i]->hide(); + player_connected_labels[i]->hide(); + } } + // Only the first player can choose handheld mode so connect the signal just to player 1 + connect(player_controllers[0], &ConfigureInputPlayer::HandheldStateChanged, + [this](bool is_handheld) { UpdateDockedState(is_handheld); }); + + advanced = new ConfigureInputAdvanced(this); + ui->tabAdvanced->setLayout(new QHBoxLayout(ui->tabAdvanced)); + ui->tabAdvanced->layout()->addWidget(advanced); + connect(advanced, &ConfigureInputAdvanced::CallDebugControllerDialog, [this, input_subsystem] { + CallConfigureDialog<ConfigureDebugController>(*this, input_subsystem, profiles.get()); + }); + connect(advanced, &ConfigureInputAdvanced::CallMouseConfigDialog, [this, input_subsystem] { + CallConfigureDialog<ConfigureMouseAdvanced>(*this, input_subsystem); + }); + connect(advanced, &ConfigureInputAdvanced::CallTouchscreenConfigDialog, + [this] { CallConfigureDialog<ConfigureTouchscreenAdvanced>(*this); }); + connect(advanced, &ConfigureInputAdvanced::CallMotionTouchConfigDialog, + [this, input_subsystem] { + CallConfigureDialog<ConfigureMotionTouch>(*this, input_subsystem); + }); - connect(ui->handheld_configure, &QPushButton::clicked, this, - [this] { CallConfigureDialog<ConfigureInputPlayer>(*this, 8, false); }); + connect(ui->vibrationButton, &QPushButton::clicked, + [this] { CallConfigureDialog<ConfigureVibration>(*this); }); - connect(ui->debug_configure, &QPushButton::clicked, this, - [this] { CallConfigureDialog<ConfigureInputPlayer>(*this, 9, true); }); + connect(ui->motionButton, &QPushButton::clicked, [this, input_subsystem] { + CallConfigureDialog<ConfigureMotionTouch>(*this, input_subsystem); + }); - connect(ui->mouse_advanced, &QPushButton::clicked, this, - [this] { CallConfigureDialog<ConfigureMouseAdvanced>(*this); }); + connect(ui->buttonClearAll, &QPushButton::clicked, [this] { ClearAll(); }); + connect(ui->buttonRestoreDefaults, &QPushButton::clicked, [this] { RestoreDefaults(); }); - connect(ui->touchscreen_advanced, &QPushButton::clicked, this, - [this] { CallConfigureDialog<ConfigureTouchscreenAdvanced>(*this); }); + RetranslateUI(); + LoadConfiguration(); } -ConfigureInput::~ConfigureInput() = default; +QList<QWidget*> ConfigureInput::GetSubTabs() const { + return { + ui->tabPlayer1, ui->tabPlayer2, ui->tabPlayer3, ui->tabPlayer4, ui->tabPlayer5, + ui->tabPlayer6, ui->tabPlayer7, ui->tabPlayer8, ui->tabAdvanced, + }; +} void ConfigureInput::ApplyConfiguration() { - for (std::size_t i = 0; i < players_controller.size(); ++i) { - const auto controller_type_index = players_controller[i]->currentIndex(); + for (auto controller : player_controllers) { + controller->ApplyConfiguration(); + } - Settings::values.players[i].connected = controller_type_index != 0; + advanced->ApplyConfiguration(); - if (controller_type_index > 0) { - Settings::values.players[i].type = - static_cast<Settings::ControllerType>(controller_type_index - 1); - } else { - Settings::values.players[i].type = Settings::ControllerType::DualJoycon; - } - } + const bool pre_docked_mode = Settings::values.use_docked_mode.GetValue(); + Settings::values.use_docked_mode.SetValue(ui->radioDocked->isChecked()); + OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode.GetValue()); - const bool pre_docked_mode = Settings::values.use_docked_mode; - Settings::values.use_docked_mode = ui->use_docked_mode->isChecked(); - OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode); - Settings::values - .players[Service::HID::Controller_NPad::NPadIdToIndex(Service::HID::NPAD_HANDHELD)] - .connected = ui->handheld_connected->isChecked(); - Settings::values.debug_pad_enabled = ui->debug_enabled->isChecked(); - Settings::values.mouse_enabled = ui->mouse_enabled->isChecked(); - Settings::values.keyboard_enabled = ui->keyboard_enabled->isChecked(); - Settings::values.touchscreen.enabled = ui->touchscreen_enabled->isChecked(); + Settings::values.vibration_enabled.SetValue(ui->vibrationGroup->isChecked()); + Settings::values.motion_enabled.SetValue(ui->motionGroup->isChecked()); } void ConfigureInput::changeEvent(QEvent* event) { @@ -146,94 +200,74 @@ void ConfigureInput::changeEvent(QEvent* event) { RetranslateUI(); } - QDialog::changeEvent(event); + QWidget::changeEvent(event); } void ConfigureInput::RetranslateUI() { ui->retranslateUi(this); - RetranslateControllerComboBoxes(); } -void ConfigureInput::RetranslateControllerComboBoxes() { - for (auto* controller_box : players_controller) { - [[maybe_unused]] const QSignalBlocker blocker(controller_box); - - controller_box->clear(); - controller_box->addItems({tr("None"), tr("Pro Controller"), tr("Dual Joycons"), - tr("Single Right Joycon"), tr("Single Left Joycon")}); - } - +void ConfigureInput::LoadConfiguration() { LoadPlayerControllerIndices(); + UpdateDockedState(Settings::values.players.GetValue()[8].connected); + + ui->vibrationGroup->setChecked(Settings::values.vibration_enabled.GetValue()); + ui->motionGroup->setChecked(Settings::values.motion_enabled.GetValue()); } -void ConfigureInput::UpdateUIEnabled() { - bool hit_disabled = false; - for (auto* player : players_controller) { - player->setDisabled(hit_disabled); - if (hit_disabled) { - player->setCurrentIndex(0); - } - if (!hit_disabled && player->currentIndex() == 0) { - hit_disabled = true; - } +void ConfigureInput::LoadPlayerControllerIndices() { + for (std::size_t i = 0; i < player_connected.size(); ++i) { + const auto connected = Settings::values.players.GetValue()[i].connected || + (i == 0 && Settings::values.players.GetValue()[8].connected); + player_connected[i]->setChecked(connected); } +} - for (std::size_t i = 0; i < players_controller.size(); ++i) { - players_configure[i]->setEnabled(players_controller[i]->currentIndex() != 0); - } +void ConfigureInput::ClearAll() { + // We don't have a good way to know what tab is active, but we can find out by getting the + // parent of the consoleInputSettings + auto* player_tab = static_cast<ConfigureInputPlayer*>(ui->consoleInputSettings->parent()); + player_tab->ClearAll(); +} - ui->handheld_connected->setChecked(ui->handheld_connected->isChecked() && - !ui->use_docked_mode->isChecked()); - ui->handheld_connected->setEnabled(!ui->use_docked_mode->isChecked()); - ui->handheld_configure->setEnabled(ui->handheld_connected->isChecked() && - !ui->use_docked_mode->isChecked()); - ui->mouse_advanced->setEnabled(ui->mouse_enabled->isChecked()); - ui->debug_configure->setEnabled(ui->debug_enabled->isChecked()); - ui->touchscreen_advanced->setEnabled(ui->touchscreen_enabled->isChecked()); +void ConfigureInput::RestoreDefaults() { + // We don't have a good way to know what tab is active, but we can find out by getting the + // parent of the consoleInputSettings + auto* player_tab = static_cast<ConfigureInputPlayer*>(ui->consoleInputSettings->parent()); + player_tab->RestoreDefaults(); + + ui->radioDocked->setChecked(true); + ui->radioUndocked->setChecked(false); + ui->vibrationGroup->setChecked(true); + ui->motionGroup->setChecked(true); } -void ConfigureInput::LoadConfiguration() { - std::stable_partition( - Settings::values.players.begin(), - Settings::values.players.begin() + - Service::HID::Controller_NPad::NPadIdToIndex(Service::HID::NPAD_HANDHELD), - [](const auto& player) { return player.connected; }); +void ConfigureInput::UpdateDockedState(bool is_handheld) { + // Disallow changing the console mode if the controller type is handheld. + ui->radioDocked->setEnabled(!is_handheld); + ui->radioUndocked->setEnabled(!is_handheld); - LoadPlayerControllerIndices(); + ui->radioDocked->setChecked(Settings::values.use_docked_mode.GetValue()); + ui->radioUndocked->setChecked(!Settings::values.use_docked_mode.GetValue()); - ui->use_docked_mode->setChecked(Settings::values.use_docked_mode); - ui->handheld_connected->setChecked( - Settings::values - .players[Service::HID::Controller_NPad::NPadIdToIndex(Service::HID::NPAD_HANDHELD)] - .connected); - ui->debug_enabled->setChecked(Settings::values.debug_pad_enabled); - ui->mouse_enabled->setChecked(Settings::values.mouse_enabled); - ui->keyboard_enabled->setChecked(Settings::values.keyboard_enabled); - ui->touchscreen_enabled->setChecked(Settings::values.touchscreen.enabled); - - UpdateUIEnabled(); + // Also force into undocked mode if the controller type is handheld. + if (is_handheld) { + ui->radioUndocked->setChecked(true); + } } -void ConfigureInput::LoadPlayerControllerIndices() { - for (std::size_t i = 0; i < players_controller.size(); ++i) { - const auto connected = Settings::values.players[i].connected; - players_controller[i]->setCurrentIndex( - connected ? static_cast<u8>(Settings::values.players[i].type) + 1 : 0); +void ConfigureInput::UpdateAllInputDevices() { + for (const auto& player : player_controllers) { + player->UpdateInputDeviceCombobox(); } } -void ConfigureInput::RestoreDefaults() { - players_controller[0]->setCurrentIndex(2); +void ConfigureInput::UpdateAllInputProfiles(std::size_t player_index) { + for (std::size_t i = 0; i < player_controllers.size(); ++i) { + if (i == player_index) { + continue; + } - for (std::size_t i = 1; i < players_controller.size(); ++i) { - players_controller[i]->setCurrentIndex(0); + player_controllers[i]->UpdateInputProfiles(); } - - ui->use_docked_mode->setCheckState(Qt::Unchecked); - ui->handheld_connected->setCheckState(Qt::Unchecked); - ui->mouse_enabled->setCheckState(Qt::Unchecked); - ui->keyboard_enabled->setCheckState(Qt::Unchecked); - ui->debug_enabled->setCheckState(Qt::Unchecked); - ui->touchscreen_enabled->setCheckState(Qt::Checked); - UpdateUIEnabled(); } diff --git a/src/yuzu/configuration/configure_input.h b/src/yuzu/configuration/configure_input.h index 2f70cb3ca..f4eb0d78b 100644 --- a/src/yuzu/configuration/configure_input.h +++ b/src/yuzu/configuration/configure_input.h @@ -7,37 +7,52 @@ #include <array> #include <memory> -#include <QDialog> #include <QKeyEvent> +#include <QList> +#include <QWidget> -#include "ui_configure_input.h" - -class QPushButton; +class QCheckBox; class QString; class QTimer; +class ConfigureInputAdvanced; +class ConfigureInputPlayer; + +class InputProfiles; + +namespace InputCommon { +class InputSubsystem; +} + namespace Ui { class ConfigureInput; } void OnDockedModeChanged(bool last_state, bool new_state); -class ConfigureInput : public QDialog { +class ConfigureInput : public QWidget { Q_OBJECT public: explicit ConfigureInput(QWidget* parent = nullptr); ~ConfigureInput() override; - /// Save all button configurations to settings file + /// Initializes the input dialog with the given input subsystem. + void Initialize(InputCommon::InputSubsystem* input_subsystem_, std::size_t max_players = 8); + + /// Save all button configurations to settings file. void ApplyConfiguration(); + QList<QWidget*> GetSubTabs() const; + private: void changeEvent(QEvent* event) override; void RetranslateUI(); - void RetranslateControllerComboBoxes(); + void ClearAll(); - void UpdateUIEnabled(); + void UpdateDockedState(bool is_handheld); + void UpdateAllInputDevices(); + void UpdateAllInputProfiles(std::size_t player_index); /// Load configuration settings. void LoadConfiguration(); @@ -48,6 +63,10 @@ private: std::unique_ptr<Ui::ConfigureInput> ui; - std::array<QComboBox*, 8> players_controller; - std::array<QPushButton*, 8> players_configure; + std::unique_ptr<InputProfiles> profiles; + + std::array<ConfigureInputPlayer*, 8> player_controllers; + std::array<QWidget*, 8> player_tabs; + std::array<QCheckBox*, 8> player_connected; + ConfigureInputAdvanced* advanced; }; diff --git a/src/yuzu/configuration/configure_input.ui b/src/yuzu/configuration/configure_input.ui index efffd8487..2707025e7 100644 --- a/src/yuzu/configuration/configure_input.ui +++ b/src/yuzu/configuration/configure_input.ui @@ -1,529 +1,548 @@ <?xml version="1.0" encoding="UTF-8"?> <ui version="4.0"> <class>ConfigureInput</class> - <widget class="QDialog" name="ConfigureInput"> + <widget class="QWidget" name="ConfigureInput"> <property name="geometry"> <rect> <x>0</x> <y>0</y> - <width>384</width> - <height>576</height> + <width>680</width> + <height>540</height> </rect> </property> <property name="windowTitle"> - <string>Custom Input Settings</string> + <string>ConfigureInput</string> </property> <layout class="QVBoxLayout" name="verticalLayout_5"> + <property name="spacing"> + <number>2</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> <item> - <layout class="QVBoxLayout" name="verticalLayout"> - <item> - <widget class="QGroupBox" name="gridGroupBox_1"> - <property name="title"> - <string>Players</string> - </property> - <layout class="QGridLayout" name="gridLayout"> - <item row="1" column="2"> - <widget class="QComboBox" name="player1_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="1" column="3"> - <widget class="QPushButton" name="player1_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="0" column="2"> - <widget class="QLabel" name="label"> - <property name="text"> - <string>Controller Type</string> - </property> - <property name="alignment"> - <set>Qt::AlignCenter</set> - </property> - </widget> - </item> - <item row="2" column="2"> - <widget class="QComboBox" name="player2_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="3" column="2"> - <widget class="QComboBox" name="player3_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="4" column="2"> - <widget class="QComboBox" name="player4_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="5" column="2"> - <widget class="QComboBox" name="player5_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="6" column="2"> - <widget class="QComboBox" name="player6_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="7" column="2"> - <widget class="QComboBox" name="player7_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="8" column="2"> - <widget class="QComboBox" name="player8_combobox"> - <property name="minimumSize"> - <size> - <width>110</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="2" column="3"> - <widget class="QPushButton" name="player2_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="3" column="3"> - <widget class="QPushButton" name="player3_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="4" column="3"> - <widget class="QPushButton" name="player4_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="5" column="3"> - <widget class="QPushButton" name="player5_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="6" column="3"> - <widget class="QPushButton" name="player6_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="7" column="3"> - <widget class="QPushButton" name="player7_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="8" column="3"> - <widget class="QPushButton" name="player8_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="0" column="0"> - <spacer name="horizontalSpacer"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="0" column="4"> - <spacer name="horizontalSpacer_2"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="1"> - <widget class="QLabel" name="label_3"> - <property name="minimumSize"> - <size> - <width>55</width> - <height>0</height> - </size> - </property> - <property name="text"> - <string>Player 1</string> - </property> - </widget> - </item> - <item row="2" column="1"> - <widget class="QLabel" name="label_4"> - <property name="text"> - <string>Player 2</string> - </property> - </widget> - </item> - <item row="3" column="1"> - <widget class="QLabel" name="label_5"> - <property name="text"> - <string>Player 3</string> - </property> - </widget> - </item> - <item row="4" column="1"> - <widget class="QLabel" name="label_6"> - <property name="text"> - <string>Player 4</string> - </property> - </widget> - </item> - <item row="5" column="1"> - <widget class="QLabel" name="label_7"> - <property name="text"> - <string>Player 5</string> - </property> - </widget> - </item> - <item row="6" column="1"> - <widget class="QLabel" name="label_8"> - <property name="text"> - <string>Player 6</string> - </property> - </widget> - </item> - <item row="7" column="1"> - <widget class="QLabel" name="label_9"> - <property name="text"> - <string>Player 7</string> - </property> - </widget> - </item> - <item row="8" column="1"> - <widget class="QLabel" name="label_10"> - <property name="text"> - <string>Player 8</string> - </property> - </widget> - </item> - </layout> - </widget> - </item> - <item> - <widget class="QGroupBox" name="gridGroupBox_2"> - <property name="title"> - <string>Handheld</string> - </property> - <layout class="QGridLayout" name="gridLayout_2"> - <item row="1" column="2"> - <spacer name="horizontalSpacer_5"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeType"> - <enum>QSizePolicy::Fixed</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>72</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="4"> - <spacer name="horizontalSpacer_4"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="3"> - <widget class="QPushButton" name="handheld_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="1" column="0"> - <spacer name="horizontalSpacer_3"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="1"> - <widget class="QCheckBox" name="handheld_connected"> - <property name="text"> - <string>Joycons Docked</string> - </property> - </widget> - </item> - <item row="0" column="1"> - <widget class="QCheckBox" name="use_docked_mode"> - <property name="text"> - <string>Use Docked Mode</string> - </property> - </widget> - </item> - </layout> - </widget> - </item> - <item> - <widget class="QGroupBox" name="gridGroupBox_3"> - <property name="title"> - <string>Other</string> - </property> - <layout class="QGridLayout" name="gridLayout_3"> - <item row="1" column="1"> - <widget class="QCheckBox" name="keyboard_enabled"> - <property name="minimumSize"> - <size> - <width>0</width> - <height>23</height> - </size> - </property> - <property name="text"> - <string>Keyboard</string> - </property> - </widget> - </item> - <item row="2" column="1"> - <widget class="QCheckBox" name="debug_enabled"> - <property name="text"> - <string>Debug Controller</string> - </property> - </widget> - </item> - <item row="3" column="1"> - <widget class="QCheckBox" name="touchscreen_enabled"> - <property name="text"> - <string>Touchscreen</string> - </property> - </widget> - </item> - <item row="0" column="1"> - <widget class="QCheckBox" name="mouse_enabled"> - <property name="minimumSize"> - <size> - <width>0</width> - <height>23</height> - </size> - </property> - <property name="text"> - <string>Mouse</string> - </property> - </widget> - </item> - <item row="0" column="4"> - <spacer name="horizontalSpacer_7"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="0" column="2"> - <spacer name="horizontalSpacer_8"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeType"> - <enum>QSizePolicy::Fixed</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>76</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="0" column="0"> - <spacer name="horizontalSpacer_6"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="3" column="3"> - <widget class="QPushButton" name="touchscreen_advanced"> - <property name="text"> - <string>Advanced</string> - </property> - </widget> - </item> - <item row="2" column="3"> - <widget class="QPushButton" name="debug_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="0" column="3"> - <widget class="QPushButton" name="mouse_advanced"> - <property name="text"> - <string>Advanced</string> - </property> - </widget> - </item> - </layout> - </widget> - </item> - <item> - <spacer name="verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout"> - <item> - <widget class="QPushButton" name="restore_defaults_button"> - <property name="text"> - <string>Restore Defaults</string> + <widget class="QTabWidget" name="tabWidget"> + <property name="currentIndex"> + <number>0</number> + </property> + <widget class="QWidget" name="tabPlayer1"> + <property name="accessibleName"> + <string>Player 1</string> + </property> + <attribute name="title"> + <string>Player 1</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer2"> + <property name="accessibleName"> + <string>Player 2</string> + </property> + <attribute name="title"> + <string>Player 2</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer3"> + <property name="accessibleName"> + <string>Player 3</string> + </property> + <attribute name="title"> + <string>Player 3</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer4"> + <property name="accessibleName"> + <string>Player 4</string> + </property> + <attribute name="title"> + <string>Player 4</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer5"> + <property name="accessibleName"> + <string>Player 5</string> + </property> + <attribute name="title"> + <string>Player 5</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer6"> + <property name="accessibleName"> + <string>Player 6</string> + </property> + <attribute name="title"> + <string>Player 6</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer7"> + <property name="accessibleName"> + <string>Player 7</string> + </property> + <attribute name="title"> + <string>Player 7</string> + </attribute> + </widget> + <widget class="QWidget" name="tabPlayer8"> + <property name="accessibleName"> + <string>Player 8</string> + </property> + <attribute name="title"> + <string>Player 8</string> + </attribute> + </widget> + <widget class="QWidget" name="tabAdvanced"> + <property name="accessibleName"> + <string>Advanced</string> + </property> + <attribute name="title"> + <string>Advanced</string> + </attribute> + </widget> + </widget> + </item> + <item alignment="Qt::AlignVCenter"> + <widget class="QWidget" name="consoleInputSettings" native="true"> + <layout class="QHBoxLayout" name="buttonsBottomRightHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignVCenter"> + <widget class="QGroupBox" name="handheldGroup"> + <property name="maximumSize"> + <size> + <width>16777215</width> + <height>16777215</height> + </size> + </property> + <property name="title"> + <string>Console Mode</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>8</number> + </property> + <property name="topMargin"> + <number>6</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QRadioButton" name="radioDocked"> + <property name="text"> + <string>Docked</string> + </property> + <property name="checked"> + <bool>true</bool> + </property> + </widget> + </item> + <item> + <widget class="QRadioButton" name="radioUndocked"> + <property name="text"> + <string>Undocked</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroup"> + <property name="title"> + <string>Vibration</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="vibrationButton"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="motionGroup"> + <property name="title"> + <string>Motion</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="motionButton"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignVCenter"> + <widget class="QWidget" name="connectedControllers" native="true"> + <layout class="QGridLayout" name="gridLayout_2"> + <property name="leftMargin"> + <number>5</number> + </property> + <property name="topMargin"> + <number>0</number> </property> - </widget> - </item> - <item> - <spacer name="horizontalSpacer_9"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> + <property name="rightMargin"> + <number>0</number> </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> + <property name="bottomMargin"> + <number>0</number> </property> - </spacer> - </item> - <item> - <widget class="QDialogButtonBox" name="buttonBox"> - <property name="standardButtons"> - <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + <property name="spacing"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - </layout> + <item row="1" column="2"> + <widget class="QCheckBox" name="checkboxPlayer2Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="0"> + <widget class="QLabel" name="label_2"> + <property name="text"> + <string>Controllers</string> + </property> + </widget> + </item> + <item row="1" column="4"> + <widget class="QCheckBox" name="checkboxPlayer4Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="3"> + <widget class="QCheckBox" name="checkboxPlayer3Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="5"> + <widget class="QCheckBox" name="checkboxPlayer5Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="0" column="1"> + <widget class="QLabel" name="label"> + <property name="text"> + <string>1</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="1" column="7"> + <widget class="QCheckBox" name="checkboxPlayer7Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="6"> + <widget class="QCheckBox" name="checkboxPlayer6Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="1" column="1"> + <widget class="QCheckBox" name="checkboxPlayer1Connected"> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + <property name="checked"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="1" column="8"> + <widget class="QCheckBox" name="checkboxPlayer8Connected"> + <property name="text"> + <string/> + </property> + </widget> + </item> + <item row="0" column="2"> + <widget class="QLabel" name="label_3"> + <property name="text"> + <string>2</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="3"> + <widget class="QLabel" name="label_4"> + <property name="text"> + <string>3</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="4"> + <widget class="QLabel" name="label_5"> + <property name="text"> + <string>4</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="5"> + <widget class="QLabel" name="label_6"> + <property name="text"> + <string>5</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="6"> + <widget class="QLabel" name="label_7"> + <property name="text"> + <string>6</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="7"> + <widget class="QLabel" name="label_8"> + <property name="text"> + <string>7</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="8"> + <widget class="QLabel" name="label_9"> + <property name="text"> + <string>8</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + </widget> + </item> + <item row="0" column="0"> + <widget class="QLabel" name="label_10"> + <property name="text"> + <string>Connected</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignBottom"> + <widget class="QPushButton" name="buttonRestoreDefaults"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="sizeIncrement"> + <size> + <width>0</width> + <height>0</height> + </size> + </property> + <property name="baseSize"> + <size> + <width>0</width> + <height>0</height> + </size> + </property> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Defaults</string> + </property> + </widget> + </item> + <item alignment="Qt::AlignBottom"> + <widget class="QPushButton" name="buttonClearAll"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="sizeIncrement"> + <size> + <width>0</width> + <height>0</height> + </size> + </property> + <property name="baseSize"> + <size> + <width>0</width> + <height>0</height> + </size> + </property> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Clear</string> + </property> + </widget> + </item> + </layout> + </widget> </item> </layout> </widget> <resources/> - <connections> - <connection> - <sender>buttonBox</sender> - <signal>accepted()</signal> - <receiver>ConfigureInput</receiver> - <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>294</x> - <y>553</y> - </hint> - <hint type="destinationlabel"> - <x>191</x> - <y>287</y> - </hint> - </hints> - </connection> - <connection> - <sender>buttonBox</sender> - <signal>rejected()</signal> - <receiver>ConfigureInput</receiver> - <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>294</x> - <y>553</y> - </hint> - <hint type="destinationlabel"> - <x>191</x> - <y>287</y> - </hint> - </hints> - </connection> - </connections> + <connections/> </ui> diff --git a/src/yuzu/configuration/configure_input_advanced.cpp b/src/yuzu/configuration/configure_input_advanced.cpp new file mode 100644 index 000000000..abaf03630 --- /dev/null +++ b/src/yuzu/configuration/configure_input_advanced.cpp @@ -0,0 +1,171 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QColorDialog> +#include "core/core.h" +#include "core/settings.h" +#include "ui_configure_input_advanced.h" +#include "yuzu/configuration/configure_input_advanced.h" + +ConfigureInputAdvanced::ConfigureInputAdvanced(QWidget* parent) + : QWidget(parent), ui(std::make_unique<Ui::ConfigureInputAdvanced>()) { + ui->setupUi(this); + + controllers_color_buttons = {{ + { + ui->player1_left_body_button, + ui->player1_left_buttons_button, + ui->player1_right_body_button, + ui->player1_right_buttons_button, + }, + { + ui->player2_left_body_button, + ui->player2_left_buttons_button, + ui->player2_right_body_button, + ui->player2_right_buttons_button, + }, + { + ui->player3_left_body_button, + ui->player3_left_buttons_button, + ui->player3_right_body_button, + ui->player3_right_buttons_button, + }, + { + ui->player4_left_body_button, + ui->player4_left_buttons_button, + ui->player4_right_body_button, + ui->player4_right_buttons_button, + }, + { + ui->player5_left_body_button, + ui->player5_left_buttons_button, + ui->player5_right_body_button, + ui->player5_right_buttons_button, + }, + { + ui->player6_left_body_button, + ui->player6_left_buttons_button, + ui->player6_right_body_button, + ui->player6_right_buttons_button, + }, + { + ui->player7_left_body_button, + ui->player7_left_buttons_button, + ui->player7_right_body_button, + ui->player7_right_buttons_button, + }, + { + ui->player8_left_body_button, + ui->player8_left_buttons_button, + ui->player8_right_body_button, + ui->player8_right_buttons_button, + }, + }}; + + for (std::size_t player_idx = 0; player_idx < controllers_color_buttons.size(); ++player_idx) { + auto& color_buttons = controllers_color_buttons[player_idx]; + for (std::size_t button_idx = 0; button_idx < color_buttons.size(); ++button_idx) { + connect(color_buttons[button_idx], &QPushButton::clicked, this, + [this, player_idx, button_idx] { + OnControllerButtonClick(player_idx, button_idx); + }); + } + } + + connect(ui->mouse_enabled, &QCheckBox::stateChanged, this, + &ConfigureInputAdvanced::UpdateUIEnabled); + connect(ui->debug_enabled, &QCheckBox::stateChanged, this, + &ConfigureInputAdvanced::UpdateUIEnabled); + connect(ui->touchscreen_enabled, &QCheckBox::stateChanged, this, + &ConfigureInputAdvanced::UpdateUIEnabled); + + connect(ui->debug_configure, &QPushButton::clicked, this, + [this] { CallDebugControllerDialog(); }); + connect(ui->mouse_advanced, &QPushButton::clicked, this, [this] { CallMouseConfigDialog(); }); + connect(ui->touchscreen_advanced, &QPushButton::clicked, this, + [this] { CallTouchscreenConfigDialog(); }); + connect(ui->buttonMotionTouch, &QPushButton::clicked, this, + &ConfigureInputAdvanced::CallMotionTouchConfigDialog); + + LoadConfiguration(); +} + +ConfigureInputAdvanced::~ConfigureInputAdvanced() = default; + +void ConfigureInputAdvanced::OnControllerButtonClick(std::size_t player_idx, + std::size_t button_idx) { + const QColor new_bg_color = QColorDialog::getColor(controllers_colors[player_idx][button_idx]); + if (!new_bg_color.isValid()) { + return; + } + controllers_colors[player_idx][button_idx] = new_bg_color; + controllers_color_buttons[player_idx][button_idx]->setStyleSheet( + QStringLiteral("background-color: %1; min-width: 60px;") + .arg(controllers_colors[player_idx][button_idx].name())); +} + +void ConfigureInputAdvanced::ApplyConfiguration() { + for (std::size_t player_idx = 0; player_idx < controllers_color_buttons.size(); ++player_idx) { + auto& player = Settings::values.players.GetValue()[player_idx]; + std::array<u32, 4> colors{}; + std::transform(controllers_colors[player_idx].begin(), controllers_colors[player_idx].end(), + colors.begin(), [](QColor color) { return color.rgb(); }); + + player.body_color_left = colors[0]; + player.button_color_left = colors[1]; + player.body_color_right = colors[2]; + player.button_color_right = colors[3]; + } + + Settings::values.debug_pad_enabled = ui->debug_enabled->isChecked(); + Settings::values.mouse_enabled = ui->mouse_enabled->isChecked(); + Settings::values.keyboard_enabled = ui->keyboard_enabled->isChecked(); + Settings::values.touchscreen.enabled = ui->touchscreen_enabled->isChecked(); +} + +void ConfigureInputAdvanced::LoadConfiguration() { + for (std::size_t player_idx = 0; player_idx < controllers_color_buttons.size(); ++player_idx) { + auto& player = Settings::values.players.GetValue()[player_idx]; + std::array<u32, 4> colors = { + player.body_color_left, + player.button_color_left, + player.body_color_right, + player.button_color_right, + }; + + std::transform(colors.begin(), colors.end(), controllers_colors[player_idx].begin(), + [](u32 rgb) { return QColor::fromRgb(rgb); }); + + for (std::size_t button_idx = 0; button_idx < colors.size(); ++button_idx) { + controllers_color_buttons[player_idx][button_idx]->setStyleSheet( + QStringLiteral("background-color: %1; min-width: 60px;") + .arg(controllers_colors[player_idx][button_idx].name())); + } + } + + ui->debug_enabled->setChecked(Settings::values.debug_pad_enabled); + ui->mouse_enabled->setChecked(Settings::values.mouse_enabled); + ui->keyboard_enabled->setChecked(Settings::values.keyboard_enabled); + ui->touchscreen_enabled->setChecked(Settings::values.touchscreen.enabled); + + UpdateUIEnabled(); +} + +void ConfigureInputAdvanced::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QWidget::changeEvent(event); +} + +void ConfigureInputAdvanced::RetranslateUI() { + ui->retranslateUi(this); +} + +void ConfigureInputAdvanced::UpdateUIEnabled() { + ui->mouse_advanced->setEnabled(ui->mouse_enabled->isChecked()); + ui->debug_configure->setEnabled(ui->debug_enabled->isChecked()); + ui->touchscreen_advanced->setEnabled(ui->touchscreen_enabled->isChecked()); +} diff --git a/src/yuzu/configuration/configure_input_advanced.h b/src/yuzu/configuration/configure_input_advanced.h new file mode 100644 index 000000000..3083d55c1 --- /dev/null +++ b/src/yuzu/configuration/configure_input_advanced.h @@ -0,0 +1,46 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <memory> +#include <QWidget> + +class QColor; +class QPushButton; + +namespace Ui { +class ConfigureInputAdvanced; +} + +class ConfigureInputAdvanced : public QWidget { + Q_OBJECT + +public: + explicit ConfigureInputAdvanced(QWidget* parent = nullptr); + ~ConfigureInputAdvanced() override; + + void ApplyConfiguration(); + +signals: + void CallDebugControllerDialog(); + void CallMouseConfigDialog(); + void CallTouchscreenConfigDialog(); + void CallMotionTouchConfigDialog(); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + void UpdateUIEnabled(); + + void OnControllerButtonClick(std::size_t player_idx, std::size_t button_idx); + + void LoadConfiguration(); + + std::unique_ptr<Ui::ConfigureInputAdvanced> ui; + + std::array<std::array<QColor, 4>, 8> controllers_colors; + std::array<std::array<QPushButton*, 4>, 8> controllers_color_buttons; +}; diff --git a/src/yuzu/configuration/configure_input_advanced.ui b/src/yuzu/configuration/configure_input_advanced.ui new file mode 100644 index 000000000..a880a7c68 --- /dev/null +++ b/src/yuzu/configuration/configure_input_advanced.ui @@ -0,0 +1,2688 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureInputAdvanced</class> + <widget class="QWidget" name="ConfigureInputAdvanced"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>710</width> + <height>580</height> + </rect> + </property> + <property name="windowTitle"> + <string>Configure Input</string> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QHBoxLayout" name="horizontalLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="mainInputAdvanced" native="true"> + <layout class="QHBoxLayout" name="main" stretch="1,1"> + <property name="spacing"> + <number>9</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="leftInputAdvanced" native="true"> + <layout class="QVBoxLayout" name="leftLayout" stretch="0"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="joyconColorsGroup"> + <property name="title"> + <string>Joycon Colors</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_3" stretch="1,1"> + <property name="leftMargin"> + <number>9</number> + </property> + <property name="topMargin"> + <number>9</number> + </property> + <property name="rightMargin"> + <number>9</number> + </property> + <property name="bottomMargin"> + <number>9</number> + </property> + <item> + <widget class="QWidget" name="topLeftInputAdvanced" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_4"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="player12Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_4"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="player1Group"> + <property name="title"> + <string>Player 1</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player1LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_14"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player1LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_66"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player1_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player1LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_67"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player1_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player1RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_14"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player1RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_64"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player1_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player1RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_65"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player1_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="player2Group"> + <property name="title"> + <string>Player 2</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_14"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player2LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_15"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player2LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_70"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player2_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player2LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_71"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player2_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player2RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_15"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player2RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_68"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player2_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player2RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_69"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player2_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player34Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_5"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="player3Group"> + <property name="title"> + <string>Player 3</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_15"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player3LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_16"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player3LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_74"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player3_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player3LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_75"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player3_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player3RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_16"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player3RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_72"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player3_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player3RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_73"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player3_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="player4Group"> + <property name="title"> + <string>Player 4</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_16"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player4LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_17"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player4LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_78"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player4_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player4LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_79"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player4_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player4RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_17"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player4RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_76"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player4_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player4RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_77"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player4_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="bottomLeftInputAdvanced" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_5"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="player56Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_6"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="player5Group"> + <property name="title"> + <string>Player 5</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_10"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player5LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_10"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player5LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_50"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player5_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player5LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_51"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player5_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player5RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_10"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player5RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_48"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player5_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player5RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_49"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player5_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="player6Group"> + <property name="title"> + <string>Player 6</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_11"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player6LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_11"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player6LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_54"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player6_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player6LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_55"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player6_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player6RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_11"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player6RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_52"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player6_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player6RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_53"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player6_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player78Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_7"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="player7Group"> + <property name="title"> + <string>Player 7</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_12"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player7LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_12"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player7LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_58"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player7_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player7LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_59"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player7_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player7RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_12"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player7RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_56"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player7_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player7RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_57"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player7_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="player8Group"> + <property name="title"> + <string>Player 8</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_13"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>6</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>6</number> + </property> + <property name="bottomMargin"> + <number>6</number> + </property> + <item> + <widget class="QWidget" name="player8LeftJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsLeftJoyconVerticalLayout_13"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player8LeftBodyGroup"> + <property name="title"> + <string>L Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_62"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player8_left_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player8LeftButtonsGroup"> + <property name="title"> + <string>L Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_63"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player8_left_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player8RightJoycon" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsRightJoyconVerticalLayout_13"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player8RightBodyGroup"> + <property name="title"> + <string>R Body</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_60"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player8_right_body_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="player8RightButtonsGroup"> + <property name="title"> + <string>R Button</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_61"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="player8_right_buttons_button"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="rightInputAdvanced" native="true"> + <layout class="QVBoxLayout" name="rightLayout" stretch="1,1"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="topRightInputAdvanced" native="true"> + <layout class="QVBoxLayout" name="verticalLayout"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="gridGroupBox_3"> + <property name="title"> + <string>Other</string> + </property> + <layout class="QGridLayout" name="gridLayout_3"> + <item row="0" column="0"> + <widget class="QCheckBox" name="keyboard_enabled"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>23</height> + </size> + </property> + <property name="text"> + <string>Keyboard</string> + </property> + </widget> + </item> + <item row="4" column="2"> + <widget class="QPushButton" name="touchscreen_advanced"> + <property name="text"> + <string>Advanced</string> + </property> + </widget> + </item> + <item row="1" column="1"> + <spacer name="horizontalSpacer_8"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeType"> + <enum>QSizePolicy::Fixed</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>76</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item row="1" column="2"> + <widget class="QPushButton" name="mouse_advanced"> + <property name="text"> + <string>Advanced</string> + </property> + </widget> + </item> + <item row="4" column="0"> + <widget class="QCheckBox" name="touchscreen_enabled"> + <property name="text"> + <string>Touchscreen</string> + </property> + </widget> + </item> + <item row="1" column="0"> + <widget class="QCheckBox" name="mouse_enabled"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>23</height> + </size> + </property> + <property name="text"> + <string>Mouse</string> + </property> + </widget> + </item> + <item row="6" column="0"> + <widget class="QLabel" name="motion_touch"> + <property name="text"> + <string>Motion / Touch</string> + </property> + </widget> + </item> + <item row="6" column="2"> + <widget class="QPushButton" name="buttonMotionTouch"> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + <item row="5" column="0"> + <widget class="QCheckBox" name="debug_enabled"> + <property name="text"> + <string>Debug Controller</string> + </property> + </widget> + </item> + <item row="5" column="2"> + <widget class="QPushButton" name="debug_configure"> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="bottomRightInputAdvanced" native="true"> + <layout class="QVBoxLayout" name="verticalLayout_2"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="mainVerticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + <resources> + </resources> + <connections/> +</ui> diff --git a/src/yuzu/configuration/configure_input_player.cpp b/src/yuzu/configuration/configure_input_player.cpp index 15ac30f12..56ab32a35 100644 --- a/src/yuzu/configuration/configure_input_player.cpp +++ b/src/yuzu/configuration/configure_input_player.cpp @@ -4,19 +4,28 @@ #include <algorithm> #include <memory> +#include <thread> #include <utility> -#include <QColorDialog> #include <QGridLayout> +#include <QInputDialog> #include <QKeyEvent> #include <QMenu> #include <QMessageBox> #include <QTimer> -#include "common/assert.h" #include "common/param_package.h" +#include "core/core.h" +#include "core/hle/service/hid/controllers/npad.h" +#include "core/hle/service/hid/hid.h" +#include "core/hle/service/sm/sm.h" +#include "input_common/gcadapter/gc_poller.h" #include "input_common/main.h" +#include "input_common/udp/udp.h" #include "ui_configure_input_player.h" #include "yuzu/configuration/config.h" #include "yuzu/configuration/configure_input_player.h" +#include "yuzu/configuration/configure_vibration.h" +#include "yuzu/configuration/input_profiles.h" +#include "yuzu/util/limitable_input_dialog.h" const std::array<std::string, ConfigureInputPlayer::ANALOG_SUB_BUTTONS_NUM> ConfigureInputPlayer::analog_sub_buttons{{ @@ -24,20 +33,74 @@ const std::array<std::string, ConfigureInputPlayer::ANALOG_SUB_BUTTONS_NUM> "down", "left", "right", - "modifier", }}; -static void LayerGridElements(QGridLayout* grid, QWidget* item, QWidget* onTopOf) { - const int index1 = grid->indexOf(item); - const int index2 = grid->indexOf(onTopOf); - int row, column, rowSpan, columnSpan; - grid->getItemPosition(index2, &row, &column, &rowSpan, &columnSpan); - grid->takeAt(index1); - grid->addWidget(item, row, column, rowSpan, columnSpan); +namespace { + +constexpr std::size_t HANDHELD_INDEX = 8; + +void UpdateController(Settings::ControllerType controller_type, std::size_t npad_index, + bool connected) { + Core::System& system{Core::System::GetInstance()}; + if (!system.IsPoweredOn()) { + return; + } + Service::SM::ServiceManager& sm = system.ServiceManager(); + + auto& npad = + sm.GetService<Service::HID::Hid>("hid") + ->GetAppletResource() + ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad); + + npad.UpdateControllerAt(npad.MapSettingsTypeToNPad(controller_type), npad_index, connected); +} + +/// Maps the controller type combobox index to Controller Type enum +constexpr Settings::ControllerType GetControllerTypeFromIndex(int index) { + switch (index) { + case 0: + default: + return Settings::ControllerType::ProController; + case 1: + return Settings::ControllerType::DualJoyconDetached; + case 2: + return Settings::ControllerType::LeftJoycon; + case 3: + return Settings::ControllerType::RightJoycon; + case 4: + return Settings::ControllerType::Handheld; + } +} + +/// Maps the Controller Type enum to controller type combobox index +constexpr int GetIndexFromControllerType(Settings::ControllerType type) { + switch (type) { + case Settings::ControllerType::ProController: + default: + return 0; + case Settings::ControllerType::DualJoyconDetached: + return 1; + case Settings::ControllerType::LeftJoycon: + return 2; + case Settings::ControllerType::RightJoycon: + return 3; + case Settings::ControllerType::Handheld: + return 4; + } } -static QString GetKeyName(int key_code) { +QString GetKeyName(int key_code) { switch (key_code) { + case Qt::LeftButton: + return QObject::tr("Click 0"); + case Qt::RightButton: + return QObject::tr("Click 1"); + case Qt::MiddleButton: + return QObject::tr("Click 2"); + case Qt::BackButton: + return QObject::tr("Click 3"); + case Qt::ForwardButton: + return QObject::tr("Click 4"); case Qt::Key_Shift: return QObject::tr("Shift"); case Qt::Key_Control: @@ -51,18 +114,24 @@ static QString GetKeyName(int key_code) { } } -static void SetAnalogButton(const Common::ParamPackage& input_param, - Common::ParamPackage& analog_param, const std::string& button_name) { - if (analog_param.Get("engine", "") != "analog_from_button") { +void SetAnalogParam(const Common::ParamPackage& input_param, Common::ParamPackage& analog_param, + const std::string& button_name) { + // The poller returned a complete axis, so set all the buttons + if (input_param.Has("axis_x") && input_param.Has("axis_y")) { + analog_param = input_param; + return; + } + // Check if the current configuration has either no engine or an axis binding. + // Clears out the old binding and adds one with analog_from_button. + if (!analog_param.Has("engine") || analog_param.Has("axis_x") || analog_param.Has("axis_y")) { analog_param = { {"engine", "analog_from_button"}, - {"modifier_scale", "0.5"}, }; } analog_param.Set(button_name, input_param.Serialize()); } -static QString ButtonToText(const Common::ParamPackage& param) { +QString ButtonToText(const Common::ParamPackage& param) { if (!param.Has("engine")) { return QObject::tr("[not set]"); } @@ -71,6 +140,28 @@ static QString ButtonToText(const Common::ParamPackage& param) { return GetKeyName(param.Get("code", 0)); } + if (param.Get("engine", "") == "gcpad") { + if (param.Has("axis")) { + const QString axis_str = QString::fromStdString(param.Get("axis", "")); + const QString direction_str = QString::fromStdString(param.Get("direction", "")); + + return QObject::tr("GC Axis %1%2").arg(axis_str, direction_str); + } + if (param.Has("button")) { + const QString button_str = QString::number(int(std::log2(param.Get("button", 0)))); + return QObject::tr("GC Button %1").arg(button_str); + } + return GetKeyName(param.Get("code", 0)); + } + + if (param.Get("engine", "") == "cemuhookudp") { + if (param.Has("pad_index")) { + const QString motion_str = QString::fromStdString(param.Get("pad_index", "")); + return QObject::tr("Motion %1").arg(motion_str); + } + return GetKeyName(param.Get("code", 0)); + } + if (param.Get("engine", "") == "sdl") { if (param.Has("hat")) { const QString hat_str = QString::fromStdString(param.Get("hat", "")); @@ -98,7 +189,7 @@ static QString ButtonToText(const Common::ParamPackage& param) { return QObject::tr("[unknown]"); } -static QString AnalogToText(const Common::ParamPackage& param, const std::string& dir) { +QString AnalogToText(const Common::ParamPackage& param, const std::string& dir) { if (!param.Has("engine")) { return QObject::tr("[not set]"); } @@ -127,24 +218,47 @@ static QString AnalogToText(const Common::ParamPackage& param, const std::string return {}; } + if (param.Get("engine", "") == "gcpad") { + if (dir == "modifier") { + return QObject::tr("[unused]"); + } + + if (dir == "left" || dir == "right") { + const QString axis_x_str = QString::fromStdString(param.Get("axis_x", "")); + + return QObject::tr("GC Axis %1").arg(axis_x_str); + } + + if (dir == "up" || dir == "down") { + const QString axis_y_str = QString::fromStdString(param.Get("axis_y", "")); + + return QObject::tr("GC Axis %1").arg(axis_y_str); + } + + return {}; + } return QObject::tr("[unknown]"); } - -ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_index, bool debug) - : QDialog(parent), ui(std::make_unique<Ui::ConfigureInputPlayer>()), player_index(player_index), - debug(debug), timeout_timer(std::make_unique<QTimer>()), - poll_timer(std::make_unique<QTimer>()) { +} // namespace + +ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_index, + QWidget* bottom_row, + InputCommon::InputSubsystem* input_subsystem_, + InputProfiles* profiles_, bool debug) + : QWidget(parent), ui(std::make_unique<Ui::ConfigureInputPlayer>()), player_index(player_index), + debug(debug), input_subsystem{input_subsystem_}, profiles(profiles_), + timeout_timer(std::make_unique<QTimer>()), poll_timer(std::make_unique<QTimer>()), + bottom_row(bottom_row) { ui->setupUi(this); + setFocusPolicy(Qt::ClickFocus); button_map = { - ui->buttonA, ui->buttonB, ui->buttonX, ui->buttonY, - ui->buttonLStick, ui->buttonRStick, ui->buttonL, ui->buttonR, - ui->buttonZL, ui->buttonZR, ui->buttonPlus, ui->buttonMinus, - ui->buttonDpadLeft, ui->buttonDpadUp, ui->buttonDpadRight, ui->buttonDpadDown, - ui->buttonLStickLeft, ui->buttonLStickUp, ui->buttonLStickRight, ui->buttonLStickDown, - ui->buttonRStickLeft, ui->buttonRStickUp, ui->buttonRStickRight, ui->buttonRStickDown, - ui->buttonSL, ui->buttonSR, ui->buttonHome, ui->buttonScreenshot, + ui->buttonA, ui->buttonB, ui->buttonX, ui->buttonY, + ui->buttonLStick, ui->buttonRStick, ui->buttonL, ui->buttonR, + ui->buttonZL, ui->buttonZR, ui->buttonPlus, ui->buttonMinus, + ui->buttonDpadLeft, ui->buttonDpadUp, ui->buttonDpadRight, ui->buttonDpadDown, + ui->buttonSL, ui->buttonSR, ui->buttonHome, ui->buttonScreenshot, }; analog_map_buttons = {{ @@ -153,219 +267,292 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i ui->buttonLStickDown, ui->buttonLStickLeft, ui->buttonLStickRight, - ui->buttonLStickMod, }, { ui->buttonRStickUp, ui->buttonRStickDown, ui->buttonRStickLeft, ui->buttonRStickRight, - ui->buttonRStickMod, }, }}; - debug_hidden = { - ui->buttonSL, ui->labelSL, - ui->buttonSR, ui->labelSR, - ui->buttonLStick, ui->labelLStickPressed, - ui->buttonRStick, ui->labelRStickPressed, - ui->buttonHome, ui->labelHome, - ui->buttonScreenshot, ui->labelScreenshot, + motion_map = { + ui->buttonMotionLeft, + ui->buttonMotionRight, }; - auto layout = Settings::values.players[player_index].type; - if (debug) - layout = Settings::ControllerType::DualJoycon; - - switch (layout) { - case Settings::ControllerType::ProController: - case Settings::ControllerType::DualJoycon: - layout_hidden = { - ui->buttonSL, - ui->labelSL, - ui->buttonSR, - ui->labelSR, - }; - break; - case Settings::ControllerType::LeftJoycon: - layout_hidden = { - ui->right_body_button, - ui->right_buttons_button, - ui->right_body_label, - ui->right_buttons_label, - ui->buttonR, - ui->labelR, - ui->buttonZR, - ui->labelZR, - ui->labelHome, - ui->buttonHome, - ui->buttonPlus, - ui->labelPlus, - ui->RStick, - ui->faceButtons, - }; - break; - case Settings::ControllerType::RightJoycon: - layout_hidden = { - ui->left_body_button, ui->left_buttons_button, - ui->left_body_label, ui->left_buttons_label, - ui->buttonL, ui->labelL, - ui->buttonZL, ui->labelZL, - ui->labelScreenshot, ui->buttonScreenshot, - ui->buttonMinus, ui->labelMinus, - ui->LStick, ui->Dpad, - }; - break; - } + analog_map_deadzone_label = {ui->labelLStickDeadzone, ui->labelRStickDeadzone}; + analog_map_deadzone_slider = {ui->sliderLStickDeadzone, ui->sliderRStickDeadzone}; + analog_map_modifier_groupbox = {ui->buttonLStickModGroup, ui->buttonRStickModGroup}; + analog_map_modifier_button = {ui->buttonLStickMod, ui->buttonRStickMod}; + analog_map_modifier_label = {ui->labelLStickModifierRange, ui->labelRStickModifierRange}; + analog_map_modifier_slider = {ui->sliderLStickModifierRange, ui->sliderRStickModifierRange}; + analog_map_range_groupbox = {ui->buttonLStickRangeGroup, ui->buttonRStickRangeGroup}; + analog_map_range_spinbox = {ui->spinboxLStickRange, ui->spinboxRStickRange}; + + const auto ConfigureButtonClick = [&](QPushButton* button, Common::ParamPackage* param, + int default_val, InputCommon::Polling::DeviceType type) { + connect(button, &QPushButton::clicked, [=, this] { + HandleClick( + button, + [=, this](Common::ParamPackage params) { + // Workaround for ZL & ZR for analog triggers like on XBOX + // controllers. Analog triggers (from controllers like the XBOX + // controller) would not work due to a different range of their + // signals (from 0 to 255 on analog triggers instead of -32768 to + // 32768 on analog joysticks). The SDL driver misinterprets analog + // triggers as analog joysticks. + // TODO: reinterpret the signal range for analog triggers to map the + // values correctly. This is required for the correct emulation of + // the analog triggers of the GameCube controller. + if (button == ui->buttonZL || button == ui->buttonZR) { + params.Set("direction", "+"); + params.Set("threshold", "0.5"); + } + *param = std::move(params); + }, + type); + }); + }; - if (debug || layout == Settings::ControllerType::ProController) { - ui->controller_color->hide(); - } else { - if (layout == Settings::ControllerType::LeftJoycon || - layout == Settings::ControllerType::RightJoycon) { - ui->horizontalSpacer_4->setGeometry({0, 0, 0, 0}); + for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; ++button_id) { + auto* const button = button_map[button_id]; - LayerGridElements(ui->buttons, ui->shoulderButtons, ui->Dpad); - LayerGridElements(ui->buttons, ui->misc, ui->RStick); - LayerGridElements(ui->buttons, ui->Dpad, ui->faceButtons); - LayerGridElements(ui->buttons, ui->RStick, ui->LStick); + if (button == nullptr) { + continue; } - } - for (auto* widget : layout_hidden) - widget->setVisible(false); + ConfigureButtonClick(button_map[button_id], &buttons_param[button_id], + Config::default_buttons[button_id], + InputCommon::Polling::DeviceType::Button); - analog_map_stick = {ui->buttonLStickAnalog, ui->buttonRStickAnalog}; - analog_map_deadzone = {ui->sliderLStickDeadzone, ui->sliderRStickDeadzone}; - analog_map_deadzone_label = {ui->labelLStickDeadzone, ui->labelRStickDeadzone}; + button->setContextMenuPolicy(Qt::CustomContextMenu); - for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; button_id++) { - auto* const button = button_map[button_id]; + connect(button, &QPushButton::customContextMenuRequested, + [=, this](const QPoint& menu_location) { + QMenu context_menu; + context_menu.addAction(tr("Clear"), [&] { + buttons_param[button_id].Clear(); + button_map[button_id]->setText(tr("[not set]")); + }); + context_menu.exec(button_map[button_id]->mapToGlobal(menu_location)); + }); + } + + for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { + auto* const button = motion_map[motion_id]; if (button == nullptr) { continue; } + ConfigureButtonClick(motion_map[motion_id], &motions_param[motion_id], + Config::default_motions[motion_id], + InputCommon::Polling::DeviceType::Motion); + button->setContextMenuPolicy(Qt::CustomContextMenu); - connect(button, &QPushButton::clicked, [=] { - HandleClick(button_map[button_id], - [=](Common::ParamPackage params) { - // Workaround for ZL & ZR for analog triggers like on XBOX controllors. - // Analog triggers (from controllers like the XBOX controller) would not - // work due to a different range of their signals (from 0 to 255 on - // analog triggers instead of -32768 to 32768 on analog joysticks). The - // SDL driver misinterprets analog triggers as analog joysticks. - // TODO: reinterpret the signal range for analog triggers to map the - // values correctly. This is required for the correct emulation of the - // analog triggers of the GameCube controller. - if (button_id == Settings::NativeButton::ZL || - button_id == Settings::NativeButton::ZR) { - params.Set("direction", "+"); - params.Set("threshold", "0.5"); - } - buttons_param[button_id] = std::move(params); - }, - InputCommon::Polling::DeviceType::Button); - }); - connect(button, &QPushButton::customContextMenuRequested, [=](const QPoint& menu_location) { - QMenu context_menu; - context_menu.addAction(tr("Clear"), [&] { - buttons_param[button_id].Clear(); - button_map[button_id]->setText(tr("[not set]")); - }); - context_menu.addAction(tr("Restore Default"), [&] { - buttons_param[button_id] = Common::ParamPackage{ - InputCommon::GenerateKeyboardParam(Config::default_buttons[button_id])}; - button_map[button_id]->setText(ButtonToText(buttons_param[button_id])); - }); - context_menu.exec(button_map[button_id]->mapToGlobal(menu_location)); - }); + + connect(button, &QPushButton::customContextMenuRequested, + [=, this](const QPoint& menu_location) { + QMenu context_menu; + context_menu.addAction(tr("Clear"), [&] { + motions_param[motion_id].Clear(); + motion_map[motion_id]->setText(tr("[not set]")); + }); + context_menu.exec(motion_map[motion_id]->mapToGlobal(menu_location)); + }); } - for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; analog_id++) { - for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; sub_button_id++) { + for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; ++analog_id) { + for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; ++sub_button_id) { auto* const analog_button = analog_map_buttons[analog_id][sub_button_id]; + if (analog_button == nullptr) { continue; } - analog_button->setContextMenuPolicy(Qt::CustomContextMenu); - connect(analog_button, &QPushButton::clicked, [=]() { - HandleClick(analog_map_buttons[analog_id][sub_button_id], - [=](const Common::ParamPackage& params) { - SetAnalogButton(params, analogs_param[analog_id], - analog_sub_buttons[sub_button_id]); - }, - InputCommon::Polling::DeviceType::Button); + connect(analog_button, &QPushButton::clicked, [=, this] { + if (!map_analog_stick_accepted) { + map_analog_stick_accepted = + QMessageBox::information( + this, tr("Map Analog Stick"), + tr("After pressing OK, first move your joystick horizontally, and then " + "vertically.\nTo invert the axes, first move your joystick " + "vertically, and then horizontally."), + QMessageBox::Ok | QMessageBox::Cancel) == QMessageBox::Ok; + if (!map_analog_stick_accepted) { + return; + } + } + HandleClick( + analog_map_buttons[analog_id][sub_button_id], + [=, this](const Common::ParamPackage& params) { + SetAnalogParam(params, analogs_param[analog_id], + analog_sub_buttons[sub_button_id]); + }, + InputCommon::Polling::DeviceType::AnalogPreferred); }); + + analog_button->setContextMenuPolicy(Qt::CustomContextMenu); + connect(analog_button, &QPushButton::customContextMenuRequested, - [=](const QPoint& menu_location) { + [=, this](const QPoint& menu_location) { QMenu context_menu; context_menu.addAction(tr("Clear"), [&] { - analogs_param[analog_id].Erase(analog_sub_buttons[sub_button_id]); + analogs_param[analog_id].Clear(); analog_map_buttons[analog_id][sub_button_id]->setText(tr("[not set]")); }); - context_menu.addAction(tr("Restore Default"), [&] { - Common::ParamPackage params{InputCommon::GenerateKeyboardParam( - Config::default_analogs[analog_id][sub_button_id])}; - SetAnalogButton(params, analogs_param[analog_id], - analog_sub_buttons[sub_button_id]); - analog_map_buttons[analog_id][sub_button_id]->setText(AnalogToText( - analogs_param[analog_id], analog_sub_buttons[sub_button_id])); - }); context_menu.exec(analog_map_buttons[analog_id][sub_button_id]->mapToGlobal( menu_location)); }); } - connect(analog_map_stick[analog_id], &QPushButton::clicked, [=] { - if (QMessageBox::information( - this, tr("Information"), - tr("After pressing OK, first move your joystick horizontally, " - "and then vertically."), - QMessageBox::Ok | QMessageBox::Cancel) == QMessageBox::Ok) { - HandleClick( - analog_map_stick[analog_id], - [=](const Common::ParamPackage& params) { analogs_param[analog_id] = params; }, - InputCommon::Polling::DeviceType::Analog); - } + + // Handle clicks for the modifier buttons as well. + connect(analog_map_modifier_button[analog_id], &QPushButton::clicked, [=, this] { + HandleClick( + analog_map_modifier_button[analog_id], + [=, this](const Common::ParamPackage& params) { + analogs_param[analog_id].Set("modifier", params.Serialize()); + }, + InputCommon::Polling::DeviceType::Button); }); - connect(analog_map_deadzone[analog_id], &QSlider::valueChanged, [=] { - const float deadzone = analog_map_deadzone[analog_id]->value() / 100.0f; - analog_map_deadzone_label[analog_id]->setText(tr("Deadzone: %1").arg(deadzone)); - analogs_param[analog_id].Set("deadzone", deadzone); + + analog_map_modifier_button[analog_id]->setContextMenuPolicy(Qt::CustomContextMenu); + + connect(analog_map_modifier_button[analog_id], &QPushButton::customContextMenuRequested, + [=, this](const QPoint& menu_location) { + QMenu context_menu; + context_menu.addAction(tr("Clear"), [&] { + analogs_param[analog_id].Set("modifier", ""); + analog_map_modifier_button[analog_id]->setText(tr("[not set]")); + }); + context_menu.exec( + analog_map_modifier_button[analog_id]->mapToGlobal(menu_location)); + }); + + connect(analog_map_range_spinbox[analog_id], qOverload<int>(&QSpinBox::valueChanged), + [=, this] { + const auto spinbox_value = analog_map_range_spinbox[analog_id]->value(); + analogs_param[analog_id].Set("range", spinbox_value / 100.0f); + }); + + connect(analog_map_deadzone_slider[analog_id], &QSlider::valueChanged, [=, this] { + const auto slider_value = analog_map_deadzone_slider[analog_id]->value(); + analog_map_deadzone_label[analog_id]->setText(tr("Deadzone: %1%").arg(slider_value)); + analogs_param[analog_id].Set("deadzone", slider_value / 100.0f); + }); + + connect(analog_map_modifier_slider[analog_id], &QSlider::valueChanged, [=, this] { + const auto slider_value = analog_map_modifier_slider[analog_id]->value(); + analog_map_modifier_label[analog_id]->setText( + tr("Modifier Range: %1%").arg(slider_value)); + analogs_param[analog_id].Set("modifier_scale", slider_value / 100.0f); }); } - connect(ui->buttonClearAll, &QPushButton::clicked, [this] { ClearAll(); }); - connect(ui->buttonRestoreDefaults, &QPushButton::clicked, [this] { RestoreDefaults(); }); + // Player Connected checkbox + connect(ui->groupConnectedController, &QGroupBox::toggled, + [this](bool checked) { emit Connected(checked); }); + + // Set up controller type. Only Player 1 can choose Handheld. + ui->comboControllerType->clear(); + + QStringList controller_types = { + tr("Pro Controller"), + tr("Dual Joycons"), + tr("Left Joycon"), + tr("Right Joycon"), + }; + + if (player_index == 0) { + controller_types.append(tr("Handheld")); + connect(ui->comboControllerType, qOverload<int>(&QComboBox::currentIndexChanged), + [this](int index) { + emit HandheldStateChanged(GetControllerTypeFromIndex(index) == + Settings::ControllerType::Handheld); + }); + } + + if (debug || player_index == 9) { + ui->groupConnectedController->setCheckable(false); + } + + // The Debug Controller can only choose the Pro Controller. + if (debug) { + ui->buttonScreenshot->setEnabled(false); + ui->buttonHome->setEnabled(false); + QStringList debug_controller_types = { + tr("Pro Controller"), + }; + ui->comboControllerType->addItems(debug_controller_types); + } else { + ui->comboControllerType->addItems(controller_types); + } + + UpdateControllerIcon(); + UpdateControllerAvailableButtons(); + UpdateMotionButtons(); + connect(ui->comboControllerType, qOverload<int>(&QComboBox::currentIndexChanged), [this](int) { + UpdateControllerIcon(); + UpdateControllerAvailableButtons(); + UpdateMotionButtons(); + }); + + connect(ui->comboDevices, qOverload<int>(&QComboBox::activated), this, + &ConfigureInputPlayer::UpdateMappingWithDefaults); + + ui->comboDevices->setCurrentIndex(-1); + + ui->buttonRefreshDevices->setIcon(QIcon::fromTheme(QStringLiteral("view-refresh"))); + connect(ui->buttonRefreshDevices, &QPushButton::clicked, + [this] { emit RefreshInputDevices(); }); timeout_timer->setSingleShot(true); connect(timeout_timer.get(), &QTimer::timeout, [this] { SetPollingResult({}, true); }); connect(poll_timer.get(), &QTimer::timeout, [this] { Common::ParamPackage params; + if (input_subsystem->GetGCButtons()->IsPolling()) { + params = input_subsystem->GetGCButtons()->GetNextInput(); + if (params.Has("engine") && IsInputAcceptable(params)) { + SetPollingResult(params, false); + return; + } + } + if (input_subsystem->GetGCAnalogs()->IsPolling()) { + params = input_subsystem->GetGCAnalogs()->GetNextInput(); + if (params.Has("engine") && IsInputAcceptable(params)) { + SetPollingResult(params, false); + return; + } + } + if (input_subsystem->GetUDPMotions()->IsPolling()) { + params = input_subsystem->GetUDPMotions()->GetNextInput(); + if (params.Has("engine")) { + SetPollingResult(params, false); + return; + } + } for (auto& poller : device_pollers) { params = poller->GetNextInput(); - if (params.Has("engine")) { + if (params.Has("engine") && IsInputAcceptable(params)) { SetPollingResult(params, false); return; } } }); - controller_color_buttons = { - ui->left_body_button, - ui->left_buttons_button, - ui->right_body_button, - ui->right_buttons_button, - }; + UpdateInputProfiles(); - for (std::size_t i = 0; i < controller_color_buttons.size(); ++i) { - connect(controller_color_buttons[i], &QPushButton::clicked, this, - [this, i] { OnControllerButtonClick(static_cast<int>(i)); }); - } + connect(ui->buttonProfilesNew, &QPushButton::clicked, this, + &ConfigureInputPlayer::CreateProfile); + connect(ui->buttonProfilesDelete, &QPushButton::clicked, this, + &ConfigureInputPlayer::DeleteProfile); + connect(ui->comboProfiles, qOverload<int>(&QComboBox::activated), this, + &ConfigureInputPlayer::LoadProfile); + connect(ui->buttonProfilesSave, &QPushButton::clicked, this, + &ConfigureInputPlayer::SaveProfile); LoadConfiguration(); - resize(0, 0); // TODO(wwylele): enable this when we actually emulate it ui->buttonHome->setEnabled(false); @@ -374,27 +561,72 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i ConfigureInputPlayer::~ConfigureInputPlayer() = default; void ConfigureInputPlayer::ApplyConfiguration() { - auto& buttons = - debug ? Settings::values.debug_pad_buttons : Settings::values.players[player_index].buttons; - auto& analogs = - debug ? Settings::values.debug_pad_analogs : Settings::values.players[player_index].analogs; + auto& player = Settings::values.players.GetValue()[player_index]; + auto& buttons = debug ? Settings::values.debug_pad_buttons : player.buttons; + auto& analogs = debug ? Settings::values.debug_pad_analogs : player.analogs; std::transform(buttons_param.begin(), buttons_param.end(), buttons.begin(), [](const Common::ParamPackage& param) { return param.Serialize(); }); std::transform(analogs_param.begin(), analogs_param.end(), analogs.begin(), [](const Common::ParamPackage& param) { return param.Serialize(); }); - if (debug) + if (debug) { + return; + } + + auto& motions = player.motions; + + std::transform(motions_param.begin(), motions_param.end(), motions.begin(), + [](const Common::ParamPackage& param) { return param.Serialize(); }); + + const auto controller_type = + GetControllerTypeFromIndex(ui->comboControllerType->currentIndex()); + const auto player_connected = ui->groupConnectedController->isChecked() && + controller_type != Settings::ControllerType::Handheld; + + if (player.controller_type == controller_type && player.connected == player_connected) { + // Set vibration devices in the event that the input device has changed. + ConfigureVibration::SetVibrationDevices(player_index); + return; + } + + // Disconnect the controller first. + UpdateController(controller_type, player_index, false); + + player.controller_type = controller_type; + player.connected = player_connected; + + ConfigureVibration::SetVibrationDevices(player_index); + + // Handheld + if (player_index == 0) { + auto& handheld = Settings::values.players.GetValue()[HANDHELD_INDEX]; + if (controller_type == Settings::ControllerType::Handheld) { + handheld = player; + } + handheld.connected = ui->groupConnectedController->isChecked() && + controller_type == Settings::ControllerType::Handheld; + UpdateController(Settings::ControllerType::Handheld, HANDHELD_INDEX, handheld.connected); + } + + if (!player.connected) { return; + } - std::array<u32, 4> colors{}; - std::transform(controller_colors.begin(), controller_colors.end(), colors.begin(), - [](QColor color) { return color.rgb(); }); + // This emulates a delay between disconnecting and reconnecting controllers as some games + // do not respond to a change in controller type if it was instantaneous. + using namespace std::chrono_literals; + std::this_thread::sleep_for(20ms); - Settings::values.players[player_index].body_color_left = colors[0]; - Settings::values.players[player_index].button_color_left = colors[1]; - Settings::values.players[player_index].body_color_right = colors[2]; - Settings::values.players[player_index].button_color_right = colors[3]; + UpdateController(controller_type, player_index, player_connected); +} + +void ConfigureInputPlayer::showEvent(QShowEvent* event) { + if (bottom_row == nullptr) { + return; + } + QWidget::showEvent(event); + ui->main->addWidget(bottom_row); } void ConfigureInputPlayer::changeEvent(QEvent* event) { @@ -402,24 +634,16 @@ void ConfigureInputPlayer::changeEvent(QEvent* event) { RetranslateUI(); } - QDialog::changeEvent(event); + QWidget::changeEvent(event); } void ConfigureInputPlayer::RetranslateUI() { ui->retranslateUi(this); - UpdateButtonLabels(); -} - -void ConfigureInputPlayer::OnControllerButtonClick(int i) { - const QColor new_bg_color = QColorDialog::getColor(controller_colors[i]); - if (!new_bg_color.isValid()) - return; - controller_colors[i] = new_bg_color; - controller_color_buttons[i]->setStyleSheet( - QStringLiteral("QPushButton { background-color: %1 }").arg(controller_colors[i].name())); + UpdateUI(); } void ConfigureInputPlayer::LoadConfiguration() { + auto& player = Settings::values.players.GetValue()[player_index]; if (debug) { std::transform(Settings::values.debug_pad_buttons.begin(), Settings::values.debug_pad_buttons.end(), buttons_param.begin(), @@ -428,66 +652,109 @@ void ConfigureInputPlayer::LoadConfiguration() { Settings::values.debug_pad_analogs.end(), analogs_param.begin(), [](const std::string& str) { return Common::ParamPackage(str); }); } else { - std::transform(Settings::values.players[player_index].buttons.begin(), - Settings::values.players[player_index].buttons.end(), buttons_param.begin(), + std::transform(player.buttons.begin(), player.buttons.end(), buttons_param.begin(), + [](const std::string& str) { return Common::ParamPackage(str); }); + std::transform(player.analogs.begin(), player.analogs.end(), analogs_param.begin(), [](const std::string& str) { return Common::ParamPackage(str); }); - std::transform(Settings::values.players[player_index].analogs.begin(), - Settings::values.players[player_index].analogs.end(), analogs_param.begin(), + std::transform(player.motions.begin(), player.motions.end(), motions_param.begin(), [](const std::string& str) { return Common::ParamPackage(str); }); } - UpdateButtonLabels(); + UpdateUI(); + UpdateInputDeviceCombobox(); - if (debug) + if (debug) { return; + } - std::array<u32, 4> colors = { - Settings::values.players[player_index].body_color_left, - Settings::values.players[player_index].button_color_left, - Settings::values.players[player_index].body_color_right, - Settings::values.players[player_index].button_color_right, - }; + ui->comboControllerType->setCurrentIndex(static_cast<int>(player.controller_type)); + ui->groupConnectedController->setChecked( + player.connected || + (player_index == 0 && Settings::values.players.GetValue()[HANDHELD_INDEX].connected)); +} - std::transform(colors.begin(), colors.end(), controller_colors.begin(), - [](u32 rgb) { return QColor::fromRgb(rgb); }); +void ConfigureInputPlayer::ConnectPlayer(bool connected) { + ui->groupConnectedController->setChecked(connected); +} - for (std::size_t i = 0; i < colors.size(); ++i) { - controller_color_buttons[i]->setStyleSheet( - QStringLiteral("QPushButton { background-color: %1 }") - .arg(controller_colors[i].name())); +void ConfigureInputPlayer::UpdateInputDeviceCombobox() { + // Skip input device persistence if "Input Devices" is set to "Any". + if (ui->comboDevices->currentIndex() == 0) { + UpdateInputDevices(); + return; } -} -void ConfigureInputPlayer::RestoreDefaults() { - for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; button_id++) { - buttons_param[button_id] = Common::ParamPackage{ - InputCommon::GenerateKeyboardParam(Config::default_buttons[button_id])}; + // Find the first button that isn't empty. + const auto button_param = + std::find_if(buttons_param.begin(), buttons_param.end(), + [](const Common::ParamPackage param) { return param.Has("engine"); }); + const bool buttons_empty = button_param == buttons_param.end(); + + const auto current_engine = button_param->Get("engine", ""); + const auto current_guid = button_param->Get("guid", ""); + const auto current_port = button_param->Get("port", ""); + + const bool is_keyboard_mouse = current_engine == "keyboard" || current_engine == "mouse"; + + UpdateInputDevices(); + + if (buttons_empty) { + return; } - for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; analog_id++) { - for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; sub_button_id++) { - Common::ParamPackage params{InputCommon::GenerateKeyboardParam( - Config::default_analogs[analog_id][sub_button_id])}; - SetAnalogButton(params, analogs_param[analog_id], analog_sub_buttons[sub_button_id]); + const bool all_one_device = + std::all_of(buttons_param.begin(), buttons_param.end(), + [current_engine, current_guid, current_port, + is_keyboard_mouse](const Common::ParamPackage param) { + if (is_keyboard_mouse) { + return !param.Has("engine") || param.Get("engine", "") == "keyboard" || + param.Get("engine", "") == "mouse"; + } + return !param.Has("engine") || (param.Get("engine", "") == current_engine && + param.Get("guid", "") == current_guid && + param.Get("port", "") == current_port); + }); + + if (all_one_device) { + if (is_keyboard_mouse) { + ui->comboDevices->setCurrentIndex(1); + return; } + const auto devices_it = std::find_if( + input_devices.begin(), input_devices.end(), + [current_engine, current_guid, current_port](const Common::ParamPackage param) { + return param.Get("class", "") == current_engine && + param.Get("guid", "") == current_guid && + param.Get("port", "") == current_port; + }); + const int device_index = + devices_it != input_devices.end() + ? static_cast<int>(std::distance(input_devices.begin(), devices_it)) + : 0; + ui->comboDevices->setCurrentIndex(device_index); + } else { + ui->comboDevices->setCurrentIndex(0); } - UpdateButtonLabels(); +} + +void ConfigureInputPlayer::RestoreDefaults() { + UpdateMappingWithDefaults(); } void ConfigureInputPlayer::ClearAll() { - for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; button_id++) { + for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; ++button_id) { const auto* const button = button_map[button_id]; - if (button == nullptr || !button->isEnabled()) { + if (button == nullptr) { continue; } buttons_param[button_id].Clear(); } - for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; analog_id++) { - for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; sub_button_id++) { + for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; ++analog_id) { + for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; ++sub_button_id) { const auto* const analog_button = analog_map_buttons[analog_id][sub_button_id]; - if (analog_button == nullptr || !analog_button->isEnabled()) { + if (analog_button == nullptr) { continue; } @@ -495,16 +762,30 @@ void ConfigureInputPlayer::ClearAll() { } } - UpdateButtonLabels(); + for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { + const auto* const motion_button = motion_map[motion_id]; + if (motion_button == nullptr) { + continue; + } + + motions_param[motion_id].Clear(); + } + + UpdateUI(); + UpdateInputDevices(); } -void ConfigureInputPlayer::UpdateButtonLabels() { - for (int button = 0; button < Settings::NativeButton::NumButtons; button++) { +void ConfigureInputPlayer::UpdateUI() { + for (int button = 0; button < Settings::NativeButton::NumButtons; ++button) { button_map[button]->setText(ButtonToText(buttons_param[button])); } - for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; analog_id++) { - for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; sub_button_id++) { + for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { + motion_map[motion_id]->setText(ButtonToText(motions_param[motion_id])); + } + + for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; ++analog_id) { + for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; ++sub_button_id) { auto* const analog_button = analog_map_buttons[analog_id][sub_button_id]; if (analog_button == nullptr) { @@ -514,80 +795,329 @@ void ConfigureInputPlayer::UpdateButtonLabels() { analog_button->setText( AnalogToText(analogs_param[analog_id], analog_sub_buttons[sub_button_id])); } - analog_map_stick[analog_id]->setText(tr("Set Analog Stick")); + analog_map_modifier_button[analog_id]->setText( + ButtonToText(Common::ParamPackage{analogs_param[analog_id].Get("modifier", "")})); + + const auto deadzone_label = analog_map_deadzone_label[analog_id]; + const auto deadzone_slider = analog_map_deadzone_slider[analog_id]; + const auto modifier_groupbox = analog_map_modifier_groupbox[analog_id]; + const auto modifier_label = analog_map_modifier_label[analog_id]; + const auto modifier_slider = analog_map_modifier_slider[analog_id]; + const auto range_groupbox = analog_map_range_groupbox[analog_id]; + const auto range_spinbox = analog_map_range_spinbox[analog_id]; + + int slider_value; auto& param = analogs_param[analog_id]; - auto* const analog_deadzone_slider = analog_map_deadzone[analog_id]; - auto* const analog_deadzone_label = analog_map_deadzone_label[analog_id]; + const bool is_controller = + param.Get("engine", "") == "sdl" || param.Get("engine", "") == "gcpad"; - if (param.Has("engine") && param.Get("engine", "") == "sdl") { + if (is_controller) { if (!param.Has("deadzone")) { param.Set("deadzone", 0.1f); } + slider_value = static_cast<int>(param.Get("deadzone", 0.1f) * 100); + deadzone_label->setText(tr("Deadzone: %1%").arg(slider_value)); + deadzone_slider->setValue(slider_value); + if (!param.Has("range")) { + param.Set("range", 1.0f); + } + range_spinbox->setValue(static_cast<int>(param.Get("range", 1.0f) * 100)); + } else { + if (!param.Has("modifier_scale")) { + param.Set("modifier_scale", 0.5f); + } + slider_value = static_cast<int>(param.Get("modifier_scale", 0.5f) * 100); + modifier_label->setText(tr("Modifier Range: %1%").arg(slider_value)); + modifier_slider->setValue(slider_value); + } - analog_deadzone_slider->setValue(static_cast<int>(param.Get("deadzone", 0.1f) * 100)); - analog_deadzone_slider->setVisible(true); - analog_deadzone_label->setVisible(true); + deadzone_label->setVisible(is_controller); + deadzone_slider->setVisible(is_controller); + modifier_groupbox->setVisible(!is_controller); + modifier_label->setVisible(!is_controller); + modifier_slider->setVisible(!is_controller); + range_groupbox->setVisible(is_controller); + } +} + +void ConfigureInputPlayer::UpdateInputDevices() { + input_devices = input_subsystem->GetInputDevices(); + ui->comboDevices->clear(); + for (auto device : input_devices) { + ui->comboDevices->addItem(QString::fromStdString(device.Get("display", "Unknown")), {}); + } +} + +void ConfigureInputPlayer::UpdateControllerIcon() { + // We aren't using Qt's built in theme support here since we aren't drawing an icon (and its + // "nonstandard" to use an image through the icon support) + const QString stylesheet = [this] { + switch (GetControllerTypeFromIndex(ui->comboControllerType->currentIndex())) { + case Settings::ControllerType::ProController: + return QStringLiteral("image: url(:/controller/pro_controller%0)"); + case Settings::ControllerType::DualJoyconDetached: + return QStringLiteral("image: url(:/controller/dual_joycon%0)"); + case Settings::ControllerType::LeftJoycon: + return QStringLiteral("image: url(:/controller/single_joycon_left_vertical%0)"); + case Settings::ControllerType::RightJoycon: + return QStringLiteral("image: url(:/controller/single_joycon_right_vertical%0)"); + case Settings::ControllerType::Handheld: + return QStringLiteral("image: url(:/controller/handheld%0)"); + default: + return QString{}; + } + }(); + + const QString theme = [] { + if (QIcon::themeName().contains(QStringLiteral("dark"))) { + return QStringLiteral("_dark"); + } else if (QIcon::themeName().contains(QStringLiteral("midnight"))) { + return QStringLiteral("_midnight"); } else { - analog_deadzone_slider->setVisible(false); - analog_deadzone_label->setVisible(false); + return QString{}; } + }(); + + ui->controllerFrame->setStyleSheet(stylesheet.arg(theme)); +} + +void ConfigureInputPlayer::UpdateControllerAvailableButtons() { + auto layout = GetControllerTypeFromIndex(ui->comboControllerType->currentIndex()); + if (debug) { + layout = Settings::ControllerType::ProController; + } + + // List of all the widgets that will be hidden by any of the following layouts that need + // "unhidden" after the controller type changes + const std::array<QWidget*, 9> layout_show = { + ui->buttonShoulderButtonsSLSR, + ui->horizontalSpacerShoulderButtonsWidget, + ui->horizontalSpacerShoulderButtonsWidget2, + ui->buttonShoulderButtonsLeft, + ui->buttonMiscButtonsMinusScreenshot, + ui->bottomLeft, + ui->buttonShoulderButtonsRight, + ui->buttonMiscButtonsPlusHome, + ui->bottomRight, + }; + + for (auto* widget : layout_show) { + widget->show(); + } + + std::vector<QWidget*> layout_hidden; + switch (layout) { + case Settings::ControllerType::ProController: + case Settings::ControllerType::DualJoyconDetached: + case Settings::ControllerType::Handheld: + layout_hidden = { + ui->buttonShoulderButtonsSLSR, + ui->horizontalSpacerShoulderButtonsWidget2, + }; + break; + case Settings::ControllerType::LeftJoycon: + layout_hidden = { + ui->horizontalSpacerShoulderButtonsWidget2, + ui->buttonShoulderButtonsRight, + ui->buttonMiscButtonsPlusHome, + ui->bottomRight, + }; + break; + case Settings::ControllerType::RightJoycon: + layout_hidden = { + ui->horizontalSpacerShoulderButtonsWidget, + ui->buttonShoulderButtonsLeft, + ui->buttonMiscButtonsMinusScreenshot, + ui->bottomLeft, + }; + break; + } + + for (auto* widget : layout_hidden) { + widget->hide(); } } +void ConfigureInputPlayer::UpdateMotionButtons() { + if (debug) { + // Motion isn't used with the debug controller, hide both groupboxes. + ui->buttonMotionLeftGroup->hide(); + ui->buttonMotionRightGroup->hide(); + return; + } + + // Show/hide the "Motion 1/2" groupboxes depending on the currently selected controller. + switch (GetControllerTypeFromIndex(ui->comboControllerType->currentIndex())) { + case Settings::ControllerType::ProController: + case Settings::ControllerType::LeftJoycon: + case Settings::ControllerType::Handheld: + // Show "Motion 1" and hide "Motion 2". + ui->buttonMotionLeftGroup->show(); + ui->buttonMotionRightGroup->hide(); + break; + case Settings::ControllerType::RightJoycon: + // Show "Motion 2" and hide "Motion 1". + ui->buttonMotionLeftGroup->hide(); + ui->buttonMotionRightGroup->show(); + break; + case Settings::ControllerType::DualJoyconDetached: + default: + // Show both "Motion 1/2". + ui->buttonMotionLeftGroup->show(); + ui->buttonMotionRightGroup->show(); + break; + } +} + +void ConfigureInputPlayer::UpdateMappingWithDefaults() { + if (ui->comboDevices->currentIndex() == 0) { + return; + } + + if (ui->comboDevices->currentIndex() == 1) { + // Reset keyboard bindings + for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; ++button_id) { + buttons_param[button_id] = Common::ParamPackage{ + InputCommon::GenerateKeyboardParam(Config::default_buttons[button_id])}; + } + for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; ++analog_id) { + for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; ++sub_button_id) { + Common::ParamPackage params{InputCommon::GenerateKeyboardParam( + Config::default_analogs[analog_id][sub_button_id])}; + SetAnalogParam(params, analogs_param[analog_id], analog_sub_buttons[sub_button_id]); + } + + analogs_param[analog_id].Set("modifier", InputCommon::GenerateKeyboardParam( + Config::default_stick_mod[analog_id])); + } + + for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { + motions_param[motion_id] = Common::ParamPackage{ + InputCommon::GenerateKeyboardParam(Config::default_motions[motion_id])}; + } + + UpdateUI(); + return; + } + + // Reset controller bindings + const auto& device = input_devices[ui->comboDevices->currentIndex()]; + auto button_mapping = input_subsystem->GetButtonMappingForDevice(device); + auto analog_mapping = input_subsystem->GetAnalogMappingForDevice(device); + for (std::size_t i = 0; i < buttons_param.size(); ++i) { + buttons_param[i] = button_mapping[static_cast<Settings::NativeButton::Values>(i)]; + } + for (std::size_t i = 0; i < analogs_param.size(); ++i) { + analogs_param[i] = analog_mapping[static_cast<Settings::NativeAnalog::Values>(i)]; + } + + UpdateUI(); +} + void ConfigureInputPlayer::HandleClick( QPushButton* button, std::function<void(const Common::ParamPackage&)> new_input_setter, InputCommon::Polling::DeviceType type) { - button->setText(tr("[press key]")); + if (button == ui->buttonMotionLeft || button == ui->buttonMotionRight) { + button->setText(tr("Shake!")); + } else { + button->setText(tr("[waiting]")); + } button->setFocus(); - // Keyboard keys can only be used as button devices - want_keyboard_keys = type == InputCommon::Polling::DeviceType::Button; - if (want_keyboard_keys) { - const auto iter = std::find(button_map.begin(), button_map.end(), button); - ASSERT(iter != button_map.end()); - const auto index = std::distance(button_map.begin(), iter); - ASSERT(index < Settings::NativeButton::NumButtons && index >= 0); - } + // The first two input devices are always Any and Keyboard/Mouse. If the user filtered to a + // controller, then they don't want keyboard/mouse input + want_keyboard_mouse = ui->comboDevices->currentIndex() < 2; input_setter = new_input_setter; - device_pollers = InputCommon::Polling::GetPollers(type); + device_pollers = input_subsystem->GetPollers(type); for (auto& poller : device_pollers) { poller->Start(); } - grabKeyboard(); - grabMouse(); - timeout_timer->start(5000); // Cancel after 5 seconds - poll_timer->start(200); // Check for new inputs every 200ms + QWidget::grabMouse(); + QWidget::grabKeyboard(); + + if (type == InputCommon::Polling::DeviceType::Button) { + input_subsystem->GetGCButtons()->BeginConfiguration(); + } else { + input_subsystem->GetGCAnalogs()->BeginConfiguration(); + } + + if (type == InputCommon::Polling::DeviceType::Motion) { + input_subsystem->GetUDPMotions()->BeginConfiguration(); + } + + timeout_timer->start(2500); // Cancel after 2.5 seconds + poll_timer->start(50); // Check for new inputs every 50ms } void ConfigureInputPlayer::SetPollingResult(const Common::ParamPackage& params, bool abort) { - releaseKeyboard(); - releaseMouse(); timeout_timer->stop(); poll_timer->stop(); for (auto& poller : device_pollers) { poller->Stop(); } + QWidget::releaseMouse(); + QWidget::releaseKeyboard(); + + input_subsystem->GetGCButtons()->EndConfiguration(); + input_subsystem->GetGCAnalogs()->EndConfiguration(); + + input_subsystem->GetUDPMotions()->EndConfiguration(); + if (!abort) { (*input_setter)(params); } - UpdateButtonLabels(); + UpdateUI(); + UpdateInputDeviceCombobox(); + input_setter = std::nullopt; } +bool ConfigureInputPlayer::IsInputAcceptable(const Common::ParamPackage& params) const { + if (ui->comboDevices->currentIndex() == 0) { + return true; + } + + // Keyboard/Mouse + if (ui->comboDevices->currentIndex() == 1) { + return params.Get("engine", "") == "keyboard" || params.Get("engine", "") == "mouse"; + } + + const auto current_input_device = input_devices[ui->comboDevices->currentIndex()]; + return params.Get("engine", "") == current_input_device.Get("class", "") && + params.Get("guid", "") == current_input_device.Get("guid", "") && + params.Get("port", "") == current_input_device.Get("port", ""); +} + +void ConfigureInputPlayer::mousePressEvent(QMouseEvent* event) { + if (!input_setter || !event) { + return; + } + + if (want_keyboard_mouse) { + SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->button())}, + false); + } else { + // We don't want any mouse buttons, so don't stop polling + return; + } + + SetPollingResult({}, true); +} + void ConfigureInputPlayer::keyPressEvent(QKeyEvent* event) { if (!input_setter || !event) { return; } if (event->key() != Qt::Key_Escape) { - if (want_keyboard_keys) { + if (want_keyboard_mouse) { SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->key())}, false); } else { @@ -595,5 +1125,105 @@ void ConfigureInputPlayer::keyPressEvent(QKeyEvent* event) { return; } } + SetPollingResult({}, true); } + +void ConfigureInputPlayer::CreateProfile() { + const auto profile_name = + LimitableInputDialog::GetText(this, tr("New Profile"), tr("Enter a profile name:"), 1, 20); + + if (profile_name.isEmpty()) { + return; + } + + if (!InputProfiles::IsProfileNameValid(profile_name.toStdString())) { + QMessageBox::critical(this, tr("Create Input Profile"), + tr("The given profile name is not valid!")); + return; + } + + ApplyConfiguration(); + + if (!profiles->CreateProfile(profile_name.toStdString(), player_index)) { + QMessageBox::critical(this, tr("Create Input Profile"), + tr("Failed to create the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } + + emit RefreshInputProfiles(player_index); + + ui->comboProfiles->addItem(profile_name); + ui->comboProfiles->setCurrentIndex(ui->comboProfiles->count() - 1); +} + +void ConfigureInputPlayer::DeleteProfile() { + const QString profile_name = ui->comboProfiles->itemText(ui->comboProfiles->currentIndex()); + + if (profile_name.isEmpty()) { + return; + } + + if (!profiles->DeleteProfile(profile_name.toStdString())) { + QMessageBox::critical(this, tr("Delete Input Profile"), + tr("Failed to delete the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } + + emit RefreshInputProfiles(player_index); + + ui->comboProfiles->removeItem(ui->comboProfiles->currentIndex()); + ui->comboProfiles->setCurrentIndex(-1); +} + +void ConfigureInputPlayer::LoadProfile() { + const QString profile_name = ui->comboProfiles->itemText(ui->comboProfiles->currentIndex()); + + if (profile_name.isEmpty()) { + return; + } + + ApplyConfiguration(); + + if (!profiles->LoadProfile(profile_name.toStdString(), player_index)) { + QMessageBox::critical(this, tr("Load Input Profile"), + tr("Failed to load the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } + + LoadConfiguration(); +} + +void ConfigureInputPlayer::SaveProfile() { + const QString profile_name = ui->comboProfiles->itemText(ui->comboProfiles->currentIndex()); + + if (profile_name.isEmpty()) { + return; + } + + ApplyConfiguration(); + + if (!profiles->SaveProfile(profile_name.toStdString(), player_index)) { + QMessageBox::critical(this, tr("Save Input Profile"), + tr("Failed to save the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } +} + +void ConfigureInputPlayer::UpdateInputProfiles() { + ui->comboProfiles->clear(); + + for (const auto& profile_name : profiles->GetInputProfileNames()) { + ui->comboProfiles->addItem(QString::fromStdString(profile_name)); + } + + ui->comboProfiles->setCurrentIndex(-1); +} diff --git a/src/yuzu/configuration/configure_input_player.h b/src/yuzu/configuration/configure_input_player.h index 045704e47..23cf6f958 100644 --- a/src/yuzu/configuration/configure_input_player.h +++ b/src/yuzu/configuration/configure_input_player.h @@ -10,16 +10,27 @@ #include <optional> #include <string> -#include <QDialog> +#include <QWidget> #include "common/param_package.h" #include "core/settings.h" #include "ui_configure_input.h" +class QCheckBox; class QKeyEvent; +class QLabel; class QPushButton; +class QSlider; +class QSpinBox; class QString; class QTimer; +class QWidget; + +class InputProfiles; + +namespace InputCommon { +class InputSubsystem; +} namespace InputCommon::Polling { class DevicePoller; @@ -30,84 +41,165 @@ namespace Ui { class ConfigureInputPlayer; } -class ConfigureInputPlayer : public QDialog { +class ConfigureInputPlayer : public QWidget { Q_OBJECT public: - explicit ConfigureInputPlayer(QWidget* parent, std::size_t player_index, bool debug = false); + explicit ConfigureInputPlayer(QWidget* parent, std::size_t player_index, QWidget* bottom_row, + InputCommon::InputSubsystem* input_subsystem_, + InputProfiles* profiles_, bool debug = false); ~ConfigureInputPlayer() override; - /// Save all button configurations to settings file + /// Save all button configurations to settings file. void ApplyConfiguration(); -private: - void changeEvent(QEvent* event) override; - void RetranslateUI(); + /// Set the connection state checkbox (used to sync state). + void ConnectPlayer(bool connected); - void OnControllerButtonClick(int i); + /// Update the input devices combobox. + void UpdateInputDeviceCombobox(); + + /// Updates the list of controller profiles. + void UpdateInputProfiles(); - /// Load configuration settings. - void LoadConfiguration(); /// Restore all buttons to their default values. void RestoreDefaults(); - /// Clear all input configuration + + /// Clear all input configuration. void ClearAll(); - /// Update UI to reflect current configuration. - void UpdateButtonLabels(); +signals: + /// Emitted when this controller is connected by the user. + void Connected(bool connected); + /// Emitted when the Handheld mode is selected (undocked with dual joycons attached). + void HandheldStateChanged(bool is_handheld); + /// Emitted when the input devices combobox is being refreshed. + void RefreshInputDevices(); + /** + * Emitted when the input profiles combobox is being refreshed. + * The player_index represents the current player's index, and the profile combobox + * will not be updated for this index as they are already updated by other mechanisms. + */ + void RefreshInputProfiles(std::size_t player_index); + +protected: + void showEvent(QShowEvent* event) override; + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + /// Load configuration settings. + void LoadConfiguration(); /// Called when the button was pressed. void HandleClick(QPushButton* button, std::function<void(const Common::ParamPackage&)> new_input_setter, InputCommon::Polling::DeviceType type); - /// Finish polling and configure input using the input_setter + /// Finish polling and configure input using the input_setter. void SetPollingResult(const Common::ParamPackage& params, bool abort); + /// Checks whether a given input can be accepted. + bool IsInputAcceptable(const Common::ParamPackage& params) const; + + /// Handle mouse button press events. + void mousePressEvent(QMouseEvent* event) override; + /// Handle key press events. void keyPressEvent(QKeyEvent* event) override; + /// Update UI to reflect current configuration. + void UpdateUI(); + + /// Update the available input devices. + void UpdateInputDevices(); + + /// Update the current controller icon. + void UpdateControllerIcon(); + + /// Hides and disables controller settings based on the current controller type. + void UpdateControllerAvailableButtons(); + + /// Shows or hides motion groupboxes based on the current controller type. + void UpdateMotionButtons(); + + /// Gets the default controller mapping for this device and auto configures the input to match. + void UpdateMappingWithDefaults(); + + /// Creates a controller profile. + void CreateProfile(); + + /// Deletes the selected controller profile. + void DeleteProfile(); + + /// Loads the selected controller profile. + void LoadProfile(); + + /// Saves the current controller configuration into a selected controller profile. + void SaveProfile(); + std::unique_ptr<Ui::ConfigureInputPlayer> ui; std::size_t player_index; bool debug; + InputCommon::InputSubsystem* input_subsystem; + + InputProfiles* profiles; + std::unique_ptr<QTimer> timeout_timer; std::unique_ptr<QTimer> poll_timer; + static constexpr int PLAYER_COUNT = 8; + std::array<QCheckBox*, PLAYER_COUNT> player_connected_checkbox; + /// This will be the the setting function when an input is awaiting configuration. std::optional<std::function<void(const Common::ParamPackage&)>> input_setter; std::array<Common::ParamPackage, Settings::NativeButton::NumButtons> buttons_param; std::array<Common::ParamPackage, Settings::NativeAnalog::NumAnalogs> analogs_param; + std::array<Common::ParamPackage, Settings::NativeMotion::NumMotions> motions_param; - static constexpr int ANALOG_SUB_BUTTONS_NUM = 5; + static constexpr int ANALOG_SUB_BUTTONS_NUM = 4; /// Each button input is represented by a QPushButton. std::array<QPushButton*, Settings::NativeButton::NumButtons> button_map; - std::vector<QWidget*> debug_hidden; - std::vector<QWidget*> layout_hidden; - - /// A group of five QPushButtons represent one analog input. The buttons each represent up, - /// down, left, right, and modifier, respectively. + /// A group of four QPushButtons represent one analog input. The buttons each represent up, + /// down, left, right, respectively. std::array<std::array<QPushButton*, ANALOG_SUB_BUTTONS_NUM>, Settings::NativeAnalog::NumAnalogs> analog_map_buttons; - /// Analog inputs are also represented each with a single button, used to configure with an - /// actual analog stick - std::array<QPushButton*, Settings::NativeAnalog::NumAnalogs> analog_map_stick; - std::array<QSlider*, Settings::NativeAnalog::NumAnalogs> analog_map_deadzone; + /// Each motion input is represented by a QPushButton. + std::array<QPushButton*, Settings::NativeMotion::NumMotions> motion_map; + std::array<QLabel*, Settings::NativeAnalog::NumAnalogs> analog_map_deadzone_label; + std::array<QSlider*, Settings::NativeAnalog::NumAnalogs> analog_map_deadzone_slider; + std::array<QGroupBox*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_groupbox; + std::array<QPushButton*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_button; + std::array<QLabel*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_label; + std::array<QSlider*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_slider; + std::array<QGroupBox*, Settings::NativeAnalog::NumAnalogs> analog_map_range_groupbox; + std::array<QSpinBox*, Settings::NativeAnalog::NumAnalogs> analog_map_range_spinbox; static const std::array<std::string, ANALOG_SUB_BUTTONS_NUM> analog_sub_buttons; std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> device_pollers; + /// A flag to indicate that the "Map Analog Stick" pop-up has been shown and accepted once. + bool map_analog_stick_accepted{}; + /// A flag to indicate if keyboard keys are okay when configuring an input. If this is false, /// keyboard events are ignored. - bool want_keyboard_keys = false; + bool want_keyboard_mouse{}; + + /// List of physical devices users can map with. If a SDL backed device is selected, then you + /// can use this device to get a default mapping. + std::vector<Common::ParamPackage> input_devices; - std::array<QPushButton*, 4> controller_color_buttons; - std::array<QColor, 4> controller_colors; + /// Bottom row is where console wide settings are held, and its "owned" by the parent + /// ConfigureInput widget. On show, add this widget to the main layout. This will change the + /// parent of the widget to this widget (but thats fine). + QWidget* bottom_row; }; diff --git a/src/yuzu/configuration/configure_input_player.ui b/src/yuzu/configuration/configure_input_player.ui index 4b37746a1..1e78b4c10 100644 --- a/src/yuzu/configuration/configure_input_player.ui +++ b/src/yuzu/configuration/configure_input_player.ui @@ -1,1243 +1,3094 @@ <?xml version="1.0" encoding="UTF-8"?> <ui version="4.0"> <class>ConfigureInputPlayer</class> - <widget class="QDialog" name="ConfigureInputPlayer"> + <widget class="QWidget" name="ConfigureInputPlayer"> <property name="geometry"> <rect> <x>0</x> <y>0</y> - <width>408</width> - <height>731</height> + <width>780</width> + <height>487</height> </rect> </property> <property name="windowTitle"> <string>Configure Input</string> </property> - <layout class="QVBoxLayout" name="verticalLayout_5"> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> <item> - <layout class="QGridLayout" name="buttons"> - <item row="1" column="1"> - <widget class="QGroupBox" name="RStick"> - <property name="title"> - <string>Right Stick</string> - </property> - <property name="alignment"> - <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignVCenter</set> - </property> - <property name="flat"> - <bool>false</bool> + <layout class="QVBoxLayout" name="main"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <layout class="QHBoxLayout" name="top" stretch="0,1,2"> + <property name="spacing"> + <number>3</number> </property> - <property name="checkable"> - <bool>false</bool> + <property name="topMargin"> + <number>0</number> </property> - <layout class="QGridLayout" name="gridLayout_5"> - <item row="1" column="1"> - <layout class="QVBoxLayout" name="buttonRStickDownVerticalLayout"> + <item> + <widget class="QGroupBox" name="groupConnectedController"> + <property name="layoutDirection"> + <enum>Qt::LeftToRight</enum> + </property> + <property name="title"> + <string>Connect Controller</string> + </property> + <property name="flat"> + <bool>false</bool> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <property name="leftMargin"> + <number>5</number> + </property> + <property name="topMargin"> + <number>5</number> + </property> + <property name="rightMargin"> + <number>5</number> + </property> + <property name="bottomMargin"> + <number>5</number> + </property> <item> - <layout class="QHBoxLayout" name="buttonRStickDownHorizontalLayout"> + <widget class="QComboBox" name="comboControllerType"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>21</height> + </size> + </property> <item> - <widget class="QLabel" name="labelRStickDown"> - <property name="text"> - <string>Down:</string> - </property> - </widget> + <property name="text"> + <string>Pro Controller</string> + </property> </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonRStickDown"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="0" column="1"> - <layout class="QVBoxLayout" name="buttonRStickRightVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonRStickRightHorizontalLayout"> <item> - <widget class="QLabel" name="labelRStickRight"> - <property name="text"> - <string>Right:</string> - </property> - </widget> + <property name="text"> + <string>Dual Joycons</string> + </property> </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonRStickRight"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="3" column="0" colspan="2"> - <widget class="QPushButton" name="buttonRStickAnalog"> - <property name="text"> - <string>Set Analog Stick</string> - </property> - </widget> - </item> - <item row="0" column="0"> - <layout class="QVBoxLayout" name="buttonRStickLeftVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonRStickLeftHorizontalLayout"> <item> - <widget class="QLabel" name="labelRStickLeft"> - <property name="text"> - <string>Left:</string> - </property> - </widget> + <property name="text"> + <string>Left Joycon</string> + </property> </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonRStickLeft"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="0"> - <layout class="QVBoxLayout" name="buttonRStickUpVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonRStickUpHorizontalLayout"> <item> - <widget class="QLabel" name="labelRStickUp"> - <property name="text"> - <string>Up:</string> - </property> - </widget> + <property name="text"> + <string>Right Joycon</string> + </property> </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonRStickUp"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="2" column="0"> - <layout class="QVBoxLayout" name="buttonRStickPressedVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonRStickPressedHorizontalLayout"> <item> - <widget class="QLabel" name="labelRStickPressed"> - <property name="text"> - <string>Pressed:</string> - </property> - </widget> + <property name="text"> + <string>Handheld</string> + </property> </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonRStick"> - <property name="text"> - <string/> - </property> </widget> </item> </layout> - </item> - <item row="2" column="1"> - <layout class="QVBoxLayout" name="buttonRStickModVerticalLayout"> + </widget> + </item> + <item> + <widget class="QGroupBox" name="devicesGroup"> + <property name="title"> + <string>Input Device</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_5"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>5</number> + </property> + <property name="topMargin"> + <number>5</number> + </property> + <property name="rightMargin"> + <number>5</number> + </property> + <property name="bottomMargin"> + <number>5</number> + </property> <item> - <layout class="QHBoxLayout" name="buttonRStickModHorizontalLayout"> + <widget class="QComboBox" name="comboDevices"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>21</height> + </size> + </property> <item> - <widget class="QLabel" name="labelRStickMod"> - <property name="text"> - <string>Modifier:</string> - </property> - </widget> + <property name="text"> + <string>Any</string> + </property> </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonRStickMod"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="4" column="0" colspan="2"> - <layout class="QVBoxLayout" name="sliderRStickDeadzoneVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="sliderRStickDeadzoneHorizontalLayout"> <item> - <widget class="QLabel" name="labelRStickDeadzone"> - <property name="text"> - <string>Deadzone: 0</string> - </property> - <property name="alignment"> - <enum>Qt::AlignHCenter</enum> - </property> - </widget> + <property name="text"> + <string>Keyboard/Mouse</string> + </property> </item> - </layout> + </widget> </item> <item> - <widget class="QSlider" name="sliderRStickDeadzone"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> + <widget class="QPushButton" name="buttonRefreshDevices"> + <property name="minimumSize"> + <size> + <width>21</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>21</width> + <height>21</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true"/> </property> </widget> </item> </layout> - </item> - <item row="5" column="0"> - <spacer name="RStick_verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> + </widget> + </item> + <item> + <widget class="QGroupBox" name="profilesGroup"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>0</height> + </size> + </property> + <property name="title"> + <string>Profile</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_4" stretch="2,0,0,0"> + <property name="spacing"> + <number>3</number> </property> - <property name="sizeHint" stdset="0"> - <size> - <width>0</width> - <height>0</height> - </size> + <property name="leftMargin"> + <number>5</number> + </property> + <property name="topMargin"> + <number>5</number> + </property> + <property name="rightMargin"> + <number>5</number> + </property> + <property name="bottomMargin"> + <number>5</number> </property> - </spacer> - </item> - </layout> - </widget> - </item> - <item row="0" column="1"> - <widget class="QGroupBox" name="Dpad"> - <property name="title"> - <string>Directional Pad</string> - </property> - <property name="flat"> - <bool>false</bool> - </property> - <property name="checkable"> - <bool>false</bool> - </property> - <layout class="QGridLayout" name="gridLayout_2"> - <item row="1" column="0"> - <layout class="QVBoxLayout" name="buttonDpadUpVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonDpadUpHorizontalLayout"> - <item> - <widget class="QLabel" name="labelDpadUp"> - <property name="text"> - <string>Up:</string> - </property> - </widget> - </item> - </layout> - </item> <item> - <widget class="QPushButton" name="buttonDpadUp"> - <property name="text"> - <string/> + <widget class="QComboBox" name="comboProfiles"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>21</height> + </size> </property> </widget> </item> - </layout> - </item> - <item row="1" column="1"> - <layout class="QVBoxLayout" name="buttonDpadDownVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonDpadDownHorizontalLayout"> - <item> - <widget class="QLabel" name="labelDpadDown"> - <property name="text"> - <string>Down:</string> - </property> - </widget> - </item> - </layout> - </item> <item> - <widget class="QPushButton" name="buttonDpadDown"> + <widget class="QPushButton" name="buttonProfilesSave"> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> <property name="text"> - <string/> + <string>Save</string> </property> </widget> </item> - </layout> - </item> - <item row="0" column="0"> - <layout class="QVBoxLayout" name="buttonDpadLeftVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonDpadLeftHorizontalLayout"> - <item> - <widget class="QLabel" name="labelDpadLeft"> - <property name="minimumSize"> - <size> - <width>80</width> - <height>0</height> - </size> - </property> - <property name="text"> - <string>Left:</string> - </property> - </widget> - </item> - </layout> - </item> <item> - <widget class="QPushButton" name="buttonDpadLeft"> + <widget class="QPushButton" name="buttonProfilesNew"> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> <property name="text"> - <string/> + <string>New</string> </property> </widget> </item> - </layout> - </item> - <item row="0" column="1"> - <layout class="QVBoxLayout" name="buttonDpadRightVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonDpadRightHorizontalLayout"> - <item> - <widget class="QLabel" name="labelDpadRight"> - <property name="minimumSize"> - <size> - <width>80</width> - <height>0</height> - </size> - </property> - <property name="text"> - <string>Right:</string> - </property> - </widget> - </item> - </layout> - </item> <item> - <widget class="QPushButton" name="buttonDpadRight"> + <widget class="QPushButton" name="buttonProfilesDelete"> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> <property name="text"> - <string/> + <string>Delete</string> </property> </widget> </item> </layout> - </item> - </layout> - </widget> + </widget> + </item> + </layout> </item> - <item row="0" column="0"> - <widget class="QGroupBox" name="faceButtons"> - <property name="title"> - <string>Face Buttons</string> - </property> - <property name="flat"> - <bool>false</bool> - </property> - <property name="checkable"> - <bool>false</bool> + <item> + <widget class="QFrame" name="bottom"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Preferred" vsizetype="MinimumExpanding"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> </property> - <layout class="QGridLayout" name="gridLayout"> - <item row="0" column="0"> - <layout class="QVBoxLayout" name="buttonFaceButtonsAVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonFaceButtonsAHorizontalLayout"> - <item> - <widget class="QLabel" name="labelA"> - <property name="minimumSize"> - <size> - <width>80</width> - <height>0</height> - </size> + <layout class="QHBoxLayout" name="_2"> + <property name="sizeConstraint"> + <enum>QLayout::SetMinimumSize</enum> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QWidget" name="bottomLeft" native="true"> + <layout class="QVBoxLayout" name="bottomLeftLayout" stretch="0,0,0,0"> + <property name="spacing"> + <number>0</number> + </property> + <property name="sizeConstraint"> + <enum>QLayout::SetDefaultConstraint</enum> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="LStick"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="title"> + <string>Left Stick</string> + </property> + <property name="alignment"> + <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignTop</set> + </property> + <layout class="QVBoxLayout" name="verticalLayout_3"> + <property name="spacing"> + <number>0</number> </property> - <property name="text"> - <string>A:</string> + <property name="sizeConstraint"> + <enum>QLayout::SetDefaultConstraint</enum> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonA"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="0" column="1"> - <layout class="QVBoxLayout" name="buttonFaceButtonsBVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonFaceButtonsBHorizontalLayout"> - <item> - <widget class="QLabel" name="labelB"> - <property name="minimumSize"> - <size> - <width>80</width> - <height>0</height> - </size> + <property name="leftMargin"> + <number>3</number> </property> - <property name="text"> - <string>B:</string> + <property name="topMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonB"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="0"> - <layout class="QVBoxLayout" name="buttonFaceButtonsXVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonFaceButtonsXHorizontalLayout"> - <item> - <widget class="QLabel" name="labelX"> - <property name="text"> - <string>X:</string> + <property name="rightMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonX"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="1"> - <layout class="QVBoxLayout" name="buttonFaceButtonsYVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonFaceButtonsYHorizontalLayout"> - <item> - <widget class="QLabel" name="labelY"> - <property name="text"> - <string>Y:</string> + <property name="bottomMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonY"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - </layout> - </widget> - </item> - <item row="5" column="0" colspan="2"> - <widget class="QGroupBox" name="controller_color"> - <property name="title"> - <string>Controller Color</string> - </property> - <layout class="QGridLayout" name="gridLayout_10" columnstretch="0,0,0,0,0,0,0"> - <item row="0" column="0"> - <spacer name="horizontalSpacer_2"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="0" column="1"> - <widget class="QLabel" name="left_body_label"> - <property name="text"> - <string>Left Body</string> - </property> - </widget> - </item> - <item row="0" column="6"> - <spacer name="horizontalSpacer_3"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="1"> - <widget class="QLabel" name="left_buttons_label"> - <property name="minimumSize"> - <size> - <width>90</width> - <height>0</height> - </size> - </property> - <property name="text"> - <string>Left Buttons</string> - </property> - </widget> - </item> - <item row="1" column="5"> - <widget class="QPushButton" name="right_buttons_button"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="minimumSize"> - <size> - <width>32</width> - <height>0</height> - </size> - </property> - <property name="maximumSize"> - <size> - <width>40</width> - <height>16777215</height> - </size> - </property> - <property name="text"> - <string/> - </property> - </widget> - </item> - <item row="0" column="4"> - <widget class="QLabel" name="right_body_label"> - <property name="text"> - <string>Right Body</string> - </property> - </widget> - </item> - <item row="1" column="4"> - <widget class="QLabel" name="right_buttons_label"> - <property name="minimumSize"> - <size> - <width>90</width> - <height>0</height> - </size> - </property> - <property name="text"> - <string>Right Buttons</string> - </property> - </widget> - </item> - <item row="1" column="2"> - <widget class="QPushButton" name="left_buttons_button"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="minimumSize"> - <size> - <width>32</width> - <height>0</height> - </size> - </property> - <property name="maximumSize"> - <size> - <width>40</width> - <height>16777215</height> - </size> - </property> - <property name="text"> - <string/> - </property> - </widget> - </item> - <item row="0" column="2"> - <widget class="QPushButton" name="left_body_button"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="minimumSize"> - <size> - <width>32</width> - <height>0</height> - </size> - </property> - <property name="maximumSize"> - <size> - <width>40</width> - <height>16777215</height> - </size> - </property> - <property name="text"> - <string/> - </property> - </widget> - </item> - <item row="0" column="5"> - <widget class="QPushButton" name="right_body_button"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="minimumSize"> - <size> - <width>32</width> - <height>0</height> - </size> - </property> - <property name="maximumSize"> - <size> - <width>40</width> - <height>16777215</height> - </size> - </property> - <property name="text"> - <string/> - </property> - </widget> - </item> - <item row="0" column="3"> - <spacer name="horizontalSpacer_4"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeType"> - <enum>QSizePolicy::Fixed</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - </layout> - </widget> - </item> - <item row="1" column="0"> - <widget class="QGroupBox" name="LStick"> - <property name="title"> - <string>Left Stick</string> - </property> - <property name="flat"> - <bool>false</bool> - </property> - <property name="checkable"> - <bool>false</bool> - </property> - <layout class="QGridLayout" name="gridLayout_4"> - <item row="1" column="1"> - <layout class="QVBoxLayout" name="buttonLStickUpVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonLStickUpHorizontalLayout"> - <item> - <widget class="QLabel" name="labelLStickUp"> - <property name="text"> - <string>Up:</string> + <item> + <widget class="QWidget" name="buttonLStickUpWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_20"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerLStickUpLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonLStickUpGroup"> + <property name="title"> + <string>Up</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <property name="flat"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="buttonLStickUpVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonLStickUp"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Up</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerLStickUpRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="buttonLStickLeftRightHorizontaLayout"> + <property name="spacing"> + <number>3</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonLStickLeftGroup"> + <property name="title"> + <string>Left</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonLStickLeftVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonLStickLeft"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Left</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonLStickRightGroup"> + <property name="title"> + <string>Right</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonLStickRightVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonLStickRight"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Right</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QWidget" name="buttonLStickDownWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_22"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerLStickDownLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonLStickDownGroup"> + <property name="title"> + <string>Down</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonLStickDownVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonLStickDown"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Down</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerLStickDownRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="buttonLStickPressedModifierHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonLStickPressedGroup"> + <property name="title"> + <string>Pressed</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonLStickPressedVerticalLayout" stretch="0"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonLStick"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Pressed</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonLStickModGroup"> + <property name="title"> + <string>Modifier</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonLStickModVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonLStickMod"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Modifier</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="buttonLStickRangeGroup"> + <property name="title"> + <string>Range</string> + </property> + <layout class="QHBoxLayout" name="buttonLStickRangeGroupHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="spinboxLStickRange"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>50</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QVBoxLayout" name="sliderLStickDeadzoneModifierRangeVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="sizeConstraint"> + <enum>QLayout::SetDefaultConstraint</enum> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>2</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <layout class="QHBoxLayout" name="sliderLStickDeadzoneHorizontalLayout"> + <item> + <widget class="QLabel" name="labelLStickDeadzone"> + <property name="text"> + <string>Deadzone: 0%</string> + </property> + <property name="alignment"> + <set>Qt::AlignHCenter</set> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QSlider" name="sliderLStickDeadzone"> + <property name="maximum"> + <number>100</number> + </property> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="sliderLStickModifierRangeHorizontalLayout"> + <item> + <widget class="QLabel" name="labelLStickModifierRange"> + <property name="text"> + <string>Modifier Range: 0%</string> + </property> + <property name="alignment"> + <set>Qt::AlignHCenter</set> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QSlider" name="sliderLStickModifierRange"> + <property name="maximum"> + <number>100</number> + </property> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="verticalSpacerBottomLeft"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>0</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QGroupBox" name="Dpad"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="title"> + <string>D-Pad</string> + </property> + <property name="flat"> + <bool>false</bool> + </property> + <property name="checkable"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_5"> + <property name="spacing"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonLStickUp"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="0" column="2"> - <layout class="QVBoxLayout" name="buttonLStickRightVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonLStickRightHorizontalLayout"> - <item> - <widget class="QLabel" name="labelLStickRight"> - <property name="text"> - <string>Right:</string> + <property name="leftMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonLStickRight"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="4" column="1" colspan="2"> - <widget class="QPushButton" name="buttonLStickAnalog"> - <property name="text"> - <string>Set Analog Stick</string> - </property> - </widget> - </item> - <item row="0" column="1"> - <layout class="QVBoxLayout" name="buttonLStickLeftVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonLStickLeftHorizontalLayout_2"> - <item> - <widget class="QLabel" name="labelLStickLeft"> - <property name="text"> - <string>Left:</string> + <property name="topMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonLStickLeft"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="2"> - <layout class="QVBoxLayout" name="buttonLStickDownVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonLStickDownHorizontalLayout"> - <item> - <widget class="QLabel" name="labelLStickDown"> - <property name="text"> - <string>Down:</string> + <property name="rightMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonLStickDown"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="3" column="2"> - <layout class="QVBoxLayout" name="buttonLStickModVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonLStickModHorizontalLayout"> - <item> - <widget class="QLabel" name="labelLStickMod"> - <property name="text"> - <string>Modifier:</string> + <property name="bottomMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonLStickMod"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> + <item> + <widget class="QWidget" name="buttonDpadUpWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_23"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerDpadUpLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonDpadUpGroup"> + <property name="title"> + <string>Up</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonDpadUpVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonDpadUp"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Up</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerDpadUpRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="buttonDpadLeftRightHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonDpadLeftGroup"> + <property name="title"> + <string>Left</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonDpadLeftVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonDpadLeft"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Left</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonDpadRightGroup"> + <property name="title"> + <string>Right</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonDpadRightVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonDpadRight"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Right</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QWidget" name="buttonDpadDownWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_24"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerDpadDownLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonDpadDownGroup"> + <property name="title"> + <string>Down</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonDpadDownVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonDpadDown"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Down</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerDpadDownRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="verticalSpacerBottomLeft_2"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>0</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> </item> - <item row="3" column="1"> - <layout class="QVBoxLayout" name="buttonLStickPressedVerticalLayout" stretch="0,0"> - <item> - <layout class="QHBoxLayout" name="buttonLStickPressedHorizontalLayout"> - <item> - <widget class="QLabel" name="labelLStickPressed"> - <property name="text"> - <string>Pressed:</string> + <item> + <widget class="QWidget" name="bottomMiddle" native="true"> + <layout class="QVBoxLayout" stretch="0,0,0"> + <property name="spacing"> + <number>6</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <layout class="QHBoxLayout" name="shoulderButtons"> + <property name="spacing"> + <number>3</number> + </property> + <item> + <widget class="QWidget" name="buttonShoulderButtonsLeft" native="true"> + <layout class="QVBoxLayout" name="buttonShoulderButtonsLeftVerticalLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="buttonShoulderButtonsButtonLGroup"> + <property name="title"> + <string>L</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonShoulderButtonsLVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonL"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>L</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="buttonShoulderButtonsButtonZLGroup"> + <property name="title"> + <string>ZL</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonShoulderButtonsZLVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonZL"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>ZL</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="horizontalSpacerShoulderButtonsWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalSpacerShoulderButtonsWidgetLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerShoulderButtons1"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="buttonMiscButtonsMinusScreenshot" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsMinusScreenshotVerticalLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonMiscButtonsMinusGroup"> + <property name="title"> + <string>Minus</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonMiscMinusVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonMinus"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Minus</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonMiscButtonsScreenshotGroup"> + <property name="title"> + <string>Capture</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonMiscScrCapVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonScreenshot"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Capture</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="buttonMiscButtonsPlusHome" native="true"> + <layout class="QVBoxLayout" name="buttonMiscButtonsPlusHomeVerticalLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonMiscButtonsPlusGroup"> + <property name="title"> + <string>Plus</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonMiscPlusVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonPlus"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Plus</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonMiscButtonsHomeGroup"> + <property name="title"> + <string>Home</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonMiscHomeVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonHome"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Home</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="horizontalSpacerShoulderButtonsWidget3" native="true"> + <layout class="QHBoxLayout" name="horizontalSpacerShoulderButtonsWidget3Layout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerShoulderButtons2"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="buttonShoulderButtonsRight" native="true"> + <layout class="QVBoxLayout" name="buttonShoulderButtonsRightVerticalLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="buttonShoulderButtonsRGroup"> + <property name="title"> + <string>R</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonShoulderButtonsRVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonR"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>R</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="buttonShoulderButtonsZRGroup"> + <property name="title"> + <string>ZR</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonShoulderButtonsZRVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonZR"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>ZR</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="horizontalSpacerShoulderButtonsWidget2" native="true"> + <layout class="QHBoxLayout" name="horizontalSpacerShoulderButtonsWidget2Layout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerShoulderButtons3"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="buttonShoulderButtonsSLSR" native="true"> + <layout class="QVBoxLayout" name="buttonShoulderButtonsSLSRVerticalLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonShoulderButtonsSLGroup"> + <property name="title"> + <string>SL</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonShoulderButtonsSLVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonSL"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>SL</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonShoulderButtonsSRGroup"> + <property name="title"> + <string>SR</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonShoulderButtonsSRVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonSR"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>SR</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QFrame" name="controllerFrame"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Expanding" vsizetype="Expanding"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="font"> + <font> + <weight>75</weight> + <bold>true</bold> + </font> + </property> + <property name="styleSheet"> + <string notr="true">image: url(:/controller/pro);</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_4"> + <property name="leftMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonLStick"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="5" column="1" colspan="2"> - <layout class="QVBoxLayout" name="sliderLStickDeadzoneVerticalLayout"> - <property name="sizeConstraint"> - <enum>QLayout::SetDefaultConstraint</enum> - </property> - <item> - <layout class="QHBoxLayout" name="sliderLStickDeadzoneHorizontalLayout"> - <item> - <widget class="QLabel" name="labelLStickDeadzone"> - <property name="text"> - <string>Deadzone: 0</string> + <property name="topMargin"> + <number>0</number> </property> - <property name="alignment"> - <enum>Qt::AlignHCenter</enum> + <property name="rightMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QSlider" name="sliderLStickDeadzone"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - </widget> - </item> - </layout> - </item> - <item row="6" column="1"> - <spacer name="LStick_verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>0</width> - <height>0</height> - </size> - </property> - </spacer> - </item> - </layout> - </widget> - </item> - <item row="3" column="0"> - <widget class="QGroupBox" name="shoulderButtons"> - <property name="title"> - <string>Shoulder Buttons</string> - </property> - <property name="flat"> - <bool>false</bool> - </property> - <property name="checkable"> - <bool>false</bool> - </property> - <layout class="QGridLayout" name="gridLayout_3"> - <item row="0" column="0"> - <layout class="QVBoxLayout" name="buttonShoulderButtonsLVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonShoulderButtonsLHorizontalLayout"> - <item> - <widget class="QLabel" name="labelL"> - <property name="text"> - <string>L:</string> + <property name="bottomMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonL"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="miscButtons"> + <property name="spacing"> + <number>3</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerMiscButtons1"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QGroupBox" name="buttonMotionLeftGroup"> + <property name="title"> + <string>Motion 1</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonDpadLeftVerticalLayout_2"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonMotionLeft"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Left</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="buttonMotionRightGroup"> + <property name="title"> + <string>Motion 2</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonDpadRightVerticalLayout_2"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonMotionRight"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Right</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerMiscButtons4"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </item> + </layout> + </widget> </item> - <item row="0" column="1"> - <layout class="QVBoxLayout" name="buttonShoulderButtonsRVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonShoulderButtonsRHorizontalLayout"> - <item> - <widget class="QLabel" name="labelR"> - <property name="text"> - <string>R:</string> + <item> + <widget class="QWidget" name="bottomRight" native="true"> + <layout class="QVBoxLayout" name="bottomRightLayout"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="faceButtons"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="title"> + <string>Face Buttons</string> + </property> + <property name="flat"> + <bool>false</bool> + </property> + <property name="checkable"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <property name="spacing"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonR"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="0"> - <layout class="QVBoxLayout" name="buttonShoulderButtonsZLVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonShoulderButtonsZLHorizontalLayout"> - <item> - <widget class="QLabel" name="labelZL"> - <property name="text"> - <string>ZL:</string> + <property name="leftMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonZL"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="1"> - <layout class="QVBoxLayout" name="buttonShoulderButtonsZRVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonShoulderButtonsZRHorizontalLayout"> - <item> - <widget class="QLabel" name="labelZR"> - <property name="text"> - <string>ZR:</string> + <property name="topMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonZR"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="0" column="2"> - <layout class="QVBoxLayout" name="buttonShoulderButtonsSLVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonShoulderButtonsSLHorizontalLayout"> - <item> - <widget class="QLabel" name="labelSL"> - <property name="text"> - <string>SL:</string> + <property name="rightMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonSL"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="2"> - <layout class="QVBoxLayout" name="buttonShoulderButtonsSRVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonShoulderButtonsSRHorizontalLayout"> - <item> - <widget class="QLabel" name="labelSR"> - <property name="text"> - <string>SR:</string> + <property name="bottomMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonSR"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - </layout> - </widget> - </item> - <item row="3" column="1"> - <widget class="QGroupBox" name="misc"> - <property name="title"> - <string>Misc.</string> - </property> - <property name="flat"> - <bool>false</bool> - </property> - <property name="checkable"> - <bool>false</bool> - </property> - <layout class="QGridLayout" name="gridLayout_6"> - <item row="1" column="0"> - <layout class="QVBoxLayout" name="buttonMiscMinusVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonMiscMinusHorizontalLayout"> - <item> - <widget class="QLabel" name="labelMinus"> - <property name="text"> - <string>Minus:</string> + <item> + <widget class="QWidget" name="buttonFaceButtonsBWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_6"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerBLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonFaceButtonsXGroup"> + <property name="title"> + <string>X</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonFaceButtonsXVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonX"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>X</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerBRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="buttonFaceButtonsYAHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonFaceButtonsYGroup"> + <property name="title"> + <string>Y</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonFaceButtonsYVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonY"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Y</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonFaceButtonsAGroup"> + <property name="title"> + <string>A</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonFaceButtonsAVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonA"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>A</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QWidget" name="buttonFaceButtonsXWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_10"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerXLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonFaceButtonsBWidget_2"> + <property name="title"> + <string>B</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonFaceButtonsBVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonB"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>B</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerXRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="verticalSpacerBottomRight"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>0</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QGroupBox" name="RStick"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="title"> + <string>Right Stick</string> + </property> + <property name="alignment"> + <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignVCenter</set> + </property> + <property name="flat"> + <bool>false</bool> + </property> + <property name="checkable"> + <bool>false</bool> + </property> + <layout class="QVBoxLayout" name="verticalLayout_2"> + <property name="spacing"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonMinus"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="3" column="1"> - <spacer name="verticalSpacer_2"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - <item row="0" column="0"> - <layout class="QVBoxLayout" name="buttonMiscPlusVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonMiscPlusHorizontalLayout"> - <item> - <widget class="QLabel" name="labelPlus"> - <property name="text"> - <string>Plus:</string> + <property name="leftMargin"> + <number>3</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonPlus"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="0" column="1"> - <layout class="QVBoxLayout" name="buttonMiscHomeVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonMiscHomeHorizontalLayout"> - <item> - <widget class="QLabel" name="labelHome"> - <property name="text"> - <string>Home:</string> + <property name="topMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonHome"> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> - </item> - <item row="1" column="1"> - <layout class="QVBoxLayout" name="buttonMiscScrCapVerticalLayout"> - <item> - <layout class="QHBoxLayout" name="buttonMiscScrCapHorizontalLayout"> - <item> - <widget class="QLabel" name="labelScreenshot"> - <property name="text"> - <string>Screen Capture:</string> + <property name="rightMargin"> + <number>3</number> </property> - <property name="wordWrap"> - <bool>false</bool> + <property name="bottomMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="buttonScreenshot"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="maximumSize"> - <size> - <width>80</width> - <height>16777215</height> - </size> - </property> - <property name="text"> - <string/> - </property> - </widget> - </item> - </layout> + <item> + <widget class="QWidget" name="buttonRStickUpWidget" native="true"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>0</height> + </size> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_9"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerRStickUpLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonRStickUpGroup"> + <property name="title"> + <string>Up</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonRStickUpVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonRStickUp"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Up</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerRStickUpRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="buttonRStickLeftRightHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonRStickLeftGroup"> + <property name="title"> + <string>Left</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonRStickLeftVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonRStickLeft"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Left</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonRStickRightGroup"> + <property name="title"> + <string>Right</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonRStickRightVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonRStickRight"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Right</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QWidget" name="buttonRStickDownWidget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_11"> + <property name="spacing"> + <number>0</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <spacer name="horizontalSpacerRStickDownLeft"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QGroupBox" name="buttonRStickDownGroup"> + <property name="title"> + <string>Down</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonRStickDownVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonRStickDown"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Down</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="horizontalSpacerRStickDownRight"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="buttonRStickPressedModifierHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="groupRStickPressed"> + <property name="title"> + <string>Pressed</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonRStickPressedVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonRStick"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Pressed</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item alignment="Qt::AlignHCenter"> + <widget class="QGroupBox" name="buttonRStickModGroup"> + <property name="title"> + <string>Modifier</string> + </property> + <property name="alignment"> + <set>Qt::AlignCenter</set> + </property> + <layout class="QVBoxLayout" name="buttonRStickModVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QPushButton" name="buttonRStickMod"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> + </property> + <property name="text"> + <string>Modifier</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="buttonRStickRangeGroup"> + <property name="title"> + <string>Range</string> + </property> + <layout class="QHBoxLayout" name="buttonRStickRangeGroupHorizontalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="spinboxRStickRange"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>50</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QVBoxLayout" name="sliderRStickDeadzoneModifierRangeVerticalLayout"> + <property name="spacing"> + <number>3</number> + </property> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>2</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <layout class="QHBoxLayout" name="sliderRStickDeadzoneHorizontalLayout"> + <item> + <widget class="QLabel" name="labelRStickDeadzone"> + <property name="text"> + <string>Deadzone: 0%</string> + </property> + <property name="alignment"> + <set>Qt::AlignHCenter</set> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QSlider" name="sliderRStickDeadzone"> + <property name="maximum"> + <number>100</number> + </property> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="sliderRStickModifierRangeHorizontalLayout"> + <item> + <widget class="QLabel" name="labelRStickModifierRange"> + <property name="text"> + <string>Modifier Range: 0%</string> + </property> + <property name="alignment"> + <set>Qt::AlignHCenter</set> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QSlider" name="sliderRStickModifierRange"> + <property name="maximum"> + <number>100</number> + </property> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="verticalSpacerBottomRight_2"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>0</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> </item> </layout> </widget> </item> </layout> </item> - <item> - <spacer name="verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout"/> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_2"> - <item> - <widget class="QPushButton" name="buttonClearAll"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="sizeIncrement"> - <size> - <width>0</width> - <height>0</height> - </size> - </property> - <property name="baseSize"> - <size> - <width>0</width> - <height>0</height> - </size> - </property> - <property name="layoutDirection"> - <enum>Qt::LeftToRight</enum> - </property> - <property name="text"> - <string>Clear All</string> - </property> - </widget> - </item> - <item> - <widget class="QPushButton" name="buttonRestoreDefaults"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="sizeIncrement"> - <size> - <width>0</width> - <height>0</height> - </size> - </property> - <property name="baseSize"> - <size> - <width>0</width> - <height>0</height> - </size> - </property> - <property name="layoutDirection"> - <enum>Qt::LeftToRight</enum> - </property> - <property name="text"> - <string>Restore Defaults</string> - </property> - </widget> - </item> - <item> - <spacer name="horizontalSpacer"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item> - <widget class="QDialogButtonBox" name="buttonBox"> - <property name="standardButtons"> - <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> - </property> - </widget> - </item> - </layout> - </item> </layout> </widget> - <resources/> - <connections> - <connection> - <sender>buttonBox</sender> - <signal>accepted()</signal> - <receiver>ConfigureInputPlayer</receiver> - <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>371</x> - <y>730</y> - </hint> - <hint type="destinationlabel"> - <x>229</x> - <y>375</y> - </hint> - </hints> - </connection> - <connection> - <sender>buttonBox</sender> - <signal>rejected()</signal> - <receiver>ConfigureInputPlayer</receiver> - <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>371</x> - <y>730</y> - </hint> - <hint type="destinationlabel"> - <x>229</x> - <y>375</y> - </hint> - </hints> - </connection> - </connections> + <resources> + <include location="../../../dist/icons/controller/controller.qrc"/> + </resources> + <connections/> </ui> diff --git a/src/yuzu/configuration/configure_input_profile_dialog.cpp b/src/yuzu/configuration/configure_input_profile_dialog.cpp new file mode 100644 index 000000000..1f5cfa75b --- /dev/null +++ b/src/yuzu/configuration/configure_input_profile_dialog.cpp @@ -0,0 +1,37 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "ui_configure_input_profile_dialog.h" +#include "yuzu/configuration/configure_input_player.h" +#include "yuzu/configuration/configure_input_profile_dialog.h" + +ConfigureInputProfileDialog::ConfigureInputProfileDialog( + QWidget* parent, InputCommon::InputSubsystem* input_subsystem, InputProfiles* profiles) + : QDialog(parent), ui(std::make_unique<Ui::ConfigureInputProfileDialog>()), + profile_widget(new ConfigureInputPlayer(this, 9, nullptr, input_subsystem, profiles, false)) { + ui->setupUi(this); + + ui->controllerLayout->addWidget(profile_widget); + + connect(ui->clear_all_button, &QPushButton::clicked, this, + [this] { profile_widget->ClearAll(); }); + connect(ui->restore_defaults_button, &QPushButton::clicked, this, + [this] { profile_widget->RestoreDefaults(); }); + + RetranslateUI(); +} + +ConfigureInputProfileDialog::~ConfigureInputProfileDialog() = default; + +void ConfigureInputProfileDialog::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QDialog::changeEvent(event); +} + +void ConfigureInputProfileDialog::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_input_profile_dialog.h b/src/yuzu/configuration/configure_input_profile_dialog.h new file mode 100644 index 000000000..e6386bdbb --- /dev/null +++ b/src/yuzu/configuration/configure_input_profile_dialog.h @@ -0,0 +1,40 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QDialog> + +class QPushButton; + +class ConfigureInputPlayer; + +class InputProfiles; + +namespace InputCommon { +class InputSubsystem; +} + +namespace Ui { +class ConfigureInputProfileDialog; +} + +class ConfigureInputProfileDialog : public QDialog { + Q_OBJECT + +public: + explicit ConfigureInputProfileDialog(QWidget* parent, + InputCommon::InputSubsystem* input_subsystem, + InputProfiles* profiles); + ~ConfigureInputProfileDialog() override; + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + std::unique_ptr<Ui::ConfigureInputProfileDialog> ui; + + ConfigureInputPlayer* profile_widget; +}; diff --git a/src/yuzu/configuration/configure_input_profile_dialog.ui b/src/yuzu/configuration/configure_input_profile_dialog.ui new file mode 100644 index 000000000..726cf6905 --- /dev/null +++ b/src/yuzu/configuration/configure_input_profile_dialog.ui @@ -0,0 +1,71 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureInputProfileDialog</class> + <widget class="QDialog" name="ConfigureInputProfileDialog"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>70</width> + <height>540</height> + </rect> + </property> + <property name="windowTitle"> + <string>Create Input Profile</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <property name="spacing"> + <number>2</number> + </property> + <property name="leftMargin"> + <number>9</number> + </property> + <property name="topMargin"> + <number>9</number> + </property> + <property name="rightMargin"> + <number>9</number> + </property> + <property name="bottomMargin"> + <number>9</number> + </property> + <item> + <layout class="QHBoxLayout" name="controllerLayout"/> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout"> + <item> + <widget class="QPushButton" name="clear_all_button"> + <property name="text"> + <string>Clear</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="restore_defaults_button"> + <property name="text"> + <string>Defaults</string> + </property> + </widget> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="standardButtons"> + <set>QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + <resources/> + <connections> + <connection> + <sender>buttonBox</sender> + <signal>accepted()</signal> + <receiver>ConfigureInputProfileDialog</receiver> + <slot>accept()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/configure_input_simple.cpp b/src/yuzu/configuration/configure_input_simple.cpp deleted file mode 100644 index 0e0e8f113..000000000 --- a/src/yuzu/configuration/configure_input_simple.cpp +++ /dev/null @@ -1,152 +0,0 @@ -// Copyright 2016 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <array> -#include <tuple> - -#include "ui_configure_input_simple.h" -#include "yuzu/configuration/configure_input.h" -#include "yuzu/configuration/configure_input_player.h" -#include "yuzu/configuration/configure_input_simple.h" -#include "yuzu/uisettings.h" - -namespace { - -template <typename Dialog, typename... Args> -void CallConfigureDialog(ConfigureInputSimple* caller, Args&&... args) { - caller->ApplyConfiguration(); - Dialog dialog(caller, std::forward<Args>(args)...); - - const auto res = dialog.exec(); - if (res == QDialog::Accepted) { - dialog.ApplyConfiguration(); - } -} - -// OnProfileSelect functions should (when applicable): -// - Set controller types -// - Set controller enabled -// - Set docked mode -// - Set advanced controller config/enabled (i.e. debug, kbd, mouse, touch) -// -// OnProfileSelect function should NOT however: -// - Reset any button mappings -// - Open any dialogs -// - Block in any way - -constexpr std::size_t PLAYER_0_INDEX = 0; -constexpr std::size_t HANDHELD_INDEX = 8; - -void HandheldOnProfileSelect() { - Settings::values.players[HANDHELD_INDEX].connected = true; - Settings::values.players[HANDHELD_INDEX].type = Settings::ControllerType::DualJoycon; - - for (std::size_t player = 0; player < HANDHELD_INDEX; ++player) { - Settings::values.players[player].connected = false; - } - - Settings::values.use_docked_mode = false; - Settings::values.keyboard_enabled = false; - Settings::values.mouse_enabled = false; - Settings::values.debug_pad_enabled = false; - Settings::values.touchscreen.enabled = true; -} - -void DualJoyconsDockedOnProfileSelect() { - Settings::values.players[PLAYER_0_INDEX].connected = true; - Settings::values.players[PLAYER_0_INDEX].type = Settings::ControllerType::DualJoycon; - - for (std::size_t player = 1; player <= HANDHELD_INDEX; ++player) { - Settings::values.players[player].connected = false; - } - - Settings::values.use_docked_mode = true; - Settings::values.keyboard_enabled = false; - Settings::values.mouse_enabled = false; - Settings::values.debug_pad_enabled = false; - Settings::values.touchscreen.enabled = true; -} - -// Name, OnProfileSelect (called when selected in drop down), OnConfigure (called when configure -// is clicked) -using InputProfile = std::tuple<const char*, void (*)(), void (*)(ConfigureInputSimple*)>; - -constexpr std::array<InputProfile, 3> INPUT_PROFILES{{ - {QT_TR_NOOP("Single Player - Handheld - Undocked"), HandheldOnProfileSelect, - [](ConfigureInputSimple* caller) { - CallConfigureDialog<ConfigureInputPlayer>(caller, HANDHELD_INDEX, false); - }}, - {QT_TR_NOOP("Single Player - Dual Joycons - Docked"), DualJoyconsDockedOnProfileSelect, - [](ConfigureInputSimple* caller) { - CallConfigureDialog<ConfigureInputPlayer>(caller, PLAYER_0_INDEX, false); - }}, - {QT_TR_NOOP("Custom"), [] {}, CallConfigureDialog<ConfigureInput>}, -}}; - -} // namespace - -void ApplyInputProfileConfiguration(int profile_index) { - std::get<1>( - INPUT_PROFILES.at(std::min(profile_index, static_cast<int>(INPUT_PROFILES.size() - 1))))(); -} - -ConfigureInputSimple::ConfigureInputSimple(QWidget* parent) - : QWidget(parent), ui(std::make_unique<Ui::ConfigureInputSimple>()) { - ui->setupUi(this); - - for (const auto& profile : INPUT_PROFILES) { - const QString label = tr(std::get<0>(profile)); - ui->profile_combobox->addItem(label, label); - } - - connect(ui->profile_combobox, QOverload<int>::of(&QComboBox::currentIndexChanged), this, - &ConfigureInputSimple::OnSelectProfile); - connect(ui->profile_configure, &QPushButton::clicked, this, &ConfigureInputSimple::OnConfigure); - - LoadConfiguration(); -} - -ConfigureInputSimple::~ConfigureInputSimple() = default; - -void ConfigureInputSimple::ApplyConfiguration() { - auto index = ui->profile_combobox->currentIndex(); - // Make the stored index for "Custom" very large so that if new profiles are added it - // doesn't change. - if (index >= static_cast<int>(INPUT_PROFILES.size() - 1)) { - index = std::numeric_limits<int>::max(); - } - - UISettings::values.profile_index = index; -} - -void ConfigureInputSimple::changeEvent(QEvent* event) { - if (event->type() == QEvent::LanguageChange) { - RetranslateUI(); - } - - QWidget::changeEvent(event); -} - -void ConfigureInputSimple::RetranslateUI() { - ui->retranslateUi(this); -} - -void ConfigureInputSimple::LoadConfiguration() { - const auto index = UISettings::values.profile_index; - if (index >= static_cast<int>(INPUT_PROFILES.size()) || index < 0) { - ui->profile_combobox->setCurrentIndex(static_cast<int>(INPUT_PROFILES.size() - 1)); - } else { - ui->profile_combobox->setCurrentIndex(index); - } -} - -void ConfigureInputSimple::OnSelectProfile(int index) { - const auto old_docked = Settings::values.use_docked_mode; - ApplyInputProfileConfiguration(index); - OnDockedModeChanged(old_docked, Settings::values.use_docked_mode); -} - -void ConfigureInputSimple::OnConfigure() { - std::get<2>(INPUT_PROFILES.at(ui->profile_combobox->currentIndex()))(this); -} diff --git a/src/yuzu/configuration/configure_input_simple.h b/src/yuzu/configuration/configure_input_simple.h deleted file mode 100644 index bb5050224..000000000 --- a/src/yuzu/configuration/configure_input_simple.h +++ /dev/null @@ -1,43 +0,0 @@ -// Copyright 2016 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <memory> - -#include <QWidget> - -class QPushButton; -class QString; -class QTimer; - -namespace Ui { -class ConfigureInputSimple; -} - -// Used by configuration loader to apply a profile if the input is invalid. -void ApplyInputProfileConfiguration(int profile_index); - -class ConfigureInputSimple : public QWidget { - Q_OBJECT - -public: - explicit ConfigureInputSimple(QWidget* parent = nullptr); - ~ConfigureInputSimple() override; - - /// Save all button configurations to settings file - void ApplyConfiguration(); - -private: - void changeEvent(QEvent* event) override; - void RetranslateUI(); - - /// Load configuration settings. - void LoadConfiguration(); - - void OnSelectProfile(int index); - void OnConfigure(); - - std::unique_ptr<Ui::ConfigureInputSimple> ui; -}; diff --git a/src/yuzu/configuration/configure_input_simple.ui b/src/yuzu/configuration/configure_input_simple.ui deleted file mode 100644 index c4889caa9..000000000 --- a/src/yuzu/configuration/configure_input_simple.ui +++ /dev/null @@ -1,97 +0,0 @@ -<?xml version="1.0" encoding="UTF-8"?> -<ui version="4.0"> - <class>ConfigureInputSimple</class> - <widget class="QWidget" name="ConfigureInputSimple"> - <property name="geometry"> - <rect> - <x>0</x> - <y>0</y> - <width>473</width> - <height>685</height> - </rect> - </property> - <property name="windowTitle"> - <string>ConfigureInputSimple</string> - </property> - <layout class="QVBoxLayout" name="verticalLayout_5"> - <item> - <layout class="QVBoxLayout" name="verticalLayout"> - <item> - <widget class="QGroupBox" name="gridGroupBox"> - <property name="title"> - <string>Profile</string> - </property> - <layout class="QGridLayout" name="gridLayout"> - <item row="1" column="2"> - <widget class="QPushButton" name="profile_configure"> - <property name="text"> - <string>Configure</string> - </property> - </widget> - </item> - <item row="1" column="0"> - <spacer name="horizontalSpacer"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="3"> - <spacer name="horizontalSpacer_2"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="1" column="1"> - <widget class="QComboBox" name="profile_combobox"> - <property name="minimumSize"> - <size> - <width>250</width> - <height>0</height> - </size> - </property> - </widget> - </item> - <item row="0" column="1" colspan="2"> - <widget class="QLabel" name="label"> - <property name="text"> - <string>Choose a controller configuration:</string> - </property> - </widget> - </item> - </layout> - </widget> - </item> - </layout> - </item> - <item> - <spacer name="verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - </layout> - </widget> - <resources/> - <connections/> -</ui> diff --git a/src/yuzu/configuration/configure_motion_touch.cpp b/src/yuzu/configuration/configure_motion_touch.cpp new file mode 100644 index 000000000..170574d9b --- /dev/null +++ b/src/yuzu/configuration/configure_motion_touch.cpp @@ -0,0 +1,314 @@ +// Copyright 2018 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> +#include <QCloseEvent> +#include <QLabel> +#include <QMessageBox> +#include <QPushButton> +#include <QVBoxLayout> +#include "common/logging/log.h" +#include "core/settings.h" +#include "input_common/main.h" +#include "input_common/udp/client.h" +#include "input_common/udp/udp.h" +#include "ui_configure_motion_touch.h" +#include "yuzu/configuration/configure_motion_touch.h" +#include "yuzu/configuration/configure_touch_from_button.h" + +CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent, + const std::string& host, u16 port, + u8 pad_index, u16 client_id) + : QDialog(parent) { + layout = new QVBoxLayout; + status_label = new QLabel(tr("Communicating with the server...")); + cancel_button = new QPushButton(tr("Cancel")); + connect(cancel_button, &QPushButton::clicked, this, [this] { + if (!completed) { + job->Stop(); + } + accept(); + }); + layout->addWidget(status_label); + layout->addWidget(cancel_button); + setLayout(layout); + + using namespace InputCommon::CemuhookUDP; + job = std::make_unique<CalibrationConfigurationJob>( + host, port, pad_index, client_id, + [this](CalibrationConfigurationJob::Status status) { + QString text; + switch (status) { + case CalibrationConfigurationJob::Status::Ready: + text = tr("Touch the top left corner <br>of your touchpad."); + break; + case CalibrationConfigurationJob::Status::Stage1Completed: + text = tr("Now touch the bottom right corner <br>of your touchpad."); + break; + case CalibrationConfigurationJob::Status::Completed: + text = tr("Configuration completed!"); + break; + } + QMetaObject::invokeMethod(this, "UpdateLabelText", Q_ARG(QString, text)); + if (status == CalibrationConfigurationJob::Status::Completed) { + QMetaObject::invokeMethod(this, "UpdateButtonText", Q_ARG(QString, tr("OK"))); + } + }, + [this](u16 min_x_, u16 min_y_, u16 max_x_, u16 max_y_) { + completed = true; + min_x = min_x_; + min_y = min_y_; + max_x = max_x_; + max_y = max_y_; + }); +} + +CalibrationConfigurationDialog::~CalibrationConfigurationDialog() = default; + +void CalibrationConfigurationDialog::UpdateLabelText(const QString& text) { + status_label->setText(text); +} + +void CalibrationConfigurationDialog::UpdateButtonText(const QString& text) { + cancel_button->setText(text); +} + +constexpr std::array<std::pair<const char*, const char*>, 2> MotionProviders = {{ + {"motion_emu", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "Mouse (Right Click)")}, + {"cemuhookudp", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "CemuhookUDP")}, +}}; + +constexpr std::array<std::pair<const char*, const char*>, 2> TouchProviders = {{ + {"emu_window", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "Emulator Window")}, + {"cemuhookudp", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "CemuhookUDP")}, +}}; + +ConfigureMotionTouch::ConfigureMotionTouch(QWidget* parent, + InputCommon::InputSubsystem* input_subsystem_) + : QDialog(parent), input_subsystem{input_subsystem_}, + ui(std::make_unique<Ui::ConfigureMotionTouch>()) { + ui->setupUi(this); + for (const auto& [provider, name] : MotionProviders) { + ui->motion_provider->addItem(tr(name), QString::fromUtf8(provider)); + } + for (const auto& [provider, name] : TouchProviders) { + ui->touch_provider->addItem(tr(name), QString::fromUtf8(provider)); + } + + ui->udp_learn_more->setOpenExternalLinks(true); + ui->udp_learn_more->setText( + tr("<a " + "href='https://yuzu-emu.org/wiki/" + "using-a-controller-or-android-phone-for-motion-or-touch-input'><span " + "style=\"text-decoration: underline; color:#039be5;\">Learn More</span></a>")); + + SetConfiguration(); + UpdateUiDisplay(); + ConnectEvents(); +} + +ConfigureMotionTouch::~ConfigureMotionTouch() = default; + +void ConfigureMotionTouch::SetConfiguration() { + const Common::ParamPackage motion_param(Settings::values.motion_device); + const Common::ParamPackage touch_param(Settings::values.touch_device); + const std::string motion_engine = motion_param.Get("engine", "motion_emu"); + const std::string touch_engine = touch_param.Get("engine", "emu_window"); + + ui->motion_provider->setCurrentIndex( + ui->motion_provider->findData(QString::fromStdString(motion_engine))); + ui->touch_provider->setCurrentIndex( + ui->touch_provider->findData(QString::fromStdString(touch_engine))); + ui->touch_from_button_checkbox->setChecked(Settings::values.use_touch_from_button); + touch_from_button_maps = Settings::values.touch_from_button_maps; + for (const auto& touch_map : touch_from_button_maps) { + ui->touch_from_button_map->addItem(QString::fromStdString(touch_map.name)); + } + ui->touch_from_button_map->setCurrentIndex(Settings::values.touch_from_button_map_index); + ui->motion_sensitivity->setValue(motion_param.Get("sensitivity", 0.01f)); + + min_x = touch_param.Get("min_x", 100); + min_y = touch_param.Get("min_y", 50); + max_x = touch_param.Get("max_x", 1800); + max_y = touch_param.Get("max_y", 850); + + ui->udp_server->setText(QString::fromStdString(Settings::values.udp_input_address)); + ui->udp_port->setText(QString::number(Settings::values.udp_input_port)); + ui->udp_pad_index->setCurrentIndex(Settings::values.udp_pad_index); +} + +void ConfigureMotionTouch::UpdateUiDisplay() { + const QString motion_engine = ui->motion_provider->currentData().toString(); + const QString touch_engine = ui->touch_provider->currentData().toString(); + const QString cemuhook_udp = QStringLiteral("cemuhookudp"); + + if (motion_engine == QStringLiteral("motion_emu")) { + ui->motion_sensitivity_label->setVisible(true); + ui->motion_sensitivity->setVisible(true); + } else { + ui->motion_sensitivity_label->setVisible(false); + ui->motion_sensitivity->setVisible(false); + } + + if (touch_engine == cemuhook_udp) { + ui->touch_calibration->setVisible(true); + ui->touch_calibration_config->setVisible(true); + ui->touch_calibration_label->setVisible(true); + ui->touch_calibration->setText( + QStringLiteral("(%1, %2) - (%3, %4)").arg(min_x).arg(min_y).arg(max_x).arg(max_y)); + } else { + ui->touch_calibration->setVisible(false); + ui->touch_calibration_config->setVisible(false); + ui->touch_calibration_label->setVisible(false); + } + + if (motion_engine == cemuhook_udp || touch_engine == cemuhook_udp) { + ui->udp_config_group_box->setVisible(true); + } else { + ui->udp_config_group_box->setVisible(false); + } +} + +void ConfigureMotionTouch::ConnectEvents() { + connect(ui->motion_provider, qOverload<int>(&QComboBox::currentIndexChanged), this, + [this](int index) { UpdateUiDisplay(); }); + connect(ui->touch_provider, qOverload<int>(&QComboBox::currentIndexChanged), this, + [this](int index) { UpdateUiDisplay(); }); + connect(ui->udp_test, &QPushButton::clicked, this, &ConfigureMotionTouch::OnCemuhookUDPTest); + connect(ui->touch_calibration_config, &QPushButton::clicked, this, + &ConfigureMotionTouch::OnConfigureTouchCalibration); + connect(ui->touch_from_button_config_btn, &QPushButton::clicked, this, + &ConfigureMotionTouch::OnConfigureTouchFromButton); + connect(ui->buttonBox, &QDialogButtonBox::rejected, this, [this] { + if (CanCloseDialog()) { + reject(); + } + }); +} + +void ConfigureMotionTouch::OnCemuhookUDPTest() { + ui->udp_test->setEnabled(false); + ui->udp_test->setText(tr("Testing")); + udp_test_in_progress = true; + InputCommon::CemuhookUDP::TestCommunication( + ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()), + static_cast<u32>(ui->udp_pad_index->currentIndex()), 24872, + [this] { + LOG_INFO(Frontend, "UDP input test success"); + QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, true)); + }, + [this] { + LOG_ERROR(Frontend, "UDP input test failed"); + QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, false)); + }); +} + +void ConfigureMotionTouch::OnConfigureTouchCalibration() { + ui->touch_calibration_config->setEnabled(false); + ui->touch_calibration_config->setText(tr("Configuring")); + CalibrationConfigurationDialog dialog( + this, ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toUInt()), + static_cast<u8>(ui->udp_pad_index->currentIndex()), 24872); + dialog.exec(); + if (dialog.completed) { + min_x = dialog.min_x; + min_y = dialog.min_y; + max_x = dialog.max_x; + max_y = dialog.max_y; + LOG_INFO(Frontend, + "UDP touchpad calibration config success: min_x={}, min_y={}, max_x={}, max_y={}", + min_x, min_y, max_x, max_y); + UpdateUiDisplay(); + } else { + LOG_ERROR(Frontend, "UDP touchpad calibration config failed"); + } + ui->touch_calibration_config->setEnabled(true); + ui->touch_calibration_config->setText(tr("Configure")); +} + +void ConfigureMotionTouch::closeEvent(QCloseEvent* event) { + if (CanCloseDialog()) { + event->accept(); + } else { + event->ignore(); + } +} + +void ConfigureMotionTouch::ShowUDPTestResult(bool result) { + udp_test_in_progress = false; + if (result) { + QMessageBox::information(this, tr("Test Successful"), + tr("Successfully received data from the server.")); + } else { + QMessageBox::warning(this, tr("Test Failed"), + tr("Could not receive valid data from the server.<br>Please verify " + "that the server is set up correctly and " + "the address and port are correct.")); + } + ui->udp_test->setEnabled(true); + ui->udp_test->setText(tr("Test")); +} + +void ConfigureMotionTouch::OnConfigureTouchFromButton() { + ConfigureTouchFromButton dialog{this, touch_from_button_maps, input_subsystem, + ui->touch_from_button_map->currentIndex()}; + if (dialog.exec() != QDialog::Accepted) { + return; + } + touch_from_button_maps = dialog.GetMaps(); + + while (ui->touch_from_button_map->count() > 0) { + ui->touch_from_button_map->removeItem(0); + } + for (const auto& touch_map : touch_from_button_maps) { + ui->touch_from_button_map->addItem(QString::fromStdString(touch_map.name)); + } + ui->touch_from_button_map->setCurrentIndex(dialog.GetSelectedIndex()); +} + +bool ConfigureMotionTouch::CanCloseDialog() { + if (udp_test_in_progress) { + QMessageBox::warning(this, tr("Citra"), + tr("UDP Test or calibration configuration is in progress.<br>Please " + "wait for them to finish.")); + return false; + } + return true; +} + +void ConfigureMotionTouch::ApplyConfiguration() { + if (!CanCloseDialog()) { + return; + } + + std::string motion_engine = ui->motion_provider->currentData().toString().toStdString(); + std::string touch_engine = ui->touch_provider->currentData().toString().toStdString(); + + Common::ParamPackage motion_param{}, touch_param{}; + motion_param.Set("engine", std::move(motion_engine)); + touch_param.Set("engine", std::move(touch_engine)); + + if (motion_engine == "motion_emu") { + motion_param.Set("sensitivity", static_cast<float>(ui->motion_sensitivity->value())); + } + + if (touch_engine == "cemuhookudp") { + touch_param.Set("min_x", min_x); + touch_param.Set("min_y", min_y); + touch_param.Set("max_x", max_x); + touch_param.Set("max_y", max_y); + } + + Settings::values.motion_device = motion_param.Serialize(); + Settings::values.touch_device = touch_param.Serialize(); + Settings::values.use_touch_from_button = ui->touch_from_button_checkbox->isChecked(); + Settings::values.touch_from_button_map_index = ui->touch_from_button_map->currentIndex(); + Settings::values.touch_from_button_maps = touch_from_button_maps; + Settings::values.udp_input_address = ui->udp_server->text().toStdString(); + Settings::values.udp_input_port = static_cast<u16>(ui->udp_port->text().toInt()); + Settings::values.udp_pad_index = static_cast<u8>(ui->udp_pad_index->currentIndex()); + input_subsystem->ReloadInputDevices(); + + accept(); +} diff --git a/src/yuzu/configuration/configure_motion_touch.h b/src/yuzu/configuration/configure_motion_touch.h new file mode 100644 index 000000000..3d4b5d659 --- /dev/null +++ b/src/yuzu/configuration/configure_motion_touch.h @@ -0,0 +1,90 @@ +// Copyright 2018 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QDialog> +#include "common/param_package.h" + +class QLabel; +class QPushButton; +class QVBoxLayout; + +namespace InputCommon { +class InputSubsystem; +} + +namespace InputCommon::CemuhookUDP { +class CalibrationConfigurationJob; +} + +namespace Ui { +class ConfigureMotionTouch; +} + +/// A dialog for touchpad calibration configuration. +class CalibrationConfigurationDialog : public QDialog { + Q_OBJECT +public: + explicit CalibrationConfigurationDialog(QWidget* parent, const std::string& host, u16 port, + u8 pad_index, u16 client_id); + ~CalibrationConfigurationDialog() override; + +private: + Q_INVOKABLE void UpdateLabelText(const QString& text); + Q_INVOKABLE void UpdateButtonText(const QString& text); + + QVBoxLayout* layout; + QLabel* status_label; + QPushButton* cancel_button; + std::unique_ptr<InputCommon::CemuhookUDP::CalibrationConfigurationJob> job; + + // Configuration results + bool completed{}; + u16 min_x{}; + u16 min_y{}; + u16 max_x{}; + u16 max_y{}; + + friend class ConfigureMotionTouch; +}; + +class ConfigureMotionTouch : public QDialog { + Q_OBJECT + +public: + explicit ConfigureMotionTouch(QWidget* parent, InputCommon::InputSubsystem* input_subsystem_); + ~ConfigureMotionTouch() override; + +public slots: + void ApplyConfiguration(); + +private slots: + void OnCemuhookUDPTest(); + void OnConfigureTouchCalibration(); + void OnConfigureTouchFromButton(); + +private: + void closeEvent(QCloseEvent* event) override; + Q_INVOKABLE void ShowUDPTestResult(bool result); + void SetConfiguration(); + void UpdateUiDisplay(); + void ConnectEvents(); + bool CanCloseDialog(); + + InputCommon::InputSubsystem* input_subsystem; + + std::unique_ptr<Ui::ConfigureMotionTouch> ui; + + // Coordinate system of the CemuhookUDP touch provider + int min_x{}; + int min_y{}; + int max_x{}; + int max_y{}; + + bool udp_test_in_progress{}; + + std::vector<Settings::TouchFromButtonMap> touch_from_button_maps; +}; diff --git a/src/yuzu/configuration/configure_motion_touch.ui b/src/yuzu/configuration/configure_motion_touch.ui new file mode 100644 index 000000000..5b78c5a4b --- /dev/null +++ b/src/yuzu/configuration/configure_motion_touch.ui @@ -0,0 +1,317 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureMotionTouch</class> + <widget class="QDialog" name="ConfigureMotionTouch"> + <property name="windowTitle"> + <string>Configure Motion / Touch</string> + </property> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>500</width> + <height>450</height> + </rect> + </property> + <layout class="QVBoxLayout"> + <item> + <widget class="QGroupBox" name="motion_group_box"> + <property name="title"> + <string>Motion</string> + </property> + <layout class="QVBoxLayout"> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="motion_provider_label"> + <property name="text"> + <string>Motion Provider:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="motion_provider"/> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="motion_sensitivity_label"> + <property name="text"> + <string>Sensitivity:</string> + </property> + </widget> + </item> + <item> + <widget class="QDoubleSpinBox" name="motion_sensitivity"> + <property name="alignment"> + <set>Qt::AlignRight|Qt::AlignTrailing|Qt::AlignVCenter</set> + </property> + <property name="decimals"> + <number>4</number> + </property> + <property name="minimum"> + <double>0.010000000000000</double> + </property> + <property name="maximum"> + <double>10.000000000000000</double> + </property> + <property name="singleStep"> + <double>0.001000000000000</double> + </property> + <property name="value"> + <double>0.010000000000000</double> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="touch_group_box"> + <property name="title"> + <string>Touch</string> + </property> + <layout class="QVBoxLayout"> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="touch_provider_label"> + <property name="text"> + <string>Touch Provider:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="touch_provider"/> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="touch_calibration_label"> + <property name="text"> + <string>Calibration:</string> + </property> + </widget> + </item> + <item> + <widget class="QLabel" name="touch_calibration"> + <property name="text"> + <string>(100, 50) - (1800, 850)</string> + </property> + <property name="alignment"> + <set>Qt::AlignRight|Qt::AlignTrailing|Qt::AlignVCenter</set> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="touch_calibration_config"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QCheckBox" name="touch_from_button_checkbox"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Use button mapping:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="touch_from_button_map"/> + </item> + <item> + <widget class="QPushButton" name="touch_from_button_config_btn"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Configure</string> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="udp_config_group_box"> + <property name="title"> + <string>CemuhookUDP Config</string> + </property> + <layout class="QVBoxLayout"> + <item> + <widget class="QLabel" name="udp_help"> + <property name="text"> + <string>You may use any Cemuhook compatible UDP input source to provide motion and touch input.</string> + </property> + <property name="alignment"> + <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignTop</set> + </property> + <property name="wordWrap"> + <bool>true</bool> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="udp_server_label"> + <property name="text"> + <string>Server:</string> + </property> + </widget> + </item> + <item> + <widget class="QLineEdit" name="udp_server"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="udp_port_label"> + <property name="text"> + <string>Port:</string> + </property> + </widget> + </item> + <item> + <widget class="QLineEdit" name="udp_port"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="udp_pad_index_label"> + <property name="text"> + <string>Pad:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="udp_pad_index"> + <item> + <property name="text"> + <string>Pad 1</string> + </property> + </item> + <item> + <property name="text"> + <string>Pad 2</string> + </property> + </item> + <item> + <property name="text"> + <string>Pad 3</string> + </property> + </item> + <item> + <property name="text"> + <string>Pad 4</string> + </property> + </item> + </widget> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout"> + <item> + <widget class="QLabel" name="udp_learn_more"> + <property name="text"> + <string>Learn More</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="udp_test"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Test</string> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <spacer> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>167</width> + <height>55</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections> + <connection> + <sender>buttonBox</sender> + <signal>accepted()</signal> + <receiver>ConfigureMotionTouch</receiver> + <slot>ApplyConfiguration()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/configure_mouse_advanced.cpp b/src/yuzu/configuration/configure_mouse_advanced.cpp index e0647ea5b..2af3afda8 100644 --- a/src/yuzu/configuration/configure_mouse_advanced.cpp +++ b/src/yuzu/configuration/configure_mouse_advanced.cpp @@ -18,6 +18,16 @@ static QString GetKeyName(int key_code) { switch (key_code) { + case Qt::LeftButton: + return QObject::tr("Click 0"); + case Qt::RightButton: + return QObject::tr("Click 1"); + case Qt::MiddleButton: + return QObject::tr("Click 2"); + case Qt::BackButton: + return QObject::tr("Click 3"); + case Qt::ForwardButton: + return QObject::tr("Click 4"); case Qt::Key_Shift: return QObject::tr("Shift"); case Qt::Key_Control: @@ -66,8 +76,10 @@ static QString ButtonToText(const Common::ParamPackage& param) { return QObject::tr("[unknown]"); } -ConfigureMouseAdvanced::ConfigureMouseAdvanced(QWidget* parent) - : QDialog(parent), ui(std::make_unique<Ui::ConfigureMouseAdvanced>()), +ConfigureMouseAdvanced::ConfigureMouseAdvanced(QWidget* parent, + InputCommon::InputSubsystem* input_subsystem_) + : QDialog(parent), + ui(std::make_unique<Ui::ConfigureMouseAdvanced>()), input_subsystem{input_subsystem_}, timeout_timer(std::make_unique<QTimer>()), poll_timer(std::make_unique<QTimer>()) { ui->setupUi(this); setFocusPolicy(Qt::ClickFocus); @@ -83,25 +95,29 @@ ConfigureMouseAdvanced::ConfigureMouseAdvanced(QWidget* parent) } button->setContextMenuPolicy(Qt::CustomContextMenu); - connect(button, &QPushButton::clicked, [=] { + connect(button, &QPushButton::clicked, [=, this] { HandleClick( button_map[button_id], - [=](const Common::ParamPackage& params) { buttons_param[button_id] = params; }, + [=, this](const Common::ParamPackage& params) { + buttons_param[button_id] = params; + }, InputCommon::Polling::DeviceType::Button); }); - connect(button, &QPushButton::customContextMenuRequested, [=](const QPoint& menu_location) { - QMenu context_menu; - context_menu.addAction(tr("Clear"), [&] { - buttons_param[button_id].Clear(); - button_map[button_id]->setText(tr("[not set]")); - }); - context_menu.addAction(tr("Restore Default"), [&] { - buttons_param[button_id] = Common::ParamPackage{ - InputCommon::GenerateKeyboardParam(Config::default_mouse_buttons[button_id])}; - button_map[button_id]->setText(ButtonToText(buttons_param[button_id])); - }); - context_menu.exec(button_map[button_id]->mapToGlobal(menu_location)); - }); + connect(button, &QPushButton::customContextMenuRequested, + [=, this](const QPoint& menu_location) { + QMenu context_menu; + context_menu.addAction(tr("Clear"), [&] { + buttons_param[button_id].Clear(); + button_map[button_id]->setText(tr("[not set]")); + }); + context_menu.addAction(tr("Restore Default"), [&] { + buttons_param[button_id] = + Common::ParamPackage{InputCommon::GenerateKeyboardParam( + Config::default_mouse_buttons[button_id])}; + button_map[button_id]->setText(ButtonToText(buttons_param[button_id])); + }); + context_menu.exec(button_map[button_id]->mapToGlobal(menu_location)); + }); } connect(ui->buttonClearAll, &QPushButton::clicked, [this] { ClearAll(); }); @@ -184,9 +200,9 @@ void ConfigureMouseAdvanced::HandleClick( button->setText(tr("[press key]")); button->setFocus(); - // Keyboard keys can only be used as button devices - want_keyboard_keys = type == InputCommon::Polling::DeviceType::Button; - if (want_keyboard_keys) { + // Keyboard keys or mouse buttons can only be used as button devices + want_keyboard_mouse = type == InputCommon::Polling::DeviceType::Button; + if (want_keyboard_mouse) { const auto iter = std::find(button_map.begin(), button_map.end(), button); ASSERT(iter != button_map.end()); const auto index = std::distance(button_map.begin(), iter); @@ -195,27 +211,29 @@ void ConfigureMouseAdvanced::HandleClick( input_setter = new_input_setter; - device_pollers = InputCommon::Polling::GetPollers(type); + device_pollers = input_subsystem->GetPollers(type); for (auto& poller : device_pollers) { poller->Start(); } - grabKeyboard(); - grabMouse(); - timeout_timer->start(5000); // Cancel after 5 seconds - poll_timer->start(200); // Check for new inputs every 200ms + QWidget::grabMouse(); + QWidget::grabKeyboard(); + + timeout_timer->start(2500); // Cancel after 2.5 seconds + poll_timer->start(50); // Check for new inputs every 50ms } void ConfigureMouseAdvanced::SetPollingResult(const Common::ParamPackage& params, bool abort) { - releaseKeyboard(); - releaseMouse(); timeout_timer->stop(); poll_timer->stop(); for (auto& poller : device_pollers) { poller->Stop(); } + QWidget::releaseMouse(); + QWidget::releaseKeyboard(); + if (!abort) { (*input_setter)(params); } @@ -224,13 +242,29 @@ void ConfigureMouseAdvanced::SetPollingResult(const Common::ParamPackage& params input_setter = std::nullopt; } +void ConfigureMouseAdvanced::mousePressEvent(QMouseEvent* event) { + if (!input_setter || !event) { + return; + } + + if (want_keyboard_mouse) { + SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->button())}, + false); + } else { + // We don't want any mouse buttons, so don't stop polling + return; + } + + SetPollingResult({}, true); +} + void ConfigureMouseAdvanced::keyPressEvent(QKeyEvent* event) { if (!input_setter || !event) { return; } if (event->key() != Qt::Key_Escape) { - if (want_keyboard_keys) { + if (want_keyboard_mouse) { SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->key())}, false); } else { diff --git a/src/yuzu/configuration/configure_mouse_advanced.h b/src/yuzu/configuration/configure_mouse_advanced.h index 342b82412..65b6fca9a 100644 --- a/src/yuzu/configuration/configure_mouse_advanced.h +++ b/src/yuzu/configuration/configure_mouse_advanced.h @@ -8,12 +8,14 @@ #include <optional> #include <QDialog> -#include "core/settings.h" - class QCheckBox; class QPushButton; class QTimer; +namespace InputCommon { +class InputSubsystem; +} + namespace Ui { class ConfigureMouseAdvanced; } @@ -22,7 +24,7 @@ class ConfigureMouseAdvanced : public QDialog { Q_OBJECT public: - explicit ConfigureMouseAdvanced(QWidget* parent); + explicit ConfigureMouseAdvanced(QWidget* parent, InputCommon::InputSubsystem* input_subsystem_); ~ConfigureMouseAdvanced() override; void ApplyConfiguration(); @@ -49,11 +51,16 @@ private: /// Finish polling and configure input using the input_setter void SetPollingResult(const Common::ParamPackage& params, bool abort); + /// Handle mouse button press events. + void mousePressEvent(QMouseEvent* event) override; + /// Handle key press events. void keyPressEvent(QKeyEvent* event) override; std::unique_ptr<Ui::ConfigureMouseAdvanced> ui; + InputCommon::InputSubsystem* input_subsystem; + /// This will be the the setting function when an input is awaiting configuration. std::optional<std::function<void(const Common::ParamPackage&)>> input_setter; @@ -67,5 +74,5 @@ private: /// A flag to indicate if keyboard keys are okay when configuring an input. If this is false, /// keyboard events are ignored. - bool want_keyboard_keys = false; + bool want_keyboard_mouse = false; }; diff --git a/src/yuzu/configuration/configure_mouse_advanced.ui b/src/yuzu/configuration/configure_mouse_advanced.ui index 08245ecf0..5b99e1c37 100644 --- a/src/yuzu/configuration/configure_mouse_advanced.ui +++ b/src/yuzu/configuration/configure_mouse_advanced.ui @@ -6,13 +6,18 @@ <rect> <x>0</x> <y>0</y> - <width>250</width> - <height>261</height> + <width>310</width> + <height>193</height> </rect> </property> <property name="windowTitle"> <string>Configure Mouse</string> </property> + <property name="styleSheet"> + <string notr="true">QPushButton { + min-width: 60px; +}</string> + </property> <layout class="QVBoxLayout" name="verticalLayout"> <item> <widget class="QGroupBox" name="gridGroupBox"> @@ -20,81 +25,33 @@ <string>Mouse Buttons</string> </property> <layout class="QGridLayout" name="gridLayout"> - <item row="0" column="4"> - <spacer name="horizontalSpacer_2"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeType"> - <enum>QSizePolicy::Fixed</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="0" column="3"> - <layout class="QVBoxLayout" name="verticalLayout_4"> + <item row="3" column="5"> + <layout class="QVBoxLayout" name="verticalLayout_6"> <item> - <layout class="QHBoxLayout" name="horizontalLayout_3"> + <layout class="QHBoxLayout" name="horizontalLayout_5"> <item> - <widget class="QLabel" name="label_3"> + <widget class="QLabel" name="label_5"> <property name="text"> - <string>Right:</string> + <string>Forward:</string> </property> </widget> </item> </layout> </item> <item> - <widget class="QPushButton" name="right_button"> + <widget class="QPushButton" name="forward_button"> <property name="minimumSize"> <size> - <width>75</width> + <width>68</width> <height>0</height> </size> </property> - <property name="text"> - <string/> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> </property> - </widget> - </item> - </layout> - </item> - <item row="0" column="0"> - <spacer name="horizontalSpacer"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeType"> - <enum>QSizePolicy::Fixed</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item row="2" column="1"> - <layout class="QVBoxLayout" name="verticalLayout_3"> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_2"> - <item> - <widget class="QLabel" name="label_2"> - <property name="text"> - <string>Middle:</string> - </property> - </widget> - </item> - </layout> - </item> - <item> - <widget class="QPushButton" name="middle_button"> <property name="text"> <string/> </property> @@ -123,6 +80,12 @@ </item> <item> <widget class="QPushButton" name="back_button"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> <property name="text"> <string/> </property> @@ -147,7 +110,7 @@ <widget class="QPushButton" name="left_button"> <property name="minimumSize"> <size> - <width>75</width> + <width>68</width> <height>0</height> </size> </property> @@ -158,21 +121,99 @@ </item> </layout> </item> - <item row="3" column="3"> - <layout class="QVBoxLayout" name="verticalLayout_6"> + <item row="0" column="3"> + <layout class="QVBoxLayout" name="verticalLayout_3"> <item> - <layout class="QHBoxLayout" name="horizontalLayout_5"> + <layout class="QHBoxLayout" name="horizontalLayout_2"> <item> - <widget class="QLabel" name="label_5"> + <widget class="QLabel" name="label_2"> <property name="text"> - <string>Forward:</string> + <string>Middle:</string> </property> </widget> </item> </layout> </item> <item> - <widget class="QPushButton" name="forward_button"> + <widget class="QPushButton" name="middle_button"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="text"> + <string/> + </property> + </widget> + </item> + </layout> + </item> + <item row="0" column="6"> + <spacer name="horizontalSpacer_2"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeType"> + <enum>QSizePolicy::Fixed</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item row="0" column="0"> + <spacer name="horizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeType"> + <enum>QSizePolicy::Fixed</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item row="0" column="5"> + <layout class="QVBoxLayout" name="verticalLayout_4"> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <item> + <widget class="QLabel" name="label_3"> + <property name="text"> + <string>Right:</string> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QPushButton" name="right_button"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> <property name="text"> <string/> </property> @@ -180,6 +221,32 @@ </item> </layout> </item> + <item row="0" column="2"> + <spacer name="horizontalSpacer_4"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item row="0" column="4"> + <spacer name="horizontalSpacer_5"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>0</width> + <height>20</height> + </size> + </property> + </spacer> + </item> </layout> </widget> </item> @@ -187,15 +254,39 @@ <layout class="QHBoxLayout" name="horizontalLayout_6"> <item> <widget class="QPushButton" name="buttonClearAll"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> <property name="text"> - <string>Clear All</string> + <string>Clear</string> </property> </widget> </item> <item> <widget class="QPushButton" name="buttonRestoreDefaults"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>0</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> <property name="text"> - <string>Restore Defaults</string> + <string>Defaults</string> </property> </widget> </item> @@ -206,21 +297,24 @@ </property> <property name="sizeHint" stdset="0"> <size> - <width>40</width> + <width>0</width> <height>20</height> </size> </property> </spacer> </item> + <item> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="styleSheet"> + <string notr="true"/> + </property> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> </layout> </item> - <item> - <widget class="QDialogButtonBox" name="buttonBox"> - <property name="standardButtons"> - <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> - </property> - </widget> - </item> </layout> </widget> <resources/> @@ -230,32 +324,12 @@ <signal>accepted()</signal> <receiver>ConfigureMouseAdvanced</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>124</x> - <y>266</y> - </hint> - <hint type="destinationlabel"> - <x>124</x> - <y>143</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureMouseAdvanced</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>124</x> - <y>266</y> - </hint> - <hint type="destinationlabel"> - <x>124</x> - <y>143</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_per_game.cpp b/src/yuzu/configuration/configure_per_game.cpp new file mode 100644 index 000000000..8eac3bd9d --- /dev/null +++ b/src/yuzu/configuration/configure_per_game.cpp @@ -0,0 +1,144 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <memory> +#include <utility> + +#include <QCheckBox> +#include <QHeaderView> +#include <QMenu> +#include <QStandardItemModel> +#include <QString> +#include <QTimer> +#include <QTreeView> + +#include "common/common_paths.h" +#include "common/file_util.h" +#include "core/core.h" +#include "core/file_sys/control_metadata.h" +#include "core/file_sys/patch_manager.h" +#include "core/file_sys/xts_archive.h" +#include "core/loader/loader.h" +#include "ui_configure_per_game.h" +#include "yuzu/configuration/config.h" +#include "yuzu/configuration/configure_input.h" +#include "yuzu/configuration/configure_per_game.h" +#include "yuzu/uisettings.h" +#include "yuzu/util/util.h" + +ConfigurePerGame::ConfigurePerGame(QWidget* parent, u64 title_id) + : QDialog(parent), ui(std::make_unique<Ui::ConfigurePerGame>()), title_id(title_id) { + game_config = std::make_unique<Config>(fmt::format("{:016X}", title_id), + Config::ConfigType::PerGameConfig); + + Settings::SetConfiguringGlobal(false); + + ui->setupUi(this); + setFocusPolicy(Qt::ClickFocus); + setWindowTitle(tr("Properties")); + + ui->addonsTab->SetTitleId(title_id); + + scene = new QGraphicsScene; + ui->icon_view->setScene(scene); + + LoadConfiguration(); +} + +ConfigurePerGame::~ConfigurePerGame() = default; + +void ConfigurePerGame::ApplyConfiguration() { + ui->addonsTab->ApplyConfiguration(); + ui->generalTab->ApplyConfiguration(); + ui->systemTab->ApplyConfiguration(); + ui->graphicsTab->ApplyConfiguration(); + ui->graphicsAdvancedTab->ApplyConfiguration(); + ui->audioTab->ApplyConfiguration(); + + Settings::Apply(); + Settings::LogSettings(); + + game_config->Save(); +} + +void ConfigurePerGame::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QDialog::changeEvent(event); +} + +void ConfigurePerGame::RetranslateUI() { + ui->retranslateUi(this); +} + +void ConfigurePerGame::LoadFromFile(FileSys::VirtualFile file) { + this->file = std::move(file); + LoadConfiguration(); +} + +void ConfigurePerGame::LoadConfiguration() { + if (file == nullptr) { + return; + } + + ui->addonsTab->LoadFromFile(file); + + ui->display_title_id->setText( + QStringLiteral("%1").arg(title_id, 16, 16, QLatin1Char{'0'}).toUpper()); + + auto& system = Core::System::GetInstance(); + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; + const auto control = pm.GetControlMetadata(); + const auto loader = Loader::GetLoader(system, file); + + if (control.first != nullptr) { + ui->display_version->setText(QString::fromStdString(control.first->GetVersionString())); + ui->display_name->setText(QString::fromStdString(control.first->GetApplicationName())); + ui->display_developer->setText(QString::fromStdString(control.first->GetDeveloperName())); + } else { + std::string title; + if (loader->ReadTitle(title) == Loader::ResultStatus::Success) + ui->display_name->setText(QString::fromStdString(title)); + + FileSys::NACP nacp; + if (loader->ReadControlData(nacp) == Loader::ResultStatus::Success) + ui->display_developer->setText(QString::fromStdString(nacp.GetDeveloperName())); + + ui->display_version->setText(QStringLiteral("1.0.0")); + } + + if (control.second != nullptr) { + scene->clear(); + + QPixmap map; + const auto bytes = control.second->ReadAllBytes(); + map.loadFromData(bytes.data(), static_cast<u32>(bytes.size())); + + scene->addPixmap(map.scaled(ui->icon_view->width(), ui->icon_view->height(), + Qt::IgnoreAspectRatio, Qt::SmoothTransformation)); + } else { + std::vector<u8> bytes; + if (loader->ReadIcon(bytes) == Loader::ResultStatus::Success) { + scene->clear(); + + QPixmap map; + map.loadFromData(bytes.data(), static_cast<u32>(bytes.size())); + + scene->addPixmap(map.scaled(ui->icon_view->width(), ui->icon_view->height(), + Qt::IgnoreAspectRatio, Qt::SmoothTransformation)); + } + } + + ui->display_filename->setText(QString::fromStdString(file->GetName())); + + ui->display_format->setText( + QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType()))); + + const auto valueText = ReadableByteSize(file->GetSize()); + ui->display_size->setText(valueText); +} diff --git a/src/yuzu/configuration/configure_per_game.h b/src/yuzu/configuration/configure_per_game.h new file mode 100644 index 000000000..5f9a08cef --- /dev/null +++ b/src/yuzu/configuration/configure_per_game.h @@ -0,0 +1,51 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <vector> + +#include <QDialog> +#include <QList> + +#include "core/file_sys/vfs_types.h" +#include "yuzu/configuration/config.h" + +class QGraphicsScene; +class QStandardItem; +class QStandardItemModel; +class QTreeView; +class QVBoxLayout; + +namespace Ui { +class ConfigurePerGame; +} + +class ConfigurePerGame : public QDialog { + Q_OBJECT + +public: + explicit ConfigurePerGame(QWidget* parent, u64 title_id); + ~ConfigurePerGame() override; + + /// Save all button configurations to settings file + void ApplyConfiguration(); + + void LoadFromFile(FileSys::VirtualFile file); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + void LoadConfiguration(); + + std::unique_ptr<Ui::ConfigurePerGame> ui; + FileSys::VirtualFile file; + u64 title_id; + + QGraphicsScene* scene; + + std::unique_ptr<Config> game_config; +}; diff --git a/src/yuzu/configuration/configure_per_game.ui b/src/yuzu/configuration/configure_per_game.ui new file mode 100644 index 000000000..25975b3b9 --- /dev/null +++ b/src/yuzu/configuration/configure_per_game.ui @@ -0,0 +1,330 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigurePerGame</class> + <widget class="QDialog" name="ConfigurePerGame"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>800</width> + <height>600</height> + </rect> + </property> + <property name="windowTitle"> + <string>Dialog</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_3"> + <item> + <layout class="QHBoxLayout" name="horizontalLayout"> + <item> + <widget class="QGroupBox" name="groupBox"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Preferred"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="title"> + <string>Info</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <item alignment="Qt::AlignHCenter"> + <widget class="QGraphicsView" name="icon_view"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Maximum"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>256</width> + <height>256</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>256</width> + <height>256</height> + </size> + </property> + <property name="verticalScrollBarPolicy"> + <enum>Qt::ScrollBarAlwaysOff</enum> + </property> + <property name="horizontalScrollBarPolicy"> + <enum>Qt::ScrollBarAlwaysOff</enum> + </property> + <property name="interactive"> + <bool>false</bool> + </property> + </widget> + </item> + <item> + <layout class="QGridLayout" name="gridLayout_2"> + <item row="6" column="1"> + <widget class="QLineEdit" name="display_size"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="3" column="1"> + <widget class="QLineEdit" name="display_version"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="1" column="0"> + <widget class="QLabel" name="label"> + <property name="text"> + <string>Name</string> + </property> + </widget> + </item> + <item row="4" column="0"> + <widget class="QLabel" name="label_4"> + <property name="text"> + <string>Title ID</string> + </property> + </widget> + </item> + <item row="4" column="1"> + <widget class="QLineEdit" name="display_title_id"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="7" column="1"> + <widget class="QLineEdit" name="display_filename"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="5" column="1"> + <widget class="QLineEdit" name="display_format"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="7" column="0"> + <widget class="QLabel" name="label_7"> + <property name="text"> + <string>Filename</string> + </property> + </widget> + </item> + <item row="1" column="1"> + <widget class="QLineEdit" name="display_name"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="2" column="1"> + <widget class="QLineEdit" name="display_developer"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="readOnly"> + <bool>true</bool> + </property> + </widget> + </item> + <item row="5" column="0"> + <widget class="QLabel" name="label_5"> + <property name="text"> + <string>Format</string> + </property> + </widget> + </item> + <item row="3" column="0"> + <widget class="QLabel" name="label_3"> + <property name="text"> + <string>Version</string> + </property> + </widget> + </item> + <item row="6" column="0"> + <widget class="QLabel" name="label_6"> + <property name="text"> + <string>Size</string> + </property> + </widget> + </item> + <item row="2" column="0"> + <widget class="QLabel" name="label_2"> + <property name="text"> + <string>Developer</string> + </property> + </widget> + </item> + </layout> + </item> + <item> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + </item> + <item> + <layout class="QVBoxLayout" name="VerticalLayout"> + <item> + <layout class="QVBoxLayout" name="verticalLayout_2"/> + </item> + <item> + <widget class="QTabWidget" name="tabWidget"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="currentIndex"> + <number>0</number> + </property> + <property name="usesScrollButtons"> + <bool>true</bool> + </property> + <property name="documentMode"> + <bool>false</bool> + </property> + <property name="tabsClosable"> + <bool>false</bool> + </property> + <widget class="ConfigurePerGameAddons" name="addonsTab"> + <attribute name="title"> + <string>Add-Ons</string> + </attribute> + </widget> + <widget class="ConfigureGeneral" name="generalTab"> + <attribute name="title"> + <string>General</string> + </attribute> + </widget> + <widget class="ConfigureSystem" name="systemTab"> + <attribute name="title"> + <string>System</string> + </attribute> + </widget> + <widget class="ConfigureGraphics" name="graphicsTab"> + <attribute name="title"> + <string>Graphics</string> + </attribute> + </widget> + <widget class="ConfigureGraphicsAdvanced" name="graphicsAdvancedTab"> + <attribute name="title"> + <string>Adv. Graphics</string> + </attribute> + </widget> + <widget class="ConfigureAudio" name="audioTab"> + <attribute name="title"> + <string>Audio</string> + </attribute> + </widget> + </widget> + </item> + </layout> + </item> + </layout> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Preferred" vsizetype="Preferred"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </widget> + <customwidgets> + <customwidget> + <class>ConfigureGeneral</class> + <extends>QWidget</extends> + <header>configuration/configure_general.h</header> + <container>1</container> + </customwidget> + <customwidget> + <class>ConfigureSystem</class> + <extends>QWidget</extends> + <header>configuration/configure_system.h</header> + <container>1</container> + </customwidget> + <customwidget> + <class>ConfigureAudio</class> + <extends>QWidget</extends> + <header>configuration/configure_audio.h</header> + <container>1</container> + </customwidget> + <customwidget> + <class>ConfigureGraphics</class> + <extends>QWidget</extends> + <header>configuration/configure_graphics.h</header> + <container>1</container> + </customwidget> + <customwidget> + <class>ConfigureGraphicsAdvanced</class> + <extends>QWidget</extends> + <header>configuration/configure_graphics_advanced.h</header> + <container>1</container> + </customwidget> + <customwidget> + <class>ConfigurePerGameAddons</class> + <extends>QWidget</extends> + <header>configuration/configure_per_game_addons.h</header> + <container>1</container> + </customwidget> + </customwidgets> + <resources/> + <connections> + <connection> + <sender>buttonBox</sender> + <signal>accepted()</signal> + <receiver>ConfigurePerGame</receiver> + <slot>accept()</slot> + </connection> + <connection> + <sender>buttonBox</sender> + <signal>rejected()</signal> + <receiver>ConfigurePerGame</receiver> + <slot>reject()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/configure_per_general.cpp b/src/yuzu/configuration/configure_per_game_addons.cpp index d7f259f12..cdeeec01c 100644 --- a/src/yuzu/configuration/configure_per_general.cpp +++ b/src/yuzu/configuration/configure_per_game_addons.cpp @@ -15,23 +15,20 @@ #include "common/common_paths.h" #include "common/file_util.h" -#include "core/file_sys/control_metadata.h" +#include "core/core.h" #include "core/file_sys/patch_manager.h" #include "core/file_sys/xts_archive.h" #include "core/loader/loader.h" -#include "ui_configure_per_general.h" +#include "ui_configure_per_game_addons.h" #include "yuzu/configuration/config.h" #include "yuzu/configuration/configure_input.h" -#include "yuzu/configuration/configure_per_general.h" +#include "yuzu/configuration/configure_per_game_addons.h" #include "yuzu/uisettings.h" #include "yuzu/util/util.h" -ConfigurePerGameGeneral::ConfigurePerGameGeneral(QWidget* parent, u64 title_id) - : QDialog(parent), ui(std::make_unique<Ui::ConfigurePerGameGeneral>()), title_id(title_id) { - +ConfigurePerGameAddons::ConfigurePerGameAddons(QWidget* parent) + : QWidget(parent), ui(new Ui::ConfigurePerGameAddons) { ui->setupUi(this); - setFocusPolicy(Qt::ClickFocus); - setWindowTitle(tr("Properties")); layout = new QVBoxLayout; tree_view = new QTreeView; @@ -52,7 +49,7 @@ ConfigurePerGameGeneral::ConfigurePerGameGeneral(QWidget* parent, u64 title_id) item_model->setHeaderData(1, Qt::Horizontal, tr("Version")); // We must register all custom types with the Qt Automoc system so that we are able to use it - // with signals/slots. In this case, QList falls under the umbrells of custom types. + // with signals/slots. In this case, QList falls under the umbrella of custom types. qRegisterMetaType<QList<QStandardItem*>>("QList<QStandardItem*>"); layout->setContentsMargins(0, 0, 0, 0); @@ -61,18 +58,15 @@ ConfigurePerGameGeneral::ConfigurePerGameGeneral(QWidget* parent, u64 title_id) ui->scrollArea->setLayout(layout); - scene = new QGraphicsScene; - ui->icon_view->setScene(scene); + ui->scrollArea->setEnabled(!Core::System::GetInstance().IsPoweredOn()); connect(item_model, &QStandardItemModel::itemChanged, [] { UISettings::values.is_game_list_reload_pending.exchange(true); }); - - LoadConfiguration(); } -ConfigurePerGameGeneral::~ConfigurePerGameGeneral() = default; +ConfigurePerGameAddons::~ConfigurePerGameAddons() = default; -void ConfigurePerGameGeneral::ApplyConfiguration() { +void ConfigurePerGameAddons::ApplyConfiguration() { std::vector<std::string> disabled_addons; for (const auto& item : list_items) { @@ -85,78 +79,43 @@ void ConfigurePerGameGeneral::ApplyConfiguration() { std::sort(disabled_addons.begin(), disabled_addons.end()); std::sort(current.begin(), current.end()); if (disabled_addons != current) { - FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + DIR_SEP + - "game_list" + DIR_SEP + fmt::format("{:016X}.pv.txt", title_id)); + Common::FS::Delete(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + DIR_SEP + + "game_list" + DIR_SEP + fmt::format("{:016X}.pv.txt", title_id)); } Settings::values.disabled_addons[title_id] = disabled_addons; } -void ConfigurePerGameGeneral::changeEvent(QEvent* event) { +void ConfigurePerGameAddons::LoadFromFile(FileSys::VirtualFile file) { + this->file = std::move(file); + LoadConfiguration(); +} + +void ConfigurePerGameAddons::SetTitleId(u64 id) { + this->title_id = id; +} + +void ConfigurePerGameAddons::changeEvent(QEvent* event) { if (event->type() == QEvent::LanguageChange) { RetranslateUI(); } - QDialog::changeEvent(event); + QWidget::changeEvent(event); } -void ConfigurePerGameGeneral::RetranslateUI() { +void ConfigurePerGameAddons::RetranslateUI() { ui->retranslateUi(this); } -void ConfigurePerGameGeneral::LoadFromFile(FileSys::VirtualFile file) { - this->file = std::move(file); - LoadConfiguration(); -} - -void ConfigurePerGameGeneral::LoadConfiguration() { +void ConfigurePerGameAddons::LoadConfiguration() { if (file == nullptr) { return; } - ui->display_title_id->setText(QString::fromStdString(fmt::format("{:016X}", title_id))); - - FileSys::PatchManager pm{title_id}; - const auto control = pm.GetControlMetadata(); - const auto loader = Loader::GetLoader(file); - - if (control.first != nullptr) { - ui->display_version->setText(QString::fromStdString(control.first->GetVersionString())); - ui->display_name->setText(QString::fromStdString(control.first->GetApplicationName())); - ui->display_developer->setText(QString::fromStdString(control.first->GetDeveloperName())); - } else { - std::string title; - if (loader->ReadTitle(title) == Loader::ResultStatus::Success) - ui->display_name->setText(QString::fromStdString(title)); - - FileSys::NACP nacp; - if (loader->ReadControlData(nacp) == Loader::ResultStatus::Success) - ui->display_developer->setText(QString::fromStdString(nacp.GetDeveloperName())); - - ui->display_version->setText(QStringLiteral("1.0.0")); - } - - if (control.second != nullptr) { - scene->clear(); - - QPixmap map; - const auto bytes = control.second->ReadAllBytes(); - map.loadFromData(bytes.data(), static_cast<u32>(bytes.size())); - - scene->addPixmap(map.scaled(ui->icon_view->width(), ui->icon_view->height(), - Qt::IgnoreAspectRatio, Qt::SmoothTransformation)); - } else { - std::vector<u8> bytes; - if (loader->ReadIcon(bytes) == Loader::ResultStatus::Success) { - scene->clear(); - - QPixmap map; - map.loadFromData(bytes.data(), static_cast<u32>(bytes.size())); - - scene->addPixmap(map.scaled(ui->icon_view->width(), ui->icon_view->height(), - Qt::IgnoreAspectRatio, Qt::SmoothTransformation)); - } - } + auto& system = Core::System::GetInstance(); + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; + const auto loader = Loader::GetLoader(system, file); FileSys::VirtualFile update_raw; loader->ReadUpdateRaw(update_raw); @@ -182,12 +141,4 @@ void ConfigurePerGameGeneral::LoadConfiguration() { } tree_view->setColumnWidth(0, 5 * tree_view->width() / 16); - - ui->display_filename->setText(QString::fromStdString(file->GetName())); - - ui->display_format->setText( - QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType()))); - - const auto valueText = ReadableByteSize(file->GetSize()); - ui->display_size->setText(valueText); } diff --git a/src/yuzu/configuration/configure_per_general.h b/src/yuzu/configuration/configure_per_game_addons.h index a3b2cdeff..a00ec3539 100644 --- a/src/yuzu/configuration/configure_per_general.h +++ b/src/yuzu/configuration/configure_per_game_addons.h @@ -1,4 +1,4 @@ -// Copyright 2016 Citra Emulator Project +// Copyright 2016 Citra Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. @@ -7,7 +7,6 @@ #include <memory> #include <vector> -#include <QDialog> #include <QList> #include "core/file_sys/vfs_types.h" @@ -19,35 +18,36 @@ class QTreeView; class QVBoxLayout; namespace Ui { -class ConfigurePerGameGeneral; +class ConfigurePerGameAddons; } -class ConfigurePerGameGeneral : public QDialog { +class ConfigurePerGameAddons : public QWidget { Q_OBJECT public: - explicit ConfigurePerGameGeneral(QWidget* parent, u64 title_id); - ~ConfigurePerGameGeneral() override; + explicit ConfigurePerGameAddons(QWidget* parent = nullptr); + ~ConfigurePerGameAddons() override; /// Save all button configurations to settings file void ApplyConfiguration(); void LoadFromFile(FileSys::VirtualFile file); + void SetTitleId(u64 id); + private: void changeEvent(QEvent* event) override; void RetranslateUI(); void LoadConfiguration(); - std::unique_ptr<Ui::ConfigurePerGameGeneral> ui; + std::unique_ptr<Ui::ConfigurePerGameAddons> ui; FileSys::VirtualFile file; u64 title_id; QVBoxLayout* layout; QTreeView* tree_view; QStandardItemModel* item_model; - QGraphicsScene* scene; std::vector<QList<QStandardItem*>> list_items; }; diff --git a/src/yuzu/configuration/configure_per_game_addons.ui b/src/yuzu/configuration/configure_per_game_addons.ui new file mode 100644 index 000000000..aefdebfcd --- /dev/null +++ b/src/yuzu/configuration/configure_per_game_addons.ui @@ -0,0 +1,38 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigurePerGameAddons</class> + <widget class="QWidget" name="ConfigurePerGameAddons"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>400</width> + <height>300</height> + </rect> + </property> + <property name="windowTitle"> + <string>Form</string> + </property> + <layout class="QGridLayout" name="gridLayout"> + <item row="0" column="0"> + <widget class="QScrollArea" name="scrollArea"> + <property name="widgetResizable"> + <bool>true</bool> + </property> + <widget class="QWidget" name="scrollAreaWidgetContents"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>380</width> + <height>280</height> + </rect> + </property> + </widget> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui> diff --git a/src/yuzu/configuration/configure_per_general.ui b/src/yuzu/configuration/configure_per_general.ui deleted file mode 100644 index 8fdd96fa4..000000000 --- a/src/yuzu/configuration/configure_per_general.ui +++ /dev/null @@ -1,276 +0,0 @@ -<?xml version="1.0" encoding="UTF-8"?> -<ui version="4.0"> - <class>ConfigurePerGameGeneral</class> - <widget class="QDialog" name="ConfigurePerGameGeneral"> - <property name="geometry"> - <rect> - <x>0</x> - <y>0</y> - <width>400</width> - <height>520</height> - </rect> - </property> - <property name="windowTitle"> - <string>ConfigurePerGameGeneral</string> - </property> - <layout class="QHBoxLayout" name="HorizontalLayout"> - <item> - <layout class="QVBoxLayout" name="VerticalLayout"> - <item> - <widget class="QGroupBox" name="GeneralGroupBox"> - <property name="title"> - <string>Info</string> - </property> - <layout class="QHBoxLayout" name="GeneralHorizontalLayout"> - <item> - <layout class="QGridLayout" name="gridLayout_2"> - <item row="6" column="1" colspan="2"> - <widget class="QLineEdit" name="display_filename"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="0" column="1"> - <widget class="QLineEdit" name="display_name"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="1" column="0"> - <widget class="QLabel" name="label_2"> - <property name="text"> - <string>Developer</string> - </property> - </widget> - </item> - <item row="5" column="1" colspan="2"> - <widget class="QLineEdit" name="display_size"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="0" column="0"> - <widget class="QLabel" name="label"> - <property name="text"> - <string>Name</string> - </property> - </widget> - </item> - <item row="6" column="0"> - <widget class="QLabel" name="label_7"> - <property name="text"> - <string>Filename</string> - </property> - </widget> - </item> - <item row="2" column="0"> - <widget class="QLabel" name="label_3"> - <property name="text"> - <string>Version</string> - </property> - </widget> - </item> - <item row="4" column="0"> - <widget class="QLabel" name="label_5"> - <property name="text"> - <string>Format</string> - </property> - </widget> - </item> - <item row="2" column="1"> - <widget class="QLineEdit" name="display_version"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="4" column="1"> - <widget class="QLineEdit" name="display_format"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="5" column="0"> - <widget class="QLabel" name="label_6"> - <property name="text"> - <string>Size</string> - </property> - </widget> - </item> - <item row="1" column="1"> - <widget class="QLineEdit" name="display_developer"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="3" column="0"> - <widget class="QLabel" name="label_4"> - <property name="text"> - <string>Title ID</string> - </property> - </widget> - </item> - <item row="3" column="1"> - <widget class="QLineEdit" name="display_title_id"> - <property name="enabled"> - <bool>true</bool> - </property> - <property name="readOnly"> - <bool>true</bool> - </property> - </widget> - </item> - <item row="0" column="2" rowspan="5"> - <widget class="QGraphicsView" name="icon_view"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Maximum" vsizetype="Maximum"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="minimumSize"> - <size> - <width>128</width> - <height>128</height> - </size> - </property> - <property name="maximumSize"> - <size> - <width>128</width> - <height>128</height> - </size> - </property> - <property name="verticalScrollBarPolicy"> - <enum>Qt::ScrollBarAlwaysOff</enum> - </property> - <property name="horizontalScrollBarPolicy"> - <enum>Qt::ScrollBarAlwaysOff</enum> - </property> - <property name="sizeAdjustPolicy"> - <enum>QAbstractScrollArea::AdjustToContents</enum> - </property> - <property name="interactive"> - <bool>false</bool> - </property> - </widget> - </item> - </layout> - </item> - </layout> - </widget> - </item> - <item> - <widget class="QGroupBox" name="PerformanceGroupBox"> - <property name="title"> - <string>Add-Ons</string> - </property> - <layout class="QHBoxLayout" name="PerformanceHorizontalLayout"> - <item> - <widget class="QScrollArea" name="scrollArea"> - <property name="widgetResizable"> - <bool>true</bool> - </property> - <widget class="QWidget" name="scrollAreaWidgetContents"> - <property name="geometry"> - <rect> - <x>0</x> - <y>0</y> - <width>350</width> - <height>169</height> - </rect> - </property> - </widget> - </widget> - </item> - <item> - <layout class="QVBoxLayout" name="PerformanceVerticalLayout"/> - </item> - </layout> - </widget> - </item> - <item> - <spacer name="verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeType"> - <enum>QSizePolicy::Fixed</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - <item> - <widget class="QDialogButtonBox" name="buttonBox"> - <property name="standardButtons"> - <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> - </property> - </widget> - </item> - </layout> - </item> - </layout> - </widget> - <resources/> - <connections> - <connection> - <sender>buttonBox</sender> - <signal>accepted()</signal> - <receiver>ConfigurePerGameGeneral</receiver> - <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>269</x> - <y>567</y> - </hint> - <hint type="destinationlabel"> - <x>269</x> - <y>294</y> - </hint> - </hints> - </connection> - <connection> - <sender>buttonBox</sender> - <signal>rejected()</signal> - <receiver>ConfigurePerGameGeneral</receiver> - <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>269</x> - <y>567</y> - </hint> - <hint type="destinationlabel"> - <x>269</x> - <y>294</y> - </hint> - </hints> - </connection> - </connections> -</ui> diff --git a/src/yuzu/configuration/configure_profile_manager.cpp b/src/yuzu/configuration/configure_profile_manager.cpp index f53423440..6334c4c50 100644 --- a/src/yuzu/configuration/configure_profile_manager.cpp +++ b/src/yuzu/configuration/configure_profile_manager.cpp @@ -34,7 +34,7 @@ constexpr std::array<u8, 107> backup_jpeg{ }; QString GetImagePath(Common::UUID uuid) { - const auto path = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + + const auto path = Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + "/system/save/8000000000000010/su/avators/" + uuid.FormatSwitch() + ".jpg"; return QString::fromStdString(path); } @@ -282,7 +282,7 @@ void ConfigureProfileManager::SetUserImage() { } const auto raw_path = QString::fromStdString( - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + "/system/save/8000000000000010"); + Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + "/system/save/8000000000000010"); const QFileInfo raw_info{raw_path}; if (raw_info.exists() && !raw_info.isDir() && !QFile::remove(raw_path)) { QMessageBox::warning(this, tr("Error deleting file"), diff --git a/src/yuzu/configuration/configure_service.cpp b/src/yuzu/configuration/configure_service.cpp index 06566e981..0de7a4f0b 100644 --- a/src/yuzu/configuration/configure_service.cpp +++ b/src/yuzu/configuration/configure_service.cpp @@ -68,6 +68,7 @@ void ConfigureService::SetConfiguration() { } std::pair<QString, QString> ConfigureService::BCATDownloadEvents() { +#ifdef YUZU_ENABLE_BOXCAT std::optional<std::string> global; std::map<std::string, Service::BCAT::EventStatus> map; const auto res = Service::BCAT::Boxcat::GetStatus(global, map); @@ -105,7 +106,10 @@ std::pair<QString, QString> ConfigureService::BCATDownloadEvents() { .arg(QString::fromStdString(key)) .arg(FormatEventStatusString(value)); } - return {QStringLiteral("Current Boxcat Events"), std::move(out)}; + return {tr("Current Boxcat Events"), std::move(out)}; +#else + return {tr("Current Boxcat Events"), tr("There are currently no events on boxcat.")}; +#endif } void ConfigureService::OnBCATImplChanged() { diff --git a/src/yuzu/configuration/configure_system.cpp b/src/yuzu/configuration/configure_system.cpp index f49cd4c8f..59a58d92c 100644 --- a/src/yuzu/configuration/configure_system.cpp +++ b/src/yuzu/configuration/configure_system.cpp @@ -12,8 +12,10 @@ #include "common/assert.h" #include "common/file_util.h" #include "core/core.h" +#include "core/hle/service/time/time.h" #include "core/settings.h" #include "ui_configure_system.h" +#include "yuzu/configuration/configuration_shared.h" #include "yuzu/configuration/configure_system.h" ConfigureSystem::ConfigureSystem(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureSystem) { @@ -21,20 +23,25 @@ ConfigureSystem::ConfigureSystem(QWidget* parent) : QWidget(parent), ui(new Ui:: connect(ui->button_regenerate_console_id, &QPushButton::clicked, this, &ConfigureSystem::RefreshConsoleID); - connect(ui->rng_seed_checkbox, &QCheckBox::stateChanged, this, [this](bool checked) { - ui->rng_seed_edit->setEnabled(checked); - if (!checked) { + connect(ui->rng_seed_checkbox, &QCheckBox::stateChanged, this, [this](int state) { + ui->rng_seed_edit->setEnabled(state == Qt::Checked); + if (state != Qt::Checked) { ui->rng_seed_edit->setText(QStringLiteral("00000000")); } }); - connect(ui->custom_rtc_checkbox, &QCheckBox::stateChanged, this, [this](bool checked) { - ui->custom_rtc_edit->setEnabled(checked); - if (!checked) { + connect(ui->custom_rtc_checkbox, &QCheckBox::stateChanged, this, [this](int state) { + ui->custom_rtc_edit->setEnabled(state == Qt::Checked); + if (state != Qt::Checked) { ui->custom_rtc_edit->setDateTime(QDateTime::currentDateTime()); } }); + ui->label_console_id->setVisible(Settings::IsConfiguringGlobal()); + ui->button_regenerate_console_id->setVisible(Settings::IsConfiguringGlobal()); + + SetupPerGameUI(); + SetConfiguration(); } @@ -54,49 +61,140 @@ void ConfigureSystem::RetranslateUI() { void ConfigureSystem::SetConfiguration() { enabled = !Core::System::GetInstance().IsPoweredOn(); + const auto rng_seed = + QStringLiteral("%1") + .arg(Settings::values.rng_seed.GetValue().value_or(0), 8, 16, QLatin1Char{'0'}) + .toUpper(); + const auto rtc_time = Settings::values.custom_rtc.GetValue().value_or( + std::chrono::seconds(QDateTime::currentSecsSinceEpoch())); - ui->combo_language->setCurrentIndex(Settings::values.language_index); - ui->combo_region->setCurrentIndex(Settings::values.region_index); - ui->combo_sound->setCurrentIndex(Settings::values.sound_index); - - ui->rng_seed_checkbox->setChecked(Settings::values.rng_seed.has_value()); - ui->rng_seed_edit->setEnabled(Settings::values.rng_seed.has_value()); - - const auto rng_seed = QStringLiteral("%1") - .arg(Settings::values.rng_seed.value_or(0), 8, 16, QLatin1Char{'0'}) - .toUpper(); + ui->rng_seed_checkbox->setChecked(Settings::values.rng_seed.GetValue().has_value()); + ui->rng_seed_edit->setEnabled(Settings::values.rng_seed.GetValue().has_value() && + Settings::values.rng_seed.UsingGlobal()); ui->rng_seed_edit->setText(rng_seed); - ui->custom_rtc_checkbox->setChecked(Settings::values.custom_rtc.has_value()); - ui->custom_rtc_edit->setEnabled(Settings::values.custom_rtc.has_value()); - - const auto rtc_time = Settings::values.custom_rtc.value_or( - std::chrono::seconds(QDateTime::currentSecsSinceEpoch())); + ui->custom_rtc_checkbox->setChecked(Settings::values.custom_rtc.GetValue().has_value()); + ui->custom_rtc_edit->setEnabled(Settings::values.custom_rtc.GetValue().has_value() && + Settings::values.rng_seed.UsingGlobal()); ui->custom_rtc_edit->setDateTime(QDateTime::fromSecsSinceEpoch(rtc_time.count())); + + if (Settings::IsConfiguringGlobal()) { + ui->combo_language->setCurrentIndex(Settings::values.language_index.GetValue()); + ui->combo_region->setCurrentIndex(Settings::values.region_index.GetValue()); + ui->combo_time_zone->setCurrentIndex(Settings::values.time_zone_index.GetValue()); + ui->combo_sound->setCurrentIndex(Settings::values.sound_index.GetValue()); + } else { + ConfigurationShared::SetPerGameSetting(ui->combo_language, + &Settings::values.language_index); + ConfigurationShared::SetPerGameSetting(ui->combo_region, &Settings::values.region_index); + ConfigurationShared::SetPerGameSetting(ui->combo_time_zone, + &Settings::values.time_zone_index); + ConfigurationShared::SetPerGameSetting(ui->combo_sound, &Settings::values.sound_index); + + ConfigurationShared::SetHighlight(ui->label_language, + !Settings::values.language_index.UsingGlobal()); + ConfigurationShared::SetHighlight(ui->label_region, + !Settings::values.region_index.UsingGlobal()); + ConfigurationShared::SetHighlight(ui->label_timezone, + !Settings::values.time_zone_index.UsingGlobal()); + ConfigurationShared::SetHighlight(ui->label_sound, + !Settings::values.sound_index.UsingGlobal()); + } } void ConfigureSystem::ReadSystemSettings() {} void ConfigureSystem::ApplyConfiguration() { + // Allow setting custom RTC even if system is powered on, to allow in-game time to be fast + // forwared + if (Settings::values.custom_rtc.UsingGlobal()) { + if (ui->custom_rtc_checkbox->isChecked()) { + Settings::values.custom_rtc.SetValue( + std::chrono::seconds(ui->custom_rtc_edit->dateTime().toSecsSinceEpoch())); + if (Core::System::GetInstance().IsPoweredOn()) { + const s64 posix_time{Settings::values.custom_rtc.GetValue()->count() + + Service::Time::TimeManager::GetExternalTimeZoneOffset()}; + Core::System::GetInstance().GetTimeManager().UpdateLocalSystemClockTime(posix_time); + } + } else { + Settings::values.custom_rtc.SetValue(std::nullopt); + } + } + if (!enabled) { return; } - Settings::values.language_index = ui->combo_language->currentIndex(); - Settings::values.region_index = ui->combo_region->currentIndex(); - Settings::values.sound_index = ui->combo_sound->currentIndex(); + if (Settings::IsConfiguringGlobal()) { + // Guard if during game and set to game-specific value + if (Settings::values.language_index.UsingGlobal()) { + Settings::values.language_index.SetValue(ui->combo_language->currentIndex()); + } + if (Settings::values.region_index.UsingGlobal()) { + Settings::values.region_index.SetValue(ui->combo_region->currentIndex()); + } + if (Settings::values.time_zone_index.UsingGlobal()) { + Settings::values.time_zone_index.SetValue(ui->combo_time_zone->currentIndex()); + } + if (Settings::values.sound_index.UsingGlobal()) { + Settings::values.sound_index.SetValue(ui->combo_sound->currentIndex()); + } - if (ui->rng_seed_checkbox->isChecked()) { - Settings::values.rng_seed = ui->rng_seed_edit->text().toULongLong(nullptr, 16); + if (Settings::values.rng_seed.UsingGlobal()) { + if (ui->rng_seed_checkbox->isChecked()) { + Settings::values.rng_seed.SetValue( + ui->rng_seed_edit->text().toULongLong(nullptr, 16)); + } else { + Settings::values.rng_seed.SetValue(std::nullopt); + } + } } else { - Settings::values.rng_seed = std::nullopt; - } + ConfigurationShared::ApplyPerGameSetting(&Settings::values.language_index, + ui->combo_language); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.region_index, ui->combo_region); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.time_zone_index, + ui->combo_time_zone); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.sound_index, ui->combo_sound); + + switch (use_rng_seed) { + case ConfigurationShared::CheckState::On: + case ConfigurationShared::CheckState::Off: + Settings::values.rng_seed.SetGlobal(false); + if (ui->rng_seed_checkbox->isChecked()) { + Settings::values.rng_seed.SetValue( + ui->rng_seed_edit->text().toULongLong(nullptr, 16)); + } else { + Settings::values.rng_seed.SetValue(std::nullopt); + } + break; + case ConfigurationShared::CheckState::Global: + Settings::values.rng_seed.SetGlobal(false); + Settings::values.rng_seed.SetValue(std::nullopt); + Settings::values.rng_seed.SetGlobal(true); + break; + case ConfigurationShared::CheckState::Count: + break; + } - if (ui->custom_rtc_checkbox->isChecked()) { - Settings::values.custom_rtc = - std::chrono::seconds(ui->custom_rtc_edit->dateTime().toSecsSinceEpoch()); - } else { - Settings::values.custom_rtc = std::nullopt; + switch (use_custom_rtc) { + case ConfigurationShared::CheckState::On: + case ConfigurationShared::CheckState::Off: + Settings::values.custom_rtc.SetGlobal(false); + if (ui->custom_rtc_checkbox->isChecked()) { + Settings::values.custom_rtc.SetValue( + std::chrono::seconds(ui->custom_rtc_edit->dateTime().toSecsSinceEpoch())); + } else { + Settings::values.custom_rtc.SetValue(std::nullopt); + } + break; + case ConfigurationShared::CheckState::Global: + Settings::values.custom_rtc.SetGlobal(false); + Settings::values.custom_rtc.SetValue(std::nullopt); + Settings::values.custom_rtc.SetGlobal(true); + break; + case ConfigurationShared::CheckState::Count: + break; + } } Settings::Apply(); @@ -118,3 +216,36 @@ void ConfigureSystem::RefreshConsoleID() { ui->label_console_id->setText( tr("Console ID: 0x%1").arg(QString::number(console_id, 16).toUpper())); } + +void ConfigureSystem::SetupPerGameUI() { + if (Settings::IsConfiguringGlobal()) { + ui->combo_language->setEnabled(Settings::values.language_index.UsingGlobal()); + ui->combo_region->setEnabled(Settings::values.region_index.UsingGlobal()); + ui->combo_time_zone->setEnabled(Settings::values.time_zone_index.UsingGlobal()); + ui->combo_sound->setEnabled(Settings::values.sound_index.UsingGlobal()); + ui->rng_seed_checkbox->setEnabled(Settings::values.rng_seed.UsingGlobal()); + ui->rng_seed_edit->setEnabled(Settings::values.rng_seed.UsingGlobal()); + ui->custom_rtc_checkbox->setEnabled(Settings::values.custom_rtc.UsingGlobal()); + ui->custom_rtc_edit->setEnabled(Settings::values.custom_rtc.UsingGlobal()); + + return; + } + + ConfigurationShared::SetColoredComboBox(ui->combo_language, ui->label_language, + Settings::values.language_index.GetValue(true)); + ConfigurationShared::SetColoredComboBox(ui->combo_region, ui->label_region, + Settings::values.region_index.GetValue(true)); + ConfigurationShared::SetColoredComboBox(ui->combo_time_zone, ui->label_timezone, + Settings::values.time_zone_index.GetValue(true)); + ConfigurationShared::SetColoredComboBox(ui->combo_sound, ui->label_sound, + Settings::values.sound_index.GetValue(true)); + + ConfigurationShared::SetColoredTristate( + ui->rng_seed_checkbox, Settings::values.rng_seed.UsingGlobal(), + Settings::values.rng_seed.GetValue().has_value(), + Settings::values.rng_seed.GetValue(true).has_value(), use_rng_seed); + ConfigurationShared::SetColoredTristate( + ui->custom_rtc_checkbox, Settings::values.custom_rtc.UsingGlobal(), + Settings::values.custom_rtc.GetValue().has_value(), + Settings::values.custom_rtc.GetValue(true).has_value(), use_custom_rtc); +} diff --git a/src/yuzu/configuration/configure_system.h b/src/yuzu/configuration/configure_system.h index d8fa2d2cc..fc5cd2945 100644 --- a/src/yuzu/configuration/configure_system.h +++ b/src/yuzu/configuration/configure_system.h @@ -9,6 +9,10 @@ #include <QList> #include <QWidget> +namespace ConfigurationShared { +enum class CheckState; +} + namespace Ui { class ConfigureSystem; } @@ -32,10 +36,16 @@ private: void RefreshConsoleID(); + void SetupPerGameUI(); + std::unique_ptr<Ui::ConfigureSystem> ui; bool enabled = false; int language_index = 0; int region_index = 0; + int time_zone_index = 0; int sound_index = 0; + + ConfigurationShared::CheckState use_rng_seed; + ConfigurationShared::CheckState use_custom_rtc; }; diff --git a/src/yuzu/configuration/configure_system.ui b/src/yuzu/configuration/configure_system.ui index 4e2c7e76e..53b95658b 100644 --- a/src/yuzu/configuration/configure_system.ui +++ b/src/yuzu/configuration/configure_system.ui @@ -21,249 +21,494 @@ <property name="title"> <string>System Settings</string> </property> - <layout class="QGridLayout" name="gridLayout"> - <item row="2" column="0"> - <widget class="QLabel" name="label_sound"> - <property name="text"> - <string>Sound output mode</string> - </property> - </widget> - </item> - <item row="3" column="0"> - <widget class="QLabel" name="label_console_id"> - <property name="text"> - <string>Console ID:</string> - </property> - </widget> - </item> - <item row="0" column="1"> - <widget class="QComboBox" name="combo_language"> - <property name="toolTip"> - <string>Note: this can be overridden when region setting is auto-select</string> - </property> - <item> - <property name="text"> - <string>Japanese (日本語)</string> - </property> - </item> - <item> - <property name="text"> - <string>English</string> - </property> - </item> - <item> - <property name="text"> - <string>French (français)</string> - </property> - </item> - <item> - <property name="text"> - <string>German (Deutsch)</string> - </property> - </item> - <item> - <property name="text"> - <string>Italian (italiano)</string> - </property> - </item> - <item> - <property name="text"> - <string>Spanish (español)</string> - </property> - </item> - <item> - <property name="text"> - <string>Chinese</string> - </property> - </item> - <item> - <property name="text"> - <string>Korean (한국어)</string> - </property> - </item> - <item> - <property name="text"> - <string>Dutch (Nederlands)</string> - </property> - </item> - <item> - <property name="text"> - <string>Portuguese (português)</string> - </property> - </item> - <item> - <property name="text"> - <string>Russian (Русский)</string> - </property> + <layout class="QVBoxLayout" name="verticalLayout_2"> + <item> + <layout class="QGridLayout" name="gridLayout_2"> + <item row="1" column="0"> + <widget class="QLabel" name="label_region"> + <property name="text"> + <string>Region:</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Taiwanese</string> - </property> + <item row="2" column="1"> + <widget class="QComboBox" name="combo_time_zone"> + <item> + <property name="text"> + <string>Auto</string> + </property> + </item> + <item> + <property name="text"> + <string>Default</string> + </property> + </item> + <item> + <property name="text"> + <string>CET</string> + </property> + </item> + <item> + <property name="text"> + <string>CST6CDT</string> + </property> + </item> + <item> + <property name="text"> + <string>Cuba</string> + </property> + </item> + <item> + <property name="text"> + <string>EET</string> + </property> + </item> + <item> + <property name="text"> + <string>Egypt</string> + </property> + </item> + <item> + <property name="text"> + <string>Eire</string> + </property> + </item> + <item> + <property name="text"> + <string>EST</string> + </property> + </item> + <item> + <property name="text"> + <string>EST5EDT</string> + </property> + </item> + <item> + <property name="text"> + <string>GB</string> + </property> + </item> + <item> + <property name="text"> + <string>GB-Eire</string> + </property> + </item> + <item> + <property name="text"> + <string>GMT</string> + </property> + </item> + <item> + <property name="text"> + <string>GMT+0</string> + </property> + </item> + <item> + <property name="text"> + <string>GMT-0</string> + </property> + </item> + <item> + <property name="text"> + <string>GMT0</string> + </property> + </item> + <item> + <property name="text"> + <string>Greenwich</string> + </property> + </item> + <item> + <property name="text"> + <string>Hongkong</string> + </property> + </item> + <item> + <property name="text"> + <string>HST</string> + </property> + </item> + <item> + <property name="text"> + <string>Iceland</string> + </property> + </item> + <item> + <property name="text"> + <string>Iran</string> + </property> + </item> + <item> + <property name="text"> + <string>Israel</string> + </property> + </item> + <item> + <property name="text"> + <string>Jamaica</string> + </property> + </item> + <item> + <property name="text"> + <string>Japan</string> + </property> + </item> + <item> + <property name="text"> + <string>Kwajalein</string> + </property> + </item> + <item> + <property name="text"> + <string>Libya</string> + </property> + </item> + <item> + <property name="text"> + <string>MET</string> + </property> + </item> + <item> + <property name="text"> + <string>MST</string> + </property> + </item> + <item> + <property name="text"> + <string>MST7MDT</string> + </property> + </item> + <item> + <property name="text"> + <string>Navajo</string> + </property> + </item> + <item> + <property name="text"> + <string>NZ</string> + </property> + </item> + <item> + <property name="text"> + <string>NZ-CHAT</string> + </property> + </item> + <item> + <property name="text"> + <string>Poland</string> + </property> + </item> + <item> + <property name="text"> + <string>Portugal</string> + </property> + </item> + <item> + <property name="text"> + <string>PRC</string> + </property> + </item> + <item> + <property name="text"> + <string>PST8PDT</string> + </property> + </item> + <item> + <property name="text"> + <string>ROC</string> + </property> + </item> + <item> + <property name="text"> + <string>ROK</string> + </property> + </item> + <item> + <property name="text"> + <string>Singapore</string> + </property> + </item> + <item> + <property name="text"> + <string>Turkey</string> + </property> + </item> + <item> + <property name="text"> + <string>UCT</string> + </property> + </item> + <item> + <property name="text"> + <string>Universal</string> + </property> + </item> + <item> + <property name="text"> + <string>UTC</string> + </property> + </item> + <item> + <property name="text"> + <string>W-SU</string> + </property> + </item> + <item> + <property name="text"> + <string>WET</string> + </property> + </item> + <item> + <property name="text"> + <string>Zulu</string> + </property> + </item> + </widget> </item> - <item> - <property name="text"> - <string>British English</string> - </property> + <item row="1" column="1"> + <widget class="QComboBox" name="combo_region"> + <item> + <property name="text"> + <string>Japan</string> + </property> + </item> + <item> + <property name="text"> + <string>USA</string> + </property> + </item> + <item> + <property name="text"> + <string>Europe</string> + </property> + </item> + <item> + <property name="text"> + <string>Australia</string> + </property> + </item> + <item> + <property name="text"> + <string>China</string> + </property> + </item> + <item> + <property name="text"> + <string>Korea</string> + </property> + </item> + <item> + <property name="text"> + <string>Taiwan</string> + </property> + </item> + </widget> </item> - <item> - <property name="text"> - <string>Canadian French</string> - </property> + <item row="2" column="0"> + <widget class="QLabel" name="label_timezone"> + <property name="text"> + <string>Time Zone:</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Latin American Spanish</string> - </property> + <item row="0" column="1"> + <widget class="QComboBox" name="combo_language"> + <property name="toolTip"> + <string>Note: this can be overridden when region setting is auto-select</string> + </property> + <item> + <property name="text"> + <string>Japanese (日本語)</string> + </property> + </item> + <item> + <property name="text"> + <string>English</string> + </property> + </item> + <item> + <property name="text"> + <string>French (français)</string> + </property> + </item> + <item> + <property name="text"> + <string>German (Deutsch)</string> + </property> + </item> + <item> + <property name="text"> + <string>Italian (italiano)</string> + </property> + </item> + <item> + <property name="text"> + <string>Spanish (español)</string> + </property> + </item> + <item> + <property name="text"> + <string>Chinese</string> + </property> + </item> + <item> + <property name="text"> + <string>Korean (한국어)</string> + </property> + </item> + <item> + <property name="text"> + <string>Dutch (Nederlands)</string> + </property> + </item> + <item> + <property name="text"> + <string>Portuguese (português)</string> + </property> + </item> + <item> + <property name="text"> + <string>Russian (Русский)</string> + </property> + </item> + <item> + <property name="text"> + <string>Taiwanese</string> + </property> + </item> + <item> + <property name="text"> + <string>British English</string> + </property> + </item> + <item> + <property name="text"> + <string>Canadian French</string> + </property> + </item> + <item> + <property name="text"> + <string>Latin American Spanish</string> + </property> + </item> + <item> + <property name="text"> + <string>Simplified Chinese</string> + </property> + </item> + <item> + <property name="text"> + <string>Traditional Chinese (正體中文)</string> + </property> + </item> + </widget> </item> - <item> - <property name="text"> - <string>Simplified Chinese</string> - </property> + <item row="5" column="0"> + <widget class="QCheckBox" name="custom_rtc_checkbox"> + <property name="text"> + <string>Custom RTC</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Traditional Chinese (正體中文)</string> - </property> + <item row="0" column="0"> + <widget class="QLabel" name="label_language"> + <property name="text"> + <string>Language</string> + </property> + </widget> </item> - </widget> - </item> - <item row="1" column="0"> - <widget class="QLabel" name="label_region"> - <property name="text"> - <string>Region:</string> - </property> - </widget> - </item> - <item row="1" column="1"> - <widget class="QComboBox" name="combo_region"> - <item> - <property name="text"> - <string>Japan</string> - </property> - </item> - <item> - <property name="text"> - <string>USA</string> - </property> + <item row="6" column="0"> + <widget class="QCheckBox" name="rng_seed_checkbox"> + <property name="text"> + <string>RNG Seed</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Europe</string> - </property> + <item row="3" column="1"> + <widget class="QComboBox" name="combo_sound"> + <item> + <property name="text"> + <string>Mono</string> + </property> + </item> + <item> + <property name="text"> + <string>Stereo</string> + </property> + </item> + <item> + <property name="text"> + <string>Surround</string> + </property> + </item> + </widget> </item> - <item> - <property name="text"> - <string>Australia</string> - </property> + <item row="4" column="0"> + <widget class="QLabel" name="label_console_id"> + <property name="text"> + <string>Console ID:</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>China</string> - </property> + <item row="3" column="0"> + <widget class="QLabel" name="label_sound"> + <property name="text"> + <string>Sound output mode</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Korea</string> - </property> - </item> - <item> - <property name="text"> - <string>Taiwan</string> - </property> - </item> - </widget> - </item> - <item row="5" column="0"> - <widget class="QCheckBox" name="rng_seed_checkbox"> - <property name="text"> - <string>RNG Seed</string> - </property> - </widget> - </item> - <item row="2" column="1"> - <widget class="QComboBox" name="combo_sound"> - <item> - <property name="text"> - <string>Mono</string> - </property> + <item row="5" column="1"> + <widget class="QDateTimeEdit" name="custom_rtc_edit"> + <property name="minimumDate"> + <date> + <year>1970</year> + <month>1</month> + <day>1</day> + </date> + </property> + <property name="displayFormat"> + <string>d MMM yyyy h:mm:ss AP</string> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Stereo</string> - </property> + <item row="6" column="1"> + <widget class="QLineEdit" name="rng_seed_edit"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="font"> + <font> + <family>Lucida Console</family> + </font> + </property> + <property name="inputMask"> + <string notr="true">HHHHHHHH</string> + </property> + <property name="maxLength"> + <number>8</number> + </property> + </widget> </item> - <item> - <property name="text"> - <string>Surround</string> - </property> + <item row="4" column="1"> + <widget class="QPushButton" name="button_regenerate_console_id"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="layoutDirection"> + <enum>Qt::RightToLeft</enum> + </property> + <property name="text"> + <string>Regenerate</string> + </property> + </widget> </item> - </widget> - </item> - <item row="0" column="0"> - <widget class="QLabel" name="label_language"> - <property name="text"> - <string>Language</string> - </property> - </widget> - </item> - <item row="3" column="1"> - <widget class="QPushButton" name="button_regenerate_console_id"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Fixed" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="layoutDirection"> - <enum>Qt::RightToLeft</enum> - </property> - <property name="text"> - <string>Regenerate</string> - </property> - </widget> - </item> - <item row="4" column="0"> - <widget class="QCheckBox" name="custom_rtc_checkbox"> - <property name="text"> - <string>Custom RTC</string> - </property> - </widget> - </item> - <item row="4" column="1"> - <widget class="QDateTimeEdit" name="custom_rtc_edit"> - <property name="minimumDate"> - <date> - <year>1970</year> - <month>1</month> - <day>1</day> - </date> - </property> - <property name="displayFormat"> - <string>d MMM yyyy h:mm:ss AP</string> - </property> - </widget> - </item> - <item row="5" column="1"> - <widget class="QLineEdit" name="rng_seed_edit"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="font"> - <font> - <family>Lucida Console</family> - </font> - </property> - <property name="inputMask"> - <string notr="true">HHHHHHHH</string> - </property> - <property name="maxLength"> - <number>8</number> - </property> - </widget> + </layout> </item> </layout> </widget> diff --git a/src/yuzu/configuration/configure_touch_from_button.cpp b/src/yuzu/configuration/configure_touch_from_button.cpp new file mode 100644 index 000000000..15557e4b8 --- /dev/null +++ b/src/yuzu/configuration/configure_touch_from_button.cpp @@ -0,0 +1,623 @@ +// Copyright 2020 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QInputDialog> +#include <QKeyEvent> +#include <QMessageBox> +#include <QMouseEvent> +#include <QResizeEvent> +#include <QStandardItemModel> +#include <QTimer> +#include "common/param_package.h" +#include "core/frontend/framebuffer_layout.h" +#include "core/settings.h" +#include "input_common/main.h" +#include "ui_configure_touch_from_button.h" +#include "yuzu/configuration/configure_touch_from_button.h" +#include "yuzu/configuration/configure_touch_widget.h" + +static QString GetKeyName(int key_code) { + switch (key_code) { + case Qt::Key_Shift: + return QObject::tr("Shift"); + case Qt::Key_Control: + return QObject::tr("Ctrl"); + case Qt::Key_Alt: + return QObject::tr("Alt"); + case Qt::Key_Meta: + return QString{}; + default: + return QKeySequence(key_code).toString(); + } +} + +static QString ButtonToText(const Common::ParamPackage& param) { + if (!param.Has("engine")) { + return QObject::tr("[not set]"); + } + + if (param.Get("engine", "") == "keyboard") { + return GetKeyName(param.Get("code", 0)); + } + + if (param.Get("engine", "") == "sdl") { + if (param.Has("hat")) { + const QString hat_str = QString::fromStdString(param.Get("hat", "")); + const QString direction_str = QString::fromStdString(param.Get("direction", "")); + + return QObject::tr("Hat %1 %2").arg(hat_str, direction_str); + } + + if (param.Has("axis")) { + const QString axis_str = QString::fromStdString(param.Get("axis", "")); + const QString direction_str = QString::fromStdString(param.Get("direction", "")); + + return QObject::tr("Axis %1%2").arg(axis_str, direction_str); + } + + if (param.Has("button")) { + const QString button_str = QString::fromStdString(param.Get("button", "")); + + return QObject::tr("Button %1").arg(button_str); + } + + return {}; + } + + return QObject::tr("[unknown]"); +} + +ConfigureTouchFromButton::ConfigureTouchFromButton( + QWidget* parent, const std::vector<Settings::TouchFromButtonMap>& touch_maps, + InputCommon::InputSubsystem* input_subsystem_, const int default_index) + : QDialog(parent), ui(std::make_unique<Ui::ConfigureTouchFromButton>()), + touch_maps(touch_maps), input_subsystem{input_subsystem_}, selected_index(default_index), + timeout_timer(std::make_unique<QTimer>()), poll_timer(std::make_unique<QTimer>()) { + ui->setupUi(this); + binding_list_model = new QStandardItemModel(0, 3, this); + binding_list_model->setHorizontalHeaderLabels( + {tr("Button"), tr("X", "X axis"), tr("Y", "Y axis")}); + ui->binding_list->setModel(binding_list_model); + ui->bottom_screen->SetCoordLabel(ui->coord_label); + + SetConfiguration(); + UpdateUiDisplay(); + ConnectEvents(); +} + +ConfigureTouchFromButton::~ConfigureTouchFromButton() = default; + +void ConfigureTouchFromButton::showEvent(QShowEvent* ev) { + QWidget::showEvent(ev); + + // width values are not valid in the constructor + const int w = + ui->binding_list->viewport()->contentsRect().width() / binding_list_model->columnCount(); + if (w <= 0) { + return; + } + ui->binding_list->setColumnWidth(0, w); + ui->binding_list->setColumnWidth(1, w); + ui->binding_list->setColumnWidth(2, w); +} + +void ConfigureTouchFromButton::SetConfiguration() { + for (const auto& touch_map : touch_maps) { + ui->mapping->addItem(QString::fromStdString(touch_map.name)); + } + + ui->mapping->setCurrentIndex(selected_index); +} + +void ConfigureTouchFromButton::UpdateUiDisplay() { + ui->button_delete->setEnabled(touch_maps.size() > 1); + ui->button_delete_bind->setEnabled(false); + + binding_list_model->removeRows(0, binding_list_model->rowCount()); + + for (const auto& button_str : touch_maps[selected_index].buttons) { + Common::ParamPackage package{button_str}; + QStandardItem* button = new QStandardItem(ButtonToText(package)); + button->setData(QString::fromStdString(button_str)); + button->setEditable(false); + QStandardItem* xcoord = new QStandardItem(QString::number(package.Get("x", 0))); + QStandardItem* ycoord = new QStandardItem(QString::number(package.Get("y", 0))); + binding_list_model->appendRow({button, xcoord, ycoord}); + + const int dot = ui->bottom_screen->AddDot(package.Get("x", 0), package.Get("y", 0)); + button->setData(dot, DataRoleDot); + } +} + +void ConfigureTouchFromButton::ConnectEvents() { + connect(ui->mapping, qOverload<int>(&QComboBox::currentIndexChanged), this, [this](int index) { + SaveCurrentMapping(); + selected_index = index; + UpdateUiDisplay(); + }); + connect(ui->button_new, &QPushButton::clicked, this, &ConfigureTouchFromButton::NewMapping); + connect(ui->button_delete, &QPushButton::clicked, this, + &ConfigureTouchFromButton::DeleteMapping); + connect(ui->button_rename, &QPushButton::clicked, this, + &ConfigureTouchFromButton::RenameMapping); + connect(ui->button_delete_bind, &QPushButton::clicked, this, + &ConfigureTouchFromButton::DeleteBinding); + connect(ui->binding_list, &QTreeView::doubleClicked, this, + &ConfigureTouchFromButton::EditBinding); + connect(ui->binding_list->selectionModel(), &QItemSelectionModel::selectionChanged, this, + &ConfigureTouchFromButton::OnBindingSelection); + connect(binding_list_model, &QStandardItemModel::itemChanged, this, + &ConfigureTouchFromButton::OnBindingChanged); + connect(ui->binding_list->model(), &QStandardItemModel::rowsAboutToBeRemoved, this, + &ConfigureTouchFromButton::OnBindingDeleted); + connect(ui->bottom_screen, &TouchScreenPreview::DotAdded, this, + &ConfigureTouchFromButton::NewBinding); + connect(ui->bottom_screen, &TouchScreenPreview::DotSelected, this, + &ConfigureTouchFromButton::SetActiveBinding); + connect(ui->bottom_screen, &TouchScreenPreview::DotMoved, this, + &ConfigureTouchFromButton::SetCoordinates); + connect(ui->buttonBox, &QDialogButtonBox::accepted, this, + &ConfigureTouchFromButton::ApplyConfiguration); + + connect(timeout_timer.get(), &QTimer::timeout, [this]() { SetPollingResult({}, true); }); + + connect(poll_timer.get(), &QTimer::timeout, [this]() { + Common::ParamPackage params; + for (auto& poller : device_pollers) { + params = poller->GetNextInput(); + if (params.Has("engine")) { + SetPollingResult(params, false); + return; + } + } + }); +} + +void ConfigureTouchFromButton::SaveCurrentMapping() { + auto& map = touch_maps[selected_index]; + map.buttons.clear(); + for (int i = 0, rc = binding_list_model->rowCount(); i < rc; ++i) { + const auto bind_str = binding_list_model->index(i, 0) + .data(Qt::ItemDataRole::UserRole + 1) + .toString() + .toStdString(); + if (bind_str.empty()) { + continue; + } + Common::ParamPackage params{bind_str}; + if (!params.Has("engine")) { + continue; + } + params.Set("x", binding_list_model->index(i, 1).data().toInt()); + params.Set("y", binding_list_model->index(i, 2).data().toInt()); + map.buttons.emplace_back(params.Serialize()); + } +} + +void ConfigureTouchFromButton::NewMapping() { + const QString name = + QInputDialog::getText(this, tr("New Profile"), tr("Enter the name for the new profile.")); + if (name.isEmpty()) { + return; + } + touch_maps.emplace_back(Settings::TouchFromButtonMap{name.toStdString(), {}}); + ui->mapping->addItem(name); + ui->mapping->setCurrentIndex(ui->mapping->count() - 1); +} + +void ConfigureTouchFromButton::DeleteMapping() { + const auto answer = QMessageBox::question( + this, tr("Delete Profile"), tr("Delete profile %1?").arg(ui->mapping->currentText())); + if (answer != QMessageBox::Yes) { + return; + } + const bool blocked = ui->mapping->blockSignals(true); + ui->mapping->removeItem(selected_index); + ui->mapping->blockSignals(blocked); + touch_maps.erase(touch_maps.begin() + selected_index); + selected_index = ui->mapping->currentIndex(); + UpdateUiDisplay(); +} + +void ConfigureTouchFromButton::RenameMapping() { + const QString new_name = QInputDialog::getText(this, tr("Rename Profile"), tr("New name:")); + if (new_name.isEmpty()) { + return; + } + ui->mapping->setItemText(selected_index, new_name); + touch_maps[selected_index].name = new_name.toStdString(); +} + +void ConfigureTouchFromButton::GetButtonInput(const int row_index, const bool is_new) { + binding_list_model->item(row_index, 0)->setText(tr("[press key]")); + + input_setter = [this, row_index, is_new](const Common::ParamPackage& params, + const bool cancel) { + auto* cell = binding_list_model->item(row_index, 0); + if (cancel) { + if (is_new) { + binding_list_model->removeRow(row_index); + } else { + cell->setText( + ButtonToText(Common::ParamPackage{cell->data().toString().toStdString()})); + } + } else { + cell->setText(ButtonToText(params)); + cell->setData(QString::fromStdString(params.Serialize())); + } + }; + + device_pollers = input_subsystem->GetPollers(InputCommon::Polling::DeviceType::Button); + + for (auto& poller : device_pollers) { + poller->Start(); + } + + grabKeyboard(); + grabMouse(); + qApp->setOverrideCursor(QCursor(Qt::CursorShape::ArrowCursor)); + timeout_timer->start(5000); // Cancel after 5 seconds + poll_timer->start(200); // Check for new inputs every 200ms +} + +void ConfigureTouchFromButton::NewBinding(const QPoint& pos) { + auto* button = new QStandardItem(); + button->setEditable(false); + auto* x_coord = new QStandardItem(QString::number(pos.x())); + auto* y_coord = new QStandardItem(QString::number(pos.y())); + + const int dot_id = ui->bottom_screen->AddDot(pos.x(), pos.y()); + button->setData(dot_id, DataRoleDot); + + binding_list_model->appendRow({button, x_coord, y_coord}); + ui->binding_list->setFocus(); + ui->binding_list->setCurrentIndex(button->index()); + + GetButtonInput(binding_list_model->rowCount() - 1, true); +} + +void ConfigureTouchFromButton::EditBinding(const QModelIndex& qi) { + if (qi.row() >= 0 && qi.column() == 0) { + GetButtonInput(qi.row(), false); + } +} + +void ConfigureTouchFromButton::DeleteBinding() { + const int row_index = ui->binding_list->currentIndex().row(); + if (row_index < 0) { + return; + } + ui->bottom_screen->RemoveDot(binding_list_model->index(row_index, 0).data(DataRoleDot).toInt()); + binding_list_model->removeRow(row_index); +} + +void ConfigureTouchFromButton::OnBindingSelection(const QItemSelection& selected, + const QItemSelection& deselected) { + ui->button_delete_bind->setEnabled(!selected.isEmpty()); + if (!selected.isEmpty()) { + const auto dot_data = selected.indexes().first().data(DataRoleDot); + if (dot_data.isValid()) { + ui->bottom_screen->HighlightDot(dot_data.toInt()); + } + } + if (!deselected.isEmpty()) { + const auto dot_data = deselected.indexes().first().data(DataRoleDot); + if (dot_data.isValid()) { + ui->bottom_screen->HighlightDot(dot_data.toInt(), false); + } + } +} + +void ConfigureTouchFromButton::OnBindingChanged(QStandardItem* item) { + if (item->column() == 0) { + return; + } + + const bool blocked = binding_list_model->blockSignals(true); + item->setText(QString::number( + std::clamp(item->text().toInt(), 0, + static_cast<int>((item->column() == 1 ? Layout::ScreenUndocked::Width + : Layout::ScreenUndocked::Height) - + 1)))); + binding_list_model->blockSignals(blocked); + + const auto dot_data = binding_list_model->index(item->row(), 0).data(DataRoleDot); + if (dot_data.isValid()) { + ui->bottom_screen->MoveDot(dot_data.toInt(), + binding_list_model->item(item->row(), 1)->text().toInt(), + binding_list_model->item(item->row(), 2)->text().toInt()); + } +} + +void ConfigureTouchFromButton::OnBindingDeleted(const QModelIndex& parent, int first, int last) { + for (int i = first; i <= last; ++i) { + const auto ix = binding_list_model->index(i, 0); + if (!ix.isValid()) { + return; + } + const auto dot_data = ix.data(DataRoleDot); + if (dot_data.isValid()) { + ui->bottom_screen->RemoveDot(dot_data.toInt()); + } + } +} + +void ConfigureTouchFromButton::SetActiveBinding(const int dot_id) { + for (int i = 0; i < binding_list_model->rowCount(); ++i) { + if (binding_list_model->index(i, 0).data(DataRoleDot) == dot_id) { + ui->binding_list->setCurrentIndex(binding_list_model->index(i, 0)); + ui->binding_list->setFocus(); + return; + } + } +} + +void ConfigureTouchFromButton::SetCoordinates(const int dot_id, const QPoint& pos) { + for (int i = 0; i < binding_list_model->rowCount(); ++i) { + if (binding_list_model->item(i, 0)->data(DataRoleDot) == dot_id) { + binding_list_model->item(i, 1)->setText(QString::number(pos.x())); + binding_list_model->item(i, 2)->setText(QString::number(pos.y())); + return; + } + } +} + +void ConfigureTouchFromButton::SetPollingResult(const Common::ParamPackage& params, + const bool cancel) { + releaseKeyboard(); + releaseMouse(); + qApp->restoreOverrideCursor(); + timeout_timer->stop(); + poll_timer->stop(); + for (auto& poller : device_pollers) { + poller->Stop(); + } + if (input_setter) { + (*input_setter)(params, cancel); + input_setter.reset(); + } +} + +void ConfigureTouchFromButton::keyPressEvent(QKeyEvent* event) { + if (!input_setter && event->key() == Qt::Key_Delete) { + DeleteBinding(); + return; + } + + if (!input_setter) { + return QDialog::keyPressEvent(event); + } + + if (event->key() != Qt::Key_Escape) { + SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->key())}, + false); + } else { + SetPollingResult({}, true); + } +} + +void ConfigureTouchFromButton::ApplyConfiguration() { + SaveCurrentMapping(); + accept(); +} + +int ConfigureTouchFromButton::GetSelectedIndex() const { + return selected_index; +} + +std::vector<Settings::TouchFromButtonMap> ConfigureTouchFromButton::GetMaps() const { + return touch_maps; +} + +TouchScreenPreview::TouchScreenPreview(QWidget* parent) : QFrame(parent) { + setBackgroundRole(QPalette::ColorRole::Base); +} + +TouchScreenPreview::~TouchScreenPreview() = default; + +void TouchScreenPreview::SetCoordLabel(QLabel* const label) { + coord_label = label; +} + +int TouchScreenPreview::AddDot(const int device_x, const int device_y) { + QFont dot_font{QStringLiteral("monospace")}; + dot_font.setStyleHint(QFont::Monospace); + dot_font.setPointSize(20); + + auto* dot = new QLabel(this); + dot->setAttribute(Qt::WA_TranslucentBackground); + dot->setFont(dot_font); + dot->setText(QChar(0xD7)); // U+00D7 Multiplication Sign + dot->setAlignment(Qt::AlignmentFlag::AlignCenter); + dot->setProperty(PropId, ++max_dot_id); + dot->setProperty(PropX, device_x); + dot->setProperty(PropY, device_y); + dot->setCursor(Qt::CursorShape::PointingHandCursor); + dot->setMouseTracking(true); + dot->installEventFilter(this); + dot->show(); + PositionDot(dot, device_x, device_y); + dots.emplace_back(max_dot_id, dot); + return max_dot_id; +} + +void TouchScreenPreview::RemoveDot(const int id) { + const auto iter = std::find_if(dots.begin(), dots.end(), + [id](const auto& entry) { return entry.first == id; }); + if (iter == dots.cend()) { + return; + } + + iter->second->deleteLater(); + dots.erase(iter); +} + +void TouchScreenPreview::HighlightDot(const int id, const bool active) const { + for (const auto& dot : dots) { + if (dot.first == id) { + // use color property from the stylesheet, or fall back to the default palette + if (dot_highlight_color.isValid()) { + dot.second->setStyleSheet( + active ? QStringLiteral("color: %1").arg(dot_highlight_color.name()) + : QString{}); + } else { + dot.second->setForegroundRole(active ? QPalette::ColorRole::LinkVisited + : QPalette::ColorRole::NoRole); + } + if (active) { + dot.second->raise(); + } + return; + } + } +} + +void TouchScreenPreview::MoveDot(const int id, const int device_x, const int device_y) const { + const auto iter = std::find_if(dots.begin(), dots.end(), + [id](const auto& entry) { return entry.first == id; }); + if (iter == dots.cend()) { + return; + } + + iter->second->setProperty(PropX, device_x); + iter->second->setProperty(PropY, device_y); + PositionDot(iter->second, device_x, device_y); +} + +void TouchScreenPreview::resizeEvent(QResizeEvent* event) { + if (ignore_resize) { + return; + } + + const int target_width = std::min(width(), height() * 4 / 3); + const int target_height = std::min(height(), width() * 3 / 4); + if (target_width == width() && target_height == height()) { + return; + } + ignore_resize = true; + setGeometry((parentWidget()->contentsRect().width() - target_width) / 2, y(), target_width, + target_height); + ignore_resize = false; + + if (event->oldSize().width() != target_width || event->oldSize().height() != target_height) { + for (const auto& dot : dots) { + PositionDot(dot.second); + } + } +} + +void TouchScreenPreview::mouseMoveEvent(QMouseEvent* event) { + if (!coord_label) { + return; + } + const auto pos = MapToDeviceCoords(event->x(), event->y()); + if (pos) { + coord_label->setText(QStringLiteral("X: %1, Y: %2").arg(pos->x()).arg(pos->y())); + } else { + coord_label->clear(); + } +} + +void TouchScreenPreview::leaveEvent(QEvent* event) { + if (coord_label) { + coord_label->clear(); + } +} + +void TouchScreenPreview::mousePressEvent(QMouseEvent* event) { + if (event->button() != Qt::MouseButton::LeftButton) { + return; + } + const auto pos = MapToDeviceCoords(event->x(), event->y()); + if (pos) { + emit DotAdded(*pos); + } +} + +bool TouchScreenPreview::eventFilter(QObject* obj, QEvent* event) { + switch (event->type()) { + case QEvent::Type::MouseButtonPress: { + const auto mouse_event = static_cast<QMouseEvent*>(event); + if (mouse_event->button() != Qt::MouseButton::LeftButton) { + break; + } + emit DotSelected(obj->property(PropId).toInt()); + + drag_state.dot = qobject_cast<QLabel*>(obj); + drag_state.start_pos = mouse_event->globalPos(); + return true; + } + case QEvent::Type::MouseMove: { + if (!drag_state.dot) { + break; + } + const auto mouse_event = static_cast<QMouseEvent*>(event); + if (!drag_state.active) { + drag_state.active = + (mouse_event->globalPos() - drag_state.start_pos).manhattanLength() >= + QApplication::startDragDistance(); + if (!drag_state.active) { + break; + } + } + auto current_pos = mapFromGlobal(mouse_event->globalPos()); + current_pos.setX(std::clamp(current_pos.x(), contentsMargins().left(), + contentsMargins().left() + contentsRect().width() - 1)); + current_pos.setY(std::clamp(current_pos.y(), contentsMargins().top(), + contentsMargins().top() + contentsRect().height() - 1)); + const auto device_coord = MapToDeviceCoords(current_pos.x(), current_pos.y()); + if (device_coord) { + drag_state.dot->setProperty(PropX, device_coord->x()); + drag_state.dot->setProperty(PropY, device_coord->y()); + PositionDot(drag_state.dot, device_coord->x(), device_coord->y()); + emit DotMoved(drag_state.dot->property(PropId).toInt(), *device_coord); + if (coord_label) { + coord_label->setText( + QStringLiteral("X: %1, Y: %2").arg(device_coord->x()).arg(device_coord->y())); + } + } + return true; + } + case QEvent::Type::MouseButtonRelease: { + drag_state.dot.clear(); + drag_state.active = false; + return true; + } + default: + break; + } + return obj->eventFilter(obj, event); +} + +std::optional<QPoint> TouchScreenPreview::MapToDeviceCoords(const int screen_x, + const int screen_y) const { + const float t_x = 0.5f + static_cast<float>(screen_x - contentsMargins().left()) * + (Layout::ScreenUndocked::Width - 1) / (contentsRect().width() - 1); + const float t_y = 0.5f + static_cast<float>(screen_y - contentsMargins().top()) * + (Layout::ScreenUndocked::Height - 1) / + (contentsRect().height() - 1); + if (t_x >= 0.5f && t_x < Layout::ScreenUndocked::Width && t_y >= 0.5f && + t_y < Layout::ScreenUndocked::Height) { + + return QPoint{static_cast<int>(t_x), static_cast<int>(t_y)}; + } + return std::nullopt; +} + +void TouchScreenPreview::PositionDot(QLabel* const dot, const int device_x, + const int device_y) const { + const float device_coord_x = + static_cast<float>(device_x >= 0 ? device_x : dot->property(PropX).toInt()); + int x_coord = static_cast<int>( + device_coord_x * (contentsRect().width() - 1) / (Layout::ScreenUndocked::Width - 1) + + contentsMargins().left() - static_cast<float>(dot->width()) / 2 + 0.5f); + + const float device_coord_y = + static_cast<float>(device_y >= 0 ? device_y : dot->property(PropY).toInt()); + const int y_coord = static_cast<int>( + device_coord_y * (contentsRect().height() - 1) / (Layout::ScreenUndocked::Height - 1) + + contentsMargins().top() - static_cast<float>(dot->height()) / 2 + 0.5f); + + dot->move(x_coord, y_coord); +} diff --git a/src/yuzu/configuration/configure_touch_from_button.h b/src/yuzu/configuration/configure_touch_from_button.h new file mode 100644 index 000000000..d9513e3bc --- /dev/null +++ b/src/yuzu/configuration/configure_touch_from_button.h @@ -0,0 +1,92 @@ +// Copyright 2020 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <functional> +#include <memory> +#include <optional> +#include <vector> +#include <QDialog> + +class QItemSelection; +class QModelIndex; +class QStandardItemModel; +class QStandardItem; +class QTimer; + +namespace Common { +class ParamPackage; +} + +namespace InputCommon { +class InputSubsystem; +} + +namespace InputCommon::Polling { +class DevicePoller; +} + +namespace Settings { +struct TouchFromButtonMap; +} + +namespace Ui { +class ConfigureTouchFromButton; +} + +class ConfigureTouchFromButton : public QDialog { + Q_OBJECT + +public: + explicit ConfigureTouchFromButton(QWidget* parent, + const std::vector<Settings::TouchFromButtonMap>& touch_maps, + InputCommon::InputSubsystem* input_subsystem_, + int default_index = 0); + ~ConfigureTouchFromButton() override; + + int GetSelectedIndex() const; + std::vector<Settings::TouchFromButtonMap> GetMaps() const; + +public slots: + void ApplyConfiguration(); + void NewBinding(const QPoint& pos); + void SetActiveBinding(int dot_id); + void SetCoordinates(int dot_id, const QPoint& pos); + +protected: + void showEvent(QShowEvent* ev) override; + void keyPressEvent(QKeyEvent* event) override; + +private slots: + void NewMapping(); + void DeleteMapping(); + void RenameMapping(); + void EditBinding(const QModelIndex& qi); + void DeleteBinding(); + void OnBindingSelection(const QItemSelection& selected, const QItemSelection& deselected); + void OnBindingChanged(QStandardItem* item); + void OnBindingDeleted(const QModelIndex& parent, int first, int last); + +private: + void SetConfiguration(); + void UpdateUiDisplay(); + void ConnectEvents(); + void GetButtonInput(int row_index, bool is_new); + void SetPollingResult(const Common::ParamPackage& params, bool cancel); + void SaveCurrentMapping(); + + std::unique_ptr<Ui::ConfigureTouchFromButton> ui; + std::vector<Settings::TouchFromButtonMap> touch_maps; + QStandardItemModel* binding_list_model; + InputCommon::InputSubsystem* input_subsystem; + int selected_index; + + std::unique_ptr<QTimer> timeout_timer; + std::unique_ptr<QTimer> poll_timer; + std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> device_pollers; + std::optional<std::function<void(const Common::ParamPackage&, bool)>> input_setter; + + static constexpr int DataRoleDot = Qt::ItemDataRole::UserRole + 2; +}; diff --git a/src/yuzu/configuration/configure_touch_from_button.ui b/src/yuzu/configuration/configure_touch_from_button.ui new file mode 100644 index 000000000..757219d54 --- /dev/null +++ b/src/yuzu/configuration/configure_touch_from_button.ui @@ -0,0 +1,221 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureTouchFromButton</class> + <widget class="QDialog" name="ConfigureTouchFromButton"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>500</width> + <height>500</height> + </rect> + </property> + <property name="windowTitle"> + <string>Configure Touchscreen Mappings</string> + </property> + <layout class="QVBoxLayout"> + <item> + <layout class="QHBoxLayout" name="horizontalLayout"> + <item> + <widget class="QLabel" name="label"> + <property name="text"> + <string>Mapping:</string> + </property> + <property name="textFormat"> + <enum>Qt::PlainText</enum> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="mapping"> + <property name="sizePolicy"> + <sizepolicy hsizetype="MinimumExpanding" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="button_new"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>New</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="button_delete"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Delete</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="button_rename"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Rename</string> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="Line" name="line"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <item> + <widget class="QLabel" name="label_2"> + <property name="text"> + <string>Click the bottom area to add a point, then press a button to bind. +Drag points to change position, or double-click table cells to edit values.</string> + </property> + <property name="textFormat"> + <enum>Qt::PlainText</enum> + </property> + </widget> + </item> + <item> + <spacer name="horizontalSpacer"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QPushButton" name="button_delete_bind"> + <property name="text"> + <string>Delete Point</string> + </property> + </widget> + </item> + </layout> + </item> + <item> + <widget class="QTreeView" name="binding_list"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Expanding" vsizetype="Expanding"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="rootIsDecorated"> + <bool>false</bool> + </property> + <property name="uniformRowHeights"> + <bool>true</bool> + </property> + <property name="itemsExpandable"> + <bool>false</bool> + </property> + </widget> + </item> + <item> + <widget class="TouchScreenPreview" name="bottom_screen"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Expanding" vsizetype="Expanding"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="minimumSize"> + <size> + <width>160</width> + <height>120</height> + </size> + </property> + <property name="baseSize"> + <size> + <width>320</width> + <height>240</height> + </size> + </property> + <property name="cursor"> + <cursorShape>CrossCursor</cursorShape> + </property> + <property name="mouseTracking"> + <bool>true</bool> + </property> + <property name="autoFillBackground"> + <bool>true</bool> + </property> + <property name="frameShape"> + <enum>QFrame::StyledPanel</enum> + </property> + <property name="frameShadow"> + <enum>QFrame::Sunken</enum> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <item> + <widget class="QLabel" name="coord_label"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Expanding" vsizetype="Preferred"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="textFormat"> + <enum>Qt::PlainText</enum> + </property> + </widget> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBox"> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + <customwidgets> + <customwidget> + <class>TouchScreenPreview</class> + <extends>QFrame</extends> + <header>yuzu/configuration/configure_touch_widget.h</header> + <container>1</container> + </customwidget> + </customwidgets> + <resources/> + <connections> + <connection> + <sender>buttonBox</sender> + <signal>rejected()</signal> + <receiver>ConfigureTouchFromButton</receiver> + <slot>reject()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/configure_touch_widget.h b/src/yuzu/configuration/configure_touch_widget.h new file mode 100644 index 000000000..347b46583 --- /dev/null +++ b/src/yuzu/configuration/configure_touch_widget.h @@ -0,0 +1,62 @@ +// Copyright 2020 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <optional> +#include <utility> +#include <vector> +#include <QFrame> +#include <QPointer> + +class QLabel; + +// Widget for representing touchscreen coordinates +class TouchScreenPreview : public QFrame { + Q_OBJECT + Q_PROPERTY(QColor dotHighlightColor MEMBER dot_highlight_color) + +public: + explicit TouchScreenPreview(QWidget* parent); + ~TouchScreenPreview() override; + + void SetCoordLabel(QLabel*); + int AddDot(int device_x, int device_y); + void RemoveDot(int id); + void HighlightDot(int id, bool active = true) const; + void MoveDot(int id, int device_x, int device_y) const; + +signals: + void DotAdded(const QPoint& pos); + void DotSelected(int dot_id); + void DotMoved(int dot_id, const QPoint& pos); + +protected: + void resizeEvent(QResizeEvent*) override; + void mouseMoveEvent(QMouseEvent*) override; + void leaveEvent(QEvent*) override; + void mousePressEvent(QMouseEvent*) override; + bool eventFilter(QObject*, QEvent*) override; + +private: + std::optional<QPoint> MapToDeviceCoords(int screen_x, int screen_y) const; + void PositionDot(QLabel* dot, int device_x = -1, int device_y = -1) const; + + bool ignore_resize = false; + QPointer<QLabel> coord_label; + + std::vector<std::pair<int, QLabel*>> dots; + int max_dot_id = 0; + QColor dot_highlight_color; + static constexpr char PropId[] = "dot_id"; + static constexpr char PropX[] = "device_x"; + static constexpr char PropY[] = "device_y"; + + struct DragState { + bool active = false; + QPointer<QLabel> dot; + QPoint start_pos; + }; + DragState drag_state; +}; diff --git a/src/yuzu/configuration/configure_touchscreen_advanced.ui b/src/yuzu/configuration/configure_touchscreen_advanced.ui index 1171c2dd1..30ceccddb 100644 --- a/src/yuzu/configuration/configure_touchscreen_advanced.ui +++ b/src/yuzu/configuration/configure_touchscreen_advanced.ui @@ -168,32 +168,12 @@ <signal>accepted()</signal> <receiver>ConfigureTouchscreenAdvanced</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>140</x> - <y>318</y> - </hint> - <hint type="destinationlabel"> - <x>140</x> - <y>169</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureTouchscreenAdvanced</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>140</x> - <y>318</y> - </hint> - <hint type="destinationlabel"> - <x>140</x> - <y>169</y> - </hint> - </hints> - </connection> + </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_ui.cpp b/src/yuzu/configuration/configure_ui.cpp index 94424ee44..dbe3f78c8 100644 --- a/src/yuzu/configuration/configure_ui.cpp +++ b/src/yuzu/configuration/configure_ui.cpp @@ -4,8 +4,11 @@ #include <array> #include <utility> +#include <QFileDialog> +#include <QDirIterator> #include "common/common_types.h" +#include "common/file_util.h" #include "core/settings.h" #include "ui_configure_ui.h" #include "yuzu/configuration/configure_ui.h" @@ -29,6 +32,8 @@ constexpr std::array row_text_names{ ConfigureUi::ConfigureUi(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureUi) { ui->setupUi(this); + InitializeLanguageComboBox(); + for (const auto& theme : UISettings::themes) { ui->theme_combobox->addItem(QString::fromUtf8(theme.first), QString::fromUtf8(theme.second)); @@ -49,9 +54,21 @@ ConfigureUi::ConfigureUi(QWidget* parent) : QWidget(parent), ui(new Ui::Configur // Update text ComboBoxes after user interaction. connect(ui->row_1_text_combobox, QOverload<int>::of(&QComboBox::activated), - [=]() { ConfigureUi::UpdateSecondRowComboBox(); }); + [this] { ConfigureUi::UpdateSecondRowComboBox(); }); connect(ui->row_2_text_combobox, QOverload<int>::of(&QComboBox::activated), - [=]() { ConfigureUi::UpdateFirstRowComboBox(); }); + [this] { ConfigureUi::UpdateFirstRowComboBox(); }); + + // Set screenshot path to user specification. + connect(ui->screenshot_path_button, &QToolButton::pressed, this, [this] { + const QString& filename = + QFileDialog::getExistingDirectory(this, tr("Select Screenshots Path..."), + QString::fromStdString(Common::FS::GetUserPath( + Common::FS::UserPath::ScreenshotsDir))) + + QDir::separator(); + if (!filename.isEmpty()) { + ui->screenshot_path_edit->setText(filename); + } + }); } ConfigureUi::~ConfigureUi() = default; @@ -63,6 +80,10 @@ void ConfigureUi::ApplyConfiguration() { UISettings::values.icon_size = ui->icon_size_combobox->currentData().toUInt(); UISettings::values.row_1_text_id = ui->row_1_text_combobox->currentData().toUInt(); UISettings::values.row_2_text_id = ui->row_2_text_combobox->currentData().toUInt(); + + UISettings::values.enable_screenshot_save_as = ui->enable_screenshot_save_as->isChecked(); + Common::FS::GetUserPath(Common::FS::UserPath::ScreenshotsDir, + ui->screenshot_path_edit->text().toStdString()); Settings::Apply(); } @@ -72,9 +93,15 @@ void ConfigureUi::RequestGameListUpdate() { void ConfigureUi::SetConfiguration() { ui->theme_combobox->setCurrentIndex(ui->theme_combobox->findData(UISettings::values.theme)); + ui->language_combobox->setCurrentIndex( + ui->language_combobox->findData(UISettings::values.language)); ui->show_add_ons->setChecked(UISettings::values.show_add_ons); ui->icon_size_combobox->setCurrentIndex( ui->icon_size_combobox->findData(UISettings::values.icon_size)); + + ui->enable_screenshot_save_as->setChecked(UISettings::values.enable_screenshot_save_as); + ui->screenshot_path_edit->setText( + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::ScreenshotsDir))); } void ConfigureUi::changeEvent(QEvent* event) { @@ -100,6 +127,25 @@ void ConfigureUi::RetranslateUI() { } } +void ConfigureUi::InitializeLanguageComboBox() { + ui->language_combobox->addItem(tr("<System>"), QString{}); + ui->language_combobox->addItem(tr("English"), QStringLiteral("en")); + QDirIterator it(QStringLiteral(":/languages"), QDirIterator::NoIteratorFlags); + while (it.hasNext()) { + QString locale = it.next(); + locale.truncate(locale.lastIndexOf(QLatin1Char{'.'})); + locale.remove(0, locale.lastIndexOf(QLatin1Char{'/'}) + 1); + const QString lang = QLocale::languageToString(QLocale(locale).language()); + ui->language_combobox->addItem(lang, locale); + } + + // Unlike other configuration changes, interface language changes need to be reflected on the + // interface immediately. This is done by passing a signal to the main window, and then + // retranslating when passing back. + connect(ui->language_combobox, QOverload<int>::of(&QComboBox::currentIndexChanged), this, + &ConfigureUi::OnLanguageChanged); +} + void ConfigureUi::InitializeIconSizeComboBox() { for (const auto& size : default_icon_sizes) { ui->icon_size_combobox->addItem(QString::fromUtf8(size.second), size.first); @@ -147,3 +193,10 @@ void ConfigureUi::UpdateSecondRowComboBox(bool init) { ui->row_2_text_combobox->removeItem( ui->row_2_text_combobox->findData(ui->row_1_text_combobox->currentData())); } + +void ConfigureUi::OnLanguageChanged(int index) { + if (index == -1) + return; + + emit LanguageChanged(ui->language_combobox->itemData(index).toString()); +} diff --git a/src/yuzu/configuration/configure_ui.h b/src/yuzu/configuration/configure_ui.h index d471afe99..c30bcf6ff 100644 --- a/src/yuzu/configuration/configure_ui.h +++ b/src/yuzu/configuration/configure_ui.h @@ -20,6 +20,12 @@ public: void ApplyConfiguration(); +private slots: + void OnLanguageChanged(int index); + +signals: + void LanguageChanged(const QString& locale); + private: void RequestGameListUpdate(); @@ -28,6 +34,7 @@ private: void changeEvent(QEvent*) override; void RetranslateUI(); + void InitializeLanguageComboBox(); void InitializeIconSizeComboBox(); void InitializeRowComboBoxes(); diff --git a/src/yuzu/configuration/configure_ui.ui b/src/yuzu/configuration/configure_ui.ui index bd5c5d3c2..d895b799f 100644 --- a/src/yuzu/configuration/configure_ui.ui +++ b/src/yuzu/configuration/configure_ui.ui @@ -6,119 +6,180 @@ <rect> <x>0</x> <y>0</y> - <width>300</width> - <height>377</height> + <width>363</width> + <height>391</height> </rect> </property> <property name="windowTitle"> <string>Form</string> </property> - <layout class="QHBoxLayout" name="HorizontalLayout"> + <layout class="QVBoxLayout" name="verticalLayout"> <item> - <layout class="QVBoxLayout" name="VerticalLayout"> - <item> - <widget class="QGroupBox" name="GeneralGroupBox"> - <property name="title"> - <string>General</string> - </property> - <layout class="QHBoxLayout" name="horizontalLayout"> + <widget class="QGroupBox" name="general_groupBox"> + <property name="title"> + <string>General</string> + </property> + <layout class="QHBoxLayout" name="horizontalLayout"> + <item> + <layout class="QVBoxLayout" name="verticalLayout_2"> <item> - <layout class="QVBoxLayout" name="verticalLayout"> + <widget class="QLabel" name="label_change_language_info"> + <property name="text"> + <string>Note: Changing language will apply your configuration.</string> + </property> + <property name="wordWrap"> + <bool>true</bool> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <item> + <widget class="QLabel" name="language_label"> + <property name="text"> + <string>Interface language:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="language_combobox"/> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <item> + <widget class="QLabel" name="theme_label"> + <property name="text"> + <string>Theme:</string> + </property> + </widget> + </item> <item> - <layout class="QHBoxLayout" name="horizontalLayout_3"> - <item> - <widget class="QLabel" name="theme_label"> - <property name="text"> - <string>Theme:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="theme_combobox"/> - </item> - </layout> + <widget class="QComboBox" name="theme_combobox"/> </item> </layout> </item> </layout> - </widget> - </item> - <item> - <widget class="QGroupBox" name="GameListGroupBox"> - <property name="title"> - <string>Game List</string> - </property> - <layout class="QHBoxLayout" name="GameListHorizontalLayout"> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="GameListGroupBox"> + <property name="title"> + <string>Game List</string> + </property> + <layout class="QHBoxLayout" name="GameListHorizontalLayout"> + <item> + <layout class="QVBoxLayout" name="GeneralVerticalLayout"> <item> - <layout class="QVBoxLayout" name="GeneralVerticalLayout"> + <widget class="QCheckBox" name="show_add_ons"> + <property name="text"> + <string>Show Add-Ons Column</string> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="icon_size_qhbox_layout_2"> <item> - <widget class="QCheckBox" name="show_add_ons"> + <widget class="QLabel" name="icon_size_label"> <property name="text"> - <string>Show Add-Ons Column</string> + <string>Icon Size:</string> </property> </widget> </item> <item> - <layout class="QHBoxLayout" name="icon_size_qhbox_layout_2"> - <item> - <widget class="QLabel" name="icon_size_label"> - <property name="text"> - <string>Icon Size:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="icon_size_combobox"/> - </item> - </layout> + <widget class="QComboBox" name="icon_size_combobox"/> </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout" name="row_1_qhbox_layout"> <item> - <layout class="QHBoxLayout" name="row_1_qhbox_layout"> - <item> - <widget class="QLabel" name="row_1_label"> - <property name="text"> - <string>Row 1 Text:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="row_1_text_combobox"/> - </item> - </layout> + <widget class="QLabel" name="row_1_label"> + <property name="text"> + <string>Row 1 Text:</string> + </property> + </widget> </item> <item> - <layout class="QHBoxLayout" name="row_2_qhbox_layout"> - <item> - <widget class="QLabel" name="row_2_label"> - <property name="text"> - <string>Row 2 Text:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="row_2_text_combobox"/> - </item> - </layout> + <widget class="QComboBox" name="row_1_text_combobox"/> + </item> + </layout> + </item> + <item> + <layout class="QHBoxLayout" name="row_2_qhbox_layout"> + <item> + <widget class="QLabel" name="row_2_label"> + <property name="text"> + <string>Row 2 Text:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="row_2_text_combobox"/> </item> </layout> </item> </layout> - </widget> - </item> - <item> - <spacer name="verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - </layout> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="screenshots_GroupBox"> + <property name="title"> + <string>Screenshots</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_4"> + <item> + <layout class="QVBoxLayout" name="verticalLayout_3"> + <item> + <widget class="QCheckBox" name="enable_screenshot_save_as"> + <property name="text"> + <string>Ask Where To Save Screenshots (Windows Only)</string> + </property> + </widget> + </item> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_4"> + <item> + <widget class="QLabel" name="label"> + <property name="text"> + <string>Screenshots Path: </string> + </property> + </widget> + </item> + <item> + <widget class="QLineEdit" name="screenshot_path_edit"/> + </item> + <item> + <widget class="QToolButton" name="screenshot_path_button"> + <property name="text"> + <string>...</string> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> </item> </layout> </widget> diff --git a/src/yuzu/configuration/configure_vibration.cpp b/src/yuzu/configuration/configure_vibration.cpp new file mode 100644 index 000000000..7dcb2c5b9 --- /dev/null +++ b/src/yuzu/configuration/configure_vibration.cpp @@ -0,0 +1,146 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <unordered_map> + +#include <fmt/format.h> + +#include "common/param_package.h" +#include "core/settings.h" +#include "ui_configure_vibration.h" +#include "yuzu/configuration/configure_vibration.h" + +ConfigureVibration::ConfigureVibration(QWidget* parent) + : QDialog(parent), ui(std::make_unique<Ui::ConfigureVibration>()) { + ui->setupUi(this); + + vibration_groupboxes = { + ui->vibrationGroupPlayer1, ui->vibrationGroupPlayer2, ui->vibrationGroupPlayer3, + ui->vibrationGroupPlayer4, ui->vibrationGroupPlayer5, ui->vibrationGroupPlayer6, + ui->vibrationGroupPlayer7, ui->vibrationGroupPlayer8, + }; + + vibration_spinboxes = { + ui->vibrationSpinPlayer1, ui->vibrationSpinPlayer2, ui->vibrationSpinPlayer3, + ui->vibrationSpinPlayer4, ui->vibrationSpinPlayer5, ui->vibrationSpinPlayer6, + ui->vibrationSpinPlayer7, ui->vibrationSpinPlayer8, + }; + + const auto& players = Settings::values.players.GetValue(); + + for (std::size_t i = 0; i < NUM_PLAYERS; ++i) { + vibration_groupboxes[i]->setChecked(players[i].vibration_enabled); + vibration_spinboxes[i]->setValue(players[i].vibration_strength); + } + + ui->checkBoxAccurateVibration->setChecked( + Settings::values.enable_accurate_vibrations.GetValue()); + + if (!Settings::IsConfiguringGlobal()) { + ui->checkBoxAccurateVibration->setDisabled(true); + } + + RetranslateUI(); +} + +ConfigureVibration::~ConfigureVibration() = default; + +void ConfigureVibration::ApplyConfiguration() { + auto& players = Settings::values.players.GetValue(); + + for (std::size_t i = 0; i < NUM_PLAYERS; ++i) { + players[i].vibration_enabled = vibration_groupboxes[i]->isChecked(); + players[i].vibration_strength = vibration_spinboxes[i]->value(); + } + + Settings::values.enable_accurate_vibrations.SetValue( + ui->checkBoxAccurateVibration->isChecked()); +} + +void ConfigureVibration::SetVibrationDevices(std::size_t player_index) { + using namespace Settings::NativeButton; + static constexpr std::array<std::array<Settings::NativeButton::Values, 6>, 2> buttons{{ + {DLeft, DUp, DRight, DDown, L, ZL}, // Left Buttons + {A, B, X, Y, R, ZR}, // Right Buttons + }}; + + auto& player = Settings::values.players.GetValue()[player_index]; + + for (std::size_t device_idx = 0; device_idx < buttons.size(); ++device_idx) { + std::unordered_map<std::string, int> params_count; + + for (const auto button_index : buttons[device_idx]) { + const auto& player_button = player.buttons[button_index]; + + if (params_count.find(player_button) != params_count.end()) { + ++params_count[player_button]; + continue; + } + + params_count.insert_or_assign(player_button, 1); + } + + const auto it = std::max_element( + params_count.begin(), params_count.end(), + [](const auto& lhs, const auto& rhs) { return lhs.second < rhs.second; }); + + auto& vibration_param_str = player.vibrations[device_idx]; + vibration_param_str.clear(); + + if (it->first.empty()) { + continue; + } + + const auto param = Common::ParamPackage(it->first); + + const auto engine = param.Get("engine", ""); + const auto guid = param.Get("guid", ""); + const auto port = param.Get("port", ""); + + if (engine.empty() || engine == "keyboard" || engine == "mouse") { + continue; + } + + vibration_param_str += fmt::format("engine:{}", engine); + + if (!port.empty()) { + vibration_param_str += fmt::format(",port:{}", port); + } + if (!guid.empty()) { + vibration_param_str += fmt::format(",guid:{}", guid); + } + } + + if (player.vibrations[0] != player.vibrations[1]) { + return; + } + + if (!player.vibrations[0].empty() && + player.controller_type != Settings::ControllerType::RightJoycon) { + player.vibrations[1].clear(); + } else if (!player.vibrations[1].empty() && + player.controller_type == Settings::ControllerType::RightJoycon) { + player.vibrations[0].clear(); + } +} + +void ConfigureVibration::SetAllVibrationDevices() { + // Set vibration devices for all player indices including handheld + for (std::size_t player_idx = 0; player_idx < NUM_PLAYERS + 1; ++player_idx) { + SetVibrationDevices(player_idx); + } +} + +void ConfigureVibration::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QDialog::changeEvent(event); +} + +void ConfigureVibration::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_vibration.h b/src/yuzu/configuration/configure_vibration.h new file mode 100644 index 000000000..07411a86f --- /dev/null +++ b/src/yuzu/configuration/configure_vibration.h @@ -0,0 +1,43 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <memory> +#include <QDialog> + +class QGroupBox; +class QSpinBox; + +namespace Ui { +class ConfigureVibration; +} + +class ConfigureVibration : public QDialog { + Q_OBJECT + +public: + explicit ConfigureVibration(QWidget* parent); + ~ConfigureVibration() override; + + void ApplyConfiguration(); + + static void SetVibrationDevices(std::size_t player_index); + static void SetAllVibrationDevices(); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + std::unique_ptr<Ui::ConfigureVibration> ui; + + static constexpr std::size_t NUM_PLAYERS = 8; + + // Groupboxes encapsulating the vibration strength spinbox. + std::array<QGroupBox*, NUM_PLAYERS> vibration_groupboxes; + + // Spinboxes representing the vibration strength percentage. + std::array<QSpinBox*, NUM_PLAYERS> vibration_spinboxes; +}; diff --git a/src/yuzu/configuration/configure_vibration.ui b/src/yuzu/configuration/configure_vibration.ui new file mode 100644 index 000000000..efdf317a9 --- /dev/null +++ b/src/yuzu/configuration/configure_vibration.ui @@ -0,0 +1,546 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureVibration</class> + <widget class="QDialog" name="ConfigureVibration"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>364</width> + <height>242</height> + </rect> + </property> + <property name="windowTitle"> + <string>Configure Vibration</string> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout"> + <item> + <widget class="QGroupBox" name="vibrationStrengthGroup"> + <property name="title"> + <string>Vibration</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_3" stretch="0,0"> + <property name="leftMargin"> + <number>9</number> + </property> + <property name="topMargin"> + <number>9</number> + </property> + <property name="rightMargin"> + <number>9</number> + </property> + <property name="bottomMargin"> + <number>9</number> + </property> + <item> + <widget class="QWidget" name="player14Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_4"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer1"> + <property name="title"> + <string>Player 1</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_8"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer1"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer2"> + <property name="title"> + <string>Player 2</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_9"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer2"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer3"> + <property name="title"> + <string>Player 3</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_10"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer3"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer4"> + <property name="title"> + <string>Player 4</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_11"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer4"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player58Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_6"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer7"> + <property name="title"> + <string>Player 5</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_14"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer7"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer8"> + <property name="title"> + <string>Player 6</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_15"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer8"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer5"> + <property name="title"> + <string>Player 7</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_12"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer5"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer6"> + <property name="title"> + <string>Player 8</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_13"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer6"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationSettingsGroup"> + <property name="title"> + <string>Settings</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <item> + <widget class="QCheckBox" name="checkBoxAccurateVibration"> + <property name="text"> + <string>Enable Accurate Vibration</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="spacerVibration"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>167</width> + <height>55</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBoxVibration"> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections> + <connection> + <sender>buttonBoxVibration</sender> + <signal>accepted()</signal> + <receiver>ConfigureVibration</receiver> + <slot>accept()</slot> + </connection> + <connection> + <sender>buttonBoxVibration</sender> + <signal>rejected()</signal> + <receiver>ConfigureVibration</receiver> + <slot>reject()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/input_profiles.cpp b/src/yuzu/configuration/input_profiles.cpp new file mode 100644 index 000000000..e87aededb --- /dev/null +++ b/src/yuzu/configuration/input_profiles.cpp @@ -0,0 +1,131 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <fmt/format.h> + +#include "common/common_paths.h" +#include "common/file_util.h" +#include "yuzu/configuration/config.h" +#include "yuzu/configuration/input_profiles.h" + +namespace FS = Common::FS; + +namespace { + +bool ProfileExistsInFilesystem(std::string_view profile_name) { + return FS::Exists(fmt::format("{}input" DIR_SEP "{}.ini", + FS::GetUserPath(FS::UserPath::ConfigDir), profile_name)); +} + +bool IsINI(std::string_view filename) { + const std::size_t index = filename.rfind('.'); + + if (index == std::string::npos) { + return false; + } + + return filename.substr(index) == ".ini"; +} + +std::string GetNameWithoutExtension(const std::string& filename) { + const std::size_t index = filename.rfind('.'); + + if (index == std::string::npos) { + return filename; + } + + return filename.substr(0, index); +} + +} // namespace + +InputProfiles::InputProfiles() { + const std::string input_profile_loc = + fmt::format("{}input", FS::GetUserPath(FS::UserPath::ConfigDir)); + + FS::ForeachDirectoryEntry( + nullptr, input_profile_loc, + [this](u64* entries_out, const std::string& directory, const std::string& filename) { + if (IsINI(filename) && IsProfileNameValid(GetNameWithoutExtension(filename))) { + map_profiles.insert_or_assign( + GetNameWithoutExtension(filename), + std::make_unique<Config>(GetNameWithoutExtension(filename), + Config::ConfigType::InputProfile)); + } + return true; + }); +} + +InputProfiles::~InputProfiles() = default; + +std::vector<std::string> InputProfiles::GetInputProfileNames() { + std::vector<std::string> profile_names; + profile_names.reserve(map_profiles.size()); + + for (const auto& [profile_name, config] : map_profiles) { + if (!ProfileExistsInFilesystem(profile_name)) { + DeleteProfile(profile_name); + continue; + } + + profile_names.push_back(profile_name); + } + + return profile_names; +} + +bool InputProfiles::IsProfileNameValid(std::string_view profile_name) { + return profile_name.find_first_of("<>:;\"/\\|,.!?*") == std::string::npos; +} + +bool InputProfiles::CreateProfile(const std::string& profile_name, std::size_t player_index) { + if (ProfileExistsInMap(profile_name)) { + return false; + } + + map_profiles.insert_or_assign( + profile_name, std::make_unique<Config>(profile_name, Config::ConfigType::InputProfile)); + + return SaveProfile(profile_name, player_index); +} + +bool InputProfiles::DeleteProfile(const std::string& profile_name) { + if (!ProfileExistsInMap(profile_name)) { + return false; + } + + if (!ProfileExistsInFilesystem(profile_name) || + FS::Delete(map_profiles[profile_name]->GetConfigFilePath())) { + map_profiles.erase(profile_name); + } + + return !ProfileExistsInMap(profile_name) && !ProfileExistsInFilesystem(profile_name); +} + +bool InputProfiles::LoadProfile(const std::string& profile_name, std::size_t player_index) { + if (!ProfileExistsInMap(profile_name)) { + return false; + } + + if (!ProfileExistsInFilesystem(profile_name)) { + map_profiles.erase(profile_name); + return false; + } + + map_profiles[profile_name]->ReadControlPlayerValue(player_index); + return true; +} + +bool InputProfiles::SaveProfile(const std::string& profile_name, std::size_t player_index) { + if (!ProfileExistsInMap(profile_name)) { + return false; + } + + map_profiles[profile_name]->SaveControlPlayerValue(player_index); + return true; +} + +bool InputProfiles::ProfileExistsInMap(const std::string& profile_name) const { + return map_profiles.find(profile_name) != map_profiles.end(); +} diff --git a/src/yuzu/configuration/input_profiles.h b/src/yuzu/configuration/input_profiles.h new file mode 100644 index 000000000..cb41fd9be --- /dev/null +++ b/src/yuzu/configuration/input_profiles.h @@ -0,0 +1,32 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <string> +#include <string_view> +#include <unordered_map> + +class Config; + +class InputProfiles { + +public: + explicit InputProfiles(); + virtual ~InputProfiles(); + + std::vector<std::string> GetInputProfileNames(); + + static bool IsProfileNameValid(std::string_view profile_name); + + bool CreateProfile(const std::string& profile_name, std::size_t player_index); + bool DeleteProfile(const std::string& profile_name); + bool LoadProfile(const std::string& profile_name, std::size_t player_index); + bool SaveProfile(const std::string& profile_name, std::size_t player_index); + +private: + bool ProfileExistsInMap(const std::string& profile_name) const; + + std::unordered_map<std::string, std::unique_ptr<Config>> map_profiles; +}; diff --git a/src/yuzu/debugger/profiler.cpp b/src/yuzu/debugger/profiler.cpp index f594ef076..0e26f765b 100644 --- a/src/yuzu/debugger/profiler.cpp +++ b/src/yuzu/debugger/profiler.cpp @@ -51,7 +51,8 @@ MicroProfileDialog::MicroProfileDialog(QWidget* parent) : QWidget(parent, Qt::Di setWindowTitle(tr("MicroProfile")); resize(1000, 600); // Remove the "?" button from the titlebar and enable the maximize button - setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint | Qt::WindowMaximizeButtonHint); + setWindowFlags((windowFlags() & ~Qt::WindowContextHelpButtonHint) | + Qt::WindowMaximizeButtonHint); #if MICROPROFILE_ENABLED @@ -108,8 +109,7 @@ MicroProfileWidget::MicroProfileWidget(QWidget* parent) : QWidget(parent) { MicroProfileSetDisplayMode(1); // Timers screen MicroProfileInitUI(); - connect(&update_timer, &QTimer::timeout, this, - static_cast<void (MicroProfileWidget::*)()>(&MicroProfileWidget::update)); + connect(&update_timer, &QTimer::timeout, this, qOverload<>(&MicroProfileWidget::update)); } void MicroProfileWidget::paintEvent(QPaintEvent* ev) { diff --git a/src/yuzu/debugger/wait_tree.cpp b/src/yuzu/debugger/wait_tree.cpp index c1ea25fb8..3439cb333 100644 --- a/src/yuzu/debugger/wait_tree.cpp +++ b/src/yuzu/debugger/wait_tree.cpp @@ -2,10 +2,15 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> +#include <fmt/format.h> + #include "yuzu/debugger/wait_tree.h" +#include "yuzu/uisettings.h" #include "yuzu/util/util.h" #include "common/assert.h" +#include "core/arm/arm_interface.h" #include "core/core.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/mutex.h" @@ -16,11 +21,40 @@ #include "core/hle/kernel/thread.h" #include "core/memory.h" +namespace { + +constexpr std::array<std::array<Qt::GlobalColor, 2>, 10> WaitTreeColors{{ + {Qt::GlobalColor::darkGreen, Qt::GlobalColor::green}, + {Qt::GlobalColor::darkGreen, Qt::GlobalColor::green}, + {Qt::GlobalColor::darkBlue, Qt::GlobalColor::cyan}, + {Qt::GlobalColor::lightGray, Qt::GlobalColor::lightGray}, + {Qt::GlobalColor::lightGray, Qt::GlobalColor::lightGray}, + {Qt::GlobalColor::darkRed, Qt::GlobalColor::red}, + {Qt::GlobalColor::darkYellow, Qt::GlobalColor::yellow}, + {Qt::GlobalColor::red, Qt::GlobalColor::red}, + {Qt::GlobalColor::darkCyan, Qt::GlobalColor::cyan}, + {Qt::GlobalColor::gray, Qt::GlobalColor::gray}, +}}; + +bool IsDarkTheme() { + const auto& theme = UISettings::values.theme; + return theme == QStringLiteral("qdarkstyle") || + theme == QStringLiteral("qdarkstyle_midnight_blue") || + theme == QStringLiteral("colorful_dark") || + theme == QStringLiteral("colorful_midnight_blue"); +} + +} // namespace + WaitTreeItem::WaitTreeItem() = default; WaitTreeItem::~WaitTreeItem() = default; QColor WaitTreeItem::GetColor() const { - return QColor(Qt::GlobalColor::black); + if (IsDarkTheme()) { + return QColor(Qt::GlobalColor::white); + } else { + return QColor(Qt::GlobalColor::black); + } } std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeItem::GetChildren() const { @@ -59,8 +93,10 @@ std::vector<std::unique_ptr<WaitTreeThread>> WaitTreeItem::MakeThreadItemList() std::size_t row = 0; auto add_threads = [&](const std::vector<std::shared_ptr<Kernel::Thread>>& threads) { for (std::size_t i = 0; i < threads.size(); ++i) { - item_list.push_back(std::make_unique<WaitTreeThread>(*threads[i])); - item_list.back()->row = row; + if (!threads[i]->IsHLEThread()) { + item_list.push_back(std::make_unique<WaitTreeThread>(*threads[i])); + item_list.back()->row = row; + } ++row; } }; @@ -114,20 +150,21 @@ QString WaitTreeCallstack::GetText() const { std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeCallstack::GetChildren() const { std::vector<std::unique_ptr<WaitTreeItem>> list; - constexpr std::size_t BaseRegister = 29; - auto& memory = Core::System::GetInstance().Memory(); - u64 base_pointer = thread.GetContext64().cpu_registers[BaseRegister]; + if (thread.IsHLEThread()) { + return list; + } - while (base_pointer != 0) { - const u64 lr = memory.Read64(base_pointer + sizeof(u64)); - if (lr == 0) { - break; - } + if (thread.GetOwnerProcess() == nullptr || !thread.GetOwnerProcess()->Is64BitProcess()) { + return list; + } - list.push_back(std::make_unique<WaitTreeText>( - tr("0x%1").arg(lr - sizeof(u32), 16, 16, QLatin1Char{'0'}))); + auto backtrace = Core::ARM_Interface::GetBacktraceFromContext(Core::System::GetInstance(), + thread.GetContext64()); - base_pointer = memory.Read64(base_pointer); + for (auto& entry : backtrace) { + std::string s = fmt::format("{:20}{:016X} {:016X} {:016X} {}", entry.module, entry.address, + entry.original_address, entry.offset, entry.name); + list.push_back(std::make_unique<WaitTreeText>(QString::fromStdString(s))); } return list; @@ -206,7 +243,15 @@ QString WaitTreeThread::GetText() const { status = tr("running"); break; case Kernel::ThreadStatus::Ready: - status = tr("ready"); + if (!thread.IsPaused()) { + if (thread.WasRunning()) { + status = tr("running"); + } else { + status = tr("ready"); + } + } else { + status = tr("paused"); + } break; case Kernel::ThreadStatus::Paused: status = tr("paused"); @@ -249,28 +294,38 @@ QString WaitTreeThread::GetText() const { } QColor WaitTreeThread::GetColor() const { + const std::size_t color_index = IsDarkTheme() ? 1 : 0; + const auto& thread = static_cast<const Kernel::Thread&>(object); switch (thread.GetStatus()) { case Kernel::ThreadStatus::Running: - return QColor(Qt::GlobalColor::darkGreen); + return QColor(WaitTreeColors[0][color_index]); case Kernel::ThreadStatus::Ready: - return QColor(Qt::GlobalColor::darkBlue); + if (!thread.IsPaused()) { + if (thread.WasRunning()) { + return QColor(WaitTreeColors[1][color_index]); + } else { + return QColor(WaitTreeColors[2][color_index]); + } + } else { + return QColor(WaitTreeColors[3][color_index]); + } case Kernel::ThreadStatus::Paused: - return QColor(Qt::GlobalColor::lightGray); + return QColor(WaitTreeColors[4][color_index]); case Kernel::ThreadStatus::WaitHLEEvent: case Kernel::ThreadStatus::WaitIPC: - return QColor(Qt::GlobalColor::darkRed); + return QColor(WaitTreeColors[5][color_index]); case Kernel::ThreadStatus::WaitSleep: - return QColor(Qt::GlobalColor::darkYellow); + return QColor(WaitTreeColors[6][color_index]); case Kernel::ThreadStatus::WaitSynch: case Kernel::ThreadStatus::WaitMutex: case Kernel::ThreadStatus::WaitCondVar: case Kernel::ThreadStatus::WaitArb: - return QColor(Qt::GlobalColor::red); + return QColor(WaitTreeColors[7][color_index]); case Kernel::ThreadStatus::Dormant: - return QColor(Qt::GlobalColor::darkCyan); + return QColor(WaitTreeColors[8][color_index]); case Kernel::ThreadStatus::Dead: - return QColor(Qt::GlobalColor::gray); + return QColor(WaitTreeColors[9][color_index]); default: return WaitTreeItem::GetColor(); } @@ -319,7 +374,7 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const { if (thread.GetStatus() == Kernel::ThreadStatus::WaitSynch) { list.push_back(std::make_unique<WaitTreeObjectList>(thread.GetSynchronizationObjects(), - thread.IsSleepingOnWait())); + thread.IsWaitingSync())); } list.push_back(std::make_unique<WaitTreeCallstack>(thread)); diff --git a/src/yuzu/discord_impl.cpp b/src/yuzu/discord_impl.cpp index ea0079353..a93733b26 100644 --- a/src/yuzu/discord_impl.cpp +++ b/src/yuzu/discord_impl.cpp @@ -18,7 +18,7 @@ DiscordImpl::DiscordImpl() { // The number is the client ID for yuzu, it's used for images and the // application name - Discord_Initialize("471872241299226636", &handlers, 1, nullptr); + Discord_Initialize("712465656758665259", &handlers, 1, nullptr); } DiscordImpl::~DiscordImpl() { diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp index dccbabcbf..70d865112 100644 --- a/src/yuzu/game_list.cpp +++ b/src/yuzu/game_list.cpp @@ -25,7 +25,8 @@ #include "yuzu/main.h" #include "yuzu/uisettings.h" -GameListSearchField::KeyReleaseEater::KeyReleaseEater(GameList* gamelist) : gamelist{gamelist} {} +GameListSearchField::KeyReleaseEater::KeyReleaseEater(GameList* gamelist, QObject* parent) + : QObject(parent), gamelist{gamelist} {} // EventFilter in order to process systemkeys while editing the searchfield bool GameListSearchField::KeyReleaseEater::eventFilter(QObject* obj, QEvent* event) { @@ -56,7 +57,7 @@ bool GameListSearchField::KeyReleaseEater::eventFilter(QObject* obj, QEvent* eve case Qt::Key_Return: case Qt::Key_Enter: { if (gamelist->search_field->visible == 1) { - QString file_path = gamelist->getLastFilterResultItem(); + const QString file_path = gamelist->GetLastFilterResultItem(); // To avoid loading error dialog loops while confirming them using enter // Also users usually want to run a different game after closing one @@ -83,22 +84,25 @@ void GameListSearchField::setFilterResult(int visible, int total) { label_filter_result->setText(tr("%1 of %n result(s)", "", total).arg(visible)); } -QString GameList::getLastFilterResultItem() const { - QStandardItem* folder; - QStandardItem* child; +QString GameList::GetLastFilterResultItem() const { QString file_path; const int folder_count = item_model->rowCount(); + for (int i = 0; i < folder_count; ++i) { - folder = item_model->item(i, 0); + const QStandardItem* folder = item_model->item(i, 0); const QModelIndex folder_index = folder->index(); const int children_count = folder->rowCount(); + for (int j = 0; j < children_count; ++j) { - if (!tree_view->isRowHidden(j, folder_index)) { - child = folder->child(j, 0); - file_path = child->data(GameListItemPath::FullPathRole).toString(); + if (tree_view->isRowHidden(j, folder_index)) { + continue; } + + const QStandardItem* child = folder->child(j, 0); + file_path = child->data(GameListItemPath::FullPathRole).toString(); } } + return file_path; } @@ -113,7 +117,7 @@ void GameListSearchField::setFocus() { } GameListSearchField::GameListSearchField(GameList* parent) : QWidget{parent} { - auto* const key_release_eater = new KeyReleaseEater(parent); + auto* const key_release_eater = new KeyReleaseEater(parent, this); layout_filter = new QHBoxLayout; layout_filter->setMargin(8); label_filter = new QLabel; @@ -123,7 +127,7 @@ GameListSearchField::GameListSearchField(GameList* parent) : QWidget{parent} { edit_filter->setPlaceholderText(tr("Enter pattern to filter")); edit_filter->installEventFilter(key_release_eater); edit_filter->setClearButtonEnabled(true); - connect(edit_filter, &QLineEdit::textChanged, parent, &GameList::onTextChanged); + connect(edit_filter, &QLineEdit::textChanged, parent, &GameList::OnTextChanged); label_filter_result = new QLabel; button_filter_close = new QToolButton(this); button_filter_close->setText(QStringLiteral("X")); @@ -133,7 +137,7 @@ GameListSearchField::GameListSearchField(GameList* parent) : QWidget{parent} { "#000000; font-weight: bold; background: #F0F0F0; }" "QToolButton:hover{ border: none; padding: 0px; color: " "#EEEEEE; font-weight: bold; background: #E81123}")); - connect(button_filter_close, &QToolButton::clicked, parent, &GameList::onFilterCloseClicked); + connect(button_filter_close, &QToolButton::clicked, parent, &GameList::OnFilterCloseClicked); layout_filter->setSpacing(10); layout_filter->addWidget(label_filter); layout_filter->addWidget(edit_filter); @@ -159,16 +163,22 @@ static bool ContainsAllWords(const QString& haystack, const QString& userinput) } // Syncs the expanded state of Game Directories with settings to persist across sessions -void GameList::onItemExpanded(const QModelIndex& item) { +void GameList::OnItemExpanded(const QModelIndex& item) { const auto type = item.data(GameListItem::TypeRole).value<GameListItemType>(); - if (type == GameListItemType::CustomDir || type == GameListItemType::SdmcDir || - type == GameListItemType::UserNandDir || type == GameListItemType::SysNandDir) - item.data(GameListDir::GameDirRole).value<UISettings::GameDir*>()->expanded = - tree_view->isExpanded(item); + const bool is_dir = type == GameListItemType::CustomDir || type == GameListItemType::SdmcDir || + type == GameListItemType::UserNandDir || + type == GameListItemType::SysNandDir; + + if (!is_dir) { + return; + } + + auto* game_dir = item.data(GameListDir::GameDirRole).value<UISettings::GameDir*>(); + game_dir->expanded = tree_view->isExpanded(item); } // Event in order to filter the gamelist after editing the searchfield -void GameList::onTextChanged(const QString& new_text) { +void GameList::OnTextChanged(const QString& new_text) { const int folder_count = tree_view->model()->rowCount(); QString edit_filter_text = new_text.toLower(); QStandardItem* folder; @@ -224,7 +234,7 @@ void GameList::onTextChanged(const QString& new_text) { } } -void GameList::onUpdateThemedIcons() { +void GameList::OnUpdateThemedIcons() { for (int i = 0; i < item_model->invisibleRootItem()->rowCount(); i++) { QStandardItem* child = item_model->invisibleRootItem()->child(i); @@ -276,7 +286,7 @@ void GameList::onUpdateThemedIcons() { } } -void GameList::onFilterCloseClicked() { +void GameList::OnFilterCloseClicked() { main_window->filterBarSetChecked(false); } @@ -317,11 +327,11 @@ GameList::GameList(FileSys::VirtualFilesystem vfs, FileSys::ManualContentProvide } item_model->setSortRole(GameListItemPath::SortRole); - connect(main_window, &GMainWindow::UpdateThemedIcons, this, &GameList::onUpdateThemedIcons); + connect(main_window, &GMainWindow::UpdateThemedIcons, this, &GameList::OnUpdateThemedIcons); connect(tree_view, &QTreeView::activated, this, &GameList::ValidateEntry); connect(tree_view, &QTreeView::customContextMenuRequested, this, &GameList::PopupContextMenu); - connect(tree_view, &QTreeView::expanded, this, &GameList::onItemExpanded); - connect(tree_view, &QTreeView::collapsed, this, &GameList::onItemExpanded); + connect(tree_view, &QTreeView::expanded, this, &GameList::OnItemExpanded); + connect(tree_view, &QTreeView::collapsed, this, &GameList::OnItemExpanded); // We must register all custom types with the Qt Automoc system so that we are able to use // it with signals/slots. In this case, QList falls under the umbrells of custom types. @@ -338,17 +348,17 @@ GameList::~GameList() { emit ShouldCancelWorker(); } -void GameList::setFilterFocus() { +void GameList::SetFilterFocus() { if (tree_view->model()->rowCount() > 0) { search_field->setFocus(); } } -void GameList::setFilterVisible(bool visibility) { +void GameList::SetFilterVisible(bool visibility) { search_field->setVisible(visibility); } -void GameList::clearFilter() { +void GameList::ClearFilter() { search_field->clear(); } @@ -397,22 +407,24 @@ void GameList::ValidateEntry(const QModelIndex& item) { } } -bool GameList::isEmpty() const { +bool GameList::IsEmpty() const { for (int i = 0; i < item_model->rowCount(); i++) { const QStandardItem* child = item_model->invisibleRootItem()->child(i); const auto type = static_cast<GameListItemType>(child->type()); + if (!child->hasChildren() && (type == GameListItemType::SdmcDir || type == GameListItemType::UserNandDir || type == GameListItemType::SysNandDir)) { item_model->invisibleRootItem()->removeRow(child->row()); i--; - }; + } } + return !item_model->invisibleRootItem()->hasChildren(); } -void GameList::DonePopulating(QStringList watch_list) { - emit ShowList(!isEmpty()); +void GameList::DonePopulating(const QStringList& watch_list) { + emit ShowList(!IsEmpty()); item_model->invisibleRootItem()->appendRow(new GameListAddDir()); @@ -472,30 +484,58 @@ void GameList::PopupContextMenu(const QPoint& menu_location) { context_menu.exec(tree_view->viewport()->mapToGlobal(menu_location)); } -void GameList::AddGamePopup(QMenu& context_menu, u64 program_id, std::string path) { +void GameList::AddGamePopup(QMenu& context_menu, u64 program_id, const std::string& path) { QAction* open_save_location = context_menu.addAction(tr("Open Save Data Location")); - QAction* open_lfs_location = context_menu.addAction(tr("Open Mod Data Location")); + QAction* open_mod_location = context_menu.addAction(tr("Open Mod Data Location")); QAction* open_transferable_shader_cache = context_menu.addAction(tr("Open Transferable Shader Cache")); context_menu.addSeparator(); + QMenu* remove_menu = context_menu.addMenu(tr("Remove")); + QAction* remove_update = remove_menu->addAction(tr("Remove Installed Update")); + QAction* remove_dlc = remove_menu->addAction(tr("Remove All Installed DLC")); + QAction* remove_shader_cache = remove_menu->addAction(tr("Remove Shader Cache")); + QAction* remove_custom_config = remove_menu->addAction(tr("Remove Custom Configuration")); + remove_menu->addSeparator(); + QAction* remove_all_content = remove_menu->addAction(tr("Remove All Installed Contents")); QAction* dump_romfs = context_menu.addAction(tr("Dump RomFS")); QAction* copy_tid = context_menu.addAction(tr("Copy Title ID to Clipboard")); QAction* navigate_to_gamedb_entry = context_menu.addAction(tr("Navigate to GameDB entry")); context_menu.addSeparator(); QAction* properties = context_menu.addAction(tr("Properties")); - open_save_location->setEnabled(program_id != 0); + open_save_location->setVisible(program_id != 0); + open_mod_location->setVisible(program_id != 0); + open_transferable_shader_cache->setVisible(program_id != 0); + remove_update->setVisible(program_id != 0); + remove_dlc->setVisible(program_id != 0); + remove_shader_cache->setVisible(program_id != 0); + remove_all_content->setVisible(program_id != 0); auto it = FindMatchingCompatibilityEntry(compatibility_list, program_id); navigate_to_gamedb_entry->setVisible(it != compatibility_list.end() && program_id != 0); - connect(open_save_location, &QAction::triggered, [this, program_id]() { - emit OpenFolderRequested(program_id, GameListOpenTarget::SaveData); + connect(open_save_location, &QAction::triggered, [this, program_id, path]() { + emit OpenFolderRequested(program_id, GameListOpenTarget::SaveData, path); }); - connect(open_lfs_location, &QAction::triggered, [this, program_id]() { - emit OpenFolderRequested(program_id, GameListOpenTarget::ModData); + connect(open_mod_location, &QAction::triggered, [this, program_id, path]() { + emit OpenFolderRequested(program_id, GameListOpenTarget::ModData, path); }); connect(open_transferable_shader_cache, &QAction::triggered, [this, program_id]() { emit OpenTransferableShaderCacheRequested(program_id); }); + connect(remove_all_content, &QAction::triggered, [this, program_id]() { + emit RemoveInstalledEntryRequested(program_id, InstalledEntryType::Game); + }); + connect(remove_update, &QAction::triggered, [this, program_id]() { + emit RemoveInstalledEntryRequested(program_id, InstalledEntryType::Update); + }); + connect(remove_dlc, &QAction::triggered, [this, program_id]() { + emit RemoveInstalledEntryRequested(program_id, InstalledEntryType::AddOnContent); + }); + connect(remove_shader_cache, &QAction::triggered, [this, program_id]() { + emit RemoveFileRequested(program_id, GameListRemoveTarget::ShaderCache); + }); + connect(remove_custom_config, &QAction::triggered, [this, program_id]() { + emit RemoveFileRequested(program_id, GameListRemoveTarget::CustomConfiguration); + }); connect(dump_romfs, &QAction::triggered, [this, program_id, path]() { emit DumpRomFSRequested(program_id, path); }); connect(copy_tid, &QAction::triggered, @@ -531,8 +571,8 @@ void GameList::AddPermDirPopup(QMenu& context_menu, QModelIndex selected) { UISettings::GameDir& game_dir = *selected.data(GameListDir::GameDirRole).value<UISettings::GameDir*>(); - QAction* move_up = context_menu.addAction(tr(u8"\U000025b2 Move Up")); - QAction* move_down = context_menu.addAction(tr(u8"\U000025bc Move Down ")); + QAction* move_up = context_menu.addAction(tr("\u25B2 Move Up")); + QAction* move_down = context_menu.addAction(tr("\u25bc Move Down")); QAction* open_directory_location = context_menu.addAction(tr("Open Directory Location")); const int row = selected.row(); @@ -662,12 +702,15 @@ void GameList::SaveInterfaceLayout() { } void GameList::LoadInterfaceLayout() { - auto header = tree_view->header(); - if (!header->restoreState(UISettings::values.gamelist_header_state)) { - // We are using the name column to display icons and titles - // so make it as large as possible as default. - header->resizeSection(COLUMN_NAME, header->width()); + auto* header = tree_view->header(); + + if (header->restoreState(UISettings::values.gamelist_header_state)) { + return; } + + // We are using the name column to display icons and titles + // so make it as large as possible as default. + header->resizeSection(COLUMN_NAME, header->width()); } const QStringList GameList::supported_file_extensions = { diff --git a/src/yuzu/game_list.h b/src/yuzu/game_list.h index 878d94413..58059a3c4 100644 --- a/src/yuzu/game_list.h +++ b/src/yuzu/game_list.h @@ -39,6 +39,17 @@ enum class GameListOpenTarget { ModData, }; +enum class GameListRemoveTarget { + ShaderCache, + CustomConfiguration, +}; + +enum class InstalledEntryType { + Game, + Update, + AddOnContent, +}; + class GameList : public QWidget { Q_OBJECT @@ -56,11 +67,11 @@ public: FileSys::ManualContentProvider* provider, GMainWindow* parent = nullptr); ~GameList() override; - QString getLastFilterResultItem() const; - void clearFilter(); - void setFilterFocus(); - void setFilterVisible(bool visibility); - bool isEmpty() const; + QString GetLastFilterResultItem() const; + void ClearFilter(); + void SetFilterFocus(); + void SetFilterVisible(bool visibility); + bool IsEmpty() const; void LoadCompatibilityList(); void PopulateAsync(QVector<UISettings::GameDir>& game_dirs); @@ -71,10 +82,13 @@ public: static const QStringList supported_file_extensions; signals: - void GameChosen(QString game_path); + void GameChosen(const QString& game_path); void ShouldCancelWorker(); - void OpenFolderRequested(u64 program_id, GameListOpenTarget target); + void OpenFolderRequested(u64 program_id, GameListOpenTarget target, + const std::string& game_path); void OpenTransferableShaderCacheRequested(u64 program_id); + void RemoveInstalledEntryRequested(u64 program_id, InstalledEntryType type); + void RemoveFileRequested(u64 program_id, GameListRemoveTarget target); void DumpRomFSRequested(u64 program_id, const std::string& game_path); void CopyTIDRequested(u64 program_id); void NavigateToGamedbEntryRequested(u64 program_id, @@ -85,21 +99,21 @@ signals: void ShowList(bool show); private slots: - void onItemExpanded(const QModelIndex& item); - void onTextChanged(const QString& new_text); - void onFilterCloseClicked(); - void onUpdateThemedIcons(); + void OnItemExpanded(const QModelIndex& item); + void OnTextChanged(const QString& new_text); + void OnFilterCloseClicked(); + void OnUpdateThemedIcons(); private: void AddDirEntry(GameListDir* entry_items); void AddEntry(const QList<QStandardItem*>& entry_items, GameListDir* parent); void ValidateEntry(const QModelIndex& item); - void DonePopulating(QStringList watch_list); + void DonePopulating(const QStringList& watch_list); void RefreshGameDirectory(); void PopupContextMenu(const QPoint& menu_location); - void AddGamePopup(QMenu& context_menu, u64 program_id, std::string path); + void AddGamePopup(QMenu& context_menu, u64 program_id, const std::string& path); void AddCustomDirPopup(QMenu& context_menu, QModelIndex selected); void AddPermDirPopup(QMenu& context_menu, QModelIndex selected); @@ -117,8 +131,6 @@ private: friend class GameListSearchField; }; -Q_DECLARE_METATYPE(GameListOpenTarget); - class GameListPlaceholder : public QWidget { Q_OBJECT public: diff --git a/src/yuzu/game_list_p.h b/src/yuzu/game_list_p.h index 3e6d5a7cd..248855aff 100644 --- a/src/yuzu/game_list_p.h +++ b/src/yuzu/game_list_p.h @@ -49,10 +49,10 @@ class GameListItem : public QStandardItem { public: // used to access type from item index - static const int TypeRole = Qt::UserRole + 1; - static const int SortRole = Qt::UserRole + 2; + static constexpr int TypeRole = Qt::UserRole + 1; + static constexpr int SortRole = Qt::UserRole + 2; GameListItem() = default; - GameListItem(const QString& string) : QStandardItem(string) { + explicit GameListItem(const QString& string) : QStandardItem(string) { setData(string, SortRole); } }; @@ -65,10 +65,10 @@ public: */ class GameListItemPath : public GameListItem { public: - static const int TitleRole = SortRole + 1; - static const int FullPathRole = SortRole + 2; - static const int ProgramIdRole = SortRole + 3; - static const int FileTypeRole = SortRole + 4; + static constexpr int TitleRole = SortRole + 1; + static constexpr int FullPathRole = SortRole + 2; + static constexpr int ProgramIdRole = SortRole + 3; + static constexpr int FileTypeRole = SortRole + 4; GameListItemPath() = default; GameListItemPath(const QString& game_path, const std::vector<u8>& picture_data, @@ -110,35 +110,32 @@ public: const auto& row1 = row_data.at(UISettings::values.row_1_text_id); const int row2_id = UISettings::values.row_2_text_id; - if (role == SortRole) + if (role == SortRole) { return row1.toLower(); + } - if (row2_id == 4) // None + // None + if (row2_id == 4) { return row1; + } const auto& row2 = row_data.at(row2_id); - if (row1 == row2) + if (row1 == row2) { return row1; + } - return QString(row1 + QStringLiteral("\n ") + row2); + return QStringLiteral("%1\n %2").arg(row1, row2); } return GameListItem::data(role); } - - /** - * Override to prevent automatic sorting. - */ - bool operator<(const QStandardItem& other) const override { - return false; - } }; class GameListItemCompat : public GameListItem { Q_DECLARE_TR_FUNCTIONS(GameListItemCompat) public: - static const int CompatNumberRole = SortRole; + static constexpr int CompatNumberRole = SortRole; GameListItemCompat() = default; explicit GameListItemCompat(const QString& compatibility) { setData(type(), TypeRole); @@ -188,7 +185,7 @@ public: */ class GameListItemSize : public GameListItem { public: - static const int SizeRole = SortRole; + static constexpr int SizeRole = SortRole; GameListItemSize() = default; explicit GameListItemSize(const qulonglong size_bytes) { @@ -224,7 +221,7 @@ public: class GameListDir : public GameListItem { public: - static const int GameDirRole = Qt::UserRole + 2; + static constexpr int GameDirRole = Qt::UserRole + 2; explicit GameListDir(UISettings::GameDir& directory, GameListItemType dir_type = GameListItemType::CustomDir) @@ -279,6 +276,13 @@ public: return static_cast<int>(dir_type); } + /** + * Override to prevent automatic sorting between folders and the addDir button. + */ + bool operator<(const QStandardItem& other) const override { + return false; + } + private: GameListItemType dir_type; }; @@ -326,7 +330,7 @@ public: private: class KeyReleaseEater : public QObject { public: - explicit KeyReleaseEater(GameList* gamelist); + explicit KeyReleaseEater(GameList* gamelist, QObject* parent = nullptr); private: GameList* gamelist = nullptr; diff --git a/src/yuzu/game_list_worker.cpp b/src/yuzu/game_list_worker.cpp index da2c27aa2..23643aea2 100644 --- a/src/yuzu/game_list_worker.cpp +++ b/src/yuzu/game_list_worker.cpp @@ -39,12 +39,12 @@ QString GetGameListCachedObject(const std::string& filename, const std::string& return generator(); } - const auto path = FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + DIR_SEP + "game_list" + - DIR_SEP + filename + '.' + ext; + const auto path = Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + DIR_SEP + + "game_list" + DIR_SEP + filename + '.' + ext; - FileUtil::CreateFullPath(path); + Common::FS::CreateFullPath(path); - if (!FileUtil::Exists(path)) { + if (!Common::FS::Exists(path)) { const auto str = generator(); QFile file{QString::fromStdString(path)}; @@ -70,14 +70,14 @@ std::pair<std::vector<u8>, std::string> GetGameListCachedObject( return generator(); } - const auto path1 = FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + DIR_SEP + "game_list" + - DIR_SEP + filename + ".jpeg"; - const auto path2 = FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + DIR_SEP + "game_list" + - DIR_SEP + filename + ".appname.txt"; + const auto path1 = Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + DIR_SEP + + "game_list" + DIR_SEP + filename + ".jpeg"; + const auto path2 = Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + DIR_SEP + + "game_list" + DIR_SEP + filename + ".appname.txt"; - FileUtil::CreateFullPath(path1); + Common::FS::CreateFullPath(path1); - if (!FileUtil::Exists(path1) || !FileUtil::Exists(path2)) { + if (!Common::FS::Exists(path1) || !Common::FS::Exists(path2)) { const auto [icon, nacp] = generator(); QFile file1{QString::fromStdString(path1)}; @@ -91,7 +91,8 @@ std::pair<std::vector<u8>, std::string> GetGameListCachedObject( return generator(); } - if (file1.write(reinterpret_cast<const char*>(icon.data()), icon.size()) != icon.size()) { + if (file1.write(reinterpret_cast<const char*>(icon.data()), icon.size()) != + s64(icon.size())) { LOG_ERROR(Frontend, "Failed to write data to cache file."); return generator(); } @@ -207,7 +208,7 @@ QList<QStandardItem*> MakeGameListEntry(const std::string& path, const std::stri file_type_string, program_id), new GameListItemCompat(compatibility), new GameListItem(file_type_string), - new GameListItemSize(FileUtil::GetSize(path)), + new GameListItemSize(Common::FS::GetSize(path)), }; if (UISettings::values.show_add_ons) { @@ -234,12 +235,11 @@ GameListWorker::~GameListWorker() = default; void GameListWorker::AddTitlesToGameList(GameListDir* parent_dir) { using namespace FileSys; - const auto& cache = - dynamic_cast<ContentProviderUnion&>(Core::System::GetInstance().GetContentProvider()); + auto& system = Core::System::GetInstance(); + const auto& cache = dynamic_cast<ContentProviderUnion&>(system.GetContentProvider()); - std::vector<std::pair<ContentProviderUnionSlot, ContentProviderEntry>> installed_games; - installed_games = cache.ListEntriesFilterOrigin(std::nullopt, TitleType::Application, - ContentRecordType::Program); + auto installed_games = cache.ListEntriesFilterOrigin(std::nullopt, TitleType::Application, + ContentRecordType::Program); if (parent_dir->type() == static_cast<int>(GameListItemType::SdmcDir)) { installed_games = cache.ListEntriesFilterOrigin( @@ -253,23 +253,27 @@ void GameListWorker::AddTitlesToGameList(GameListDir* parent_dir) { } for (const auto& [slot, game] : installed_games) { - if (slot == ContentProviderUnionSlot::FrontendManual) + if (slot == ContentProviderUnionSlot::FrontendManual) { continue; + } const auto file = cache.GetEntryUnparsed(game.title_id, game.type); - std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(file); - if (!loader) + std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(system, file); + if (!loader) { continue; + } std::vector<u8> icon; std::string name; u64 program_id = 0; loader->ReadProgramId(program_id); - const PatchManager patch{program_id}; + const PatchManager patch{program_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto control = cache.GetEntry(game.title_id, ContentRecordType::Control); - if (control != nullptr) + if (control != nullptr) { GetMetadataFromControlNCA(patch, *control, icon, name); + } emit EntryReady(MakeGameListEntry(file->GetFullPath(), name, icon, *loader, program_id, compatibility_list, patch), @@ -279,20 +283,22 @@ void GameListWorker::AddTitlesToGameList(GameListDir* parent_dir) { void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_path, unsigned int recursion, GameListDir* parent_dir) { - const auto callback = [this, target, recursion, - parent_dir](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { + auto& system = Core::System::GetInstance(); + + const auto callback = [this, target, recursion, parent_dir, + &system](u64* num_entries_out, const std::string& directory, + const std::string& virtual_name) -> bool { if (stop_processing) { // Breaks the callback loop. return false; } const std::string physical_name = directory + DIR_SEP + virtual_name; - const bool is_dir = FileUtil::IsDirectory(physical_name); + const bool is_dir = Common::FS::IsDirectory(physical_name); if (!is_dir && (HasSupportedFileExtension(physical_name) || IsExtractedNCAMain(physical_name))) { const auto file = vfs->OpenFile(physical_name, FileSys::Mode::Read); - auto loader = Loader::GetLoader(file); + auto loader = Loader::GetLoader(system, file); if (!loader) { return true; } @@ -330,7 +336,8 @@ void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_pa std::string name = " "; [[maybe_unused]] const auto res3 = loader->ReadTitle(name); - const FileSys::PatchManager patch{program_id}; + const FileSys::PatchManager patch{program_id, system.GetFileSystemController(), + system.GetContentProvider()}; emit EntryReady(MakeGameListEntry(physical_name, name, icon, *loader, program_id, compatibility_list, patch), @@ -344,11 +351,12 @@ void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_pa return true; }; - FileUtil::ForeachDirectoryEntry(nullptr, dir_path, callback); + Common::FS::ForeachDirectoryEntry(nullptr, dir_path, callback); } void GameListWorker::run() { stop_processing = false; + provider->ClearAllEntries(); for (UISettings::GameDir& game_dir : game_dirs) { if (game_dir.path == QStringLiteral("SDMC")) { @@ -367,13 +375,12 @@ void GameListWorker::run() { watch_list.append(game_dir.path); auto* const game_list_dir = new GameListDir(game_dir); emit DirEntryReady(game_list_dir); - provider->ClearAllEntries(); - ScanFileSystem(ScanTarget::FillManualContentProvider, game_dir.path.toStdString(), 2, - game_list_dir); + ScanFileSystem(ScanTarget::FillManualContentProvider, game_dir.path.toStdString(), + game_dir.deep_scan ? 256 : 0, game_list_dir); ScanFileSystem(ScanTarget::PopulateGameList, game_dir.path.toStdString(), game_dir.deep_scan ? 256 : 0, game_list_dir); } - }; + } emit Finished(watch_list); } diff --git a/src/yuzu/install_dialog.cpp b/src/yuzu/install_dialog.cpp new file mode 100644 index 000000000..06b0b1874 --- /dev/null +++ b/src/yuzu/install_dialog.cpp @@ -0,0 +1,72 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QCheckBox> +#include <QDialogButtonBox> +#include <QFileInfo> +#include <QHBoxLayout> +#include <QLabel> +#include <QListWidget> +#include <QVBoxLayout> +#include "yuzu/install_dialog.h" +#include "yuzu/uisettings.h" + +InstallDialog::InstallDialog(QWidget* parent, const QStringList& files) : QDialog(parent) { + file_list = new QListWidget(this); + + for (const QString& file : files) { + QListWidgetItem* item = new QListWidgetItem(QFileInfo(file).fileName(), file_list); + item->setData(Qt::UserRole, file); + item->setFlags(item->flags() | Qt::ItemIsUserCheckable); + item->setCheckState(Qt::Checked); + } + + file_list->setMinimumWidth((file_list->sizeHintForColumn(0) * 11) / 10); + + vbox_layout = new QVBoxLayout; + + hbox_layout = new QHBoxLayout; + + description = new QLabel(tr("Please confirm these are the files you wish to install.")); + + update_description = + new QLabel(tr("Installing an Update or DLC will overwrite the previously installed one.")); + + buttons = new QDialogButtonBox; + buttons->addButton(QDialogButtonBox::Cancel); + buttons->addButton(tr("Install"), QDialogButtonBox::AcceptRole); + + connect(buttons, &QDialogButtonBox::accepted, this, &InstallDialog::accept); + connect(buttons, &QDialogButtonBox::rejected, this, &InstallDialog::reject); + + hbox_layout->addWidget(buttons); + + vbox_layout->addWidget(description); + vbox_layout->addWidget(update_description); + vbox_layout->addWidget(file_list); + vbox_layout->addLayout(hbox_layout); + + setLayout(vbox_layout); + setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint); + setWindowTitle(tr("Install Files to NAND")); +} + +InstallDialog::~InstallDialog() = default; + +QStringList InstallDialog::GetFiles() const { + QStringList files; + + for (int i = 0; i < file_list->count(); ++i) { + const QListWidgetItem* item = file_list->item(i); + if (item->checkState() == Qt::Checked) { + files.append(item->data(Qt::UserRole).toString()); + } + } + + return files; +} + +int InstallDialog::GetMinimumWidth() const { + return file_list->width(); +} diff --git a/src/yuzu/install_dialog.h b/src/yuzu/install_dialog.h new file mode 100644 index 000000000..68e03fe4e --- /dev/null +++ b/src/yuzu/install_dialog.h @@ -0,0 +1,35 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <QDialog> + +class QCheckBox; +class QDialogButtonBox; +class QHBoxLayout; +class QLabel; +class QListWidget; +class QVBoxLayout; + +class InstallDialog : public QDialog { + Q_OBJECT + +public: + explicit InstallDialog(QWidget* parent, const QStringList& files); + ~InstallDialog() override; + + [[nodiscard]] QStringList GetFiles() const; + [[nodiscard]] int GetMinimumWidth() const; + +private: + QListWidget* file_list; + + QVBoxLayout* vbox_layout; + QHBoxLayout* hbox_layout; + + QLabel* description; + QLabel* update_description; + QDialogButtonBox* buttons; +}; diff --git a/src/yuzu/loading_screen.cpp b/src/yuzu/loading_screen.cpp index 2a6483370..ae842306c 100644 --- a/src/yuzu/loading_screen.cpp +++ b/src/yuzu/loading_screen.cpp @@ -19,6 +19,7 @@ #include <QTime> #include <QtConcurrent/QtConcurrentRun> #include "common/logging/log.h" +#include "core/frontend/framebuffer_layout.h" #include "core/loader/loader.h" #include "ui_loading_screen.h" #include "video_core/rasterizer_interface.h" @@ -61,7 +62,7 @@ LoadingScreen::LoadingScreen(QWidget* parent) : QWidget(parent), ui(std::make_unique<Ui::LoadingScreen>()), previous_stage(VideoCore::LoadCallbackStage::Complete) { ui->setupUi(this); - setMinimumSize(1280, 720); + setMinimumSize(Layout::MinimumSize::Width, Layout::MinimumSize::Height); // Create a fade out effect to hide this loading screen widget. // When fading opacity, it will fade to the parent widgets background color, which is why we diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index 1717e06f9..e704cc656 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -11,15 +11,19 @@ #endif // VFS includes must be before glad as they will conflict with Windows file api, which uses defines. +#include "applets/controller.h" #include "applets/error.h" #include "applets/profile_select.h" #include "applets/software_keyboard.h" #include "applets/web_browser.h" #include "configuration/configure_input.h" -#include "configuration/configure_per_general.h" +#include "configuration/configure_per_game.h" +#include "configuration/configure_vibration.h" #include "core/file_sys/vfs.h" #include "core/file_sys/vfs_real.h" +#include "core/frontend/applets/controller.h" #include "core/frontend/applets/general_frontend.h" +#include "core/frontend/applets/software_keyboard.h" #include "core/hle/service/acc/profile_manager.h" #include "core/hle/service/am/applet_ae.h" #include "core/hle/service/am/applet_oe.h" @@ -47,15 +51,18 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include <QDesktopServices> #include <QDesktopWidget> #include <QDialogButtonBox> +#include <QDir> #include <QFile> #include <QFileDialog> #include <QInputDialog> #include <QMessageBox> #include <QProgressBar> #include <QProgressDialog> +#include <QPushButton> #include <QShortcut> #include <QStatusBar> #include <QSysInfo> +#include <QUrl> #include <QtConcurrent/QtConcurrent> #include <fmt/format.h> @@ -65,6 +72,7 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "common/logging/backend.h" #include "common/logging/filter.h" #include "common/logging/log.h" +#include "common/memory_detect.h" #include "common/microprofile.h" #include "common/scm_rev.h" #include "common/scope_exit.h" @@ -82,7 +90,6 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "core/file_sys/romfs.h" #include "core/file_sys/savedata_factory.h" #include "core/file_sys/submission_package.h" -#include "core/frontend/applets/software_keyboard.h" #include "core/hle/kernel/process.h" #include "core/hle/service/am/am.h" #include "core/hle/service/filesystem/filesystem.h" @@ -92,6 +99,9 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "core/perf_stats.h" #include "core/settings.h" #include "core/telemetry_session.h" +#include "input_common/main.h" +#include "video_core/gpu.h" +#include "video_core/shader_notify.h" #include "yuzu/about_dialog.h" #include "yuzu/bootmanager.h" #include "yuzu/compatdb.h" @@ -105,6 +115,7 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "yuzu/game_list.h" #include "yuzu/game_list_p.h" #include "yuzu/hotkeys.h" +#include "yuzu/install_dialog.h" #include "yuzu/loading_screen.h" #include "yuzu/main.h" #include "yuzu/uisettings.h" @@ -135,6 +146,8 @@ __declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1; } #endif +constexpr int default_mouse_timeout = 2500; + constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; /** @@ -170,8 +183,8 @@ static void InitializeLogging() { log_filter.ParseFilterString(Settings::values.log_filter); Log::SetGlobalFilter(log_filter); - const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir); - FileUtil::CreateFullPath(log_dir); + const std::string& log_dir = Common::FS::GetUserPath(Common::FS::UserPath::LogDir); + Common::FS::CreateFullPath(log_dir); Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE)); #ifdef _WIN32 Log::AddBackend(std::make_unique<Log::DebuggerBackend>()); @@ -179,11 +192,13 @@ static void InitializeLogging() { } GMainWindow::GMainWindow() - : config(new Config()), emu_thread(nullptr), - vfs(std::make_shared<FileSys::RealVfsFilesystem>()), - provider(std::make_unique<FileSys::ManualContentProvider>()) { + : input_subsystem{std::make_shared<InputCommon::InputSubsystem>()}, + config{std::make_unique<Config>()}, vfs{std::make_shared<FileSys::RealVfsFilesystem>()}, + provider{std::make_unique<FileSys::ManualContentProvider>()} { InitializeLogging(); + LoadTranslation(); + setAcceptDrops(true); ui.setupUi(this); statusBar()->hide(); @@ -214,9 +229,26 @@ GMainWindow::GMainWindow() LOG_INFO(Frontend, "yuzu Version: {} | {}-{}", yuzu_build_version, Common::g_scm_branch, Common::g_scm_desc); #ifdef ARCHITECTURE_x86_64 - LOG_INFO(Frontend, "Host CPU: {}", Common::GetCPUCaps().cpu_string); + const auto& caps = Common::GetCPUCaps(); + std::string cpu_string = caps.cpu_string; + if (caps.avx || caps.avx2 || caps.avx512) { + cpu_string += " | AVX"; + if (caps.avx512) { + cpu_string += "512"; + } else if (caps.avx2) { + cpu_string += '2'; + } + if (caps.fma || caps.fma4) { + cpu_string += " | FMA"; + } + } + LOG_INFO(Frontend, "Host CPU: {}", cpu_string); #endif LOG_INFO(Frontend, "Host OS: {}", QSysInfo::prettyProductName().toStdString()); + LOG_INFO(Frontend, "Host RAM: {:.2f} GB", + Common::GetMemInfo().TotalPhysicalMemory / 1024.0f / 1024 / 1024); + LOG_INFO(Frontend, "Host Swap: {:.2f} GB", + Common::GetMemInfo().TotalSwapMemory / 1024.0f / 1024 / 1024); UpdateWindowTitle(); show(); @@ -236,10 +268,20 @@ GMainWindow::GMainWindow() // Show one-time "callout" messages to the user ShowTelemetryCallout(); + // make sure menubar has the arrow cursor instead of inheriting from this + ui.menubar->setCursor(QCursor()); + statusBar()->setCursor(QCursor()); + + mouse_hide_timer.setInterval(default_mouse_timeout); + connect(&mouse_hide_timer, &QTimer::timeout, this, &GMainWindow::HideMouseCursor); + connect(ui.menubar, &QMenuBar::hovered, this, &GMainWindow::ShowMouseCursor); + QStringList args = QApplication::arguments(); if (args.length() >= 2) { BootGame(args[1]); } + + MigrateConfigFiles(); } GMainWindow::~GMainWindow() { @@ -248,17 +290,36 @@ GMainWindow::~GMainWindow() { delete render_window; } +void GMainWindow::ControllerSelectorReconfigureControllers( + const Core::Frontend::ControllerParameters& parameters) { + QtControllerSelectorDialog dialog(this, parameters, input_subsystem.get()); + + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowStaysOnTopHint | + Qt::WindowTitleHint | Qt::WindowSystemMenuHint); + dialog.setWindowModality(Qt::WindowModal); + dialog.exec(); + + emit ControllerSelectorReconfigureFinished(); + + // Don't forget to apply settings. + Settings::Apply(); + config->Save(); + + UpdateStatusButtons(); +} + void GMainWindow::ProfileSelectorSelectProfile() { QtProfileSelectionDialog dialog(this); - dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | - Qt::WindowSystemMenuHint | Qt::WindowCloseButtonHint); + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowStaysOnTopHint | + Qt::WindowTitleHint | Qt::WindowSystemMenuHint | + Qt::WindowCloseButtonHint); dialog.setWindowModality(Qt::WindowModal); if (dialog.exec() == QDialog::Rejected) { emit ProfileSelectorFinishedSelection(std::nullopt); return; } - Service::Account::ProfileManager manager; + const Service::Account::ProfileManager manager; const auto uuid = manager.GetUser(static_cast<std::size_t>(dialog.GetIndex())); if (!uuid.has_value()) { emit ProfileSelectorFinishedSelection(std::nullopt); @@ -271,8 +332,9 @@ void GMainWindow::ProfileSelectorSelectProfile() { void GMainWindow::SoftwareKeyboardGetText( const Core::Frontend::SoftwareKeyboardParameters& parameters) { QtSoftwareKeyboardDialog dialog(this, parameters); - dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | - Qt::WindowSystemMenuHint | Qt::WindowCloseButtonHint); + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowStaysOnTopHint | + Qt::WindowTitleHint | Qt::WindowSystemMenuHint | + Qt::WindowCloseButtonHint); dialog.setWindowModality(Qt::WindowModal); if (dialog.exec() == QDialog::Rejected) { @@ -435,7 +497,7 @@ void GMainWindow::InitializeWidgets() { #ifdef YUZU_ENABLE_COMPATIBILITY_REPORTING ui.action_Report_Compatibility->setVisible(true); #endif - render_window = new GRenderWindow(this, emu_thread.get()); + render_window = new GRenderWindow(this, emu_thread.get(), input_subsystem); render_window->hide(); game_list = new GameList(vfs, provider.get(), this); @@ -464,6 +526,8 @@ void GMainWindow::InitializeWidgets() { message_label->setAlignment(Qt::AlignLeft); statusBar()->addPermanentWidget(message_label, 1); + shader_building_label = new QLabel(); + shader_building_label->setToolTip(tr("The amount of shaders currently being built")); emu_speed_label = new QLabel(); emu_speed_label->setToolTip( tr("Current emulation speed. Values higher or lower than 100% " @@ -476,7 +540,8 @@ void GMainWindow::InitializeWidgets() { tr("Time taken to emulate a Switch frame, not counting framelimiting or v-sync. For " "full-speed emulation this should be at most 16.67 ms.")); - for (auto& label : {emu_speed_label, game_fps_label, emu_frametime_label}) { + for (auto& label : + {shader_building_label, emu_speed_label, game_fps_label, emu_frametime_label}) { label->setVisible(false); label->setFrameStyle(QFrame::NoFrame); label->setContentsMargins(4, 0, 4, 0); @@ -488,13 +553,14 @@ void GMainWindow::InitializeWidgets() { dock_status_button->setObjectName(QStringLiteral("TogglableStatusBarButton")); dock_status_button->setFocusPolicy(Qt::NoFocus); connect(dock_status_button, &QPushButton::clicked, [&] { - Settings::values.use_docked_mode = !Settings::values.use_docked_mode; - dock_status_button->setChecked(Settings::values.use_docked_mode); - OnDockedModeChanged(!Settings::values.use_docked_mode, Settings::values.use_docked_mode); + Settings::values.use_docked_mode.SetValue(!Settings::values.use_docked_mode.GetValue()); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); + OnDockedModeChanged(!Settings::values.use_docked_mode.GetValue(), + Settings::values.use_docked_mode.GetValue()); }); dock_status_button->setText(tr("DOCK")); dock_status_button->setCheckable(true); - dock_status_button->setChecked(Settings::values.use_docked_mode); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); statusBar()->insertPermanentWidget(0, dock_status_button); // Setup ASync button @@ -505,14 +571,36 @@ void GMainWindow::InitializeWidgets() { if (emulation_running) { return; } - Settings::values.use_asynchronous_gpu_emulation = - !Settings::values.use_asynchronous_gpu_emulation; - async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation); + bool is_async = !Settings::values.use_asynchronous_gpu_emulation.GetValue() || + Settings::values.use_multi_core.GetValue(); + Settings::values.use_asynchronous_gpu_emulation.SetValue(is_async); + async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); Settings::Apply(); }); async_status_button->setText(tr("ASYNC")); async_status_button->setCheckable(true); - async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation); + async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); + + // Setup Multicore button + multicore_status_button = new QPushButton(); + multicore_status_button->setObjectName(QStringLiteral("TogglableStatusBarButton")); + multicore_status_button->setFocusPolicy(Qt::NoFocus); + connect(multicore_status_button, &QPushButton::clicked, [&] { + if (emulation_running) { + return; + } + Settings::values.use_multi_core.SetValue(!Settings::values.use_multi_core.GetValue()); + bool is_async = Settings::values.use_asynchronous_gpu_emulation.GetValue() || + Settings::values.use_multi_core.GetValue(); + Settings::values.use_asynchronous_gpu_emulation.SetValue(is_async); + async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); + multicore_status_button->setChecked(Settings::values.use_multi_core.GetValue()); + Settings::Apply(); + }); + multicore_status_button->setText(tr("MULTICORE")); + multicore_status_button->setCheckable(true); + multicore_status_button->setChecked(Settings::values.use_multi_core.GetValue()); + statusBar()->insertPermanentWidget(0, multicore_status_button); statusBar()->insertPermanentWidget(0, async_status_button); // Setup Renderer API button @@ -520,7 +608,7 @@ void GMainWindow::InitializeWidgets() { renderer_status_button->setObjectName(QStringLiteral("RendererStatusBarButton")); renderer_status_button->setCheckable(true); renderer_status_button->setFocusPolicy(Qt::NoFocus); - connect(renderer_status_button, &QPushButton::toggled, [=](bool checked) { + connect(renderer_status_button, &QPushButton::toggled, [this](bool checked) { renderer_status_button->setText(checked ? tr("VULKAN") : tr("OPENGL")); }); renderer_status_button->toggle(); @@ -530,16 +618,16 @@ void GMainWindow::InitializeWidgets() { renderer_status_button->setCheckable(false); renderer_status_button->setDisabled(true); #else - renderer_status_button->setChecked(Settings::values.renderer_backend == + renderer_status_button->setChecked(Settings::values.renderer_backend.GetValue() == Settings::RendererBackend::Vulkan); - connect(renderer_status_button, &QPushButton::clicked, [=] { + connect(renderer_status_button, &QPushButton::clicked, [this] { if (emulation_running) { return; } if (renderer_status_button->isChecked()) { - Settings::values.renderer_backend = Settings::RendererBackend::Vulkan; + Settings::values.renderer_backend.SetValue(Settings::RendererBackend::Vulkan); } else { - Settings::values.renderer_backend = Settings::RendererBackend::OpenGL; + Settings::values.renderer_backend.SetValue(Settings::RendererBackend::OpenGL); } Settings::Apply(); @@ -638,6 +726,11 @@ void GMainWindow::InitializeHotkeys() { ui.action_Capture_Screenshot->setShortcutContext( hotkey_registry.GetShortcutContext(main_window, capture_screenshot)); + ui.action_Fullscreen->setShortcut( + hotkey_registry.GetHotkey(main_window, fullscreen, this)->key()); + ui.action_Fullscreen->setShortcutContext( + hotkey_registry.GetShortcutContext(main_window, fullscreen)); + connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Load File"), this), &QShortcut::activated, this, &GMainWindow::OnMenuLoadFile); connect( @@ -671,24 +764,24 @@ void GMainWindow::InitializeHotkeys() { }); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Toggle Speed Limit"), this), &QShortcut::activated, this, [&] { - Settings::values.use_frame_limit = !Settings::values.use_frame_limit; + Settings::values.use_frame_limit.SetValue( + !Settings::values.use_frame_limit.GetValue()); UpdateStatusBar(); }); - // TODO: Remove this comment/static whenever the next major release of - // MSVC occurs and we make it a requirement (see: - // https://developercommunity.visualstudio.com/content/problem/93922/constexprs-are-trying-to-be-captured-in-lambda-fun.html) - static constexpr u16 SPEED_LIMIT_STEP = 5; + constexpr u16 SPEED_LIMIT_STEP = 5; connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Increase Speed Limit"), this), &QShortcut::activated, this, [&] { - if (Settings::values.frame_limit < 9999 - SPEED_LIMIT_STEP) { - Settings::values.frame_limit += SPEED_LIMIT_STEP; + if (Settings::values.frame_limit.GetValue() < 9999 - SPEED_LIMIT_STEP) { + Settings::values.frame_limit.SetValue(SPEED_LIMIT_STEP + + Settings::values.frame_limit.GetValue()); UpdateStatusBar(); } }); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Decrease Speed Limit"), this), &QShortcut::activated, this, [&] { - if (Settings::values.frame_limit > SPEED_LIMIT_STEP) { - Settings::values.frame_limit -= SPEED_LIMIT_STEP; + if (Settings::values.frame_limit.GetValue() > SPEED_LIMIT_STEP) { + Settings::values.frame_limit.SetValue(Settings::values.frame_limit.GetValue() - + SPEED_LIMIT_STEP); UpdateStatusBar(); } }); @@ -700,27 +793,31 @@ void GMainWindow::InitializeHotkeys() { }); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Capture Screenshot"), this), &QShortcut::activated, this, [&] { - if (emu_thread->IsRunning()) { + if (emu_thread != nullptr && emu_thread->IsRunning()) { OnCaptureScreenshot(); } }); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Change Docked Mode"), this), &QShortcut::activated, this, [&] { - Settings::values.use_docked_mode = !Settings::values.use_docked_mode; - OnDockedModeChanged(!Settings::values.use_docked_mode, - Settings::values.use_docked_mode); - dock_status_button->setChecked(Settings::values.use_docked_mode); + Settings::values.use_docked_mode.SetValue( + !Settings::values.use_docked_mode.GetValue()); + OnDockedModeChanged(!Settings::values.use_docked_mode.GetValue(), + Settings::values.use_docked_mode.GetValue()); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); }); + connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Mute Audio"), this), + &QShortcut::activated, this, + [] { Settings::values.audio_muted = !Settings::values.audio_muted; }); } void GMainWindow::SetDefaultUIGeometry() { - // geometry: 55% of the window contents are in the upper screen half, 45% in the lower half + // geometry: 53% of the window contents are in the upper screen half, 47% in the lower half const QRect screenRect = QApplication::desktop()->screenGeometry(this); const int w = screenRect.width() * 2 / 3; - const int h = screenRect.height() / 2; + const int h = screenRect.height() * 2 / 3; const int x = (screenRect.x() + screenRect.width()) / 2 - w / 2; - const int y = (screenRect.y() + screenRect.height()) / 2 - h * 55 / 100; + const int y = (screenRect.y() + screenRect.height()) / 2 - h * 53 / 100; setGeometry(x, y, w, h); } @@ -745,7 +842,7 @@ void GMainWindow::RestoreUIState() { OnDisplayTitleBars(ui.action_Display_Dock_Widget_Headers->isChecked()); ui.action_Show_Filter_Bar->setChecked(UISettings::values.show_filter_bar); - game_list->setFilterVisible(ui.action_Show_Filter_Bar->isChecked()); + game_list->SetFilterVisible(ui.action_Show_Filter_Bar->isChecked()); ui.action_Show_Status_Bar->setChecked(UISettings::values.show_status_bar); statusBar()->setVisible(ui.action_Show_Status_Bar->isChecked()); @@ -776,6 +873,9 @@ void GMainWindow::ConnectWidgetEvents() { connect(game_list, &GameList::OpenFolderRequested, this, &GMainWindow::OnGameListOpenFolder); connect(game_list, &GameList::OpenTransferableShaderCacheRequested, this, &GMainWindow::OnTransferableShaderCacheOpenFile); + connect(game_list, &GameList::RemoveInstalledEntryRequested, this, + &GMainWindow::OnGameListRemoveInstalledEntry); + connect(game_list, &GameList::RemoveFileRequested, this, &GMainWindow::OnGameListRemoveFile); connect(game_list, &GameList::DumpRomFSRequested, this, &GMainWindow::OnGameListDumpRomFS); connect(game_list, &GameList::CopyTIDRequested, this, &GMainWindow::OnGameListCopyTID); connect(game_list, &GameList::NavigateToGamedbEntryRequested, this, @@ -788,6 +888,9 @@ void GMainWindow::ConnectWidgetEvents() { connect(game_list, &GameList::OpenPerGameGeneralRequested, this, &GMainWindow::OnGameListOpenPerGameProperties); + connect(this, &GMainWindow::UpdateInstallProgress, this, + &GMainWindow::IncrementInstallProgress); + connect(this, &GMainWindow::EmulationStarting, render_window, &GRenderWindow::OnEmulationStarting); connect(this, &GMainWindow::EmulationStopping, render_window, @@ -802,10 +905,6 @@ void GMainWindow::ConnectMenuEvents() { connect(ui.action_Load_Folder, &QAction::triggered, this, &GMainWindow::OnMenuLoadFolder); connect(ui.action_Install_File_NAND, &QAction::triggered, this, &GMainWindow::OnMenuInstallToNAND); - connect(ui.action_Select_NAND_Directory, &QAction::triggered, this, - [this] { OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget::NAND); }); - connect(ui.action_Select_SDMC_Directory, &QAction::triggered, this, - [this] { OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget::SDMC); }); connect(ui.action_Exit, &QAction::triggered, this, &QMainWindow::close); connect(ui.action_Load_Amiibo, &QAction::triggered, this, &GMainWindow::OnLoadAmiibo); @@ -815,8 +914,14 @@ void GMainWindow::ConnectMenuEvents() { connect(ui.action_Stop, &QAction::triggered, this, &GMainWindow::OnStopGame); connect(ui.action_Report_Compatibility, &QAction::triggered, this, &GMainWindow::OnMenuReportCompatibility); + connect(ui.action_Open_Mods_Page, &QAction::triggered, this, &GMainWindow::OnOpenModsPage); + connect(ui.action_Open_Quickstart_Guide, &QAction::triggered, this, + &GMainWindow::OnOpenQuickstartGuide); + connect(ui.action_Open_FAQ, &QAction::triggered, this, &GMainWindow::OnOpenFAQ); connect(ui.action_Restart, &QAction::triggered, this, [this] { BootGame(QString(game_path)); }); connect(ui.action_Configure, &QAction::triggered, this, &GMainWindow::OnConfigure); + connect(ui.action_Configure_Current_Game, &QAction::triggered, this, + &GMainWindow::OnConfigurePerGame); // View connect(ui.action_Single_Window_Mode, &QAction::triggered, this, @@ -825,12 +930,9 @@ void GMainWindow::ConnectMenuEvents() { &GMainWindow::OnDisplayTitleBars); connect(ui.action_Show_Filter_Bar, &QAction::triggered, this, &GMainWindow::OnToggleFilterBar); connect(ui.action_Show_Status_Bar, &QAction::triggered, statusBar(), &QStatusBar::setVisible); + connect(ui.action_Reset_Window_Size, &QAction::triggered, this, &GMainWindow::ResetWindowSize); // Fullscreen - ui.action_Fullscreen->setShortcut( - hotkey_registry - .GetHotkey(QStringLiteral("Main Window"), QStringLiteral("Fullscreen"), this) - ->key()); connect(ui.action_Fullscreen, &QAction::triggered, this, &GMainWindow::ToggleFullscreen); // Movie @@ -889,15 +991,18 @@ bool GMainWindow::LoadROM(const QString& filename) { system.SetFilesystem(vfs); system.SetAppletFrontendSet({ - nullptr, // Parental Controls - std::make_unique<QtErrorDisplay>(*this), // - nullptr, // Photo Viewer - std::make_unique<QtProfileSelector>(*this), // - std::make_unique<QtSoftwareKeyboard>(*this), // - std::make_unique<QtWebBrowser>(*this), // - nullptr, // E-Commerce + std::make_unique<QtControllerSelector>(*this), // Controller Selector + nullptr, // E-Commerce + std::make_unique<QtErrorDisplay>(*this), // Error Display + nullptr, // Parental Controls + nullptr, // Photo Viewer + std::make_unique<QtProfileSelector>(*this), // Profile Selector + std::make_unique<QtSoftwareKeyboard>(*this), // Software Keyboard + std::make_unique<QtWebBrowser>(*this), // Web Browser }); + system.RegisterHostThread(); + const Core::System::ResultStatus result{system.Load(*render_window, filename.toStdString())}; const auto drd_callout = @@ -940,16 +1045,18 @@ bool GMainWindow::LoadROM(const QString& filename) { default: if (static_cast<u32>(result) > static_cast<u32>(Core::System::ResultStatus::ErrorLoader)) { - LOG_CRITICAL(Frontend, "Failed to load ROM!"); const u16 loader_id = static_cast<u16>(Core::System::ResultStatus::ErrorLoader); const u16 error_id = static_cast<u16>(result) - loader_id; + const std::string error_code = fmt::format("({:04X}-{:04X})", loader_id, error_id); + LOG_CRITICAL(Frontend, "Failed to load ROM! {}", error_code); QMessageBox::critical( - this, tr("Error while loading ROM!"), + this, + tr("Error while loading ROM! ").append(QString::fromStdString(error_code)), QString::fromStdString(fmt::format( - "While attempting to load the ROM requested, an error occured. Please " - "refer to the yuzu wiki for more information or the yuzu discord for " - "additional help.\n\nError Code: {:04X}-{:04X}\nError Description: {}", - loader_id, error_id, static_cast<Loader::ResultStatus>(error_id)))); + "{}<br>Please follow <a href='https://yuzu-emu.org/help/quickstart/'>the " + "yuzu quickstart guide</a> to redump your files.<br>You can refer " + "to the yuzu wiki</a> or the yuzu Discord</a> for help.", + static_cast<Loader::ResultStatus>(error_id)))); } else { QMessageBox::critical( this, tr("Error while loading ROM!"), @@ -961,7 +1068,7 @@ bool GMainWindow::LoadROM(const QString& filename) { } game_path = filename; - system.TelemetrySession().AddField(Telemetry::FieldType::App, "Frontend", "Qt"); + system.TelemetrySession().AddField(Common::Telemetry::FieldType::App, "Frontend", "Qt"); return true; } @@ -982,6 +1089,19 @@ void GMainWindow::BootGame(const QString& filename) { LOG_INFO(Frontend, "yuzu starting..."); StoreRecentFile(filename); // Put the filename on top of the list + u64 title_id{0}; + auto& system = Core::System::GetInstance(); + const auto v_file = Core::GetGameFileFromPath(vfs, filename.toUtf8().constData()); + const auto loader = Loader::GetLoader(system, v_file); + if (!(loader == nullptr || loader->ReadProgramId(title_id) != Loader::ResultStatus::Success)) { + // Load per game settings + Config per_game_config(fmt::format("{:016X}", title_id), Config::ConfigType::PerGameConfig); + } + + ConfigureVibration::SetAllVibrationDevices(); + + Settings::LogSettings(); + if (UISettings::values.select_user_on_boot) { SelectAndSetCurrentUser(); } @@ -1006,30 +1126,42 @@ void GMainWindow::BootGame(const QString& filename) { &LoadingScreen::OnLoadProgress, Qt::QueuedConnection); // Update the GUI + UpdateStatusButtons(); if (ui.action_Single_Window_Mode->isChecked()) { game_list->hide(); game_list_placeholder->hide(); } status_bar_update_timer.start(2000); async_status_button->setDisabled(true); + multicore_status_button->setDisabled(true); renderer_status_button->setDisabled(true); - const u64 title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); + if (UISettings::values.hide_mouse) { + mouse_hide_timer.start(); + setMouseTracking(true); + ui.centralwidget->setMouseTracking(true); + } std::string title_name; - const auto res = Core::System::GetInstance().GetGameName(title_name); - if (res != Loader::ResultStatus::Success) { - const auto [nacp, icon_file] = FileSys::PatchManager(title_id).GetControlMetadata(); - if (nacp != nullptr) - title_name = nacp->GetApplicationName(); + std::string title_version; + const auto res = system.GetGameName(title_name); - if (title_name.empty()) - title_name = FileUtil::GetFilename(filename.toStdString()); + const auto metadata = [&system, title_id] { + const FileSys::PatchManager pm(title_id, system.GetFileSystemController(), + system.GetContentProvider()); + return pm.GetControlMetadata(); + }(); + if (metadata.first != nullptr) { + title_version = metadata.first->GetVersionString(); + title_name = metadata.first->GetApplicationName(); + } + if (res != Loader::ResultStatus::Success || title_name.empty()) { + title_name = Common::FS::GetFilename(filename.toStdString()); } - LOG_INFO(Frontend, "Booting game: {:016X} | {}", title_id, title_name); - UpdateWindowTitle(QString::fromStdString(title_name)); + LOG_INFO(Frontend, "Booting game: {:016X} | {} | {}", title_id, title_name, title_version); + UpdateWindowTitle(title_name, title_version); - loading_screen->Prepare(Core::System::GetInstance().GetAppLoader()); + loading_screen->Prepare(system.GetAppLoader()); loading_screen->show(); emulation_running = true; @@ -1070,26 +1202,33 @@ void GMainWindow::ShutdownGame() { ui.action_Pause->setEnabled(false); ui.action_Stop->setEnabled(false); ui.action_Restart->setEnabled(false); + ui.action_Configure_Current_Game->setEnabled(false); ui.action_Report_Compatibility->setEnabled(false); ui.action_Load_Amiibo->setEnabled(false); ui.action_Capture_Screenshot->setEnabled(false); render_window->hide(); loading_screen->hide(); loading_screen->Clear(); - if (game_list->isEmpty()) + if (game_list->IsEmpty()) { game_list_placeholder->show(); - else + } else { game_list->show(); - game_list->setFilterFocus(); + } + game_list->SetFilterFocus(); + + setMouseTracking(false); + ui.centralwidget->setMouseTracking(false); UpdateWindowTitle(); // Disable status bar updates status_bar_update_timer.stop(); + shader_building_label->setVisible(false); emu_speed_label->setVisible(false); game_fps_label->setVisible(false); emu_frametime_label->setVisible(false); async_status_button->setEnabled(true); + multicore_status_button->setEnabled(true); #ifdef HAS_VULKAN renderer_status_button->setEnabled(true); #endif @@ -1137,50 +1276,82 @@ void GMainWindow::OnGameListLoadFile(QString game_path) { BootGame(game_path); } -void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target) { +void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target, + const std::string& game_path) { std::string path; QString open_target; + auto& system = Core::System::GetInstance(); + + const auto [user_save_size, device_save_size] = [this, &game_path, &program_id, &system] { + const FileSys::PatchManager pm{program_id, system.GetFileSystemController(), + system.GetContentProvider()}; + const auto control = pm.GetControlMetadata().first; + if (control != nullptr) { + return std::make_pair(control->GetDefaultNormalSaveSize(), + control->GetDeviceSaveDataSize()); + } else { + const auto file = Core::GetGameFileFromPath(vfs, game_path); + const auto loader = Loader::GetLoader(system, file); + + FileSys::NACP nacp{}; + loader->ReadControlData(nacp); + return std::make_pair(nacp.GetDefaultNormalSaveSize(), nacp.GetDeviceSaveDataSize()); + } + }(); + + const bool has_user_save{user_save_size > 0}; + const bool has_device_save{device_save_size > 0}; + + ASSERT_MSG(has_user_save != has_device_save, "Game uses both user and device savedata?"); + switch (target) { case GameListOpenTarget::SaveData: { open_target = tr("Save Data"); - const std::string nand_dir = FileUtil::GetUserPath(FileUtil::UserPath::NANDDir); - ASSERT(program_id != 0); + const std::string nand_dir = Common::FS::GetUserPath(Common::FS::UserPath::NANDDir); + + if (has_user_save) { + // User save data + const auto select_profile = [this] { + QtProfileSelectionDialog dialog(this); + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | + Qt::WindowSystemMenuHint | Qt::WindowCloseButtonHint); + dialog.setWindowModality(Qt::WindowModal); + + if (dialog.exec() == QDialog::Rejected) { + return -1; + } - const auto select_profile = [this] { - QtProfileSelectionDialog dialog(this); - dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | - Qt::WindowSystemMenuHint | Qt::WindowCloseButtonHint); - dialog.setWindowModality(Qt::WindowModal); + return dialog.GetIndex(); + }; - if (dialog.exec() == QDialog::Rejected) { - return -1; + const auto index = select_profile(); + if (index == -1) { + return; } - return dialog.GetIndex(); - }; - - const auto index = select_profile(); - if (index == -1) { - return; + Service::Account::ProfileManager manager; + const auto user_id = manager.GetUser(static_cast<std::size_t>(index)); + ASSERT(user_id); + path = nand_dir + FileSys::SaveDataFactory::GetFullPath( + FileSys::SaveDataSpaceId::NandUser, + FileSys::SaveDataType::SaveData, program_id, user_id->uuid, 0); + } else { + // Device save data + path = nand_dir + FileSys::SaveDataFactory::GetFullPath( + FileSys::SaveDataSpaceId::NandUser, + FileSys::SaveDataType::SaveData, program_id, {}, 0); } - Service::Account::ProfileManager manager; - const auto user_id = manager.GetUser(static_cast<std::size_t>(index)); - ASSERT(user_id); - path = nand_dir + FileSys::SaveDataFactory::GetFullPath(FileSys::SaveDataSpaceId::NandUser, - FileSys::SaveDataType::SaveData, - program_id, user_id->uuid, 0); - - if (!FileUtil::Exists(path)) { - FileUtil::CreateFullPath(path); - FileUtil::CreateDir(path); + if (!Common::FS::Exists(path)) { + Common::FS::CreateFullPath(path); + Common::FS::CreateDir(path); } break; } case GameListOpenTarget::ModData: { open_target = tr("Mod Data"); - const auto load_dir = FileUtil::GetUserPath(FileUtil::UserPath::LoadDir); + const auto load_dir = Common::FS::GetUserPath(Common::FS::UserPath::LoadDir); path = fmt::format("{}{:016X}", load_dir, program_id); break; } @@ -1201,14 +1372,12 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target } void GMainWindow::OnTransferableShaderCacheOpenFile(u64 program_id) { - ASSERT(program_id != 0); - const QString shader_dir = - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::ShaderDir)); - const QString tranferable_shader_cache_folder_path = + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::ShaderDir)); + const QString transferable_shader_cache_folder_path = shader_dir + QStringLiteral("opengl") + QDir::separator() + QStringLiteral("transferable"); const QString transferable_shader_cache_file_path = - tranferable_shader_cache_folder_path + QDir::separator() + + transferable_shader_cache_folder_path + QDir::separator() + QString::fromStdString(fmt::format("{:016X}.bin", program_id)); if (!QFile::exists(transferable_shader_cache_file_path)) { @@ -1229,7 +1398,7 @@ void GMainWindow::OnTransferableShaderCacheOpenFile(u64 program_id) { param << QDir::toNativeSeparators(transferable_shader_cache_file_path); QProcess::startDetached(explorer, param); #else - QDesktopServices::openUrl(QUrl::fromLocalFile(tranferable_shader_cache_folder_path)); + QDesktopServices::openUrl(QUrl::fromLocalFile(transferable_shader_cache_folder_path)); #endif } @@ -1273,6 +1442,175 @@ static bool RomFSRawCopy(QProgressDialog& dialog, const FileSys::VirtualDir& src return true; } +void GMainWindow::OnGameListRemoveInstalledEntry(u64 program_id, InstalledEntryType type) { + const QString entry_type = [this, type] { + switch (type) { + case InstalledEntryType::Game: + return tr("Contents"); + case InstalledEntryType::Update: + return tr("Update"); + case InstalledEntryType::AddOnContent: + return tr("DLC"); + default: + return QString{}; + } + }(); + + if (QMessageBox::question( + this, tr("Remove Entry"), tr("Remove Installed Game %1?").arg(entry_type), + QMessageBox::Yes | QMessageBox::No, QMessageBox::No) != QMessageBox::Yes) { + return; + } + + switch (type) { + case InstalledEntryType::Game: + RemoveBaseContent(program_id, entry_type); + [[fallthrough]]; + case InstalledEntryType::Update: + RemoveUpdateContent(program_id, entry_type); + if (type != InstalledEntryType::Game) { + break; + } + [[fallthrough]]; + case InstalledEntryType::AddOnContent: + RemoveAddOnContent(program_id, entry_type); + break; + } + Common::FS::DeleteDirRecursively(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + + DIR_SEP + "game_list"); + game_list->PopulateAsync(UISettings::values.game_dirs); +} + +void GMainWindow::RemoveBaseContent(u64 program_id, const QString& entry_type) { + const auto& fs_controller = Core::System::GetInstance().GetFileSystemController(); + const auto res = fs_controller.GetUserNANDContents()->RemoveExistingEntry(program_id) || + fs_controller.GetSDMCContents()->RemoveExistingEntry(program_id); + + if (res) { + QMessageBox::information(this, tr("Successfully Removed"), + tr("Successfully removed the installed base game.")); + } else { + QMessageBox::warning( + this, tr("Error Removing %1").arg(entry_type), + tr("The base game is not installed in the NAND and cannot be removed.")); + } +} + +void GMainWindow::RemoveUpdateContent(u64 program_id, const QString& entry_type) { + const auto update_id = program_id | 0x800; + const auto& fs_controller = Core::System::GetInstance().GetFileSystemController(); + const auto res = fs_controller.GetUserNANDContents()->RemoveExistingEntry(update_id) || + fs_controller.GetSDMCContents()->RemoveExistingEntry(update_id); + + if (res) { + QMessageBox::information(this, tr("Successfully Removed"), + tr("Successfully removed the installed update.")); + } else { + QMessageBox::warning(this, tr("Error Removing %1").arg(entry_type), + tr("There is no update installed for this title.")); + } +} + +void GMainWindow::RemoveAddOnContent(u64 program_id, const QString& entry_type) { + u32 count{}; + const auto& fs_controller = Core::System::GetInstance().GetFileSystemController(); + const auto dlc_entries = Core::System::GetInstance().GetContentProvider().ListEntriesFilter( + FileSys::TitleType::AOC, FileSys::ContentRecordType::Data); + + for (const auto& entry : dlc_entries) { + if ((entry.title_id & DLC_BASE_TITLE_ID_MASK) == program_id) { + const auto res = + fs_controller.GetUserNANDContents()->RemoveExistingEntry(entry.title_id) || + fs_controller.GetSDMCContents()->RemoveExistingEntry(entry.title_id); + if (res) { + ++count; + } + } + } + + if (count == 0) { + QMessageBox::warning(this, tr("Error Removing %1").arg(entry_type), + tr("There are no DLC installed for this title.")); + return; + } + + QMessageBox::information(this, tr("Successfully Removed"), + tr("Successfully removed %1 installed DLC.").arg(count)); +} + +void GMainWindow::OnGameListRemoveFile(u64 program_id, GameListRemoveTarget target) { + const QString question = [this, target] { + switch (target) { + case GameListRemoveTarget::ShaderCache: + return tr("Delete Transferable Shader Cache?"); + case GameListRemoveTarget::CustomConfiguration: + return tr("Remove Custom Game Configuration?"); + default: + return QString{}; + } + }(); + + if (QMessageBox::question(this, tr("Remove File"), question, QMessageBox::Yes | QMessageBox::No, + QMessageBox::No) != QMessageBox::Yes) { + return; + } + + switch (target) { + case GameListRemoveTarget::ShaderCache: + RemoveTransferableShaderCache(program_id); + break; + case GameListRemoveTarget::CustomConfiguration: + RemoveCustomConfiguration(program_id); + break; + } +} + +void GMainWindow::RemoveTransferableShaderCache(u64 program_id) { + const QString shader_dir = + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::ShaderDir)); + const QString transferable_shader_cache_folder_path = + shader_dir + QStringLiteral("opengl") + QDir::separator() + QStringLiteral("transferable"); + const QString transferable_shader_cache_file_path = + transferable_shader_cache_folder_path + QDir::separator() + + QString::fromStdString(fmt::format("{:016X}.bin", program_id)); + + if (!QFile::exists(transferable_shader_cache_file_path)) { + QMessageBox::warning(this, tr("Error Removing Transferable Shader Cache"), + tr("A shader cache for this title does not exist.")); + return; + } + + if (QFile::remove(transferable_shader_cache_file_path)) { + QMessageBox::information(this, tr("Successfully Removed"), + tr("Successfully removed the transferable shader cache.")); + } else { + QMessageBox::warning(this, tr("Error Removing Transferable Shader Cache"), + tr("Failed to remove the transferable shader cache.")); + } +} + +void GMainWindow::RemoveCustomConfiguration(u64 program_id) { + const QString config_dir = + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::ConfigDir)); + const QString custom_config_file_path = + config_dir + QStringLiteral("custom") + QDir::separator() + + QString::fromStdString(fmt::format("{:016X}.ini", program_id)); + + if (!QFile::exists(custom_config_file_path)) { + QMessageBox::warning(this, tr("Error Removing Custom Configuration"), + tr("A custom configuration for this title does not exist.")); + return; + } + + if (QFile::remove(custom_config_file_path)) { + QMessageBox::information(this, tr("Successfully Removed"), + tr("Successfully removed the custom game configuration.")); + } else { + QMessageBox::warning(this, tr("Error Removing Custom Configuration"), + tr("Failed to remove the custom game configuration.")); + } +} + void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_path) { const auto failed = [this] { QMessageBox::warning(this, tr("RomFS Extraction Failed!"), @@ -1280,7 +1618,8 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa "cancelled the operation.")); }; - const auto loader = Loader::GetLoader(vfs->OpenFile(game_path, FileSys::Mode::Read)); + auto& system = Core::System::GetInstance(); + const auto loader = Loader::GetLoader(system, vfs->OpenFile(game_path, FileSys::Mode::Read)); if (loader == nullptr) { failed(); return; @@ -1292,7 +1631,7 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa return; } - const auto& installed = Core::System::GetInstance().GetContentProvider(); + const auto& installed = system.GetContentProvider(); const auto romfs_title_id = SelectRomFSDumpTarget(installed, program_id); if (!romfs_title_id) { @@ -1301,12 +1640,14 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa } const auto path = fmt::format( - "{}{:016X}/romfs", FileUtil::GetUserPath(FileUtil::UserPath::DumpDir), *romfs_title_id); + "{}{:016X}/romfs", Common::FS::GetUserPath(Common::FS::UserPath::DumpDir), *romfs_title_id); FileSys::VirtualFile romfs; if (*romfs_title_id == program_id) { - romfs = file; + const u64 ivfc_offset = loader->ReadRomFSIVFCOffset(); + const FileSys::PatchManager pm{program_id, system.GetFileSystemController(), installed}; + romfs = pm.PatchRomFS(file, ivfc_offset, FileSys::ContentRecordType::Program); } else { romfs = installed.GetEntry(*romfs_title_id, FileSys::ContentRecordType::Data)->GetRomFS(); } @@ -1379,13 +1720,13 @@ void GMainWindow::OnGameListNavigateToGamedbEntry(u64 program_id, void GMainWindow::OnGameListOpenDirectory(const QString& directory) { QString path; if (directory == QStringLiteral("SDMC")) { - path = QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir) + + path = QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir) + "Nintendo/Contents/registered"); } else if (directory == QStringLiteral("UserNAND")) { - path = QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + + path = QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + "user/Contents/registered"); } else if (directory == QStringLiteral("SysNAND")) { - path = QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) + + path = QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::NANDDir) + "system/Contents/registered"); } else { path = directory; @@ -1399,8 +1740,10 @@ void GMainWindow::OnGameListOpenDirectory(const QString& directory) { void GMainWindow::OnGameListAddDirectory() { const QString dir_path = QFileDialog::getExistingDirectory(this, tr("Select Directory")); - if (dir_path.isEmpty()) + if (dir_path.isEmpty()) { return; + } + UISettings::GameDir game_dir{dir_path, false, true}; if (!UISettings::values.game_dirs.contains(game_dir)) { UISettings::values.game_dirs.append(game_dir); @@ -1420,26 +1763,15 @@ void GMainWindow::OnGameListShowList(bool show) { void GMainWindow::OnGameListOpenPerGameProperties(const std::string& file) { u64 title_id{}; const auto v_file = Core::GetGameFileFromPath(vfs, file); - const auto loader = Loader::GetLoader(v_file); + const auto loader = Loader::GetLoader(Core::System::GetInstance(), v_file); + if (loader == nullptr || loader->ReadProgramId(title_id) != Loader::ResultStatus::Success) { QMessageBox::information(this, tr("Properties"), tr("The game properties could not be loaded.")); return; } - ConfigurePerGameGeneral dialog(this, title_id); - dialog.LoadFromFile(v_file); - auto result = dialog.exec(); - if (result == QDialog::Accepted) { - dialog.ApplyConfiguration(); - - const auto reload = UISettings::values.is_game_list_reload_pending.exchange(false); - if (reload) { - game_list->PopulateAsync(UISettings::values.game_dirs); - } - - config->Save(); - } + OpenPerGameConfiguration(title_id, file); } void GMainWindow::OnMenuLoadFile() { @@ -1479,209 +1811,255 @@ void GMainWindow::OnMenuLoadFolder() { } } +void GMainWindow::IncrementInstallProgress() { + install_progress->setValue(install_progress->value() + 1); +} + void GMainWindow::OnMenuInstallToNAND() { const QString file_filter = tr("Installable Switch File (*.nca *.nsp *.xci);;Nintendo Content Archive " - "(*.nca);;Nintendo Submissions Package (*.nsp);;NX Cartridge " + "(*.nca);;Nintendo Submission Package (*.nsp);;NX Cartridge " "Image (*.xci)"); - QString filename = QFileDialog::getOpenFileName(this, tr("Install File"), - UISettings::values.roms_path, file_filter); - if (filename.isEmpty()) { + QStringList filenames = QFileDialog::getOpenFileNames( + this, tr("Install Files"), UISettings::values.roms_path, file_filter); + + if (filenames.isEmpty()) { + return; + } + + InstallDialog installDialog(this, filenames); + if (installDialog.exec() == QDialog::Rejected) { + return; + } + + const QStringList files = installDialog.GetFiles(); + + if (files.isEmpty()) { return; } + int remaining = filenames.size(); + + // This would only overflow above 2^43 bytes (8.796 TB) + int total_size = 0; + for (const QString& file : files) { + total_size += static_cast<int>(QFile(file).size() / 0x1000); + } + if (total_size < 0) { + LOG_CRITICAL(Frontend, "Attempting to install too many files, aborting."); + return; + } + + QStringList new_files{}; // Newly installed files that do not yet exist in the NAND + QStringList overwritten_files{}; // Files that overwrote those existing in the NAND + QStringList failed_files{}; // Files that failed to install due to errors + + ui.action_Install_File_NAND->setEnabled(false); + + install_progress = new QProgressDialog(QString{}, tr("Cancel"), 0, total_size, this); + install_progress->setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint & + ~Qt::WindowMaximizeButtonHint); + install_progress->setAttribute(Qt::WA_DeleteOnClose, true); + install_progress->setFixedWidth(installDialog.GetMinimumWidth() + 40); + install_progress->show(); + + for (const QString& file : files) { + install_progress->setWindowTitle(tr("%n file(s) remaining", "", remaining)); + install_progress->setLabelText( + tr("Installing file \"%1\"...").arg(QFileInfo(file).fileName())); + + QFuture<InstallResult> future; + InstallResult result; + + if (file.endsWith(QStringLiteral("xci"), Qt::CaseInsensitive) || + file.endsWith(QStringLiteral("nsp"), Qt::CaseInsensitive)) { + + future = QtConcurrent::run([this, &file] { return InstallNSPXCI(file); }); + + while (!future.isFinished()) { + QCoreApplication::processEvents(); + } + + result = future.result(); + + } else { + result = InstallNCA(file); + } + + std::this_thread::sleep_for(std::chrono::milliseconds(10)); + + switch (result) { + case InstallResult::Success: + new_files.append(QFileInfo(file).fileName()); + break; + case InstallResult::Overwrite: + overwritten_files.append(QFileInfo(file).fileName()); + break; + case InstallResult::Failure: + failed_files.append(QFileInfo(file).fileName()); + break; + } + + --remaining; + } + + install_progress->close(); + + const QString install_results = + (new_files.isEmpty() ? QString{} + : tr("%n file(s) were newly installed\n", "", new_files.size())) + + (overwritten_files.isEmpty() + ? QString{} + : tr("%n file(s) were overwritten\n", "", overwritten_files.size())) + + (failed_files.isEmpty() ? QString{} + : tr("%n file(s) failed to install\n", "", failed_files.size())); + + QMessageBox::information(this, tr("Install Results"), install_results); + Common::FS::DeleteDirRecursively(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + + DIR_SEP + "game_list"); + game_list->PopulateAsync(UISettings::values.game_dirs); + ui.action_Install_File_NAND->setEnabled(true); +} + +InstallResult GMainWindow::InstallNSPXCI(const QString& filename) { const auto qt_raw_copy = [this](const FileSys::VirtualFile& src, const FileSys::VirtualFile& dest, std::size_t block_size) { - if (src == nullptr || dest == nullptr) + if (src == nullptr || dest == nullptr) { return false; - if (!dest->Resize(src->GetSize())) + } + if (!dest->Resize(src->GetSize())) { return false; + } std::array<u8, 0x1000> buffer{}; - const int progress_maximum = static_cast<int>(src->GetSize() / buffer.size()); - - QProgressDialog progress( - tr("Installing file \"%1\"...").arg(QString::fromStdString(src->GetName())), - tr("Cancel"), 0, progress_maximum, this); - progress.setWindowModality(Qt::WindowModal); for (std::size_t i = 0; i < src->GetSize(); i += buffer.size()) { - if (progress.wasCanceled()) { + if (install_progress->wasCanceled()) { dest->Resize(0); return false; } - const int progress_value = static_cast<int>(i / buffer.size()); - progress.setValue(progress_value); + emit UpdateInstallProgress(); const auto read = src->Read(buffer.data(), buffer.size(), i); dest->Write(buffer.data(), read, i); } - return true; }; - const auto success = [this]() { - QMessageBox::information(this, tr("Successfully Installed"), - tr("The file was successfully installed.")); - game_list->PopulateAsync(UISettings::values.game_dirs); - FileUtil::DeleteDirRecursively(FileUtil::GetUserPath(FileUtil::UserPath::CacheDir) + - DIR_SEP + "game_list"); - }; - - const auto failed = [this]() { - QMessageBox::warning( - this, tr("Failed to Install"), - tr("There was an error while attempting to install the provided file. It " - "could have an incorrect format or be missing metadata. Please " - "double-check your file and try again.")); - }; - - const auto overwrite = [this]() { - return QMessageBox::question(this, tr("Failed to Install"), - tr("The file you are attempting to install already exists " - "in the cache. Would you like to overwrite it?")) == - QMessageBox::Yes; - }; - - if (filename.endsWith(QStringLiteral("xci"), Qt::CaseInsensitive) || - filename.endsWith(QStringLiteral("nsp"), Qt::CaseInsensitive)) { - std::shared_ptr<FileSys::NSP> nsp; - if (filename.endsWith(QStringLiteral("nsp"), Qt::CaseInsensitive)) { - nsp = std::make_shared<FileSys::NSP>( - vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); - if (nsp->IsExtractedType()) - failed(); - } else { - const auto xci = std::make_shared<FileSys::XCI>( - vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); - nsp = xci->GetSecurePartitionNSP(); - } - - if (nsp->GetStatus() != Loader::ResultStatus::Success) { - failed(); - return; - } - const auto res = Core::System::GetInstance() - .GetFileSystemController() - .GetUserNANDContents() - ->InstallEntry(*nsp, false, qt_raw_copy); - if (res == FileSys::InstallResult::Success) { - success(); - } else { - if (res == FileSys::InstallResult::ErrorAlreadyExists) { - if (overwrite()) { - const auto res2 = Core::System::GetInstance() - .GetFileSystemController() - .GetUserNANDContents() - ->InstallEntry(*nsp, true, qt_raw_copy); - if (res2 == FileSys::InstallResult::Success) { - success(); - } else { - failed(); - } - } - } else { - failed(); - } + std::shared_ptr<FileSys::NSP> nsp; + if (filename.endsWith(QStringLiteral("nsp"), Qt::CaseInsensitive)) { + nsp = std::make_shared<FileSys::NSP>( + vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); + if (nsp->IsExtractedType()) { + return InstallResult::Failure; } } else { - const auto nca = std::make_shared<FileSys::NCA>( + const auto xci = std::make_shared<FileSys::XCI>( vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); - const auto id = nca->GetStatus(); + nsp = xci->GetSecurePartitionNSP(); + } - // Game updates necessary are missing base RomFS - if (id != Loader::ResultStatus::Success && - id != Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { - failed(); - return; - } + if (nsp->GetStatus() != Loader::ResultStatus::Success) { + return InstallResult::Failure; + } + const auto res = + Core::System::GetInstance().GetFileSystemController().GetUserNANDContents()->InstallEntry( + *nsp, true, qt_raw_copy); + if (res == FileSys::InstallResult::Success) { + return InstallResult::Success; + } else if (res == FileSys::InstallResult::OverwriteExisting) { + return InstallResult::Overwrite; + } else { + return InstallResult::Failure; + } +} - const QStringList tt_options{tr("System Application"), - tr("System Archive"), - tr("System Application Update"), - tr("Firmware Package (Type A)"), - tr("Firmware Package (Type B)"), - tr("Game"), - tr("Game Update"), - tr("Game DLC"), - tr("Delta Title")}; - bool ok; - const auto item = QInputDialog::getItem( - this, tr("Select NCA Install Type..."), - tr("Please select the type of title you would like to install this NCA as:\n(In " - "most instances, the default 'Game' is fine.)"), - tt_options, 5, false, &ok); - - auto index = tt_options.indexOf(item); - if (!ok || index == -1) { - QMessageBox::warning(this, tr("Failed to Install"), - tr("The title type you selected for the NCA is invalid.")); - return; +InstallResult GMainWindow::InstallNCA(const QString& filename) { + const auto qt_raw_copy = [this](const FileSys::VirtualFile& src, + const FileSys::VirtualFile& dest, std::size_t block_size) { + if (src == nullptr || dest == nullptr) { + return false; } - - // If index is equal to or past Game, add the jump in TitleType. - if (index >= 5) { - index += static_cast<size_t>(FileSys::TitleType::Application) - - static_cast<size_t>(FileSys::TitleType::FirmwarePackageB); + if (!dest->Resize(src->GetSize())) { + return false; } - FileSys::InstallResult res; - if (index >= static_cast<size_t>(FileSys::TitleType::Application)) { - res = Core::System::GetInstance() - .GetFileSystemController() - .GetUserNANDContents() - ->InstallEntry(*nca, static_cast<FileSys::TitleType>(index), false, - qt_raw_copy); - } else { - res = Core::System::GetInstance() - .GetFileSystemController() - .GetSystemNANDContents() - ->InstallEntry(*nca, static_cast<FileSys::TitleType>(index), false, - qt_raw_copy); - } + std::array<u8, 0x1000> buffer{}; - if (res == FileSys::InstallResult::Success) { - success(); - } else if (res == FileSys::InstallResult::ErrorAlreadyExists) { - if (overwrite()) { - const auto res2 = Core::System::GetInstance() - .GetFileSystemController() - .GetUserNANDContents() - ->InstallEntry(*nca, static_cast<FileSys::TitleType>(index), - true, qt_raw_copy); - if (res2 == FileSys::InstallResult::Success) { - success(); - } else { - failed(); - } + for (std::size_t i = 0; i < src->GetSize(); i += buffer.size()) { + if (install_progress->wasCanceled()) { + dest->Resize(0); + return false; } - } else { - failed(); - } - } -} -void GMainWindow::OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget target) { - const auto res = QMessageBox::information( - this, tr("Changing Emulated Directory"), - tr("You are about to change the emulated %1 directory of the system. Please note " - "that this does not also move the contents of the previous directory to the " - "new one and you will have to do that yourself.") - .arg(target == EmulatedDirectoryTarget::SDMC ? tr("SD card") : tr("NAND")), - QMessageBox::StandardButtons{QMessageBox::Ok, QMessageBox::Cancel}); + emit UpdateInstallProgress(); - if (res == QMessageBox::Cancel) - return; + const auto read = src->Read(buffer.data(), buffer.size(), i); + dest->Write(buffer.data(), read, i); + } + return true; + }; - QString dir_path = QFileDialog::getExistingDirectory(this, tr("Select Directory")); - if (!dir_path.isEmpty()) { - FileUtil::GetUserPath(target == EmulatedDirectoryTarget::SDMC ? FileUtil::UserPath::SDMCDir - : FileUtil::UserPath::NANDDir, - dir_path.toStdString()); - Core::System::GetInstance().GetFileSystemController().CreateFactories(*vfs); - game_list->PopulateAsync(UISettings::values.game_dirs); + const auto nca = + std::make_shared<FileSys::NCA>(vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); + const auto id = nca->GetStatus(); + + // Game updates necessary are missing base RomFS + if (id != Loader::ResultStatus::Success && + id != Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { + return InstallResult::Failure; + } + + const QStringList tt_options{tr("System Application"), + tr("System Archive"), + tr("System Application Update"), + tr("Firmware Package (Type A)"), + tr("Firmware Package (Type B)"), + tr("Game"), + tr("Game Update"), + tr("Game DLC"), + tr("Delta Title")}; + bool ok; + const auto item = QInputDialog::getItem( + this, tr("Select NCA Install Type..."), + tr("Please select the type of title you would like to install this NCA as:\n(In " + "most instances, the default 'Game' is fine.)"), + tt_options, 5, false, &ok); + + auto index = tt_options.indexOf(item); + if (!ok || index == -1) { + QMessageBox::warning(this, tr("Failed to Install"), + tr("The title type you selected for the NCA is invalid.")); + return InstallResult::Failure; + } + + // If index is equal to or past Game, add the jump in TitleType. + if (index >= 5) { + index += static_cast<size_t>(FileSys::TitleType::Application) - + static_cast<size_t>(FileSys::TitleType::FirmwarePackageB); + } + + FileSys::InstallResult res; + if (index >= static_cast<s32>(FileSys::TitleType::Application)) { + res = Core::System::GetInstance() + .GetFileSystemController() + .GetUserNANDContents() + ->InstallEntry(*nca, static_cast<FileSys::TitleType>(index), true, qt_raw_copy); + } else { + res = Core::System::GetInstance() + .GetFileSystemController() + .GetSystemNANDContents() + ->InstallEntry(*nca, static_cast<FileSys::TitleType>(index), true, qt_raw_copy); + } + + if (res == FileSys::InstallResult::Success) { + return InstallResult::Success; + } else if (res == FileSys::InstallResult::OverwriteExisting) { + return InstallResult::Overwrite; + } else { + return InstallResult::Failure; } } @@ -1707,6 +2085,7 @@ void GMainWindow::OnStartGame() { emu_thread->SetRunning(true); + qRegisterMetaType<Core::Frontend::ControllerParameters>("Core::Frontend::ControllerParameters"); qRegisterMetaType<Core::Frontend::SoftwareKeyboardParameters>( "Core::Frontend::SoftwareKeyboardParameters"); qRegisterMetaType<Core::System::ResultStatus>("Core::System::ResultStatus"); @@ -1722,6 +2101,7 @@ void GMainWindow::OnStartGame() { ui.action_Pause->setEnabled(true); ui.action_Stop->setEnabled(true); ui.action_Restart->setEnabled(true); + ui.action_Configure_Current_Game->setEnabled(true); ui.action_Report_Compatibility->setEnabled(true); discord_rpc->Update(); @@ -1747,6 +2127,9 @@ void GMainWindow::OnStopGame() { } ShutdownGame(); + + Settings::RestoreGlobalState(); + UpdateStatusButtons(); } void GMainWindow::OnLoadComplete() { @@ -1772,6 +2155,26 @@ void GMainWindow::OnMenuReportCompatibility() { } } +void GMainWindow::OpenURL(const QUrl& url) { + const bool open = QDesktopServices::openUrl(url); + if (!open) { + QMessageBox::warning(this, tr("Error opening URL"), + tr("Unable to open the URL \"%1\".").arg(url.toString())); + } +} + +void GMainWindow::OnOpenModsPage() { + OpenURL(QUrl(QStringLiteral("https://github.com/yuzu-emu/yuzu/wiki/Switch-Mods"))); +} + +void GMainWindow::OnOpenQuickstartGuide() { + OpenURL(QUrl(QStringLiteral("https://yuzu-emu.org/help/quickstart/"))); +} + +void GMainWindow::OnOpenFAQ() { + OpenURL(QUrl(QStringLiteral("https://yuzu-emu.org/wiki/faq/"))); +} + void GMainWindow::ToggleFullscreen() { if (!emulation_running) { return; @@ -1832,11 +2235,28 @@ void GMainWindow::ToggleWindowMode() { } } +void GMainWindow::ResetWindowSize() { + const auto aspect_ratio = Layout::EmulationAspectRatio( + static_cast<Layout::AspectRatio>(Settings::values.aspect_ratio.GetValue()), + static_cast<float>(Layout::ScreenUndocked::Height) / Layout::ScreenUndocked::Width); + if (!ui.action_Single_Window_Mode->isChecked()) { + render_window->resize(Layout::ScreenUndocked::Height / aspect_ratio, + Layout::ScreenUndocked::Height); + } else { + resize(Layout::ScreenUndocked::Height / aspect_ratio, + Layout::ScreenUndocked::Height + menuBar()->height() + + (ui.action_Show_Status_Bar->isChecked() ? statusBar()->height() : 0)); + } +} + void GMainWindow::OnConfigure() { const auto old_theme = UISettings::values.theme; const bool old_discord_presence = UISettings::values.enable_discord_presence; - ConfigureDialog configure_dialog(this, hotkey_registry); + ConfigureDialog configure_dialog(this, hotkey_registry, input_subsystem.get()); + connect(&configure_dialog, &ConfigureDialog::LanguageChanged, this, + &GMainWindow::OnLanguageChanged); + const auto result = configure_dialog.exec(); if (result != QDialog::Accepted) { return; @@ -1859,12 +2279,46 @@ void GMainWindow::OnConfigure() { config->Save(); - dock_status_button->setChecked(Settings::values.use_docked_mode); - async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation); -#ifdef HAS_VULKAN - renderer_status_button->setChecked(Settings::values.renderer_backend == - Settings::RendererBackend::Vulkan); -#endif + if (UISettings::values.hide_mouse && emulation_running) { + setMouseTracking(true); + ui.centralwidget->setMouseTracking(true); + mouse_hide_timer.start(); + } else { + setMouseTracking(false); + ui.centralwidget->setMouseTracking(false); + } + + UpdateStatusButtons(); +} + +void GMainWindow::OnConfigurePerGame() { + const u64 title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); + OpenPerGameConfiguration(title_id, game_path.toStdString()); +} + +void GMainWindow::OpenPerGameConfiguration(u64 title_id, const std::string& file_name) { + const auto v_file = Core::GetGameFileFromPath(vfs, file_name); + + ConfigurePerGame dialog(this, title_id); + dialog.LoadFromFile(v_file); + auto result = dialog.exec(); + if (result == QDialog::Accepted) { + dialog.ApplyConfiguration(); + + const auto reload = UISettings::values.is_game_list_reload_pending.exchange(false); + if (reload) { + game_list->PopulateAsync(UISettings::values.game_dirs); + } + + // Do not cause the global config to write local settings into the config file + Settings::RestoreGlobalState(); + + if (!Core::System::GetInstance().IsPoweredOn()) { + config->Save(); + } + } else { + Settings::RestoreGlobalState(); + } } void GMainWindow::OnLoadAmiibo() { @@ -1914,7 +2368,7 @@ void GMainWindow::LoadAmiibo(const QString& filename) { void GMainWindow::OnOpenYuzuFolder() { QDesktopServices::openUrl(QUrl::fromLocalFile( - QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::UserDir)))); + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::UserDir)))); } void GMainWindow::OnAbout() { @@ -1923,31 +2377,66 @@ void GMainWindow::OnAbout() { } void GMainWindow::OnToggleFilterBar() { - game_list->setFilterVisible(ui.action_Show_Filter_Bar->isChecked()); + game_list->SetFilterVisible(ui.action_Show_Filter_Bar->isChecked()); if (ui.action_Show_Filter_Bar->isChecked()) { - game_list->setFilterFocus(); + game_list->SetFilterFocus(); } else { - game_list->clearFilter(); + game_list->ClearFilter(); } } void GMainWindow::OnCaptureScreenshot() { OnPauseGame(); - QFileDialog png_dialog(this, tr("Capture Screenshot"), UISettings::values.screenshot_path, - tr("PNG Image (*.png)")); - png_dialog.setAcceptMode(QFileDialog::AcceptSave); - png_dialog.setDefaultSuffix(QStringLiteral("png")); - if (png_dialog.exec()) { - const QString path = png_dialog.selectedFiles().first(); - if (!path.isEmpty()) { - UISettings::values.screenshot_path = QFileInfo(path).path(); - render_window->CaptureScreenshot(UISettings::values.screenshot_resolution_factor, path); + + const u64 title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); + const auto screenshot_path = + QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::ScreenshotsDir)); + const auto date = + QDateTime::currentDateTime().toString(QStringLiteral("yyyy-MM-dd_hh-mm-ss-zzz")); + QString filename = QStringLiteral("%1%2_%3.png") + .arg(screenshot_path) + .arg(title_id, 16, 16, QLatin1Char{'0'}) + .arg(date); + +#ifdef _WIN32 + if (UISettings::values.enable_screenshot_save_as) { + filename = QFileDialog::getSaveFileName(this, tr("Capture Screenshot"), filename, + tr("PNG Image (*.png)")); + if (filename.isEmpty()) { + OnStartGame(); + return; } } +#endif + render_window->CaptureScreenshot(UISettings::values.screenshot_resolution_factor, filename); OnStartGame(); } -void GMainWindow::UpdateWindowTitle(const QString& title_name) { +// TODO: Written 2020-10-01: Remove per-game config migration code when it is irrelevant +void GMainWindow::MigrateConfigFiles() { + const std::string& config_dir_str = Common::FS::GetUserPath(Common::FS::UserPath::ConfigDir); + const QDir config_dir = QDir(QString::fromStdString(config_dir_str)); + const QStringList config_dir_list = config_dir.entryList(QStringList(QStringLiteral("*.ini"))); + + Common::FS::CreateFullPath(fmt::format("{}custom" DIR_SEP, config_dir_str)); + for (QStringList::const_iterator it = config_dir_list.constBegin(); + it != config_dir_list.constEnd(); ++it) { + const auto filename = it->toStdString(); + if (filename.find_first_not_of("0123456789abcdefACBDEF", 0) < 16) { + continue; + } + const auto origin = fmt::format("{}{}", config_dir_str, filename); + const auto destination = fmt::format("{}custom" DIR_SEP "{}", config_dir_str, filename); + LOG_INFO(Frontend, "Migrating config file from {} to {}", origin, destination); + if (!Common::FS::Rename(origin, destination)) { + // Delete the old config file if one already exists in the new location. + Common::FS::Delete(origin); + } + } +} + +void GMainWindow::UpdateWindowTitle(const std::string& title_name, + const std::string& title_version) { const auto full_name = std::string(Common::g_build_fullname); const auto branch_name = std::string(Common::g_scm_branch); const auto description = std::string(Common::g_scm_desc); @@ -1956,7 +2445,7 @@ void GMainWindow::UpdateWindowTitle(const QString& title_name) { const auto date = QDateTime::currentDateTime().toString(QStringLiteral("yyyy-MM-dd")).toStdString(); - if (title_name.isEmpty()) { + if (title_name.empty()) { const auto fmt = std::string(Common::g_title_bar_format_idle); setWindowTitle(QString::fromStdString(fmt::format(fmt.empty() ? "yuzu {0}| {1}-{2}" : fmt, full_name, branch_name, description, @@ -1964,8 +2453,8 @@ void GMainWindow::UpdateWindowTitle(const QString& title_name) { } else { const auto fmt = std::string(Common::g_title_bar_format_running); setWindowTitle(QString::fromStdString( - fmt::format(fmt.empty() ? "yuzu {0}| {3} | {1}-{2}" : fmt, full_name, branch_name, - description, title_name.toStdString(), date, build_id))); + fmt::format(fmt.empty() ? "yuzu {0}| {3} | {6} | {1}-{2}" : fmt, full_name, branch_name, + description, title_name, date, build_id, title_version))); } } @@ -1976,22 +2465,69 @@ void GMainWindow::UpdateStatusBar() { } auto results = Core::System::GetInstance().GetAndResetPerfStats(); + auto& shader_notify = Core::System::GetInstance().GPU().ShaderNotify(); + const auto shaders_building = shader_notify.GetShadersBuilding(); - if (Settings::values.use_frame_limit) { + if (shaders_building != 0) { + shader_building_label->setText( + tr("Building: %n shader(s)", "", static_cast<int>(shaders_building))); + shader_building_label->setVisible(true); + } else { + shader_building_label->setVisible(false); + } + + if (Settings::values.use_frame_limit.GetValue()) { emu_speed_label->setText(tr("Speed: %1% / %2%") .arg(results.emulation_speed * 100.0, 0, 'f', 0) - .arg(Settings::values.frame_limit)); + .arg(Settings::values.frame_limit.GetValue())); } else { emu_speed_label->setText(tr("Speed: %1%").arg(results.emulation_speed * 100.0, 0, 'f', 0)); } game_fps_label->setText(tr("Game: %1 FPS").arg(results.game_fps, 0, 'f', 0)); emu_frametime_label->setText(tr("Frame: %1 ms").arg(results.frametime * 1000.0, 0, 'f', 2)); - emu_speed_label->setVisible(true); + emu_speed_label->setVisible(!Settings::values.use_multi_core.GetValue()); game_fps_label->setVisible(true); emu_frametime_label->setVisible(true); } +void GMainWindow::UpdateStatusButtons() { + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); + multicore_status_button->setChecked(Settings::values.use_multi_core.GetValue()); + Settings::values.use_asynchronous_gpu_emulation.SetValue( + Settings::values.use_asynchronous_gpu_emulation.GetValue() || + Settings::values.use_multi_core.GetValue()); + async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); +#ifdef HAS_VULKAN + renderer_status_button->setChecked(Settings::values.renderer_backend.GetValue() == + Settings::RendererBackend::Vulkan); +#endif +} + +void GMainWindow::HideMouseCursor() { + if (emu_thread == nullptr || UISettings::values.hide_mouse == false) { + mouse_hide_timer.stop(); + ShowMouseCursor(); + return; + } + setCursor(QCursor(Qt::BlankCursor)); +} + +void GMainWindow::ShowMouseCursor() { + unsetCursor(); + if (emu_thread != nullptr && UISettings::values.hide_mouse) { + mouse_hide_timer.start(); + } +} + +void GMainWindow::mouseMoveEvent(QMouseEvent* event) { + ShowMouseCursor(); +} + +void GMainWindow::mousePressEvent(QMouseEvent* event) { + ShowMouseCursor(); +} + void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string details) { QMessageBox::StandardButton answer; QString status_message; @@ -2051,6 +2587,9 @@ void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string det if (answer == QMessageBox::Yes) { if (emu_thread) { ShutdownGame(); + + Settings::RestoreGlobalState(); + UpdateStatusButtons(); } } else { // Only show the message if the game is still running. @@ -2065,57 +2604,60 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) { if (behavior == ReinitializeKeyBehavior::Warning) { const auto res = QMessageBox::information( this, tr("Confirm Key Rederivation"), - tr("You are about to force rederive all of your keys. \nIf you do not know what this " - "means or what you are doing, \nthis is a potentially destructive action. \nPlease " - "make sure this is what you want \nand optionally make backups.\n\nThis will delete " + tr("You are about to force rederive all of your keys. \nIf you do not know what " + "this " + "means or what you are doing, \nthis is a potentially destructive action. " + "\nPlease " + "make sure this is what you want \nand optionally make backups.\n\nThis will " + "delete " "your autogenerated key files and re-run the key derivation module."), QMessageBox::StandardButtons{QMessageBox::Ok, QMessageBox::Cancel}); if (res == QMessageBox::Cancel) return; - FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::KeysDir) + - "prod.keys_autogenerated"); - FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::KeysDir) + - "console.keys_autogenerated"); - FileUtil::Delete(FileUtil::GetUserPath(FileUtil::UserPath::KeysDir) + - "title.keys_autogenerated"); + Common::FS::Delete(Common::FS::GetUserPath(Common::FS::UserPath::KeysDir) + + "prod.keys_autogenerated"); + Common::FS::Delete(Common::FS::GetUserPath(Common::FS::UserPath::KeysDir) + + "console.keys_autogenerated"); + Common::FS::Delete(Common::FS::GetUserPath(Common::FS::UserPath::KeysDir) + + "title.keys_autogenerated"); } - Core::Crypto::KeyManager keys{}; + Core::Crypto::KeyManager& keys = Core::Crypto::KeyManager::Instance(); if (keys.BaseDeriveNecessary()) { Core::Crypto::PartitionDataManager pdm{vfs->OpenDirectory( - FileUtil::GetUserPath(FileUtil::UserPath::SysDataDir), FileSys::Mode::Read)}; + Common::FS::GetUserPath(Common::FS::UserPath::SysDataDir), FileSys::Mode::Read)}; const auto function = [this, &keys, &pdm] { keys.PopulateFromPartitionData(pdm); - Core::System::GetInstance().GetFileSystemController().CreateFactories(*vfs); - keys.DeriveETicket(pdm); + + auto& system = Core::System::GetInstance(); + system.GetFileSystemController().CreateFactories(*vfs); + keys.DeriveETicket(pdm, system.GetContentProvider()); }; QString errors; if (!pdm.HasFuses()) { - errors += tr("- Missing fuses - Cannot derive SBK\n"); + errors += tr("Missing fuses"); } if (!pdm.HasBoot0()) { - errors += tr("- Missing BOOT0 - Cannot derive master keys\n"); + errors += tr(" - Missing BOOT0"); } if (!pdm.HasPackage2()) { - errors += tr("- Missing BCPKG2-1-Normal-Main - Cannot derive general keys\n"); + errors += tr(" - Missing BCPKG2-1-Normal-Main"); } if (!pdm.HasProdInfo()) { - errors += tr("- Missing PRODINFO - Cannot derive title keys\n"); + errors += tr(" - Missing PRODINFO"); } if (!errors.isEmpty()) { QMessageBox::warning( - this, tr("Warning Missing Derivation Components"), - tr("The following are missing from your configuration that may hinder key " - "derivation. It will be attempted but may not complete.<br><br>") + - errors + - tr("<br><br>You can get all of these and dump all of your games easily by " - "following <a href='https://yuzu-emu.org/help/quickstart/'>the " - "quickstart guide</a>. Alternatively, you can use another method of dumping " - "to obtain all of your keys.")); + this, tr("Derivation Components Missing"), + tr("Components are missing that may hinder key derivation from completing. " + "<br>Please follow <a href='https://yuzu-emu.org/help/quickstart/'>the yuzu " + "quickstart guide</a> to get all your keys and " + "games.<br><br><small>(%1)</small>") + .arg(errors)); } QProgressDialog prog; @@ -2215,9 +2757,13 @@ void GMainWindow::closeEvent(QCloseEvent* event) { hotkey_registry.SaveHotkeys(); // Shutdown session if the emu thread is active... - if (emu_thread != nullptr) + if (emu_thread != nullptr) { ShutdownGame(); + Settings::RestoreGlobalState(); + UpdateStatusButtons(); + } + render_window->close(); QWidget::closeEvent(event); @@ -2348,6 +2894,43 @@ void GMainWindow::UpdateUITheme() { QIcon::setThemeSearchPaths(theme_paths); } +void GMainWindow::LoadTranslation() { + // If the selected language is English, no need to install any translation + if (UISettings::values.language == QStringLiteral("en")) { + return; + } + + bool loaded; + + if (UISettings::values.language.isEmpty()) { + // If the selected language is empty, use system locale + loaded = translator.load(QLocale(), {}, {}, QStringLiteral(":/languages/")); + } else { + // Otherwise load from the specified file + loaded = translator.load(UISettings::values.language, QStringLiteral(":/languages/")); + } + + if (loaded) { + qApp->installTranslator(&translator); + } else { + UISettings::values.language = QStringLiteral("en"); + } +} + +void GMainWindow::OnLanguageChanged(const QString& locale) { + if (UISettings::values.language != QStringLiteral("en")) { + qApp->removeTranslator(&translator); + } + + UISettings::values.language = locale; + LoadTranslation(); + ui.retranslateUi(this); + UpdateWindowTitle(); + + if (emulation_running) + ui.action_Start->setText(tr("Continue")); +} + void GMainWindow::SetDiscordEnabled([[maybe_unused]] bool state) { #ifdef USE_DISCORD_PRESENCE if (state) { @@ -2376,9 +2959,9 @@ int main(int argc, char* argv[]) { #ifdef __APPLE__ // If you start a bundle (binary) on OSX without the Terminal, the working directory is "/". - // But since we require the working directory to be the executable path for the location of the - // user folder in the Qt Frontend, we need to cd into that working directory - const std::string bin_path = FileUtil::GetBundleDirectory() + DIR_SEP + ".."; + // But since we require the working directory to be the executable path for the location of + // the user folder in the Qt Frontend, we need to cd into that working directory + const std::string bin_path = Common::FS::GetBundleDirectory() + DIR_SEP + ".."; chdir(bin_path.c_str()); #endif @@ -2397,8 +2980,6 @@ int main(int argc, char* argv[]) { QObject::connect(&app, &QGuiApplication::applicationStateChanged, &main_window, &GMainWindow::OnAppFocusStateChanged); - Settings::LogSettings(); - int result = app.exec(); detached_tasks.WaitForAllTasks(); return result; diff --git a/src/yuzu/main.h b/src/yuzu/main.h index a67125567..b380a66f3 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -10,6 +10,7 @@ #include <QMainWindow> #include <QTimer> +#include <QTranslator> #include "common/common_types.h" #include "core/core.h" @@ -28,34 +29,48 @@ class MicroProfileDialog; class ProfilerWidget; class QLabel; class QPushButton; +class QProgressDialog; class WaitTreeWidget; enum class GameListOpenTarget; +enum class GameListRemoveTarget; +enum class InstalledEntryType; class GameListPlaceholder; namespace Core::Frontend { +struct ControllerParameters; struct SoftwareKeyboardParameters; } // namespace Core::Frontend +namespace DiscordRPC { +class DiscordInterface; +} + namespace FileSys { class ContentProvider; class ManualContentProvider; class VfsFilesystem; } // namespace FileSys +namespace InputCommon { +class InputSubsystem; +} + enum class EmulatedDirectoryTarget { NAND, SDMC, }; +enum class InstallResult { + Success, + Overwrite, + Failure, +}; + enum class ReinitializeKeyBehavior { NoWarning, Warning, }; -namespace DiscordRPC { -class DiscordInterface; -} - class GMainWindow : public QMainWindow { Q_OBJECT @@ -76,8 +91,6 @@ public: GMainWindow(); ~GMainWindow() override; - std::unique_ptr<DiscordRPC::DiscordInterface> discord_rpc; - bool DropAction(QDropEvent* event); void AcceptDropEvent(QDropEvent* event); @@ -102,9 +115,14 @@ signals: // Signal that tells widgets to update icons to use the current theme void UpdateThemedIcons(); + void UpdateInstallProgress(); + + void ControllerSelectorReconfigureFinished(); + void ErrorDisplayFinished(); void ProfileSelectorFinishedSelection(std::optional<Common::UUID> uuid); + void SoftwareKeyboardFinishedText(std::optional<std::u16string> text); void SoftwareKeyboardFinishedCheckDialog(); @@ -113,6 +131,8 @@ signals: public slots: void OnLoadComplete(); + void ControllerSelectorReconfigureControllers( + const Core::Frontend::ControllerParameters& parameters); void ErrorDisplayDisplayError(QString body); void ProfileSelectorSelectProfile(); void SoftwareKeyboardGetText(const Core::Frontend::SoftwareKeyboardParameters& parameters); @@ -181,10 +201,16 @@ private slots: void OnPauseGame(); void OnStopGame(); void OnMenuReportCompatibility(); + void OnOpenModsPage(); + void OnOpenQuickstartGuide(); + void OnOpenFAQ(); /// Called whenever a user selects a game in the game list widget. void OnGameListLoadFile(QString game_path); - void OnGameListOpenFolder(u64 program_id, GameListOpenTarget target); + void OnGameListOpenFolder(u64 program_id, GameListOpenTarget target, + const std::string& game_path); void OnTransferableShaderCacheOpenFile(u64 program_id); + void OnGameListRemoveInstalledEntry(u64 program_id, InstalledEntryType type); + void OnGameListRemoveFile(u64 program_id, GameListRemoveTarget target); void OnGameListDumpRomFS(u64 program_id, const std::string& game_path); void OnGameListCopyTID(u64 program_id); void OnGameListNavigateToGamedbEntry(u64 program_id, @@ -195,11 +221,11 @@ private slots: void OnGameListOpenPerGameProperties(const std::string& file); void OnMenuLoadFile(); void OnMenuLoadFolder(); + void IncrementInstallProgress(); void OnMenuInstallToNAND(); - /// Called whenever a user select the "File->Select -- Directory" where -- is NAND or SD Card - void OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget target); void OnMenuRecentFile(); void OnConfigure(); + void OnConfigurePerGame(); void OnLoadAmiibo(); void OnOpenYuzuFolder(); void OnAbout(); @@ -210,17 +236,37 @@ private slots: void ShowFullscreen(); void HideFullscreen(); void ToggleWindowMode(); + void ResetWindowSize(); void OnCaptureScreenshot(); void OnCoreError(Core::System::ResultStatus, std::string); void OnReinitializeKeys(ReinitializeKeyBehavior behavior); + void OnLanguageChanged(const QString& locale); private: + void RemoveBaseContent(u64 program_id, const QString& entry_type); + void RemoveUpdateContent(u64 program_id, const QString& entry_type); + void RemoveAddOnContent(u64 program_id, const QString& entry_type); + void RemoveTransferableShaderCache(u64 program_id); + void RemoveCustomConfiguration(u64 program_id); std::optional<u64> SelectRomFSDumpTarget(const FileSys::ContentProvider&, u64 program_id); - void UpdateWindowTitle(const QString& title_name = {}); + InstallResult InstallNSPXCI(const QString& filename); + InstallResult InstallNCA(const QString& filename); + void MigrateConfigFiles(); + void UpdateWindowTitle(const std::string& title_name = {}, + const std::string& title_version = {}); void UpdateStatusBar(); + void UpdateStatusButtons(); + void HideMouseCursor(); + void ShowMouseCursor(); + void OpenURL(const QUrl& url); + void LoadTranslation(); + void OpenPerGameConfiguration(u64 title_id, const std::string& file_name); Ui::MainWindow ui; + std::unique_ptr<DiscordRPC::DiscordInterface> discord_rpc; + std::shared_ptr<InputCommon::InputSubsystem> input_subsystem; + GRenderWindow* render_window; GameList* game_list; LoadingScreen* loading_screen; @@ -229,10 +275,12 @@ private: // Status bar elements QLabel* message_label = nullptr; + QLabel* shader_building_label = nullptr; QLabel* emu_speed_label = nullptr; QLabel* game_fps_label = nullptr; QLabel* emu_frametime_label = nullptr; QPushButton* async_status_button = nullptr; + QPushButton* multicore_status_button = nullptr; QPushButton* renderer_status_button = nullptr; QPushButton* dock_status_button = nullptr; QTimer status_bar_update_timer; @@ -246,6 +294,7 @@ private: QString game_path; bool auto_paused = false; + QTimer mouse_hide_timer; // FS std::shared_ptr<FileSys::VfsFilesystem> vfs; @@ -263,8 +312,15 @@ private: HotkeyRegistry hotkey_registry; + QTranslator translator; + + // Install progress dialog + QProgressDialog* install_progress; + protected: void dropEvent(QDropEvent* event) override; void dragEnterEvent(QDragEnterEvent* event) override; void dragMoveEvent(QDragMoveEvent* event) override; + void mouseMoveEvent(QMouseEvent* event) override; + void mousePressEvent(QMouseEvent* event) override; }; diff --git a/src/yuzu/main.ui b/src/yuzu/main.ui index a2c9e4547..2f3792247 100644 --- a/src/yuzu/main.ui +++ b/src/yuzu/main.ui @@ -6,8 +6,8 @@ <rect> <x>0</x> <y>0</y> - <width>1081</width> - <height>730</height> + <width>1280</width> + <height>720</height> </rect> </property> <property name="windowTitle"> @@ -44,7 +44,7 @@ <rect> <x>0</x> <y>0</y> - <width>1081</width> + <width>1280</width> <height>21</height> </rect> </property> @@ -64,8 +64,6 @@ <addaction name="separator"/> <addaction name="menu_recent_files"/> <addaction name="separator"/> - <addaction name="action_Select_NAND_Directory"/> - <addaction name="action_Select_SDMC_Directory"/> <addaction name="separator"/> <addaction name="action_Load_Amiibo"/> <addaction name="separator"/> @@ -83,6 +81,7 @@ <addaction name="action_Restart"/> <addaction name="separator"/> <addaction name="action_Configure"/> + <addaction name="action_Configure_Current_Game"/> </widget> <widget class="QMenu" name="menu_View"> <property name="title"> @@ -98,6 +97,7 @@ <addaction name="action_Display_Dock_Widget_Headers"/> <addaction name="action_Show_Filter_Bar"/> <addaction name="action_Show_Status_Bar"/> + <addaction name="action_Reset_Window_Size"/> <addaction name="separator"/> <addaction name="menu_View_Debugging"/> </widget> @@ -114,6 +114,9 @@ <string>&Help</string> </property> <addaction name="action_Report_Compatibility"/> + <addaction name="action_Open_Mods_Page"/> + <addaction name="action_Open_Quickstart_Guide"/> + <addaction name="action_Open_FAQ"/> <addaction name="separator"/> <addaction name="action_About"/> </widget> @@ -128,7 +131,7 @@ <bool>true</bool> </property> <property name="text"> - <string>Install File to NAND...</string> + <string>Install Files to NAND...</string> </property> </action> <action name="action_Load_File"> @@ -217,20 +220,9 @@ <string>Show Status Bar</string> </property> </action> - <action name="action_Select_NAND_Directory"> + <action name="action_Reset_Window_Size"> <property name="text"> - <string>Select NAND Directory...</string> - </property> - <property name="toolTip"> - <string>Selects a folder to use as the root of the emulated NAND</string> - </property> - </action> - <action name="action_Select_SDMC_Directory"> - <property name="text"> - <string>Select SD Card Directory...</string> - </property> - <property name="toolTip"> - <string>Selects a folder to use as the root of the emulated SD card</string> + <string>Reset Window Size</string> </property> </action> <action name="action_Fullscreen"> @@ -268,6 +260,21 @@ <bool>false</bool> </property> </action> + <action name="action_Open_Mods_Page"> + <property name="text"> + <string>Open Mods Page</string> + </property> + </action> + <action name="action_Open_Quickstart_Guide"> + <property name="text"> + <string>Open Quickstart Guide</string> + </property> + </action> + <action name="action_Open_FAQ"> + <property name="text"> + <string>FAQ</string> + </property> + </action> <action name="action_Open_yuzu_Folder"> <property name="text"> <string>Open yuzu Folder</string> @@ -281,6 +288,14 @@ <string>Capture Screenshot</string> </property> </action> + <action name="action_Configure_Current_Game"> + <property name="enabled"> + <bool>false</bool> + </property> + <property name="text"> + <string>Configure Current Game...</string> + </property> + </action> </widget> <resources/> <connections/> diff --git a/src/yuzu/uisettings.cpp b/src/yuzu/uisettings.cpp index 738c4b2fc..37499fc85 100644 --- a/src/yuzu/uisettings.cpp +++ b/src/yuzu/uisettings.cpp @@ -11,7 +11,10 @@ const Themes themes{{ {"Light Colorful", "colorful"}, {"Dark", "qdarkstyle"}, {"Dark Colorful", "colorful_dark"}, + {"Midnight Blue", "qdarkstyle_midnight_blue"}, + {"Midnight Blue Colorful", "colorful_midnight_blue"}, }}; Values values = {}; + } // namespace UISettings diff --git a/src/yuzu/uisettings.h b/src/yuzu/uisettings.h index a675ecf4d..ce3945485 100644 --- a/src/yuzu/uisettings.h +++ b/src/yuzu/uisettings.h @@ -24,19 +24,19 @@ struct Shortcut { ContextualShortcut shortcut; }; -using Themes = std::array<std::pair<const char*, const char*>, 4>; +using Themes = std::array<std::pair<const char*, const char*>, 6>; extern const Themes themes; struct GameDir { QString path; - bool deep_scan; - bool expanded; + bool deep_scan = false; + bool expanded = false; bool operator==(const GameDir& rhs) const { return path == rhs.path; - }; + } bool operator!=(const GameDir& rhs) const { return !operator==(rhs); - }; + } }; struct Values { @@ -59,21 +59,23 @@ struct Values { bool confirm_before_closing; bool first_start; bool pause_when_in_background; + bool hide_mouse; bool select_user_on_boot; // Discord RPC bool enable_discord_presence; + bool enable_screenshot_save_as; u16 screenshot_resolution_factor; QString roms_path; QString symbols_path; - QString screenshot_path; QString game_dir_deprecated; bool game_dir_deprecated_deepscan; QVector<UISettings::GameDir> game_dirs; QStringList recent_files; + QString language; QString theme; @@ -85,9 +87,6 @@ struct Values { // logging bool show_console; - // Controllers - int profile_index; - // Game List bool show_add_ons; uint32_t icon_size; @@ -98,6 +97,7 @@ struct Values { }; extern Values values; + } // namespace UISettings Q_DECLARE_METATYPE(UISettings::GameDir*); diff --git a/src/yuzu/yuzu.rc b/src/yuzu/yuzu.rc index 1b253653f..4a3645a71 100644 --- a/src/yuzu/yuzu.rc +++ b/src/yuzu/yuzu.rc @@ -16,4 +16,4 @@ IDI_ICON1 ICON "../../dist/yuzu.ico" // RT_MANIFEST // -1 RT_MANIFEST "../../dist/yuzu.manifest" +0 RT_MANIFEST "../../dist/yuzu.manifest" diff --git a/src/yuzu_cmd/CMakeLists.txt b/src/yuzu_cmd/CMakeLists.txt index a15719a0f..57f9916f6 100644 --- a/src/yuzu_cmd/CMakeLists.txt +++ b/src/yuzu_cmd/CMakeLists.txt @@ -39,7 +39,5 @@ endif() if (MSVC) include(CopyYuzuSDLDeps) - include(CopyYuzuUnicornDeps) copy_yuzu_SDL_deps(yuzu-cmd) - copy_yuzu_unicorn_deps(yuzu-cmd) endif() diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp index f4cd905c9..e1adbbf2b 100644 --- a/src/yuzu_cmd/config.cpp +++ b/src/yuzu_cmd/config.cpp @@ -16,9 +16,11 @@ #include "yuzu_cmd/config.h" #include "yuzu_cmd/default_ini.h" +namespace FS = Common::FS; + Config::Config() { // TODO: Don't hardcode the path; let the frontend decide where to put the config files. - sdl2_config_loc = FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + "sdl2-config.ini"; + sdl2_config_loc = FS::GetUserPath(FS::UserPath::ConfigDir) + "sdl2-config.ini"; sdl2_config = std::make_unique<INIReader>(sdl2_config_loc); Reload(); @@ -31,8 +33,8 @@ bool Config::LoadINI(const std::string& default_contents, bool retry) { if (sdl2_config->ParseError() < 0) { if (retry) { LOG_WARNING(Config, "Failed to load {}. Creating file from defaults...", location); - FileUtil::CreateFullPath(location); - FileUtil::WriteStringToFile(true, location, default_contents); + FS::CreateFullPath(location); + FS::WriteStringToFile(true, location, default_contents); sdl2_config = std::make_unique<INIReader>(location); // Reopen file return LoadINI(default_contents, false); @@ -226,24 +228,24 @@ static const std::array<int, 8> keyboard_mods{ void Config::ReadValues() { // Controls - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { const auto group = fmt::format("ControlsP{}", p); for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { std::string default_param = InputCommon::GenerateKeyboardParam(default_buttons[i]); - Settings::values.players[p].buttons[i] = + Settings::values.players.GetValue()[p].buttons[i] = sdl2_config->Get(group, Settings::NativeButton::mapping[i], default_param); - if (Settings::values.players[p].buttons[i].empty()) - Settings::values.players[p].buttons[i] = default_param; + if (Settings::values.players.GetValue()[p].buttons[i].empty()) + Settings::values.players.GetValue()[p].buttons[i] = default_param; } for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { std::string default_param = InputCommon::GenerateAnalogParamFromKeys( default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], default_analogs[i][3], default_analogs[i][4], 0.5f); - Settings::values.players[p].analogs[i] = + Settings::values.players.GetValue()[p].analogs[i] = sdl2_config->Get(group, Settings::NativeAnalog::mapping[i], default_param); - if (Settings::values.players[p].analogs[i].empty()) - Settings::values.players[p].analogs[i] = default_param; + if (Settings::values.players.GetValue()[p].analogs[i].empty()) + Settings::values.players.GetValue()[p].analogs[i] = default_param; } } @@ -286,6 +288,12 @@ void Config::ReadValues() { Settings::values.debug_pad_analogs[i] = default_param; } + Settings::values.vibration_enabled.SetValue( + sdl2_config->GetBoolean("ControlsGeneral", "vibration_enabled", true)); + Settings::values.enable_accurate_vibrations.SetValue( + sdl2_config->GetBoolean("ControlsGeneral", "enable_accurate_vibrations", false)); + Settings::values.motion_enabled.SetValue( + sdl2_config->GetBoolean("ControlsGeneral", "motion_enabled", true)); Settings::values.touchscreen.enabled = sdl2_config->GetBoolean("ControlsGeneral", "touch_enabled", true); Settings::values.touchscreen.device = @@ -315,38 +323,30 @@ void Config::ReadValues() { // Data Storage Settings::values.use_virtual_sd = sdl2_config->GetBoolean("Data Storage", "use_virtual_sd", true); - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir, - sdl2_config->Get("Data Storage", "nand_directory", - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir))); - FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir, - sdl2_config->Get("Data Storage", "sdmc_directory", - FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir))); - FileUtil::GetUserPath(FileUtil::UserPath::LoadDir, - sdl2_config->Get("Data Storage", "load_directory", - FileUtil::GetUserPath(FileUtil::UserPath::LoadDir))); - FileUtil::GetUserPath(FileUtil::UserPath::DumpDir, - sdl2_config->Get("Data Storage", "dump_directory", - FileUtil::GetUserPath(FileUtil::UserPath::DumpDir))); - FileUtil::GetUserPath(FileUtil::UserPath::CacheDir, - sdl2_config->Get("Data Storage", "cache_directory", - FileUtil::GetUserPath(FileUtil::UserPath::CacheDir))); + FS::GetUserPath( + FS::UserPath::NANDDir, + sdl2_config->Get("Data Storage", "nand_directory", FS::GetUserPath(FS::UserPath::NANDDir))); + FS::GetUserPath( + FS::UserPath::SDMCDir, + sdl2_config->Get("Data Storage", "sdmc_directory", FS::GetUserPath(FS::UserPath::SDMCDir))); + FS::GetUserPath( + FS::UserPath::LoadDir, + sdl2_config->Get("Data Storage", "load_directory", FS::GetUserPath(FS::UserPath::LoadDir))); + FS::GetUserPath( + FS::UserPath::DumpDir, + sdl2_config->Get("Data Storage", "dump_directory", FS::GetUserPath(FS::UserPath::DumpDir))); + FS::GetUserPath(FS::UserPath::CacheDir, + sdl2_config->Get("Data Storage", "cache_directory", + FS::GetUserPath(FS::UserPath::CacheDir))); Settings::values.gamecard_inserted = sdl2_config->GetBoolean("Data Storage", "gamecard_inserted", false); Settings::values.gamecard_current_game = sdl2_config->GetBoolean("Data Storage", "gamecard_current_game", false); Settings::values.gamecard_path = sdl2_config->Get("Data Storage", "gamecard_path", ""); - Settings::values.nand_total_size = static_cast<Settings::NANDTotalSize>(sdl2_config->GetInteger( - "Data Storage", "nand_total_size", static_cast<long>(Settings::NANDTotalSize::S29_1GB))); - Settings::values.nand_user_size = static_cast<Settings::NANDUserSize>(sdl2_config->GetInteger( - "Data Storage", "nand_user_size", static_cast<long>(Settings::NANDUserSize::S26GB))); - Settings::values.nand_system_size = static_cast<Settings::NANDSystemSize>( - sdl2_config->GetInteger("Data Storage", "nand_system_size", - static_cast<long>(Settings::NANDSystemSize::S2_5GB))); - Settings::values.sdmc_size = static_cast<Settings::SDMCSize>(sdl2_config->GetInteger( - "Data Storage", "sdmc_size", static_cast<long>(Settings::SDMCSize::S16GB))); // System - Settings::values.use_docked_mode = sdl2_config->GetBoolean("System", "use_docked_mode", false); + Settings::values.use_docked_mode.SetValue( + sdl2_config->GetBoolean("System", "use_docked_mode", false)); const auto size = sdl2_config->GetInteger("System", "users_size", 0); Settings::values.current_user = std::clamp<int>( @@ -354,60 +354,76 @@ void Config::ReadValues() { const auto rng_seed_enabled = sdl2_config->GetBoolean("System", "rng_seed_enabled", false); if (rng_seed_enabled) { - Settings::values.rng_seed = sdl2_config->GetInteger("System", "rng_seed", 0); + Settings::values.rng_seed.SetValue(sdl2_config->GetInteger("System", "rng_seed", 0)); } else { - Settings::values.rng_seed = std::nullopt; + Settings::values.rng_seed.SetValue(std::nullopt); } const auto custom_rtc_enabled = sdl2_config->GetBoolean("System", "custom_rtc_enabled", false); if (custom_rtc_enabled) { - Settings::values.custom_rtc = - std::chrono::seconds(sdl2_config->GetInteger("System", "custom_rtc", 0)); + Settings::values.custom_rtc.SetValue( + std::chrono::seconds(sdl2_config->GetInteger("System", "custom_rtc", 0))); } else { - Settings::values.custom_rtc = std::nullopt; + Settings::values.custom_rtc.SetValue(std::nullopt); } + Settings::values.language_index.SetValue( + sdl2_config->GetInteger("System", "language_index", 1)); + Settings::values.time_zone_index.SetValue( + sdl2_config->GetInteger("System", "time_zone_index", 0)); + // Core - Settings::values.use_multi_core = sdl2_config->GetBoolean("Core", "use_multi_core", false); + Settings::values.use_multi_core.SetValue( + sdl2_config->GetBoolean("Core", "use_multi_core", true)); // Renderer const int renderer_backend = sdl2_config->GetInteger( "Renderer", "backend", static_cast<int>(Settings::RendererBackend::OpenGL)); - Settings::values.renderer_backend = static_cast<Settings::RendererBackend>(renderer_backend); + Settings::values.renderer_backend.SetValue( + static_cast<Settings::RendererBackend>(renderer_backend)); Settings::values.renderer_debug = sdl2_config->GetBoolean("Renderer", "debug", false); - Settings::values.vulkan_device = sdl2_config->GetInteger("Renderer", "vulkan_device", 0); - - Settings::values.resolution_factor = - static_cast<float>(sdl2_config->GetReal("Renderer", "resolution_factor", 1.0)); - Settings::values.aspect_ratio = - static_cast<int>(sdl2_config->GetInteger("Renderer", "aspect_ratio", 0)); - Settings::values.max_anisotropy = - static_cast<int>(sdl2_config->GetInteger("Renderer", "max_anisotropy", 0)); - Settings::values.use_frame_limit = sdl2_config->GetBoolean("Renderer", "use_frame_limit", true); - Settings::values.frame_limit = - static_cast<u16>(sdl2_config->GetInteger("Renderer", "frame_limit", 100)); - Settings::values.use_disk_shader_cache = - sdl2_config->GetBoolean("Renderer", "use_disk_shader_cache", false); - Settings::values.use_accurate_gpu_emulation = - sdl2_config->GetBoolean("Renderer", "use_accurate_gpu_emulation", false); - Settings::values.use_asynchronous_gpu_emulation = - sdl2_config->GetBoolean("Renderer", "use_asynchronous_gpu_emulation", false); - Settings::values.use_vsync = - static_cast<u16>(sdl2_config->GetInteger("Renderer", "use_vsync", 1)); - - Settings::values.bg_red = static_cast<float>(sdl2_config->GetReal("Renderer", "bg_red", 0.0)); - Settings::values.bg_green = - static_cast<float>(sdl2_config->GetReal("Renderer", "bg_green", 0.0)); - Settings::values.bg_blue = static_cast<float>(sdl2_config->GetReal("Renderer", "bg_blue", 0.0)); + Settings::values.vulkan_device.SetValue( + sdl2_config->GetInteger("Renderer", "vulkan_device", 0)); + + Settings::values.aspect_ratio.SetValue( + static_cast<int>(sdl2_config->GetInteger("Renderer", "aspect_ratio", 0))); + Settings::values.max_anisotropy.SetValue( + static_cast<int>(sdl2_config->GetInteger("Renderer", "max_anisotropy", 0))); + Settings::values.use_frame_limit.SetValue( + sdl2_config->GetBoolean("Renderer", "use_frame_limit", true)); + Settings::values.frame_limit.SetValue( + static_cast<u16>(sdl2_config->GetInteger("Renderer", "frame_limit", 100))); + Settings::values.use_disk_shader_cache.SetValue( + sdl2_config->GetBoolean("Renderer", "use_disk_shader_cache", false)); + const int gpu_accuracy_level = sdl2_config->GetInteger("Renderer", "gpu_accuracy", 0); + Settings::values.gpu_accuracy.SetValue(static_cast<Settings::GPUAccuracy>(gpu_accuracy_level)); + Settings::values.use_asynchronous_gpu_emulation.SetValue( + sdl2_config->GetBoolean("Renderer", "use_asynchronous_gpu_emulation", true)); + Settings::values.use_vsync.SetValue( + static_cast<u16>(sdl2_config->GetInteger("Renderer", "use_vsync", 1))); + Settings::values.use_assembly_shaders.SetValue( + sdl2_config->GetBoolean("Renderer", "use_assembly_shaders", true)); + Settings::values.use_asynchronous_shaders.SetValue( + sdl2_config->GetBoolean("Renderer", "use_asynchronous_shaders", false)); + Settings::values.use_asynchronous_shaders.SetValue( + sdl2_config->GetBoolean("Renderer", "use_asynchronous_shaders", false)); + Settings::values.use_fast_gpu_time.SetValue( + sdl2_config->GetBoolean("Renderer", "use_fast_gpu_time", true)); + + Settings::values.bg_red.SetValue( + static_cast<float>(sdl2_config->GetReal("Renderer", "bg_red", 0.0))); + Settings::values.bg_green.SetValue( + static_cast<float>(sdl2_config->GetReal("Renderer", "bg_green", 0.0))); + Settings::values.bg_blue.SetValue( + static_cast<float>(sdl2_config->GetReal("Renderer", "bg_blue", 0.0))); // Audio Settings::values.sink_id = sdl2_config->Get("Audio", "output_engine", "auto"); - Settings::values.enable_audio_stretching = - sdl2_config->GetBoolean("Audio", "enable_audio_stretching", true); + Settings::values.enable_audio_stretching.SetValue( + sdl2_config->GetBoolean("Audio", "enable_audio_stretching", true)); Settings::values.audio_device_id = sdl2_config->Get("Audio", "output_device", "auto"); - Settings::values.volume = static_cast<float>(sdl2_config->GetReal("Audio", "volume", 1)); - - Settings::values.language_index = sdl2_config->GetInteger("System", "language_index", 1); + Settings::values.volume.SetValue( + static_cast<float>(sdl2_config->GetReal("Audio", "volume", 1))); // Miscellaneous Settings::values.log_filter = sdl2_config->Get("Miscellaneous", "log_filter", "*:Trace"); @@ -425,6 +441,8 @@ void Config::ReadValues() { Settings::values.reporting_services = sdl2_config->GetBoolean("Debugging", "reporting_services", false); Settings::values.quest_flag = sdl2_config->GetBoolean("Debugging", "quest_flag", false); + Settings::values.disable_macro_jit = + sdl2_config->GetBoolean("Debugging", "disable_macro_jit", false); const auto title_list = sdl2_config->Get("AddOns", "title_ids", ""); std::stringstream ss(title_list); diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h index d63d7a58e..bcbbcd4ca 100644 --- a/src/yuzu_cmd/default_ini.h +++ b/src/yuzu_cmd/default_ini.h @@ -65,6 +65,14 @@ button_screenshot= lstick= rstick= +# Whether to enable or disable vibration +# 0: Disabled, 1 (default): Enabled +vibration_enabled= + +# Whether to enable or disable accurate vibrations +# 0 (default): Disabled, 1: Enabled +enable_accurate_vibrations= + # for motion input, the following devices are available: # - "motion_emu" (default) for emulating motion input from mouse input. Required parameters: # - "update_period": update period in milliseconds (default to 100) @@ -94,9 +102,42 @@ udp_pad_index= [Core] # Whether to use multi-core for CPU emulation -# 0 (default): Disabled, 1: Enabled +# 0: Disabled, 1 (default): Enabled use_multi_core= +[Cpu] +# Enable inline page tables optimization (faster guest memory access) +# 0: Disabled, 1 (default): Enabled +cpuopt_page_tables = + +# Enable block linking CPU optimization (reduce block dispatcher use during predictable jumps) +# 0: Disabled, 1 (default): Enabled +cpuopt_block_linking = + +# Enable return stack buffer CPU optimization (reduce block dispatcher use during predictable returns) +# 0: Disabled, 1 (default): Enabled +cpuopt_return_stack_buffer = + +# Enable fast dispatcher CPU optimization (use a two-tiered dispatcher architecture) +# 0: Disabled, 1 (default): Enabled +cpuopt_fast_dispatcher = + +# Enable context elimination CPU Optimization (reduce host memory use for guest context) +# 0: Disabled, 1 (default): Enabled +cpuopt_context_elimination = + +# Enable constant propagation CPU optimization (basic IR optimization) +# 0: Disabled, 1 (default): Enabled +cpuopt_const_prop = + +# Enable miscellaneous CPU optimizations (basic IR optimization) +# 0: Disabled, 1 (default): Enabled +cpuopt_misc_ir = + +# Enable reduction of memory misalignment checks (reduce memory fallbacks for misaligned access) +# 0: Disabled, 1 (default): Enabled +cpuopt_reduce_misalign_checks = + [Renderer] # Which backend API to use. # 0 (default): OpenGL, 1: Vulkan @@ -117,11 +158,6 @@ use_hw_renderer = # 0: Interpreter (slow), 1 (default): JIT (fast) use_shader_jit = -# Resolution scale factor -# 0: Auto (scales resolution to window size), 1: Native Switch screen resolution, Otherwise a scale -# factor for the Switch resolution -resolution_factor = - # Aspect ratio # 0: Default (16:9), 1: Force 4:3, 2: Force 21:9, 3: Stretch to Window aspect_ratio = @@ -134,6 +170,14 @@ max_anisotropy = # 0 (default): Off, 1: On use_vsync = +# Whether to use OpenGL assembly shaders or not. NV_gpu_program5 is required. +# 0: Off, 1 (default): On +use_assembly_shaders = + +# Whether to allow asynchronous shader building. +# 0 (default): Off, 1: On +use_asynchronous_shaders = + # Turns on the frame limiter, which will limit frames output to the target game speed # 0: Off, 1: On (default) use_frame_limit = @@ -146,9 +190,9 @@ frame_limit = # 0 (default): Off, 1 : On use_disk_shader_cache = -# Whether to use accurate GPU emulation -# 0 (default): Off (fast), 1 : On (slow) -use_accurate_gpu_emulation = +# Which gpu accuracy level to use +# 0 (Normal), 1 (High), 2 (Extreme) +gpu_accuracy = # Whether to use asynchronous GPU emulation # 0 : Off (slow), 1 (default): On (fast) @@ -262,6 +306,10 @@ language_index = # -1: Auto-select (default), 0: Japan, 1: USA, 2: Europe, 3: Australia, 4: China, 5: Korea, 6: Taiwan region_value = +# The system time zone that yuzu will use during emulation +# 0: Auto-select (default), 1: Default (system archive value), Others: Index for specified time zone +time_zone_index = + [Miscellaneous] # A filter which removes logs below a certain logging level. # Examples: *:Debug Kernel.SVC:Trace Service.*:Critical @@ -280,6 +328,8 @@ dump_nso=false # Determines whether or not yuzu will report to the game that the emulated console is in Kiosk Mode # false: Retail/Normal Mode (default), true: Kiosk Mode quest_flag = +# Enables/Disables the macro JIT compiler +disable_macro_jit=false [WebService] # Whether or not to enable telemetry diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp index 19584360c..521209622 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp @@ -13,23 +13,24 @@ #include "input_common/sdl/sdl.h" #include "yuzu_cmd/emu_window/emu_window_sdl2.h" -EmuWindow_SDL2::EmuWindow_SDL2(Core::System& system, bool fullscreen) : system{system} { +EmuWindow_SDL2::EmuWindow_SDL2(InputCommon::InputSubsystem* input_subsystem_) + : input_subsystem{input_subsystem_} { if (SDL_Init(SDL_INIT_VIDEO | SDL_INIT_JOYSTICK) < 0) { LOG_CRITICAL(Frontend, "Failed to initialize SDL2! Exiting..."); exit(1); } - InputCommon::Init(); + input_subsystem->Initialize(); SDL_SetMainReady(); } EmuWindow_SDL2::~EmuWindow_SDL2() { - InputCommon::Shutdown(); + input_subsystem->Shutdown(); SDL_Quit(); } void EmuWindow_SDL2::OnMouseMotion(s32 x, s32 y) { TouchMoved((unsigned)std::max(x, 0), (unsigned)std::max(y, 0)); - InputCommon::GetMotionEmu()->Tilt(x, y); + input_subsystem->GetMotionEmu()->Tilt(x, y); } void EmuWindow_SDL2::OnMouseButton(u32 button, u8 state, s32 x, s32 y) { @@ -41,9 +42,9 @@ void EmuWindow_SDL2::OnMouseButton(u32 button, u8 state, s32 x, s32 y) { } } else if (button == SDL_BUTTON_RIGHT) { if (state == SDL_PRESSED) { - InputCommon::GetMotionEmu()->BeginTilt(x, y); + input_subsystem->GetMotionEmu()->BeginTilt(x, y); } else { - InputCommon::GetMotionEmu()->EndTilt(); + input_subsystem->GetMotionEmu()->EndTilt(); } } } @@ -79,9 +80,9 @@ void EmuWindow_SDL2::OnFingerUp() { void EmuWindow_SDL2::OnKeyEvent(int key, u8 state) { if (state == SDL_PRESSED) { - InputCommon::GetKeyboard()->PressKey(key); + input_subsystem->GetKeyboard()->PressKey(key); } else if (state == SDL_RELEASED) { - InputCommon::GetKeyboard()->ReleaseKey(key); + input_subsystem->GetKeyboard()->ReleaseKey(key); } } @@ -181,9 +182,10 @@ void EmuWindow_SDL2::PollEvents() { const u32 current_time = SDL_GetTicks(); if (current_time > last_time + 2000) { const auto results = Core::System::GetInstance().GetAndResetPerfStats(); - const auto title = fmt::format( - "yuzu {} | {}-{} | FPS: {:.0f} ({:.0%})", Common::g_build_fullname, - Common::g_scm_branch, Common::g_scm_desc, results.game_fps, results.emulation_speed); + const auto title = + fmt::format("yuzu {} | {}-{} | FPS: {:.0f} ({:.0f}%)", Common::g_build_fullname, + Common::g_scm_branch, Common::g_scm_desc, results.game_fps, + results.emulation_speed * 100.0); SDL_SetWindowTitle(render_window, title.c_str()); last_time = current_time; } diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.h b/src/yuzu_cmd/emu_window/emu_window_sdl2.h index fffac4252..53d756c3c 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.h +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.h @@ -14,9 +14,13 @@ namespace Core { class System; } +namespace InputCommon { +class InputSubsystem; +} + class EmuWindow_SDL2 : public Core::Frontend::EmuWindow { public: - explicit EmuWindow_SDL2(Core::System& system, bool fullscreen); + explicit EmuWindow_SDL2(InputCommon::InputSubsystem* input_subsystem); ~EmuWindow_SDL2(); /// Polls window events @@ -28,9 +32,6 @@ public: /// Returns if window is shown (not minimized) bool IsShown() const override; - /// Presents the next frame - virtual void Present() = 0; - protected: /// Called by PollEvents when a key is pressed or released. void OnKeyEvent(int key, u8 state); @@ -62,9 +63,6 @@ protected: /// Called when a configuration change affects the minimal size of the window void OnMinimalClientAreaChangeRequest(std::pair<unsigned, unsigned> minimal_size) override; - /// Instance of the system, used to access renderer for the presentation thread - Core::System& system; - /// Is the window still open? bool is_open = true; @@ -76,4 +74,7 @@ protected: /// Keeps track of how often to update the title bar during gameplay u32 last_time = 0; + + /// Input subsystem to use with this window. + InputCommon::InputSubsystem* input_subsystem; }; diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp index 411e7e647..5f35233b5 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp @@ -87,8 +87,8 @@ bool EmuWindow_SDL2_GL::SupportsRequiredGLExtensions() { return unsupported_ext.empty(); } -EmuWindow_SDL2_GL::EmuWindow_SDL2_GL(Core::System& system, bool fullscreen) - : EmuWindow_SDL2{system, fullscreen} { +EmuWindow_SDL2_GL::EmuWindow_SDL2_GL(InputCommon::InputSubsystem* input_subsystem, bool fullscreen) + : EmuWindow_SDL2{input_subsystem} { SDL_GL_SetAttribute(SDL_GL_CONTEXT_MAJOR_VERSION, 4); SDL_GL_SetAttribute(SDL_GL_CONTEXT_MINOR_VERSION, 3); SDL_GL_SetAttribute(SDL_GL_CONTEXT_PROFILE_MASK, SDL_GL_CONTEXT_PROFILE_COMPATIBILITY); @@ -98,6 +98,9 @@ EmuWindow_SDL2_GL::EmuWindow_SDL2_GL(Core::System& system, bool fullscreen) SDL_GL_SetAttribute(SDL_GL_BLUE_SIZE, 8); SDL_GL_SetAttribute(SDL_GL_ALPHA_SIZE, 0); SDL_GL_SetAttribute(SDL_GL_SHARE_WITH_CURRENT_CONTEXT, 1); + if (Settings::values.renderer_debug) { + SDL_GL_SetAttribute(SDL_GL_CONTEXT_FLAGS, SDL_GL_CONTEXT_DEBUG_FLAG); + } SDL_GL_SetSwapInterval(0); std::string window_title = fmt::format("yuzu {} | {}-{}", Common::g_build_fullname, @@ -159,13 +162,3 @@ EmuWindow_SDL2_GL::~EmuWindow_SDL2_GL() { std::unique_ptr<Core::Frontend::GraphicsContext> EmuWindow_SDL2_GL::CreateSharedContext() const { return std::make_unique<SDLGLContext>(); } - -void EmuWindow_SDL2_GL::Present() { - SDL_GL_MakeCurrent(render_window, window_context); - SDL_GL_SetSwapInterval(Settings::values.use_vsync ? 1 : 0); - while (IsOpen()) { - system.Renderer().TryPresent(100); - SDL_GL_SwapWindow(render_window); - } - SDL_GL_MakeCurrent(render_window, nullptr); -} diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.h b/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.h index 48bb41683..dba5c293c 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.h +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.h @@ -8,13 +8,15 @@ #include "core/frontend/emu_window.h" #include "yuzu_cmd/emu_window/emu_window_sdl2.h" +namespace InputCommon { +class InputSubsystem; +} + class EmuWindow_SDL2_GL final : public EmuWindow_SDL2 { public: - explicit EmuWindow_SDL2_GL(Core::System& system, bool fullscreen); + explicit EmuWindow_SDL2_GL(InputCommon::InputSubsystem* input_subsystem, bool fullscreen); ~EmuWindow_SDL2_GL(); - void Present() override; - std::unique_ptr<Core::Frontend::GraphicsContext> CreateSharedContext() const override; private: diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.cpp index f2990910e..3ba657c00 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.cpp @@ -19,8 +19,8 @@ #include <SDL.h> #include <SDL_syswm.h> -EmuWindow_SDL2_VK::EmuWindow_SDL2_VK(Core::System& system, bool fullscreen) - : EmuWindow_SDL2{system, fullscreen} { +EmuWindow_SDL2_VK::EmuWindow_SDL2_VK(InputCommon::InputSubsystem* input_subsystem) + : EmuWindow_SDL2{input_subsystem} { const std::string window_title = fmt::format("yuzu {} | {}-{} (Vulkan)", Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc); render_window = @@ -29,6 +29,7 @@ EmuWindow_SDL2_VK::EmuWindow_SDL2_VK(Core::System& system, bool fullscreen) SDL_WINDOW_RESIZABLE | SDL_WINDOW_ALLOW_HIGHDPI); SDL_SysWMinfo wm; + SDL_VERSION(&wm.version); if (SDL_GetWindowWMInfo(render_window, &wm) == SDL_FALSE) { LOG_CRITICAL(Frontend, "Failed to get information from the window manager"); std::exit(EXIT_FAILURE); @@ -70,9 +71,5 @@ EmuWindow_SDL2_VK::EmuWindow_SDL2_VK(Core::System& system, bool fullscreen) EmuWindow_SDL2_VK::~EmuWindow_SDL2_VK() = default; std::unique_ptr<Core::Frontend::GraphicsContext> EmuWindow_SDL2_VK::CreateSharedContext() const { - return nullptr; -} - -void EmuWindow_SDL2_VK::Present() { - // TODO (bunnei): ImplementMe + return std::make_unique<DummyContext>(); } diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.h b/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.h index b8021ebea..bdfdc3c6f 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.h +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2_vk.h @@ -13,12 +13,16 @@ namespace Core { class System; } +namespace InputCommon { +class InputSubsystem; +} + class EmuWindow_SDL2_VK final : public EmuWindow_SDL2 { public: - explicit EmuWindow_SDL2_VK(Core::System& system, bool fullscreen); - ~EmuWindow_SDL2_VK(); - - void Present() override; + explicit EmuWindow_SDL2_VK(InputCommon::InputSubsystem* input_subsystem); + ~EmuWindow_SDL2_VK() override; std::unique_ptr<Core::Frontend::GraphicsContext> CreateSharedContext() const override; }; + +class DummyContext : public Core::Frontend::GraphicsContext {}; diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp index 4d2ea7e9e..14a23c71b 100644 --- a/src/yuzu_cmd/yuzu.cpp +++ b/src/yuzu_cmd/yuzu.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <chrono> #include <iostream> #include <memory> #include <string> @@ -22,12 +23,15 @@ #include "common/telemetry.h" #include "core/core.h" #include "core/crypto/key_manager.h" +#include "core/file_sys/registered_cache.h" #include "core/file_sys/vfs_real.h" #include "core/gdbstub/gdbstub.h" +#include "core/hle/kernel/process.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/loader.h" #include "core/settings.h" #include "core/telemetry_session.h" +#include "input_common/main.h" #include "video_core/renderer_base.h" #include "yuzu_cmd/config.h" #include "yuzu_cmd/emu_window/emu_window_sdl2.h" @@ -36,8 +40,6 @@ #include "yuzu_cmd/emu_window/emu_window_sdl2_vk.h" #endif -#include "core/file_sys/registered_cache.h" - #ifdef _WIN32 // windows.h needs to be included before shellapi.h #include <windows.h> @@ -81,8 +83,8 @@ static void InitializeLogging() { Log::AddBackend(std::make_unique<Log::ColorConsoleBackend>()); - const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir); - FileUtil::CreateFullPath(log_dir); + const std::string& log_dir = Common::FS::GetUserPath(Common::FS::UserPath::LogDir); + Common::FS::CreateFullPath(log_dir); Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE)); #ifdef _WIN32 Log::AddBackend(std::make_unique<Log::DebuggerBackend>()); @@ -178,15 +180,16 @@ int main(int argc, char** argv) { Settings::Apply(); Core::System& system{Core::System::GetInstance()}; + InputCommon::InputSubsystem input_subsystem; std::unique_ptr<EmuWindow_SDL2> emu_window; - switch (Settings::values.renderer_backend) { + switch (Settings::values.renderer_backend.GetValue()) { case Settings::RendererBackend::OpenGL: - emu_window = std::make_unique<EmuWindow_SDL2_GL>(system, fullscreen); + emu_window = std::make_unique<EmuWindow_SDL2_GL>(&input_subsystem, fullscreen); break; case Settings::RendererBackend::Vulkan: #ifdef HAS_VULKAN - emu_window = std::make_unique<EmuWindow_SDL2_VK>(system, fullscreen); + emu_window = std::make_unique<EmuWindow_SDL2_VK>(&input_subsystem); break; #else LOG_CRITICAL(Frontend, "Vulkan backend has not been compiled!"); @@ -228,19 +231,20 @@ int main(int argc, char** argv) { } } - system.TelemetrySession().AddField(Telemetry::FieldType::App, "Frontend", "SDL"); + system.TelemetrySession().AddField(Common::Telemetry::FieldType::App, "Frontend", "SDL"); // Core is loaded, start the GPU (makes the GPU contexts current to this thread) system.GPU().Start(); - system.Renderer().Rasterizer().LoadDiskResources(); + system.Renderer().Rasterizer().LoadDiskResources( + system.CurrentProcess()->GetTitleID(), false, + [](VideoCore::LoadCallbackStage, size_t value, size_t total) {}); - std::thread render_thread([&emu_window] { emu_window->Present(); }); + void(system.Run()); while (emu_window->IsOpen()) { - system.RunLoop(); + std::this_thread::sleep_for(std::chrono::milliseconds(1)); } - render_thread.join(); - + void(system.Pause()); system.Shutdown(); detached_tasks.WaitForAllTasks(); diff --git a/src/yuzu_cmd/yuzu.rc b/src/yuzu_cmd/yuzu.rc index 7de8ef3d9..0cde75e2f 100644 --- a/src/yuzu_cmd/yuzu.rc +++ b/src/yuzu_cmd/yuzu.rc @@ -14,4 +14,4 @@ YUZU_ICON ICON "../../dist/yuzu.ico" // RT_MANIFEST // -1 RT_MANIFEST "../../dist/yuzu.manifest" +0 RT_MANIFEST "../../dist/yuzu.manifest" diff --git a/src/yuzu_tester/CMakeLists.txt b/src/yuzu_tester/CMakeLists.txt index 06c2ee011..d8a2a1511 100644 --- a/src/yuzu_tester/CMakeLists.txt +++ b/src/yuzu_tester/CMakeLists.txt @@ -28,7 +28,5 @@ endif() if (MSVC) include(CopyYuzuSDLDeps) - include(CopyYuzuUnicornDeps) copy_yuzu_SDL_deps(yuzu-tester) - copy_yuzu_unicorn_deps(yuzu-tester) endif() diff --git a/src/yuzu_tester/config.cpp b/src/yuzu_tester/config.cpp index ee2591c8f..b6cdc7c1c 100644 --- a/src/yuzu_tester/config.cpp +++ b/src/yuzu_tester/config.cpp @@ -15,10 +15,11 @@ #include "yuzu_tester/config.h" #include "yuzu_tester/default_ini.h" +namespace FS = Common::FS; + Config::Config() { // TODO: Don't hardcode the path; let the frontend decide where to put the config files. - sdl2_config_loc = - FileUtil::GetUserPath(FileUtil::UserPath::ConfigDir) + "sdl2-tester-config.ini"; + sdl2_config_loc = FS::GetUserPath(FS::UserPath::ConfigDir) + "sdl2-tester-config.ini"; sdl2_config = std::make_unique<INIReader>(sdl2_config_loc); Reload(); @@ -31,8 +32,8 @@ bool Config::LoadINI(const std::string& default_contents, bool retry) { if (sdl2_config->ParseError() < 0) { if (retry) { LOG_WARNING(Config, "Failed to load {}. Creating file from defaults...", location); - FileUtil::CreateFullPath(location); - FileUtil::WriteStringToFile(true, default_contents, location); + FS::CreateFullPath(location); + FS::WriteStringToFile(true, default_contents, location); sdl2_config = std::make_unique<INIReader>(location); // Reopen file return LoadINI(default_contents, false); @@ -46,13 +47,13 @@ bool Config::LoadINI(const std::string& default_contents, bool retry) { void Config::ReadValues() { // Controls - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { - Settings::values.players[p].buttons[i] = ""; + Settings::values.players.GetValue()[p].buttons[i] = ""; } for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { - Settings::values.players[p].analogs[i] = ""; + Settings::values.players.GetValue()[p].analogs[i] = ""; } } @@ -74,6 +75,9 @@ void Config::ReadValues() { Settings::values.debug_pad_analogs[i] = ""; } + Settings::values.vibration_enabled.SetValue(true); + Settings::values.enable_accurate_vibrations.SetValue(false); + Settings::values.motion_enabled.SetValue(true); Settings::values.touchscreen.enabled = ""; Settings::values.touchscreen.device = ""; Settings::values.touchscreen.finger = 0; @@ -81,68 +85,73 @@ void Config::ReadValues() { Settings::values.touchscreen.diameter_x = 15; Settings::values.touchscreen.diameter_y = 15; + Settings::values.use_docked_mode.SetValue( + sdl2_config->GetBoolean("Controls", "use_docked_mode", false)); + // Data Storage Settings::values.use_virtual_sd = sdl2_config->GetBoolean("Data Storage", "use_virtual_sd", true); - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir, - sdl2_config->Get("Data Storage", "nand_directory", - FileUtil::GetUserPath(FileUtil::UserPath::NANDDir))); - FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir, - sdl2_config->Get("Data Storage", "sdmc_directory", - FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir))); + FS::GetUserPath(Common::FS::UserPath::NANDDir, + sdl2_config->Get("Data Storage", "nand_directory", + Common::FS::GetUserPath(Common::FS::UserPath::NANDDir))); + FS::GetUserPath(Common::FS::UserPath::SDMCDir, + sdl2_config->Get("Data Storage", "sdmc_directory", + Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir))); // System - Settings::values.use_docked_mode = sdl2_config->GetBoolean("System", "use_docked_mode", false); - Settings::values.current_user = std::clamp<int>( sdl2_config->GetInteger("System", "current_user", 0), 0, Service::Account::MAX_USERS - 1); const auto rng_seed_enabled = sdl2_config->GetBoolean("System", "rng_seed_enabled", false); if (rng_seed_enabled) { - Settings::values.rng_seed = sdl2_config->GetInteger("System", "rng_seed", 0); + Settings::values.rng_seed.SetValue(sdl2_config->GetInteger("System", "rng_seed", 0)); } else { - Settings::values.rng_seed = std::nullopt; + Settings::values.rng_seed.SetValue(std::nullopt); } const auto custom_rtc_enabled = sdl2_config->GetBoolean("System", "custom_rtc_enabled", false); if (custom_rtc_enabled) { - Settings::values.custom_rtc = - std::chrono::seconds(sdl2_config->GetInteger("System", "custom_rtc", 0)); + Settings::values.custom_rtc.SetValue( + std::chrono::seconds(sdl2_config->GetInteger("System", "custom_rtc", 0))); } else { - Settings::values.custom_rtc = std::nullopt; + Settings::values.custom_rtc.SetValue(std::nullopt); } // Core - Settings::values.use_multi_core = sdl2_config->GetBoolean("Core", "use_multi_core", false); + Settings::values.use_multi_core.SetValue( + sdl2_config->GetBoolean("Core", "use_multi_core", false)); // Renderer - Settings::values.resolution_factor = - static_cast<float>(sdl2_config->GetReal("Renderer", "resolution_factor", 1.0)); - Settings::values.aspect_ratio = - static_cast<int>(sdl2_config->GetInteger("Renderer", "aspect_ratio", 0)); - Settings::values.max_anisotropy = - static_cast<int>(sdl2_config->GetInteger("Renderer", "max_anisotropy", 0)); - Settings::values.use_frame_limit = false; - Settings::values.frame_limit = 100; - Settings::values.use_disk_shader_cache = - sdl2_config->GetBoolean("Renderer", "use_disk_shader_cache", false); - Settings::values.use_accurate_gpu_emulation = - sdl2_config->GetBoolean("Renderer", "use_accurate_gpu_emulation", false); - Settings::values.use_asynchronous_gpu_emulation = - sdl2_config->GetBoolean("Renderer", "use_asynchronous_gpu_emulation", false); - - Settings::values.bg_red = static_cast<float>(sdl2_config->GetReal("Renderer", "bg_red", 0.0)); - Settings::values.bg_green = - static_cast<float>(sdl2_config->GetReal("Renderer", "bg_green", 0.0)); - Settings::values.bg_blue = static_cast<float>(sdl2_config->GetReal("Renderer", "bg_blue", 0.0)); + Settings::values.aspect_ratio.SetValue( + static_cast<int>(sdl2_config->GetInteger("Renderer", "aspect_ratio", 0))); + Settings::values.max_anisotropy.SetValue( + static_cast<int>(sdl2_config->GetInteger("Renderer", "max_anisotropy", 0))); + Settings::values.use_frame_limit.SetValue(false); + Settings::values.frame_limit.SetValue(100); + Settings::values.use_disk_shader_cache.SetValue( + sdl2_config->GetBoolean("Renderer", "use_disk_shader_cache", false)); + const int gpu_accuracy_level = sdl2_config->GetInteger("Renderer", "gpu_accuracy", 0); + Settings::values.gpu_accuracy.SetValue(static_cast<Settings::GPUAccuracy>(gpu_accuracy_level)); + Settings::values.use_asynchronous_gpu_emulation.SetValue( + sdl2_config->GetBoolean("Renderer", "use_asynchronous_gpu_emulation", false)); + Settings::values.use_fast_gpu_time.SetValue( + sdl2_config->GetBoolean("Renderer", "use_fast_gpu_time", true)); + + Settings::values.bg_red.SetValue( + static_cast<float>(sdl2_config->GetReal("Renderer", "bg_red", 0.0))); + Settings::values.bg_green.SetValue( + static_cast<float>(sdl2_config->GetReal("Renderer", "bg_green", 0.0))); + Settings::values.bg_blue.SetValue( + static_cast<float>(sdl2_config->GetReal("Renderer", "bg_blue", 0.0))); // Audio Settings::values.sink_id = "null"; - Settings::values.enable_audio_stretching = false; + Settings::values.enable_audio_stretching.SetValue(false); Settings::values.audio_device_id = "auto"; - Settings::values.volume = 0; + Settings::values.volume.SetValue(0); - Settings::values.language_index = sdl2_config->GetInteger("System", "language_index", 1); + Settings::values.language_index.SetValue( + sdl2_config->GetInteger("System", "language_index", 1)); // Miscellaneous Settings::values.log_filter = sdl2_config->Get("Miscellaneous", "log_filter", "*:Trace"); diff --git a/src/yuzu_tester/default_ini.h b/src/yuzu_tester/default_ini.h index ca203b64d..3eb64e9d7 100644 --- a/src/yuzu_tester/default_ini.h +++ b/src/yuzu_tester/default_ini.h @@ -12,6 +12,39 @@ const char* sdl2_config_file = R"( # 0 (default): Disabled, 1: Enabled use_multi_core= +[Cpu] +# Enable inline page tables optimization (faster guest memory access) +# 0: Disabled, 1 (default): Enabled +cpuopt_page_tables = + +# Enable block linking CPU optimization (reduce block dispatcher use during predictable jumps) +# 0: Disabled, 1 (default): Enabled +cpuopt_block_linking = + +# Enable return stack buffer CPU optimization (reduce block dispatcher use during predictable returns) +# 0: Disabled, 1 (default): Enabled +cpuopt_return_stack_buffer = + +# Enable fast dispatcher CPU optimization (use a two-tiered dispatcher architecture) +# 0: Disabled, 1 (default): Enabled +cpuopt_fast_dispatcher = + +# Enable context elimination CPU Optimization (reduce host memory use for guest context) +# 0: Disabled, 1 (default): Enabled +cpuopt_context_elimination = + +# Enable constant propagation CPU optimization (basic IR optimization) +# 0: Disabled, 1 (default): Enabled +cpuopt_const_prop = + +# Enable miscellaneous CPU optimizations (basic IR optimization) +# 0: Disabled, 1 (default): Enabled +cpuopt_misc_ir = + +# Enable reduction of memory misalignment checks (reduce memory fallbacks for misaligned access) +# 0: Disabled, 1 (default): Enabled +cpuopt_reduce_misalign_checks = + [Renderer] # Whether to use software or hardware rendering. # 0: Software, 1 (default): Hardware @@ -21,11 +54,6 @@ use_hw_renderer = # 0: Interpreter (slow), 1 (default): JIT (fast) use_shader_jit = -# Resolution scale factor -# 0: Auto (scales resolution to window size), 1: Native Switch screen resolution, Otherwise a scale -# factor for the Switch resolution -resolution_factor = - # Aspect ratio # 0: Default (16:9), 1: Force 4:3, 2: Force 21:9, 3: Stretch to Window aspect_ratio = diff --git a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp index 8584f6671..78f75fb38 100644 --- a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp +++ b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp @@ -13,7 +13,6 @@ #include <glad/glad.h> -#include "common/assert.h" #include "common/logging/log.h" #include "common/scm_rev.h" #include "core/settings.h" @@ -53,7 +52,7 @@ EmuWindow_SDL2_Hide::EmuWindow_SDL2_Hide() { exit(1); } - InputCommon::Init(); + input_subsystem->Initialize(); SDL_SetMainReady(); @@ -105,7 +104,7 @@ EmuWindow_SDL2_Hide::EmuWindow_SDL2_Hide() { } EmuWindow_SDL2_Hide::~EmuWindow_SDL2_Hide() { - InputCommon::Shutdown(); + input_subsystem->Shutdown(); SDL_GL_DeleteContext(gl_context); SDL_Quit(); } diff --git a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h index c13a82df2..a553b4b95 100644 --- a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h +++ b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h @@ -8,6 +8,10 @@ struct SDL_Window; +namespace InputCommon { +class InputSubsystem; +} + class EmuWindow_SDL2_Hide : public Core::Frontend::EmuWindow { public: explicit EmuWindow_SDL2_Hide(); @@ -25,6 +29,8 @@ private: /// Whether the GPU and driver supports the OpenGL extension required bool SupportsRequiredGLExtensions(); + std::unique_ptr<InputCommon::InputSubsystem> input_subsystem; + /// Internal SDL2 render window SDL_Window* render_window; diff --git a/src/yuzu_tester/service/yuzutest.cpp b/src/yuzu_tester/service/yuzutest.cpp index 85d3f436b..2d3f6e3a7 100644 --- a/src/yuzu_tester/service/yuzutest.cpp +++ b/src/yuzu_tester/service/yuzutest.cpp @@ -53,7 +53,7 @@ private: IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(write_size); + rb.Push<u32>(static_cast<u32>(write_size)); } void StartIndividual(Kernel::HLERequestContext& ctx) { diff --git a/src/yuzu_tester/yuzu.cpp b/src/yuzu_tester/yuzu.cpp index 676e70ebd..88e4bd1f7 100644 --- a/src/yuzu_tester/yuzu.cpp +++ b/src/yuzu_tester/yuzu.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <chrono> #include <iostream> #include <memory> #include <string> @@ -78,8 +79,8 @@ static void InitializeLogging(bool console) { if (console) Log::AddBackend(std::make_unique<Log::ColorConsoleBackend>()); - const std::string& log_dir = FileUtil::GetUserPath(FileUtil::UserPath::LogDir); - FileUtil::CreateFullPath(log_dir); + const std::string& log_dir = Common::FS::GetUserPath(Common::FS::UserPath::LogDir); + Common::FS::CreateFullPath(log_dir); Log::AddBackend(std::make_unique<Log::FileBackend>(log_dir + LOG_FILE)); #ifdef _WIN32 Log::AddBackend(std::make_unique<Log::DebuggerBackend>()); @@ -250,14 +251,16 @@ int main(int argc, char** argv) { Service::Yuzu::InstallInterfaces(system.ServiceManager(), datastring, callback); - system.TelemetrySession().AddField(Telemetry::FieldType::App, "Frontend", "SDLHideTester"); + system.TelemetrySession().AddField(Common::Telemetry::FieldType::App, "Frontend", + "SDLHideTester"); system.GPU().Start(); - system.Renderer().Rasterizer().LoadDiskResources(); + void(system.Run()); while (!finished) { - system.RunLoop(); + std::this_thread::sleep_for(std::chrono::milliseconds(1)); } + void(system.Pause()); detached_tasks.WaitForAllTasks(); return return_value; diff --git a/src/yuzu_tester/yuzu.rc b/src/yuzu_tester/yuzu.rc index 7de8ef3d9..0cde75e2f 100644 --- a/src/yuzu_tester/yuzu.rc +++ b/src/yuzu_tester/yuzu.rc @@ -14,4 +14,4 @@ YUZU_ICON ICON "../../dist/yuzu.ico" // RT_MANIFEST // -1 RT_MANIFEST "../../dist/yuzu.manifest" +0 RT_MANIFEST "../../dist/yuzu.manifest" |