diff options
68 files changed, 1244 insertions, 348 deletions
diff --git a/src/audio_core/CMakeLists.txt b/src/audio_core/CMakeLists.txt index c08ce69e0..5a2747e78 100644 --- a/src/audio_core/CMakeLists.txt +++ b/src/audio_core/CMakeLists.txt @@ -4,6 +4,8 @@ set(SRCS hle/dsp.cpp hle/filter.cpp hle/pipe.cpp + interpolate.cpp + sink_details.cpp ) set(HEADERS @@ -13,7 +15,10 @@ set(HEADERS hle/dsp.h hle/filter.h hle/pipe.h + interpolate.h + null_sink.h sink.h + sink_details.h ) include_directories(../../externals/soundtouch/include) diff --git a/src/audio_core/audio_core.cpp b/src/audio_core/audio_core.cpp index b512b0f9b..d42249ebd 100644 --- a/src/audio_core/audio_core.cpp +++ b/src/audio_core/audio_core.cpp @@ -2,9 +2,15 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <memory> +#include <string> + #include "audio_core/audio_core.h" #include "audio_core/hle/dsp.h" #include "audio_core/hle/pipe.h" +#include "audio_core/null_sink.h" +#include "audio_core/sink.h" +#include "audio_core/sink_details.h" #include "core/core_timing.h" #include "core/hle/kernel/vm_manager.h" @@ -28,7 +34,6 @@ static void AudioTickCallback(u64 /*userdata*/, int cycles_late) { CoreTiming::ScheduleEvent(audio_frame_ticks - cycles_late, tick_event); } -/// Initialise Audio void Init() { DSP::HLE::Init(); @@ -36,19 +41,39 @@ void Init() { CoreTiming::ScheduleEvent(audio_frame_ticks, tick_event); } -/// Add DSP address spaces to Process's address space. void AddAddressSpace(Kernel::VMManager& address_space) { - auto r0_vma = address_space.MapBackingMemory(DSP::HLE::region0_base, reinterpret_cast<u8*>(&DSP::HLE::g_region0), sizeof(DSP::HLE::SharedMemory), Kernel::MemoryState::IO).MoveFrom(); + auto r0_vma = address_space.MapBackingMemory(DSP::HLE::region0_base, reinterpret_cast<u8*>(&DSP::HLE::g_regions[0]), sizeof(DSP::HLE::SharedMemory), Kernel::MemoryState::IO).MoveFrom(); address_space.Reprotect(r0_vma, Kernel::VMAPermission::ReadWrite); - auto r1_vma = address_space.MapBackingMemory(DSP::HLE::region1_base, reinterpret_cast<u8*>(&DSP::HLE::g_region1), sizeof(DSP::HLE::SharedMemory), Kernel::MemoryState::IO).MoveFrom(); + auto r1_vma = address_space.MapBackingMemory(DSP::HLE::region1_base, reinterpret_cast<u8*>(&DSP::HLE::g_regions[1]), sizeof(DSP::HLE::SharedMemory), Kernel::MemoryState::IO).MoveFrom(); address_space.Reprotect(r1_vma, Kernel::VMAPermission::ReadWrite); } -/// Shutdown Audio +void SelectSink(std::string sink_id) { + if (sink_id == "auto") { + // Auto-select. + // g_sink_details is ordered in terms of desirability, with the best choice at the front. + const auto& sink_detail = g_sink_details.front(); + DSP::HLE::SetSink(sink_detail.factory()); + return; + } + + auto iter = std::find_if(g_sink_details.begin(), g_sink_details.end(), [sink_id](const auto& sink_detail) { + return sink_detail.id == sink_id; + }); + + if (iter == g_sink_details.end()) { + LOG_ERROR(Audio, "AudioCore::SelectSink given invalid sink_id"); + DSP::HLE::SetSink(std::make_unique<NullSink>()); + return; + } + + DSP::HLE::SetSink(iter->factory()); +} + void Shutdown() { CoreTiming::UnscheduleEvent(tick_event, 0); DSP::HLE::Shutdown(); } -} //namespace +} // namespace AudioCore diff --git a/src/audio_core/audio_core.h b/src/audio_core/audio_core.h index b349895ea..f618361f3 100644 --- a/src/audio_core/audio_core.h +++ b/src/audio_core/audio_core.h @@ -4,6 +4,8 @@ #pragma once +#include <string> + namespace Kernel { class VMManager; } @@ -18,6 +20,9 @@ void Init(); /// Add DSP address spaces to a Process. void AddAddressSpace(Kernel::VMManager& vm_manager); +/// Select the sink to use based on sink id. +void SelectSink(std::string sink_id); + /// Shutdown Audio Core void Shutdown(); diff --git a/src/audio_core/hle/dsp.cpp b/src/audio_core/hle/dsp.cpp index c89356edc..4d44bd2d9 100644 --- a/src/audio_core/hle/dsp.cpp +++ b/src/audio_core/hle/dsp.cpp @@ -2,14 +2,43 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <memory> + #include "audio_core/hle/dsp.h" #include "audio_core/hle/pipe.h" +#include "audio_core/sink.h" namespace DSP { namespace HLE { -SharedMemory g_region0; -SharedMemory g_region1; +std::array<SharedMemory, 2> g_regions; + +static size_t CurrentRegionIndex() { + // The region with the higher frame counter is chosen unless there is wraparound. + // This function only returns a 0 or 1. + + if (g_regions[0].frame_counter == 0xFFFFu && g_regions[1].frame_counter != 0xFFFEu) { + // Wraparound has occured. + return 1; + } + + if (g_regions[1].frame_counter == 0xFFFFu && g_regions[0].frame_counter != 0xFFFEu) { + // Wraparound has occured. + return 0; + } + + return (g_regions[0].frame_counter > g_regions[1].frame_counter) ? 0 : 1; +} + +static SharedMemory& ReadRegion() { + return g_regions[CurrentRegionIndex()]; +} + +static SharedMemory& WriteRegion() { + return g_regions[1 - CurrentRegionIndex()]; +} + +static std::unique_ptr<AudioCore::Sink> sink; void Init() { DSP::HLE::ResetPipes(); @@ -22,20 +51,8 @@ bool Tick() { return true; } -SharedMemory& CurrentRegion() { - // The region with the higher frame counter is chosen unless there is wraparound. - - if (g_region0.frame_counter == 0xFFFFu && g_region1.frame_counter != 0xFFFEu) { - // Wraparound has occured. - return g_region1; - } - - if (g_region1.frame_counter == 0xFFFFu && g_region0.frame_counter != 0xFFFEu) { - // Wraparound has occured. - return g_region0; - } - - return (g_region0.frame_counter > g_region1.frame_counter) ? g_region0 : g_region1; +void SetSink(std::unique_ptr<AudioCore::Sink> sink_) { + sink = std::move(sink_); } } // namespace HLE diff --git a/src/audio_core/hle/dsp.h b/src/audio_core/hle/dsp.h index c76350bdd..4f2410c27 100644 --- a/src/audio_core/hle/dsp.h +++ b/src/audio_core/hle/dsp.h @@ -4,7 +4,9 @@ #pragma once +#include <array> #include <cstddef> +#include <memory> #include <type_traits> #include "audio_core/hle/common.h" @@ -14,6 +16,10 @@ #include "common/common_types.h" #include "common/swap.h" +namespace AudioCore { +class Sink; +} + namespace DSP { namespace HLE { @@ -30,10 +36,9 @@ namespace HLE { struct SharedMemory; constexpr VAddr region0_base = 0x1FF50000; -extern SharedMemory g_region0; - constexpr VAddr region1_base = 0x1FF70000; -extern SharedMemory g_region1; + +extern std::array<SharedMemory, 2> g_regions; /** * The DSP is native 16-bit. The DSP also appears to be big-endian. When reading 32-bit numbers from @@ -535,8 +540,11 @@ void Shutdown(); */ bool Tick(); -/// Returns a mutable reference to the current region. Current region is selected based on the frame counter. -SharedMemory& CurrentRegion(); +/** + * Set the output sink. This must be called before calling Tick(). + * @param sink The sink to which audio will be output to. + */ +void SetSink(std::unique_ptr<AudioCore::Sink> sink); } // namespace HLE } // namespace DSP diff --git a/src/audio_core/interpolate.cpp b/src/audio_core/interpolate.cpp new file mode 100644 index 000000000..fcd3aa066 --- /dev/null +++ b/src/audio_core/interpolate.cpp @@ -0,0 +1,85 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "audio_core/interpolate.h" + +#include "common/assert.h" +#include "common/math_util.h" + +namespace AudioInterp { + +// Calculations are done in fixed point with 24 fractional bits. +// (This is not verified. This was chosen for minimal error.) +constexpr u64 scale_factor = 1 << 24; +constexpr u64 scale_mask = scale_factor - 1; + +/// Here we step over the input in steps of rate_multiplier, until we consume all of the input. +/// Three adjacent samples are passed to fn each step. +template <typename Function> +static StereoBuffer16 StepOverSamples(State& state, const StereoBuffer16& input, float rate_multiplier, Function fn) { + ASSERT(rate_multiplier > 0); + + if (input.size() < 2) + return {}; + + StereoBuffer16 output; + output.reserve(static_cast<size_t>(input.size() / rate_multiplier)); + + u64 step_size = static_cast<u64>(rate_multiplier * scale_factor); + + u64 fposition = 0; + const u64 max_fposition = input.size() * scale_factor; + + while (fposition < 1 * scale_factor) { + u64 fraction = fposition & scale_mask; + + output.push_back(fn(fraction, state.xn2, state.xn1, input[0])); + + fposition += step_size; + } + + while (fposition < 2 * scale_factor) { + u64 fraction = fposition & scale_mask; + + output.push_back(fn(fraction, state.xn1, input[0], input[1])); + + fposition += step_size; + } + + while (fposition < max_fposition) { + u64 fraction = fposition & scale_mask; + + size_t index = static_cast<size_t>(fposition / scale_factor); + output.push_back(fn(fraction, input[index - 2], input[index - 1], input[index])); + + fposition += step_size; + } + + state.xn2 = input[input.size() - 2]; + state.xn1 = input[input.size() - 1]; + + return output; +} + +StereoBuffer16 None(State& state, const StereoBuffer16& input, float rate_multiplier) { + return StepOverSamples(state, input, rate_multiplier, [](u64 fraction, const auto& x0, const auto& x1, const auto& x2) { + return x0; + }); +} + +StereoBuffer16 Linear(State& state, const StereoBuffer16& input, float rate_multiplier) { + // Note on accuracy: Some values that this produces are +/- 1 from the actual firmware. + return StepOverSamples(state, input, rate_multiplier, [](u64 fraction, const auto& x0, const auto& x1, const auto& x2) { + // This is a saturated subtraction. (Verified by black-box fuzzing.) + s64 delta0 = MathUtil::Clamp<s64>(x1[0] - x0[0], -32768, 32767); + s64 delta1 = MathUtil::Clamp<s64>(x1[1] - x0[1], -32768, 32767); + + return std::array<s16, 2> { + static_cast<s16>(x0[0] + fraction * delta0 / scale_factor), + static_cast<s16>(x0[1] + fraction * delta1 / scale_factor) + }; + }); +} + +} // namespace AudioInterp diff --git a/src/audio_core/interpolate.h b/src/audio_core/interpolate.h new file mode 100644 index 000000000..a4c0a453d --- /dev/null +++ b/src/audio_core/interpolate.h @@ -0,0 +1,41 @@ +// Copyright 2016 Citra 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" + +namespace AudioInterp { + +/// A variable length buffer of signed PCM16 stereo samples. +using StereoBuffer16 = std::vector<std::array<s16, 2>>; + +struct State { + // Two historical samples. + std::array<s16, 2> xn1 = {}; ///< x[n-1] + std::array<s16, 2> xn2 = {}; ///< x[n-2] +}; + +/** + * No interpolation. This is equivalent to a zero-order hold. There is a two-sample predelay. + * @param input Input buffer. + * @param rate_multiplier Stretch factor. Must be a positive non-zero value. + * rate_multiplier > 1.0 performs decimation and rate_multipler < 1.0 performs upsampling. + * @return The resampled audio buffer. + */ +StereoBuffer16 None(State& state, const StereoBuffer16& input, float rate_multiplier); + +/** + * Linear interpolation. This is equivalent to a first-order hold. There is a two-sample predelay. + * @param input Input buffer. + * @param rate_multiplier Stretch factor. Must be a positive non-zero value. + * rate_multiplier > 1.0 performs decimation and rate_multipler < 1.0 performs upsampling. + * @return The resampled audio buffer. + */ +StereoBuffer16 Linear(State& state, const StereoBuffer16& input, float rate_multiplier); + +} // namespace AudioInterp diff --git a/src/audio_core/null_sink.h b/src/audio_core/null_sink.h new file mode 100644 index 000000000..faf0ee4e1 --- /dev/null +++ b/src/audio_core/null_sink.h @@ -0,0 +1,29 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstddef> + +#include "audio_core/audio_core.h" +#include "audio_core/sink.h" + +namespace AudioCore { + +class NullSink final : public Sink { +public: + ~NullSink() override = default; + + unsigned int GetNativeSampleRate() const override { + return native_sample_rate; + } + + void EnqueueSamples(const std::vector<s16>&) override {} + + size_t SamplesInQueue() const override { + return 0; + } +}; + +} // namespace AudioCore diff --git a/src/audio_core/sink_details.cpp b/src/audio_core/sink_details.cpp new file mode 100644 index 000000000..d2cc74103 --- /dev/null +++ b/src/audio_core/sink_details.cpp @@ -0,0 +1,18 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <memory> +#include <vector> + +#include "audio_core/null_sink.h" +#include "audio_core/sink_details.h" + +namespace AudioCore { + +// g_sink_details is ordered in terms of desirability, with the best choice at the top. +const std::vector<SinkDetails> g_sink_details = { + { "null", []() { return std::make_unique<NullSink>(); } }, +}; + +} // namespace AudioCore diff --git a/src/audio_core/sink_details.h b/src/audio_core/sink_details.h new file mode 100644 index 000000000..4b30cf835 --- /dev/null +++ b/src/audio_core/sink_details.h @@ -0,0 +1,27 @@ +// Copyright 2016 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <functional> +#include <memory> +#include <vector> + +namespace AudioCore { + +class Sink; + +struct SinkDetails { + SinkDetails(const char* id_, std::function<std::unique_ptr<Sink>()> factory_) + : id(id_), factory(factory_) {} + + /// Name for this sink. + const char* id; + /// A method to call to construct an instance of this type of sink. + std::function<std::unique_ptr<Sink>()> factory; +}; + +extern const std::vector<SinkDetails> g_sink_details; + +} // namespace AudioCore diff --git a/src/citra/config.cpp b/src/citra/config.cpp index 9e2ecd307..0d17c80bf 100644 --- a/src/citra/config.cpp +++ b/src/citra/config.cpp @@ -71,6 +71,9 @@ void Config::ReadValues() { Settings::values.bg_green = (float)sdl2_config->GetReal("Renderer", "bg_green", 1.0); Settings::values.bg_blue = (float)sdl2_config->GetReal("Renderer", "bg_blue", 1.0); + // Audio + Settings::values.sink_id = sdl2_config->Get("Audio", "output_engine", "auto"); + // Data Storage Settings::values.use_virtual_sd = sdl2_config->GetBoolean("Data Storage", "use_virtual_sd", true); diff --git a/src/citra/default_ini.h b/src/citra/default_ini.h index 1f1aa716b..0e6171736 100644 --- a/src/citra/default_ini.h +++ b/src/citra/default_ini.h @@ -56,6 +56,11 @@ bg_red = bg_blue = bg_green = +[Audio] +# Which audio output engine to use. +# auto (default): Auto-select, null: No audio output +output_engine = + [Data Storage] # Whether to create a virtual SD card. # 1 (default): Yes, 0: No diff --git a/src/citra_qt/config.cpp b/src/citra_qt/config.cpp index 7dc61fe40..b5bb75537 100644 --- a/src/citra_qt/config.cpp +++ b/src/citra_qt/config.cpp @@ -52,6 +52,10 @@ void Config::ReadValues() { Settings::values.bg_blue = qt_config->value("bg_blue", 1.0).toFloat(); qt_config->endGroup(); + qt_config->beginGroup("Audio"); + Settings::values.sink_id = qt_config->value("output_engine", "auto").toString().toStdString(); + qt_config->endGroup(); + qt_config->beginGroup("Data Storage"); Settings::values.use_virtual_sd = qt_config->value("use_virtual_sd", true).toBool(); qt_config->endGroup(); @@ -138,6 +142,10 @@ void Config::SaveValues() { qt_config->setValue("bg_blue", (double)Settings::values.bg_blue); qt_config->endGroup(); + qt_config->beginGroup("Audio"); + qt_config->setValue("output_engine", QString::fromStdString(Settings::values.sink_id)); + qt_config->endGroup(); + qt_config->beginGroup("Data Storage"); qt_config->setValue("use_virtual_sd", Settings::values.use_virtual_sd); qt_config->endGroup(); diff --git a/src/citra_qt/debugger/graphics_tracing.cpp b/src/citra_qt/debugger/graphics_tracing.cpp index e06498744..1402f8e79 100644 --- a/src/citra_qt/debugger/graphics_tracing.cpp +++ b/src/citra_qt/debugger/graphics_tracing.cpp @@ -2,6 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> +#include <array> +#include <iterator> #include <memory> #include <boost/range/algorithm/copy.hpp> @@ -18,6 +21,7 @@ #include "core/hw/gpu.h" #include "core/hw/lcd.h" +#include "core/tracer/recorder.h" #include "nihstro/float24.h" diff --git a/src/common/bit_field.h b/src/common/bit_field.h index 371eb17a1..4748999ed 100644 --- a/src/common/bit_field.h +++ b/src/common/bit_field.h @@ -186,5 +186,5 @@ private: #pragma pack() #if (__GNUC__ >= 5) || defined(__clang__) || defined(_MSC_VER) -static_assert(std::is_trivially_copyable<BitField<0, 1, u32>>::value, "BitField must be trivially copyable"); +static_assert(std::is_trivially_copyable<BitField<0, 1, unsigned>>::value, "BitField must be trivially copyable"); #endif diff --git a/src/common/bit_set.h b/src/common/bit_set.h index 85f91e786..7f5de8df2 100644 --- a/src/common/bit_set.h +++ b/src/common/bit_set.h @@ -7,6 +7,7 @@ #include <intrin.h> #endif #include <initializer_list> +#include <new> #include <type_traits> #include "common/common_types.h" @@ -186,4 +187,4 @@ public: typedef Common::BitSet<u8> BitSet8; typedef Common::BitSet<u16> BitSet16; typedef Common::BitSet<u32> BitSet32; -typedef Common::BitSet<u64> BitSet64;
\ No newline at end of file +typedef Common::BitSet<u64> BitSet64; diff --git a/src/common/code_block.h b/src/common/code_block.h index 9ef7296d3..2fa4a0090 100644 --- a/src/common/code_block.h +++ b/src/common/code_block.h @@ -4,8 +4,10 @@ #pragma once -#include "common_types.h" -#include "memory_util.h" +#include <cstddef> + +#include "common/common_types.h" +#include "common/memory_util.h" // Everything that needs to generate code should inherit from this. // You get memory management for free, plus, you can use all emitter functions without diff --git a/src/common/common_funcs.h b/src/common/common_funcs.h index aa6aff7b9..ab3515683 100644 --- a/src/common/common_funcs.h +++ b/src/common/common_funcs.h @@ -4,6 +4,10 @@ #pragma once +#if !defined(ARCHITECTURE_x86_64) && !defined(_M_ARM) +#include <cstdlib> // for exit +#endif + #include "common_types.h" #define ARRAY_SIZE(a) (sizeof(a) / sizeof(a[0])) diff --git a/src/common/file_util.h b/src/common/file_util.h index 3aac4fa46..c6a8694ce 100644 --- a/src/common/file_util.h +++ b/src/common/file_util.h @@ -7,9 +7,9 @@ #include <array> #include <fstream> #include <functional> -#include <cstddef> #include <cstdio> #include <string> +#include <type_traits> #include <vector> #include "common/common_types.h" diff --git a/src/common/x64/emitter.h b/src/common/x64/emitter.h index a33724146..60a77dfe1 100644 --- a/src/common/x64/emitter.h +++ b/src/common/x64/emitter.h @@ -17,6 +17,8 @@ #pragma once +#include <cstddef> + #include "common/assert.h" #include "common/bit_set.h" #include "common/common_types.h" diff --git a/src/core/hle/result.h b/src/core/hle/result.h index 53931a106..3fc1ab4ee 100644 --- a/src/core/hle/result.h +++ b/src/core/hle/result.h @@ -5,7 +5,6 @@ #pragma once #include <new> -#include <type_traits> #include <utility> #include "common/assert.h" diff --git a/src/core/hle/service/ac_u.cpp b/src/core/hle/service/ac_u.cpp index d67325506..5241dd3e7 100644 --- a/src/core/hle/service/ac_u.cpp +++ b/src/core/hle/service/ac_u.cpp @@ -3,6 +3,8 @@ // Refer to the license.txt file included. #include "common/logging/log.h" + +#include "core/hle/kernel/event.h" #include "core/hle/service/ac_u.h" //////////////////////////////////////////////////////////////////////////////////////////////////// @@ -11,6 +13,28 @@ namespace AC_U { /** + * AC_U::CloseAsync service function + * Inputs: + * 1 : Always 0x20 + * 3 : Always 0 + * 4 : Event handle, should be signaled when AC connection is closed + * Outputs: + * 1 : Result of function, 0 on success, otherwise error code + */ +static void CloseAsync(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + auto evt = Kernel::g_handle_table.Get<Kernel::Event>(cmd_buff[4]); + + if (evt) { + evt->name = "AC_U:close_event"; + evt->Signal(); + } + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + + LOG_WARNING(Service_AC, "(STUBBED) called"); +} +/** * AC_U::GetWifiStatus service function * Outputs: * 1 : Result of function, 0 on success, otherwise error code @@ -47,7 +71,7 @@ const Interface::FunctionInfo FunctionTable[] = { {0x00010000, nullptr, "CreateDefaultConfig"}, {0x00040006, nullptr, "ConnectAsync"}, {0x00050002, nullptr, "GetConnectResult"}, - {0x00080004, nullptr, "CloseAsync"}, + {0x00080004, CloseAsync, "CloseAsync"}, {0x00090002, nullptr, "GetCloseResult"}, {0x000A0000, nullptr, "GetLastErrorCode"}, {0x000D0000, GetWifiStatus, "GetWifiStatus"}, diff --git a/src/core/hle/service/gsp_gpu.cpp b/src/core/hle/service/gsp_gpu.cpp index 233592d7f..b4c146e08 100644 --- a/src/core/hle/service/gsp_gpu.cpp +++ b/src/core/hle/service/gsp_gpu.cpp @@ -43,6 +43,8 @@ Kernel::SharedPtr<Kernel::SharedMemory> g_shared_memory; /// Thread index into interrupt relay queue u32 g_thread_id = 0; +static bool gpu_right_acquired = false; + /// Gets a pointer to a thread command buffer in GSP shared memory static inline u8* GetCommandBuffer(u32 thread_id) { return g_shared_memory->GetPointer(0x800 + (thread_id * sizeof(CommandBuffer))); @@ -370,6 +372,9 @@ static void UnregisterInterruptRelayQueue(Service::Interface* self) { * @todo This probably does not belong in the GSP module, instead move to video_core */ void SignalInterrupt(InterruptId interrupt_id) { + if (!gpu_right_acquired) { + return; + } if (nullptr == g_interrupt_event) { LOG_WARNING(Service_GSP, "cannot synchronize until GSP event has been created!"); return; @@ -624,6 +629,35 @@ static void ImportDisplayCaptureInfo(Service::Interface* self) { LOG_WARNING(Service_GSP, "called"); } +/** + * GSP_GPU::AcquireRight service function + * Outputs: + * 1: Result code + */ +static void AcquireRight(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + gpu_right_acquired = true; + + cmd_buff[1] = RESULT_SUCCESS.raw; + + LOG_WARNING(Service_GSP, "called"); +} + +/** + * GSP_GPU::ReleaseRight service function + * Outputs: + * 1: Result code + */ +static void ReleaseRight(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + gpu_right_acquired = false; + + cmd_buff[1] = RESULT_SUCCESS.raw; + + LOG_WARNING(Service_GSP, "called"); +} const Interface::FunctionInfo FunctionTable[] = { {0x00010082, WriteHWRegs, "WriteHWRegs"}, @@ -647,8 +681,8 @@ const Interface::FunctionInfo FunctionTable[] = { {0x00130042, RegisterInterruptRelayQueue, "RegisterInterruptRelayQueue"}, {0x00140000, UnregisterInterruptRelayQueue, "UnregisterInterruptRelayQueue"}, {0x00150002, nullptr, "TryAcquireRight"}, - {0x00160042, nullptr, "AcquireRight"}, - {0x00170000, nullptr, "ReleaseRight"}, + {0x00160042, AcquireRight, "AcquireRight"}, + {0x00170000, ReleaseRight, "ReleaseRight"}, {0x00180000, ImportDisplayCaptureInfo, "ImportDisplayCaptureInfo"}, {0x00190000, nullptr, "SaveVramSysArea"}, {0x001A0000, nullptr, "RestoreVramSysArea"}, @@ -669,11 +703,13 @@ Interface::Interface() { g_shared_memory = nullptr; g_thread_id = 0; + gpu_right_acquired = false; } Interface::~Interface() { g_interrupt_event = nullptr; g_shared_memory = nullptr; + gpu_right_acquired = false; } } // namespace diff --git a/src/core/hle/service/gsp_gpu.h b/src/core/hle/service/gsp_gpu.h index 55a993bb8..3b4b678a3 100644 --- a/src/core/hle/service/gsp_gpu.h +++ b/src/core/hle/service/gsp_gpu.h @@ -10,6 +10,7 @@ #include "common/bit_field.h" #include "common/common_types.h" +#include "core/hle/result.h" #include "core/hle/service/service.h" //////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/src/core/hle/service/ndm/ndm.cpp b/src/core/hle/service/ndm/ndm.cpp index 47076a7b8..bc9c3413d 100644 --- a/src/core/hle/service/ndm/ndm.cpp +++ b/src/core/hle/service/ndm/ndm.cpp @@ -11,28 +11,217 @@ namespace Service { namespace NDM { -void SuspendDaemons(Service::Interface* self) { +enum : u32 { + DEFAULT_RETRY_INTERVAL = 10, + DEFAULT_SCAN_INTERVAL = 30 +}; + +static DaemonMask daemon_bit_mask = DaemonMask::Default; +static DaemonMask default_daemon_bit_mask = DaemonMask::Default; +static std::array<DaemonStatus, 4> daemon_status = { DaemonStatus::Idle, DaemonStatus::Idle, DaemonStatus::Idle, DaemonStatus::Idle }; +static ExclusiveState exclusive_state = ExclusiveState::None; +static u32 scan_interval = DEFAULT_SCAN_INTERVAL; +static u32 retry_interval = DEFAULT_RETRY_INTERVAL; +static bool daemon_lock_enabled = false; + +void EnterExclusiveState(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + exclusive_state = static_cast<ExclusiveState>(cmd_buff[1]); + + cmd_buff[0] = IPC::MakeHeader(0x1, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) exclusive_state=0x%08X ", exclusive_state); +} + +void LeaveExclusiveState(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + exclusive_state = ExclusiveState::None; + + cmd_buff[0] = IPC::MakeHeader(0x2, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) exclusive_state=0x%08X ", exclusive_state); +} + +void QueryExclusiveMode(Service::Interface* self) { u32* cmd_buff = Kernel::GetCommandBuffer(); - LOG_WARNING(Service_NDM, "(STUBBED) bit_mask=0x%08X ", cmd_buff[1]); + cmd_buff[0] = IPC::MakeHeader(0x3, 2, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = static_cast<u32>(exclusive_state); + LOG_WARNING(Service_NDM, "(STUBBED) exclusive_state=0x%08X ", exclusive_state); +} + +void LockState(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + daemon_lock_enabled = true; + + cmd_buff[0] = IPC::MakeHeader(0x4, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) daemon_lock_enabled=0x%08X ", daemon_lock_enabled); +} + +void UnlockState(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + daemon_lock_enabled = false; + cmd_buff[0] = IPC::MakeHeader(0x5, 1, 0); cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) daemon_lock_enabled=0x%08X ", daemon_lock_enabled); +} + +void SuspendDaemons(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + u32 bit_mask = cmd_buff[1] & 0xF; + daemon_bit_mask = static_cast<DaemonMask>(static_cast<u32>(default_daemon_bit_mask) & ~bit_mask); + for (size_t index = 0; index < daemon_status.size(); ++index) { + if (bit_mask & (1 << index)) { + daemon_status[index] = DaemonStatus::Suspended; + } + } + + cmd_buff[0] = IPC::MakeHeader(0x6, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) daemon_bit_mask=0x%08X ", daemon_bit_mask); } void ResumeDaemons(Service::Interface* self) { u32* cmd_buff = Kernel::GetCommandBuffer(); + u32 bit_mask = cmd_buff[1] & 0xF; + daemon_bit_mask = static_cast<DaemonMask>(static_cast<u32>(daemon_bit_mask) | bit_mask); + for (size_t index = 0; index < daemon_status.size(); ++index) { + if (bit_mask & (1 << index)) { + daemon_status[index] = DaemonStatus::Idle; + } + } + + cmd_buff[0] = IPC::MakeHeader(0x7, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) daemon_bit_mask=0x%08X ", daemon_bit_mask); +} + +void SuspendScheduler(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[0] = IPC::MakeHeader(0x8, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) called"); +} + +void ResumeScheduler(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[0] = IPC::MakeHeader(0x9, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) called"); +} + +void QueryStatus(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + u32 daemon = cmd_buff[1] & 0xF; - LOG_WARNING(Service_NDM, "(STUBBED) bit_mask=0x%08X ", cmd_buff[1]); + cmd_buff[0] = IPC::MakeHeader(0xD, 2, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = static_cast<u32>(daemon_status.at(daemon)); + LOG_WARNING(Service_NDM, "(STUBBED) daemon=0x%08X, daemon_status=0x%08X", daemon, cmd_buff[2]); +} + +void GetDaemonDisableCount(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + u32 daemon = cmd_buff[1] & 0xF; + + cmd_buff[0] = IPC::MakeHeader(0xE, 3, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = 0; + cmd_buff[3] = 0; + LOG_WARNING(Service_NDM, "(STUBBED) daemon=0x%08X", daemon); +} + +void GetSchedulerDisableCount(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[0] = IPC::MakeHeader(0xF, 3, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = 0; + cmd_buff[3] = 0; + LOG_WARNING(Service_NDM, "(STUBBED) called"); +} + +void SetScanInterval(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + scan_interval = cmd_buff[1]; + cmd_buff[0] = IPC::MakeHeader(0x10, 1, 0); cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) scan_interval=0x%08X ", scan_interval); +} + +void GetScanInterval(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[0] = IPC::MakeHeader(0x11, 2, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = scan_interval; + LOG_WARNING(Service_NDM, "(STUBBED) scan_interval=0x%08X ", scan_interval); +} + +void SetRetryInterval(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + retry_interval = cmd_buff[1]; + + cmd_buff[0] = IPC::MakeHeader(0x12, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) retry_interval=0x%08X ", retry_interval); +} + +void GetRetryInterval(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[0] = IPC::MakeHeader(0x13, 2, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = retry_interval; + LOG_WARNING(Service_NDM, "(STUBBED) retry_interval=0x%08X ", retry_interval); } void OverrideDefaultDaemons(Service::Interface* self) { u32* cmd_buff = Kernel::GetCommandBuffer(); + u32 bit_mask = cmd_buff[1] & 0xF; + default_daemon_bit_mask = static_cast<DaemonMask>(bit_mask); + daemon_bit_mask = default_daemon_bit_mask; + for (size_t index = 0; index < daemon_status.size(); ++index) { + if (bit_mask & (1 << index)) { + daemon_status[index] = DaemonStatus::Idle; + } + } - LOG_WARNING(Service_NDM, "(STUBBED) bit_mask=0x%08X ", cmd_buff[1]); + cmd_buff[0] = IPC::MakeHeader(0x14, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) default_daemon_bit_mask=0x%08X ", default_daemon_bit_mask); +} + +void ResetDefaultDaemons(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + default_daemon_bit_mask = DaemonMask::Default; + + cmd_buff[0] = IPC::MakeHeader(0x15, 1, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) default_daemon_bit_mask=0x%08X ", default_daemon_bit_mask); +} + +void GetDefaultDaemons(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[0] = IPC::MakeHeader(0x16, 2, 0); + cmd_buff[1] = RESULT_SUCCESS.raw; // No error + cmd_buff[2] = static_cast<u32>(default_daemon_bit_mask); + LOG_WARNING(Service_NDM, "(STUBBED) default_daemon_bit_mask=0x%08X ", default_daemon_bit_mask); +} + +void ClearHalfAwakeMacFilter(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + cmd_buff[0] = IPC::MakeHeader(0x17, 1, 0); cmd_buff[1] = RESULT_SUCCESS.raw; // No error + LOG_WARNING(Service_NDM, "(STUBBED) called"); } void Init() { diff --git a/src/core/hle/service/ndm/ndm.h b/src/core/hle/service/ndm/ndm.h index 734730f8c..5c2b968dc 100644 --- a/src/core/hle/service/ndm/ndm.h +++ b/src/core/hle/service/ndm/ndm.h @@ -12,10 +12,91 @@ class Interface; namespace NDM { +enum class Daemon : u32 { + Cec = 0, + Boss = 1, + Nim = 2, + Friend = 3 +}; + +enum class DaemonMask : u32 { + None = 0, + Cec = (1 << static_cast<u32>(Daemon::Cec)), + Boss = (1 << static_cast<u32>(Daemon::Boss)), + Nim = (1 << static_cast<u32>(Daemon::Nim)), + Friend = (1 << static_cast<u32>(Daemon::Friend)), + Default = Cec | Friend, + All = Cec | Boss | Nim | Friend +}; + +enum class DaemonStatus : u32 { + Busy = 0, + Idle = 1, + Suspending = 2, + Suspended = 3 +}; + +enum class ExclusiveState : u32 { + None = 0, + Infrastructure = 1, + LocalCommunications = 2, + Streetpass = 3, + StreetpassData = 4, +}; + +/** + * NDM::EnterExclusiveState service function + * Inputs: + * 0 : Header code [0x00010042] + * 1 : Exclusive State + * 2 : 0x20 + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void EnterExclusiveState(Service::Interface* self); + +/** + * NDM::LeaveExclusiveState service function + * Inputs: + * 0 : Header code [0x00020002] + * 1 : 0x20 + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void LeaveExclusiveState(Service::Interface* self); + +/** + * NDM::QueryExclusiveMode service function + * Inputs: + * 0 : Header code [0x00030000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Current Exclusive State + */ +void QueryExclusiveMode(Service::Interface* self); + +/** + * NDM::LockState service function + * Inputs: + * 0 : Header code [0x00040002] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void LockState(Service::Interface* self); + +/** + * NDM::UnlockState service function + * Inputs: + * 0 : Header code [0x00050002] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void UnlockState(Service::Interface* self); + /** - * SuspendDaemons + * NDM::SuspendDaemons service function * Inputs: - * 0 : Command header (0x00020082) + * 0 : Header code [0x00060040] * 1 : Daemon bit mask * Outputs: * 1 : Result, 0 on success, otherwise error code @@ -23,9 +104,9 @@ namespace NDM { void SuspendDaemons(Service::Interface* self); /** - * ResumeDaemons + * NDM::ResumeDaemons service function * Inputs: - * 0 : Command header (0x00020082) + * 0 : Header code [0x00070040] * 1 : Daemon bit mask * Outputs: * 1 : Result, 0 on success, otherwise error code @@ -33,15 +114,138 @@ void SuspendDaemons(Service::Interface* self); void ResumeDaemons(Service::Interface* self); /** - * OverrideDefaultDaemons + * NDM::SuspendScheduler service function * Inputs: - * 0 : Command header (0x00020082) + * 0 : Header code [0x00080040] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void SuspendScheduler(Service::Interface* self); + +/** + * NDM::ResumeScheduler service function + * Inputs: + * 0 : Header code [0x00090000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void ResumeScheduler(Service::Interface* self); + +/** + * NDM::QueryStatus service function + * Inputs: + * 0 : Header code [0x000D0040] + * 1 : Daemon + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Daemon status + */ +void QueryStatus(Service::Interface* self); + +/** + * NDM::GetDaemonDisableCount service function + * Inputs: + * 0 : Header code [0x000E0040] + * 1 : Daemon + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Current process disable count + * 3 : Total disable count + */ +void GetDaemonDisableCount(Service::Interface* self); + +/** + * NDM::GetSchedulerDisableCount service function + * Inputs: + * 0 : Header code [0x000F0000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Current process disable count + * 3 : Total disable count + */ +void GetSchedulerDisableCount(Service::Interface* self); + +/** + * NDM::SetScanInterval service function + * Inputs: + * 0 : Header code [0x00100040] + * 1 : Interval (default = 30) + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void SetScanInterval(Service::Interface* self); + +/** + * NDM::GetScanInterval service function + * Inputs: + * 0 : Header code [0x00110000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Interval (default = 30) + */ +void GetScanInterval(Service::Interface* self); + +/** + * NDM::SetRetryInterval service function + * Inputs: + * 0 : Header code [0x00120040] + * 1 : Interval (default = 10) + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void SetRetryInterval(Service::Interface* self); + +/** + * NDM::GetRetryInterval service function + * Inputs: + * 0 : Header code [0x00130000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Interval (default = 10) + */ +void GetRetryInterval(Service::Interface* self); + + +/** + * NDM::OverrideDefaultDaemons service function + * Inputs: + * 0 : Header code [0x00140040] * 1 : Daemon bit mask * Outputs: * 1 : Result, 0 on success, otherwise error code */ void OverrideDefaultDaemons(Service::Interface* self); +/** + * NDM::ResetDefaultDaemons service function + * Inputs: + * 0 : Header code [0x00150000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void ResetDefaultDaemons(Service::Interface* self); + +/** + * NDM::GetDefaultDaemons service function + * Inputs: + * 0 : Header code [0x00160000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + * 2 : Daemon bit mask + * Note: + * Gets the current default daemon bit mask. The default value is (DAEMONMASK_CEC | DAEMONMASK_FRIENDS) + */ +void GetDefaultDaemons(Service::Interface* self); + +/** + * NDM::ClearHalfAwakeMacFilter service function + * Inputs: + * 0 : Header code [0x00170000] + * Outputs: + * 1 : Result, 0 on success, otherwise error code + */ +void ClearHalfAwakeMacFilter(Service::Interface* self); + /// Initialize NDM service void Init(); diff --git a/src/core/hle/service/ndm/ndm_u.cpp b/src/core/hle/service/ndm/ndm_u.cpp index bf95cc7aa..3ff0744ee 100644 --- a/src/core/hle/service/ndm/ndm_u.cpp +++ b/src/core/hle/service/ndm/ndm_u.cpp @@ -9,29 +9,29 @@ namespace Service { namespace NDM { const Interface::FunctionInfo FunctionTable[] = { - {0x00010042, nullptr, "EnterExclusiveState"}, - {0x00020002, nullptr, "LeaveExclusiveState"}, - {0x00030000, nullptr, "QueryExclusiveMode"}, - {0x00040002, nullptr, "LockState"}, - {0x00050002, nullptr, "UnlockState"}, + {0x00010042, EnterExclusiveState, "EnterExclusiveState"}, + {0x00020002, LeaveExclusiveState, "LeaveExclusiveState"}, + {0x00030000, QueryExclusiveMode, "QueryExclusiveMode"}, + {0x00040002, LockState, "LockState"}, + {0x00050002, UnlockState, "UnlockState"}, {0x00060040, SuspendDaemons, "SuspendDaemons"}, {0x00070040, ResumeDaemons, "ResumeDaemons"}, - {0x00080040, nullptr, "DisableWifiUsage"}, - {0x00090000, nullptr, "EnableWifiUsage"}, + {0x00080040, SuspendScheduler, "SuspendScheduler"}, + {0x00090000, ResumeScheduler, "ResumeScheduler"}, {0x000A0000, nullptr, "GetCurrentState"}, {0x000B0000, nullptr, "GetTargetState"}, {0x000C0000, nullptr, "<Stubbed>"}, - {0x000D0040, nullptr, "QueryStatus"}, - {0x000E0040, nullptr, "GetDaemonDisableCount"}, - {0x000F0000, nullptr, "GetSchedulerDisableCount"}, - {0x00100040, nullptr, "SetScanInterval"}, - {0x00110000, nullptr, "GetScanInterval"}, - {0x00120040, nullptr, "SetRetryInterval"}, - {0x00130000, nullptr, "GetRetryInterval"}, + {0x000D0040, QueryStatus, "QueryStatus"}, + {0x000E0040, GetDaemonDisableCount, "GetDaemonDisableCount"}, + {0x000F0000, GetSchedulerDisableCount,"GetSchedulerDisableCount"}, + {0x00100040, SetScanInterval, "SetScanInterval"}, + {0x00110000, GetScanInterval, "GetScanInterval"}, + {0x00120040, SetRetryInterval, "SetRetryInterval"}, + {0x00130000, GetRetryInterval, "GetRetryInterval"}, {0x00140040, OverrideDefaultDaemons, "OverrideDefaultDaemons"}, - {0x00150000, nullptr, "ResetDefaultDaemons"}, - {0x00160000, nullptr, "GetDefaultDaemons"}, - {0x00170000, nullptr, "ClearHalfAwakeMacFilter"}, + {0x00150000, ResetDefaultDaemons, "ResetDefaultDaemons"}, + {0x00160000, GetDefaultDaemons, "GetDefaultDaemons"}, + {0x00170000, ClearHalfAwakeMacFilter, "ClearHalfAwakeMacFilter"}, }; NDM_U_Interface::NDM_U_Interface() { diff --git a/src/core/hle/svc.cpp b/src/core/hle/svc.cpp index a9a1a3244..fb2aecbf2 100644 --- a/src/core/hle/svc.cpp +++ b/src/core/hle/svc.cpp @@ -859,6 +859,10 @@ static ResultCode GetProcessInfo(s64* out, Handle process_handle, u32 type) { // TODO(yuriks): Type 0 returns a slightly higher number than type 2, but I'm not sure // what's the difference between them. *out = process->heap_used + process->linear_heap_used + process->misc_memory_used; + if(*out % Memory::PAGE_SIZE != 0) { + LOG_ERROR(Kernel_SVC, "called, memory size not page-aligned"); + return ERR_MISALIGNED_SIZE; + } break; case 1: case 3: diff --git a/src/core/hw/lcd.h b/src/core/hw/lcd.h index 3dd877fbf..57029c5e8 100644 --- a/src/core/hw/lcd.h +++ b/src/core/hw/lcd.h @@ -52,8 +52,6 @@ struct Regs { return content[index]; } -#undef ASSERT_MEMBER_SIZE - }; static_assert(std::is_standard_layout<Regs>::value, "Structure does not use standard layout"); diff --git a/src/core/settings.cpp b/src/core/settings.cpp index eaf5c8461..77261eafe 100644 --- a/src/core/settings.cpp +++ b/src/core/settings.cpp @@ -4,6 +4,8 @@ #include "settings.h" +#include "audio_core/audio_core.h" + #include "core/gdbstub/gdbstub.h" #include "video_core/video_core.h" @@ -20,6 +22,9 @@ void Apply() { VideoCore::g_hw_renderer_enabled = values.use_hw_renderer; VideoCore::g_shader_jit_enabled = values.use_shader_jit; VideoCore::g_scaled_resolution_enabled = values.use_scaled_resolution; + + AudioCore::SelectSink(values.sink_id); + } } // namespace diff --git a/src/core/settings.h b/src/core/settings.h index d620d8461..ce2a31164 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -6,7 +6,8 @@ #include <string> #include <array> -#include <common/file_util.h> + +#include "common/common_types.h" namespace Settings { @@ -63,6 +64,9 @@ struct Values { std::string log_filter; + // Audio + std::string sink_id; + // Debugging bool use_gdbstub; u16 gdbstub_port; diff --git a/src/core/tracer/recorder.h b/src/core/tracer/recorder.h index a42ccc45f..febf883c8 100644 --- a/src/core/tracer/recorder.h +++ b/src/core/tracer/recorder.h @@ -4,6 +4,7 @@ #pragma once +#include <string> #include <unordered_map> #include <vector> diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index 76cfd4f7d..581a37897 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -15,7 +15,7 @@ set(SRCS shader/shader.cpp shader/shader_interpreter.cpp swrasterizer.cpp - utils.cpp + vertex_loader.cpp video_core.cpp ) @@ -43,6 +43,7 @@ set(HEADERS shader/shader_interpreter.h swrasterizer.h utils.h + vertex_loader.h video_core.h ) diff --git a/src/video_core/clipper.cpp b/src/video_core/clipper.cpp index 3d503486e..2bc747102 100644 --- a/src/video_core/clipper.cpp +++ b/src/video_core/clipper.cpp @@ -2,13 +2,24 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> +#include <array> +#include <cstddef> + #include <boost/container/static_vector.hpp> +#include <boost/container/vector.hpp> + +#include "common/bit_field.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "common/vector_math.h" #include "video_core/clipper.h" #include "video_core/pica.h" #include "video_core/pica_state.h" +#include "video_core/pica_types.h" #include "video_core/rasterizer.h" -#include "video_core/shader/shader_interpreter.h" +#include "video_core/shader/shader.h" namespace Pica { diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp index 97ba8214e..be1a936b2 100644 --- a/src/video_core/command_processor.cpp +++ b/src/video_core/command_processor.cpp @@ -2,25 +2,32 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cmath> -#include <boost/range/algorithm/fill.hpp> +#include <array> +#include <cstddef> +#include <memory> +#include <utility> -#include "common/alignment.h" +#include "common/assert.h" +#include "common/logging/log.h" #include "common/microprofile.h" +#include "common/vector_math.h" -#include "core/settings.h" #include "core/hle/service/gsp_gpu.h" #include "core/hw/gpu.h" +#include "core/memory.h" +#include "core/tracer/recorder.h" -#include "video_core/clipper.h" #include "video_core/command_processor.h" +#include "video_core/debug_utils/debug_utils.h" #include "video_core/pica.h" #include "video_core/pica_state.h" +#include "video_core/pica_types.h" #include "video_core/primitive_assembly.h" +#include "video_core/rasterizer_interface.h" #include "video_core/renderer_base.h" +#include "video_core/shader/shader.h" +#include "video_core/vertex_loader.h" #include "video_core/video_core.h" -#include "video_core/debug_utils/debug_utils.h" -#include "video_core/shader/shader_interpreter.h" namespace Pica { @@ -188,54 +195,14 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { #if PICA_LOG_TEV DebugUtils::DumpTevStageConfig(regs.GetTevStages()); #endif - if (g_debug_context) g_debug_context->OnEvent(DebugContext::Event::IncomingPrimitiveBatch, nullptr); - const auto& attribute_config = regs.vertex_attributes; - const u32 base_address = attribute_config.GetPhysicalBaseAddress(); - - // Information about internal vertex attributes - u32 vertex_attribute_sources[16]; - boost::fill(vertex_attribute_sources, 0xdeadbeef); - u32 vertex_attribute_strides[16] = {}; - Regs::VertexAttributeFormat vertex_attribute_formats[16] = {}; - - u32 vertex_attribute_elements[16] = {}; - u32 vertex_attribute_element_size[16] = {}; - - // Setup attribute data from loaders - for (int loader = 0; loader < 12; ++loader) { - const auto& loader_config = attribute_config.attribute_loaders[loader]; - - u32 offset = 0; - - // TODO: What happens if a loader overwrites a previous one's data? - for (unsigned component = 0; component < loader_config.component_count; ++component) { - if (component >= 12) { - LOG_ERROR(HW_GPU, "Overflow in the vertex attribute loader %u trying to load component %u", loader, component); - continue; - } - - u32 attribute_index = loader_config.GetComponent(component); - if (attribute_index < 12) { - int element_size = attribute_config.GetElementSizeInBytes(attribute_index); - offset = Common::AlignUp(offset, element_size); - vertex_attribute_sources[attribute_index] = base_address + loader_config.data_offset + offset; - vertex_attribute_strides[attribute_index] = static_cast<u32>(loader_config.byte_count); - vertex_attribute_formats[attribute_index] = attribute_config.GetFormat(attribute_index); - vertex_attribute_elements[attribute_index] = attribute_config.GetNumElements(attribute_index); - vertex_attribute_element_size[attribute_index] = element_size; - offset += attribute_config.GetStride(attribute_index); - } else if (attribute_index < 16) { - // Attribute ids 12, 13, 14 and 15 signify 4, 8, 12 and 16-byte paddings, respectively - offset = Common::AlignUp(offset, 4); - offset += (attribute_index - 11) * 4; - } else { - UNREACHABLE(); // This is truly unreachable due to the number of bits for each component - } - } - } + // Processes information about internal vertex attributes to figure out how a vertex is loaded. + // Later, these can be compiled and cached. + VertexLoader loader; + const u32 base_address = regs.vertex_attributes.GetPhysicalBaseAddress(); + loader.Setup(regs); // Load vertices bool is_indexed = (id == PICA_REG_INDEX(trigger_draw_indexed)); @@ -259,32 +226,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { } } - class { - /// Combine overlapping and close ranges - void SimplifyRanges() { - for (auto it = ranges.begin(); it != ranges.end(); ++it) { - // NOTE: We add 32 to the range end address to make sure "close" ranges are combined, too - auto it2 = std::next(it); - while (it2 != ranges.end() && it->first + it->second + 32 >= it2->first) { - it->second = std::max(it->second, it2->first + it2->second - it->first); - it2 = ranges.erase(it2); - } - } - } - - public: - /// Record a particular memory access in the list - void AddAccess(u32 paddr, u32 size) { - // Create new range or extend existing one - ranges[paddr] = std::max(ranges[paddr], size); - - // Simplify ranges... - SimplifyRanges(); - } - - /// Map of accessed ranges (mapping start address to range size) - std::map<u32, u32> ranges; - } memory_accesses; + DebugUtils::MemoryAccessTracker memory_accesses; // Simple circular-replacement vertex cache // The size has been tuned for optimal balance between hit-rate and the cost of lookup @@ -328,60 +270,13 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { if (!vertex_cache_hit) { // Initialize data for the current vertex Shader::InputVertex input; - - for (int i = 0; i < attribute_config.GetNumTotalAttributes(); ++i) { - if (vertex_attribute_elements[i] != 0) { - // Default attribute values set if array elements have < 4 components. This - // is *not* carried over from the default attribute settings even if they're - // enabled for this attribute. - static const float24 zero = float24::FromFloat32(0.0f); - static const float24 one = float24::FromFloat32(1.0f); - input.attr[i] = Math::Vec4<float24>(zero, zero, zero, one); - - // Load per-vertex data from the loader arrays - for (unsigned int comp = 0; comp < vertex_attribute_elements[i]; ++comp) { - u32 source_addr = vertex_attribute_sources[i] + vertex_attribute_strides[i] * vertex + comp * vertex_attribute_element_size[i]; - const u8* srcdata = Memory::GetPhysicalPointer(source_addr); - - if (g_debug_context && Pica::g_debug_context->recorder) { - memory_accesses.AddAccess(source_addr, - (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::FLOAT) ? 4 - : (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::SHORT) ? 2 : 1); - } - - const float srcval = - (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::BYTE) ? *reinterpret_cast<const s8*>(srcdata) : - (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::UBYTE) ? *reinterpret_cast<const u8*>(srcdata) : - (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::SHORT) ? *reinterpret_cast<const s16*>(srcdata) : - *reinterpret_cast<const float*>(srcdata); - - input.attr[i][comp] = float24::FromFloat32(srcval); - LOG_TRACE(HW_GPU, "Loaded component %x of attribute %x for vertex %x (index %x) from 0x%08x + 0x%08x + 0x%04x: %f", - comp, i, vertex, index, - attribute_config.GetPhysicalBaseAddress(), - vertex_attribute_sources[i] - base_address, - vertex_attribute_strides[i] * vertex + comp * vertex_attribute_element_size[i], - input.attr[i][comp].ToFloat32()); - } - } else if (attribute_config.IsDefaultAttribute(i)) { - // Load the default attribute if we're configured to do so - input.attr[i] = g_state.vs.default_attributes[i]; - LOG_TRACE(HW_GPU, "Loaded default attribute %x for vertex %x (index %x): (%f, %f, %f, %f)", - i, vertex, index, - input.attr[i][0].ToFloat32(), input.attr[i][1].ToFloat32(), - input.attr[i][2].ToFloat32(), input.attr[i][3].ToFloat32()); - } else { - // TODO(yuriks): In this case, no data gets loaded and the vertex - // remains with the last value it had. This isn't currently maintained - // as global state, however, and so won't work in Citra yet. - } - } + loader.LoadVertex(base_address, index, vertex, input, memory_accesses); if (g_debug_context) g_debug_context->OnEvent(DebugContext::Event::VertexLoaded, (void*)&input); // Send to vertex shader - output = Shader::Run(shader_unit, input, attribute_config.GetNumTotalAttributes()); + output = Shader::Run(shader_unit, input, loader.GetNumTotalAttributes()); if (is_indexed) { vertex_cache[vertex_cache_pos] = output; diff --git a/src/video_core/debug_utils/debug_utils.cpp b/src/video_core/debug_utils/debug_utils.cpp index 178a566f7..fb20f81dd 100644 --- a/src/video_core/debug_utils/debug_utils.cpp +++ b/src/video_core/debug_utils/debug_utils.cpp @@ -4,35 +4,41 @@ #include <algorithm> #include <condition_variable> +#include <cstdint> #include <cstring> #include <fstream> -#include <list> #include <map> #include <mutex> +#include <stdexcept> #include <string> #ifdef HAVE_PNG #include <png.h> +#include <setjmp.h> #endif +#include <nihstro/bit_field.h> #include <nihstro/float24.h> #include <nihstro/shader_binary.h> #include "common/assert.h" +#include "common/bit_field.h" #include "common/color.h" #include "common/common_types.h" #include "common/file_util.h" +#include "common/logging/log.h" #include "common/math_util.h" #include "common/vector_math.h" -#include "core/settings.h" - +#include "video_core/debug_utils/debug_utils.h" #include "video_core/pica.h" #include "video_core/pica_state.h" +#include "video_core/pica_types.h" +#include "video_core/rasterizer_interface.h" #include "video_core/renderer_base.h" +#include "video_core/shader/shader.h" #include "video_core/utils.h" #include "video_core/video_core.h" -#include "video_core/debug_utils/debug_utils.h" using nihstro::DVLBHeader; using nihstro::DVLEHeader; diff --git a/src/video_core/debug_utils/debug_utils.h b/src/video_core/debug_utils/debug_utils.h index 56f9bd958..be2d0301a 100644 --- a/src/video_core/debug_utils/debug_utils.h +++ b/src/video_core/debug_utils/debug_utils.h @@ -4,23 +4,33 @@ #pragma once +#include <algorithm> #include <array> #include <condition_variable> +#include <iterator> #include <list> #include <map> #include <memory> #include <mutex> +#include <string> +#include <utility> #include <vector> +#include "common/common_types.h" #include "common/vector_math.h" -#include "core/tracer/recorder.h" - #include "video_core/pica.h" -#include "video_core/shader/shader.h" + +namespace CiTrace { +class Recorder; +} namespace Pica { +namespace Shader { +struct ShaderSetup; +} + class DebugContext { public: enum class Event { @@ -216,6 +226,36 @@ void DumpTexture(const Pica::Regs::TextureConfig& texture_config, u8* data); void DumpTevStageConfig(const std::array<Pica::Regs::TevStageConfig,6>& stages); +/** + * Used in the vertex loader to merge access records. TODO: Investigate if actually useful. + */ +class MemoryAccessTracker { + /// Combine overlapping and close ranges + void SimplifyRanges() { + for (auto it = ranges.begin(); it != ranges.end(); ++it) { + // NOTE: We add 32 to the range end address to make sure "close" ranges are combined, too + auto it2 = std::next(it); + while (it2 != ranges.end() && it->first + it->second + 32 >= it2->first) { + it->second = std::max(it->second, it2->first + it2->second - it->first); + it2 = ranges.erase(it2); + } + } + } + +public: + /// Record a particular memory access in the list + void AddAccess(u32 paddr, u32 size) { + // Create new range or extend existing one + ranges[paddr] = std::max(ranges[paddr], size); + + // Simplify ranges... + SimplifyRanges(); + } + + /// Map of accessed ranges (mapping start address to range size) + std::map<u32, u32> ranges; +}; + } // namespace } // namespace diff --git a/src/video_core/pica.cpp b/src/video_core/pica.cpp index ccbaf071b..be82cf4b5 100644 --- a/src/video_core/pica.cpp +++ b/src/video_core/pica.cpp @@ -3,10 +3,13 @@ // Refer to the license.txt file included. #include <cstring> +#include <iterator> #include <unordered_map> +#include <utility> #include "video_core/pica.h" #include "video_core/pica_state.h" +#include "video_core/primitive_assembly.h" #include "video_core/shader/shader.h" namespace Pica { @@ -480,7 +483,7 @@ std::string Regs::GetCommandName(int index) { static std::unordered_map<u32, const char*> map; if (map.empty()) { - map.insert(begin(register_names), end(register_names)); + map.insert(std::begin(register_names), std::end(register_names)); } // Return empty string if no match is found diff --git a/src/video_core/pica.h b/src/video_core/pica.h index cf130d7f8..5891fb72a 100644 --- a/src/video_core/pica.h +++ b/src/video_core/pica.h @@ -5,10 +5,13 @@ #pragma once #include <array> -#include <cmath> #include <cstddef> #include <string> +#ifndef _MSC_VER +#include <type_traits> // for std::enable_if +#endif + #include "common/assert.h" #include "common/bit_field.h" #include "common/common_funcs.h" @@ -16,8 +19,6 @@ #include "common/vector_math.h" #include "common/logging/log.h" -#include "pica_types.h" - namespace Pica { // Returns index corresponding to the Regs member labeled by field_name diff --git a/src/video_core/pica_state.h b/src/video_core/pica_state.h index 323290054..bbecad850 100644 --- a/src/video_core/pica_state.h +++ b/src/video_core/pica_state.h @@ -4,6 +4,11 @@ #pragma once +#include <array> + +#include "common/bit_field.h" +#include "common/common_types.h" + #include "video_core/pica.h" #include "video_core/primitive_assembly.h" #include "video_core/shader/shader.h" diff --git a/src/video_core/pica_types.h b/src/video_core/pica_types.h index ecf45654b..3b7bfbdca 100644 --- a/src/video_core/pica_types.h +++ b/src/video_core/pica_types.h @@ -4,6 +4,7 @@ #pragma once +#include <cmath> #include <cstring> #include "common/common_types.h" diff --git a/src/video_core/primitive_assembly.cpp b/src/video_core/primitive_assembly.cpp index ff3e2b862..68ea3c08a 100644 --- a/src/video_core/primitive_assembly.cpp +++ b/src/video_core/primitive_assembly.cpp @@ -6,8 +6,7 @@ #include "video_core/pica.h" #include "video_core/primitive_assembly.h" -#include "video_core/debug_utils/debug_utils.h" -#include "video_core/shader/shader_interpreter.h" +#include "video_core/shader/shader.h" namespace Pica { diff --git a/src/video_core/rasterizer.cpp b/src/video_core/rasterizer.cpp index 9cf77b1f2..df67b9081 100644 --- a/src/video_core/rasterizer.cpp +++ b/src/video_core/rasterizer.cpp @@ -3,22 +3,28 @@ // Refer to the license.txt file included. #include <algorithm> +#include <array> #include <cmath> +#include "common/assert.h" +#include "common/bit_field.h" #include "common/color.h" #include "common/common_types.h" +#include "common/logging/log.h" #include "common/math_util.h" #include "common/microprofile.h" +#include "common/vector_math.h" #include "core/memory.h" #include "core/hw/gpu.h" +#include "video_core/debug_utils/debug_utils.h" #include "video_core/pica.h" #include "video_core/pica_state.h" +#include "video_core/pica_types.h" #include "video_core/rasterizer.h" #include "video_core/utils.h" -#include "video_core/debug_utils/debug_utils.h" -#include "video_core/shader/shader_interpreter.h" +#include "video_core/shader/shader.h" namespace Pica { diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp index ccd497de0..3f451e062 100644 --- a/src/video_core/renderer_base.cpp +++ b/src/video_core/renderer_base.cpp @@ -2,10 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <atomic> #include <memory> -#include "core/settings.h" - #include "video_core/renderer_base.h" #include "video_core/video_core.h" #include "video_core/swrasterizer.h" diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index a8c775c80..519d81aeb 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -2,27 +2,28 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> #include <memory> +#include <string> +#include <tuple> +#include <utility> #include <glad/glad.h> +#include "common/assert.h" #include "common/color.h" -#include "common/file_util.h" +#include "common/logging/log.h" #include "common/math_util.h" -#include "common/microprofile.h" +#include "common/vector_math.h" -#include "core/memory.h" -#include "core/settings.h" #include "core/hw/gpu.h" #include "video_core/pica.h" #include "video_core/pica_state.h" -#include "video_core/utils.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_shader_gen.h" #include "video_core/renderer_opengl/gl_shader_util.h" #include "video_core/renderer_opengl/pica_to_gl.h" +#include "video_core/renderer_opengl/renderer_opengl.h" static bool IsPassThroughTevStage(const Pica::Regs::TevStageConfig& stage) { return (stage.color_op == Pica::Regs::TevStageConfig::Operation::Replace && @@ -812,6 +813,7 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Pica::Regs::TextureConf if (wrap_s == TextureConfig::ClampToBorder || wrap_t == TextureConfig::ClampToBorder) { if (border_color != config.border_color.raw) { + border_color = config.border_color.raw; auto gl_color = PicaToGL::ColorRGBA8(border_color); glSamplerParameterfv(s, GL_TEXTURE_BORDER_COLOR, gl_color.data()); } diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 8d6177e88..63ff7716d 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -4,23 +4,33 @@ #pragma once +#include <array> #include <cstddef> #include <cstring> #include <memory> #include <vector> #include <unordered_map> +#include <glad/glad.h> + +#include "common/bit_field.h" #include "common/common_types.h" #include "common/hash.h" +#include "common/vector_math.h" + +#include "core/hw/gpu.h" #include "video_core/pica.h" #include "video_core/pica_state.h" +#include "video_core/pica_types.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_opengl/gl_rasterizer_cache.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_state.h" #include "video_core/renderer_opengl/pica_to_gl.h" -#include "video_core/renderer_opengl/renderer_opengl.h" -#include "video_core/shader/shader_interpreter.h" +#include "video_core/shader/shader.h" + +struct ScreenInfo; /** * This struct contains all state used to generate the GLSL shader program that emulates the current @@ -39,36 +49,17 @@ struct PicaShaderConfig { res.alpha_test_func = regs.output_merger.alpha_test.enable ? regs.output_merger.alpha_test.func.Value() : Pica::Regs::CompareFunc::Always; - // Copy relevant TevStageConfig fields only. We're doing this manually (instead of calling - // the GetTevStages() function) because BitField explicitly disables copies. - - res.tev_stages[0].sources_raw = regs.tev_stage0.sources_raw; - res.tev_stages[1].sources_raw = regs.tev_stage1.sources_raw; - res.tev_stages[2].sources_raw = regs.tev_stage2.sources_raw; - res.tev_stages[3].sources_raw = regs.tev_stage3.sources_raw; - res.tev_stages[4].sources_raw = regs.tev_stage4.sources_raw; - res.tev_stages[5].sources_raw = regs.tev_stage5.sources_raw; - - res.tev_stages[0].modifiers_raw = regs.tev_stage0.modifiers_raw; - res.tev_stages[1].modifiers_raw = regs.tev_stage1.modifiers_raw; - res.tev_stages[2].modifiers_raw = regs.tev_stage2.modifiers_raw; - res.tev_stages[3].modifiers_raw = regs.tev_stage3.modifiers_raw; - res.tev_stages[4].modifiers_raw = regs.tev_stage4.modifiers_raw; - res.tev_stages[5].modifiers_raw = regs.tev_stage5.modifiers_raw; - - res.tev_stages[0].ops_raw = regs.tev_stage0.ops_raw; - res.tev_stages[1].ops_raw = regs.tev_stage1.ops_raw; - res.tev_stages[2].ops_raw = regs.tev_stage2.ops_raw; - res.tev_stages[3].ops_raw = regs.tev_stage3.ops_raw; - res.tev_stages[4].ops_raw = regs.tev_stage4.ops_raw; - res.tev_stages[5].ops_raw = regs.tev_stage5.ops_raw; - - res.tev_stages[0].scales_raw = regs.tev_stage0.scales_raw; - res.tev_stages[1].scales_raw = regs.tev_stage1.scales_raw; - res.tev_stages[2].scales_raw = regs.tev_stage2.scales_raw; - res.tev_stages[3].scales_raw = regs.tev_stage3.scales_raw; - res.tev_stages[4].scales_raw = regs.tev_stage4.scales_raw; - res.tev_stages[5].scales_raw = regs.tev_stage5.scales_raw; + // Copy tev stages + const auto& tev_stages = regs.GetTevStages(); + DEBUG_ASSERT(res.tev_stages.size() == tev_stages.size()); + for (size_t i = 0; i < tev_stages.size(); i++) { + const auto& tev_stage = tev_stages[i]; + res.tev_stages[i].sources_raw = tev_stage.sources_raw; + res.tev_stages[i].modifiers_raw = tev_stage.modifiers_raw; + res.tev_stages[i].ops_raw = tev_stage.ops_raw; + res.tev_stages[i].const_color = tev_stage.const_color; + res.tev_stages[i].scales_raw = tev_stage.scales_raw; + } res.combiner_buffer_input = regs.tev_combiner_buffer_input.update_mask_rgb.Value() | diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index 55c2fb283..7efd0038a 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -2,10 +2,19 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> +#include <atomic> +#include <cstring> +#include <iterator> #include <unordered_set> +#include <utility> +#include <vector> +#include <glad/glad.h> + +#include "common/bit_field.h" #include "common/emu_window.h" -#include "common/hash.h" +#include "common/logging/log.h" #include "common/math_util.h" #include "common/microprofile.h" #include "common/vector_math.h" @@ -15,7 +24,7 @@ #include "video_core/debug_utils/debug_utils.h" #include "video_core/pica_state.h" #include "video_core/renderer_opengl/gl_rasterizer_cache.h" -#include "video_core/renderer_opengl/pica_to_gl.h" +#include "video_core/renderer_opengl/gl_state.h" #include "video_core/utils.h" #include "video_core/video_core.h" diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h index 893d51138..225596415 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -4,20 +4,26 @@ #pragma once -#include <map> +#include <array> #include <memory> #include <set> +#include <tuple> #include <boost/icl/interval_map.hpp> +#include <glad/glad.h> -#include "common/math_util.h" +#include "common/assert.h" +#include "common/common_funcs.h" +#include "common/common_types.h" #include "core/hw/gpu.h" #include "video_core/pica.h" -#include "video_core/debug_utils/debug_utils.h" #include "video_core/renderer_opengl/gl_resource_manager.h" -#include "video_core/renderer_opengl/gl_state.h" + +namespace MathUtil { +template <class T> struct Rectangle; +} struct CachedSurface; diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index 646b4eaaf..9011caa39 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -2,9 +2,17 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> +#include <cstddef> + +#include "common/assert.h" +#include "common/bit_field.h" +#include "common/logging/log.h" + #include "video_core/pica.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_shader_gen.h" +#include "video_core/renderer_opengl/gl_shader_util.h" using Pica::Regs; using TevStageConfig = Regs::TevStageConfig; diff --git a/src/video_core/renderer_opengl/gl_shader_gen.h b/src/video_core/renderer_opengl/gl_shader_gen.h index 0ca9d2879..3eb07d57a 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.h +++ b/src/video_core/renderer_opengl/gl_shader_gen.h @@ -6,7 +6,7 @@ #include <string> -#include "video_core/renderer_opengl/gl_rasterizer.h" +struct PicaShaderConfig; namespace GLShader { diff --git a/src/video_core/renderer_opengl/gl_shader_util.cpp b/src/video_core/renderer_opengl/gl_shader_util.cpp index e3f7a5868..dded3db46 100644 --- a/src/video_core/renderer_opengl/gl_shader_util.cpp +++ b/src/video_core/renderer_opengl/gl_shader_util.cpp @@ -2,9 +2,10 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <algorithm> #include <vector> +#include <glad/glad.h> + #include "common/logging/log.h" #include "video_core/renderer_opengl/gl_shader_util.h" diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp index f04bdd8c5..02cd9f417 100644 --- a/src/video_core/renderer_opengl/gl_state.cpp +++ b/src/video_core/renderer_opengl/gl_state.cpp @@ -2,8 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "video_core/pica.h" -#include "video_core/renderer_opengl/gl_resource_manager.h" +#include <glad/glad.h> + +#include "common/common_funcs.h" +#include "common/logging/log.h" + #include "video_core/renderer_opengl/gl_state.h" OpenGLState OpenGLState::cur_state; diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h index 0f72e9004..24f20e47c 100644 --- a/src/video_core/renderer_opengl/gl_state.h +++ b/src/video_core/renderer_opengl/gl_state.h @@ -5,7 +5,6 @@ #pragma once #include <glad/glad.h> -#include <memory> class OpenGLState { public: diff --git a/src/video_core/renderer_opengl/pica_to_gl.h b/src/video_core/renderer_opengl/pica_to_gl.h index fd3617d77..976d1f364 100644 --- a/src/video_core/renderer_opengl/pica_to_gl.h +++ b/src/video_core/renderer_opengl/pica_to_gl.h @@ -4,9 +4,16 @@ #pragma once +#include <array> +#include <cstddef> + #include <glad/glad.h> +#include "common/assert.h" +#include "common/bit_field.h" +#include "common/common_funcs.h" #include "common/common_types.h" +#include "common/logging/log.h" #include "video_core/pica.h" diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index 8f907593f..0e9a0be8b 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -5,23 +5,28 @@ #include <algorithm> #include <cstddef> #include <cstdlib> +#include <memory> + +#include <glad/glad.h> #include "common/assert.h" +#include "common/bit_field.h" #include "common/emu_window.h" #include "common/logging/log.h" #include "common/profiler_reporting.h" +#include "common/synchronized_wrapper.h" -#include "core/memory.h" -#include "core/settings.h" #include "core/hw/gpu.h" #include "core/hw/hw.h" #include "core/hw/lcd.h" +#include "core/memory.h" +#include "core/settings.h" +#include "core/tracer/recorder.h" -#include "video_core/video_core.h" #include "video_core/debug_utils/debug_utils.h" -#include "video_core/renderer_opengl/gl_rasterizer.h" -#include "video_core/renderer_opengl/gl_shader_util.h" +#include "video_core/rasterizer_interface.h" #include "video_core/renderer_opengl/renderer_opengl.h" +#include "video_core/video_core.h" static const char vertex_shader[] = R"( #version 150 core diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h index 5ca5255ac..00e1044ab 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.h +++ b/src/video_core/renderer_opengl/renderer_opengl.h @@ -8,6 +8,9 @@ #include <glad/glad.h> +#include "common/common_types.h" +#include "common/math_util.h" + #include "core/hw/gpu.h" #include "video_core/renderer_base.h" diff --git a/src/video_core/shader/shader.cpp b/src/video_core/shader/shader.cpp index 043e99190..356d87f13 100644 --- a/src/video_core/shader/shader.cpp +++ b/src/video_core/shader/shader.cpp @@ -2,26 +2,30 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <memory> +#include <atomic> +#include <cmath> +#include <cstring> #include <unordered_map> +#include <utility> #include <boost/range/algorithm/fill.hpp> +#include "common/bit_field.h" #include "common/hash.h" +#include "common/logging/log.h" #include "common/microprofile.h" -#include "video_core/debug_utils/debug_utils.h" #include "video_core/pica.h" #include "video_core/pica_state.h" -#include "video_core/video_core.h" - -#include "shader.h" -#include "shader_interpreter.h" +#include "video_core/shader/shader.h" +#include "video_core/shader/shader_interpreter.h" #ifdef ARCHITECTURE_x86_64 -#include "shader_jit_x64.h" +#include "video_core/shader/shader_jit_x64.h" #endif // ARCHITECTURE_x86_64 +#include "video_core/video_core.h" + namespace Pica { namespace Shader { diff --git a/src/video_core/shader/shader.h b/src/video_core/shader/shader.h index 9c5bd97bd..56b83bfeb 100644 --- a/src/video_core/shader/shader.h +++ b/src/video_core/shader/shader.h @@ -4,17 +4,23 @@ #pragma once +#include <array> +#include <cstddef> +#include <memory> +#include <type_traits> #include <vector> #include <boost/container/static_vector.hpp> -#include <nihstro/shader_binary.h> +#include <nihstro/shader_bytecode.h> +#include "common/assert.h" #include "common/common_funcs.h" #include "common/common_types.h" #include "common/vector_math.h" #include "video_core/pica.h" +#include "video_core/pica_types.h" using nihstro::RegisterType; using nihstro::SourceRegister; @@ -25,7 +31,7 @@ namespace Pica { namespace Shader { struct InputVertex { - Math::Vec4<float24> attr[16]; + alignas(16) Math::Vec4<float24> attr[16]; }; struct OutputVertex { diff --git a/src/video_core/shader/shader_interpreter.cpp b/src/video_core/shader/shader_interpreter.cpp index 9b978583e..7710f7fbc 100644 --- a/src/video_core/shader/shader_interpreter.cpp +++ b/src/video_core/shader/shader_interpreter.cpp @@ -2,12 +2,20 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <algorithm> +#include <array> +#include <cmath> #include <numeric> + #include <nihstro/shader_bytecode.h> -#include "common/file_util.h" -#include "video_core/pica.h" +#include "common/assert.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "common/vector_math.h" + #include "video_core/pica_state.h" +#include "video_core/pica_types.h" #include "video_core/shader/shader.h" #include "video_core/shader/shader_interpreter.h" diff --git a/src/video_core/shader/shader_interpreter.h b/src/video_core/shader/shader_interpreter.h index 294bca50e..6048cdf3a 100644 --- a/src/video_core/shader/shader_interpreter.h +++ b/src/video_core/shader/shader_interpreter.h @@ -4,12 +4,12 @@ #pragma once -#include "video_core/shader/shader.h" - namespace Pica { namespace Shader { +template <bool Debug> struct UnitState; + template<bool Debug> void RunInterpreter(UnitState<Debug>& state); diff --git a/src/video_core/shader/shader_jit_x64.cpp b/src/video_core/shader/shader_jit_x64.cpp index b7747fa42..99f6c51eb 100644 --- a/src/video_core/shader/shader_jit_x64.cpp +++ b/src/video_core/shader/shader_jit_x64.cpp @@ -3,8 +3,15 @@ // Refer to the license.txt file included. #include <algorithm> -#include <smmintrin.h> +#include <cmath> +#include <cstdint> +#include <xmmintrin.h> +#include <nihstro/shader_bytecode.h> + +#include "common/assert.h" +#include "common/logging/log.h" +#include "common/vector_math.h" #include "common/x64/abi.h" #include "common/x64/cpu_detect.h" #include "common/x64/emitter.h" @@ -13,6 +20,7 @@ #include "shader_jit_x64.h" #include "video_core/pica_state.h" +#include "video_core/pica_types.h" namespace Pica { diff --git a/src/video_core/shader/shader_jit_x64.h b/src/video_core/shader/shader_jit_x64.h index cd6280ade..30aa7ff30 100644 --- a/src/video_core/shader/shader_jit_x64.h +++ b/src/video_core/shader/shader_jit_x64.h @@ -4,14 +4,17 @@ #pragma once +#include <array> +#include <cstddef> #include <utility> #include <vector> #include <nihstro/shader_bytecode.h> +#include "common/bit_set.h" +#include "common/common_types.h" #include "common/x64/emitter.h" -#include "video_core/pica.h" #include "video_core/shader/shader.h" using nihstro::Instruction; diff --git a/src/video_core/swrasterizer.h b/src/video_core/swrasterizer.h index 090f899bc..0a028b774 100644 --- a/src/video_core/swrasterizer.h +++ b/src/video_core/swrasterizer.h @@ -8,6 +8,12 @@ #include "video_core/rasterizer_interface.h" +namespace Pica { +namespace Shader { +struct OutputVertex; +} +} + namespace VideoCore { class SWRasterizer : public RasterizerInterface { diff --git a/src/video_core/utils.cpp b/src/video_core/utils.cpp deleted file mode 100644 index 6e1ff5cf4..000000000 --- a/src/video_core/utils.cpp +++ /dev/null @@ -1,36 +0,0 @@ -// Copyright 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <cstdio> -#include <cstring> - -#include "video_core/utils.h" - -namespace VideoCore { - -/** - * Dumps a texture to TGA - * @param filename String filename to dump texture to - * @param width Width of texture in pixels - * @param height Height of texture in pixels - * @param raw_data Raw RGBA8 texture data to dump - * @todo This should be moved to some general purpose/common code - */ -void DumpTGA(std::string filename, short width, short height, u8* raw_data) { - TGAHeader hdr = {0, 0, 2, 0, 0, 0, 0, width, height, 24, 0}; - FILE* fout = fopen(filename.c_str(), "wb"); - - fwrite(&hdr, sizeof(TGAHeader), 1, fout); - - for (int y = 0; y < height; y++) { - for (int x = 0; x < width; x++) { - putc(raw_data[(3 * (y * width)) + (3 * x) + 0], fout); // b - putc(raw_data[(3 * (y * width)) + (3 * x) + 1], fout); // g - putc(raw_data[(3 * (y * width)) + (3 * x) + 2], fout); // r - } - } - - fclose(fout); -} -} // namespace diff --git a/src/video_core/utils.h b/src/video_core/utils.h index 4fa60a10e..7ce83a055 100644 --- a/src/video_core/utils.h +++ b/src/video_core/utils.h @@ -4,37 +4,10 @@ #pragma once -#include <string> - #include "common/common_types.h" namespace VideoCore { -/// Structure for the TGA texture format (for dumping) -struct TGAHeader { - char idlength; - char colormaptype; - char datatypecode; - short int colormaporigin; - short int colormaplength; - short int x_origin; - short int y_origin; - short width; - short height; - char bitsperpixel; - char imagedescriptor; -}; - -/** - * Dumps a texture to TGA - * @param filename String filename to dump texture to - * @param width Width of texture in pixels - * @param height Height of texture in pixels - * @param raw_data Raw RGBA8 texture data to dump - * @todo This should be moved to some general purpose/common code - */ -void DumpTGA(std::string filename, short width, short height, u8* raw_data); - /** * Interleave the lower 3 bits of each coordinate to get the intra-block offsets, which are * arranged in a Z-order curve. More details on the bit manipulation at: diff --git a/src/video_core/vertex_loader.cpp b/src/video_core/vertex_loader.cpp new file mode 100644 index 000000000..21ae52949 --- /dev/null +++ b/src/video_core/vertex_loader.cpp @@ -0,0 +1,140 @@ +#include <memory> + +#include <boost/range/algorithm/fill.hpp> + +#include "common/assert.h" +#include "common/alignment.h" +#include "common/bit_field.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "common/vector_math.h" + +#include "core/memory.h" + +#include "video_core/debug_utils/debug_utils.h" +#include "video_core/pica.h" +#include "video_core/pica_state.h" +#include "video_core/pica_types.h" +#include "video_core/shader/shader.h" +#include "video_core/vertex_loader.h" + +namespace Pica { + +void VertexLoader::Setup(const Pica::Regs& regs) { + const auto& attribute_config = regs.vertex_attributes; + num_total_attributes = attribute_config.GetNumTotalAttributes(); + + boost::fill(vertex_attribute_sources, 0xdeadbeef); + + for (int i = 0; i < 16; i++) { + vertex_attribute_is_default[i] = attribute_config.IsDefaultAttribute(i); + } + + // Setup attribute data from loaders + for (int loader = 0; loader < 12; ++loader) { + const auto& loader_config = attribute_config.attribute_loaders[loader]; + + u32 offset = 0; + + // TODO: What happens if a loader overwrites a previous one's data? + for (unsigned component = 0; component < loader_config.component_count; ++component) { + if (component >= 12) { + LOG_ERROR(HW_GPU, "Overflow in the vertex attribute loader %u trying to load component %u", loader, component); + continue; + } + + u32 attribute_index = loader_config.GetComponent(component); + if (attribute_index < 12) { + offset = Common::AlignUp(offset, attribute_config.GetElementSizeInBytes(attribute_index)); + vertex_attribute_sources[attribute_index] = loader_config.data_offset + offset; + vertex_attribute_strides[attribute_index] = static_cast<u32>(loader_config.byte_count); + vertex_attribute_formats[attribute_index] = attribute_config.GetFormat(attribute_index); + vertex_attribute_elements[attribute_index] = attribute_config.GetNumElements(attribute_index); + offset += attribute_config.GetStride(attribute_index); + } else if (attribute_index < 16) { + // Attribute ids 12, 13, 14 and 15 signify 4, 8, 12 and 16-byte paddings, respectively + offset = Common::AlignUp(offset, 4); + offset += (attribute_index - 11) * 4; + } else { + UNREACHABLE(); // This is truly unreachable due to the number of bits for each component + } + } + } +} + +void VertexLoader::LoadVertex(u32 base_address, int index, int vertex, Shader::InputVertex& input, DebugUtils::MemoryAccessTracker& memory_accesses) { + for (int i = 0; i < num_total_attributes; ++i) { + if (vertex_attribute_elements[i] != 0) { + // Load per-vertex data from the loader arrays + u32 source_addr = base_address + vertex_attribute_sources[i] + vertex_attribute_strides[i] * vertex; + + if (g_debug_context && Pica::g_debug_context->recorder) { + memory_accesses.AddAccess(source_addr, vertex_attribute_elements[i] * ( + (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::FLOAT) ? 4 + : (vertex_attribute_formats[i] == Regs::VertexAttributeFormat::SHORT) ? 2 : 1)); + } + + switch (vertex_attribute_formats[i]) { + case Regs::VertexAttributeFormat::BYTE: + { + const s8* srcdata = reinterpret_cast<const s8*>(Memory::GetPhysicalPointer(source_addr)); + for (unsigned int comp = 0; comp < vertex_attribute_elements[i]; ++comp) { + input.attr[i][comp] = float24::FromFloat32(srcdata[comp]); + } + break; + } + case Regs::VertexAttributeFormat::UBYTE: + { + const u8* srcdata = reinterpret_cast<const u8*>(Memory::GetPhysicalPointer(source_addr)); + for (unsigned int comp = 0; comp < vertex_attribute_elements[i]; ++comp) { + input.attr[i][comp] = float24::FromFloat32(srcdata[comp]); + } + break; + } + case Regs::VertexAttributeFormat::SHORT: + { + const s16* srcdata = reinterpret_cast<const s16*>(Memory::GetPhysicalPointer(source_addr)); + for (unsigned int comp = 0; comp < vertex_attribute_elements[i]; ++comp) { + input.attr[i][comp] = float24::FromFloat32(srcdata[comp]); + } + break; + } + case Regs::VertexAttributeFormat::FLOAT: + { + const float* srcdata = reinterpret_cast<const float*>(Memory::GetPhysicalPointer(source_addr)); + for (unsigned int comp = 0; comp < vertex_attribute_elements[i]; ++comp) { + input.attr[i][comp] = float24::FromFloat32(srcdata[comp]); + } + break; + } + } + + // Default attribute values set if array elements have < 4 components. This + // is *not* carried over from the default attribute settings even if they're + // enabled for this attribute. + for (unsigned int comp = vertex_attribute_elements[i]; comp < 4; ++comp) { + input.attr[i][comp] = comp == 3 ? float24::FromFloat32(1.0f) : float24::FromFloat32(0.0f); + } + + LOG_TRACE(HW_GPU, "Loaded %d components of attribute %x for vertex %x (index %x) from 0x%08x + 0x%08x + 0x%04x: %f %f %f %f", + vertex_attribute_elements[i], i, vertex, index, + base_address, + vertex_attribute_sources[i], + vertex_attribute_strides[i] * vertex, + input.attr[i][0].ToFloat32(), input.attr[i][1].ToFloat32(), input.attr[i][2].ToFloat32(), input.attr[i][3].ToFloat32()); + } else if (vertex_attribute_is_default[i]) { + // Load the default attribute if we're configured to do so + input.attr[i] = g_state.vs.default_attributes[i]; + LOG_TRACE(HW_GPU, "Loaded default attribute %x for vertex %x (index %x): (%f, %f, %f, %f)", + i, vertex, index, + input.attr[i][0].ToFloat32(), input.attr[i][1].ToFloat32(), + input.attr[i][2].ToFloat32(), input.attr[i][3].ToFloat32()); + } else { + // TODO(yuriks): In this case, no data gets loaded and the vertex + // remains with the last value it had. This isn't currently maintained + // as global state, however, and so won't work in Citra yet. + } + } +} + +} // namespace Pica diff --git a/src/video_core/vertex_loader.h b/src/video_core/vertex_loader.h new file mode 100644 index 000000000..becf5a403 --- /dev/null +++ b/src/video_core/vertex_loader.h @@ -0,0 +1,33 @@ +#pragma once + +#include "common/common_types.h" + +#include "video_core/pica.h" + +namespace Pica { + +namespace DebugUtils { +class MemoryAccessTracker; +} + +namespace Shader { +class InputVertex; +} + +class VertexLoader { +public: + void Setup(const Pica::Regs& regs); + void LoadVertex(u32 base_address, int index, int vertex, Shader::InputVertex& input, DebugUtils::MemoryAccessTracker& memory_accesses); + + int GetNumTotalAttributes() const { return num_total_attributes; } + +private: + u32 vertex_attribute_sources[16]; + u32 vertex_attribute_strides[16] = {}; + Regs::VertexAttributeFormat vertex_attribute_formats[16] = {}; + u32 vertex_attribute_elements[16] = {}; + bool vertex_attribute_is_default[16]; + int num_total_attributes; +}; + +} // namespace Pica diff --git a/src/video_core/video_core.cpp b/src/video_core/video_core.cpp index 855286173..c9975876d 100644 --- a/src/video_core/video_core.cpp +++ b/src/video_core/video_core.cpp @@ -4,12 +4,8 @@ #include <memory> -#include "common/emu_window.h" #include "common/logging/log.h" -#include "core/core.h" -#include "core/settings.h" - #include "video_core/pica.h" #include "video_core/renderer_base.h" #include "video_core/video_core.h" |