diff options
Diffstat (limited to 'src')
131 files changed, 3775 insertions, 1203 deletions
diff --git a/src/audio_core/CMakeLists.txt b/src/audio_core/CMakeLists.txt index ec71524a3..82e4850f7 100644 --- a/src/audio_core/CMakeLists.txt +++ b/src/audio_core/CMakeLists.txt @@ -1,4 +1,8 @@ add_library(audio_core STATIC + algorithm/filter.cpp + algorithm/filter.h + algorithm/interpolate.cpp + algorithm/interpolate.h audio_out.cpp audio_out.h audio_renderer.cpp @@ -7,12 +11,12 @@ add_library(audio_core STATIC codec.cpp codec.h null_sink.h - stream.cpp - stream.h sink.h sink_details.cpp sink_details.h sink_stream.h + stream.cpp + stream.h $<$<BOOL:${ENABLE_CUBEB}>:cubeb_sink.cpp cubeb_sink.h> ) diff --git a/src/audio_core/algorithm/filter.cpp b/src/audio_core/algorithm/filter.cpp new file mode 100644 index 000000000..403b8503f --- /dev/null +++ b/src/audio_core/algorithm/filter.cpp @@ -0,0 +1,79 @@ +// Copyright 2018 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#define _USE_MATH_DEFINES + +#include <algorithm> +#include <array> +#include <cmath> +#include <vector> +#include "audio_core/algorithm/filter.h" +#include "common/common_types.h" + +namespace AudioCore { + +Filter Filter::LowPass(double cutoff, double Q) { + const double w0 = 2.0 * M_PI * cutoff; + const double sin_w0 = std::sin(w0); + const double cos_w0 = std::cos(w0); + const double alpha = sin_w0 / (2 * Q); + + const double a0 = 1 + alpha; + const double a1 = -2.0 * cos_w0; + const double a2 = 1 - alpha; + const double b0 = 0.5 * (1 - cos_w0); + const double b1 = 1.0 * (1 - cos_w0); + const double b2 = 0.5 * (1 - cos_w0); + + return {a0, a1, a2, b0, b1, b2}; +} + +Filter::Filter() : Filter(1.0, 0.0, 0.0, 1.0, 0.0, 0.0) {} + +Filter::Filter(double a0, double a1, double a2, double b0, double b1, double b2) + : a1(a1 / a0), a2(a2 / a0), b0(b0 / a0), b1(b1 / a0), b2(b2 / a0) {} + +void Filter::Process(std::vector<s16>& signal) { + const size_t num_frames = signal.size() / 2; + for (size_t i = 0; i < num_frames; i++) { + std::rotate(in.begin(), in.end() - 1, in.end()); + std::rotate(out.begin(), out.end() - 1, out.end()); + + for (size_t ch = 0; ch < channel_count; ch++) { + in[0][ch] = signal[i * channel_count + ch]; + + out[0][ch] = b0 * in[0][ch] + b1 * in[1][ch] + b2 * in[2][ch] - a1 * out[1][ch] - + a2 * out[2][ch]; + + signal[i * 2 + ch] = std::clamp(out[0][ch], -32768.0, 32767.0); + } + } +} + +/// Calculates the appropriate Q for each biquad in a cascading filter. +/// @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(size_t total_count, size_t index) { + const double pole = M_PI * (2 * index + 1) / (4.0 * total_count); + return 1.0 / (2.0 * std::cos(pole)); +} + +CascadingFilter CascadingFilter::LowPass(double cutoff, size_t cascade_size) { + std::vector<Filter> cascade(cascade_size); + for (size_t i = 0; i < cascade_size; i++) { + cascade[i] = Filter::LowPass(cutoff, CascadingBiquadQ(cascade_size, i)); + } + return CascadingFilter{std::move(cascade)}; +} + +CascadingFilter::CascadingFilter() = default; +CascadingFilter::CascadingFilter(std::vector<Filter> filters) : filters(std::move(filters)) {} + +void CascadingFilter::Process(std::vector<s16>& signal) { + for (auto& filter : filters) { + filter.Process(signal); + } +} + +} // namespace AudioCore diff --git a/src/audio_core/algorithm/filter.h b/src/audio_core/algorithm/filter.h new file mode 100644 index 000000000..a41beef98 --- /dev/null +++ b/src/audio_core/algorithm/filter.h @@ -0,0 +1,62 @@ +// Copyright 2018 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" + +namespace AudioCore { + +/// Digital biquad filter: +/// +/// b0 + b1 z^-1 + b2 z^-2 +/// H(z) = ------------------------ +/// a0 + a1 z^-1 + b2 z^-2 +class Filter { +public: + /// Creates a low-pass filter. + /// @param cutoff Determines the cutoff frequency. A value from 0.0 to 1.0. + /// @param Q Determines the quality factor of this filter. + static Filter LowPass(double cutoff, double Q = 0.7071); + + /// Passthrough filter. + Filter(); + + Filter(double a0, double a1, double a2, double b0, double b1, double b2); + + void Process(std::vector<s16>& signal); + +private: + static constexpr size_t channel_count = 2; + + /// Coefficients are in normalized form (a0 = 1.0). + double a1, a2, b0, b1, b2; + /// Input History + std::array<std::array<double, channel_count>, 3> in; + /// Output History + std::array<std::array<double, channel_count>, 3> out; +}; + +/// Cascade filters to build up higher-order filters from lower-order ones. +class CascadingFilter { +public: + /// Creates a cascading low-pass filter. + /// @param cutoff Determines the cutoff frequency. A value from 0.0 to 1.0. + /// @param cascade_size Number of biquads in cascade. + static CascadingFilter LowPass(double cutoff, size_t cascade_size); + + /// Passthrough. + CascadingFilter(); + + explicit CascadingFilter(std::vector<Filter> filters); + + void Process(std::vector<s16>& signal); + +private: + std::vector<Filter> filters; +}; + +} // namespace AudioCore diff --git a/src/audio_core/algorithm/interpolate.cpp b/src/audio_core/algorithm/interpolate.cpp new file mode 100644 index 000000000..11459821f --- /dev/null +++ b/src/audio_core/algorithm/interpolate.cpp @@ -0,0 +1,71 @@ +// Copyright 2018 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#define _USE_MATH_DEFINES + +#include <algorithm> +#include <cmath> +#include <vector> +#include "audio_core/algorithm/interpolate.h" +#include "common/common_types.h" +#include "common/logging/log.h" + +namespace AudioCore { + +/// The Lanczos kernel +static double Lanczos(size_t a, double x) { + if (x == 0.0) + return 1.0; + const double px = M_PI * x; + return a * std::sin(px) * std::sin(px / a) / (px * px); +} + +std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, double ratio) { + if (input.size() < 2) + return {}; + + if (ratio <= 0) { + LOG_CRITICAL(Audio, "Nonsensical interpolation ratio {}", ratio); + ratio = 1.0; + } + + if (ratio != state.current_ratio) { + const double cutoff_frequency = std::min(0.5 / ratio, 0.5 * ratio); + state.nyquist = CascadingFilter::LowPass(std::clamp(cutoff_frequency, 0.0, 0.4), 3); + state.current_ratio = ratio; + } + state.nyquist.Process(input); + + constexpr size_t taps = InterpolationState::lanczos_taps; + const size_t num_frames = input.size() / 2; + + std::vector<s16> output; + output.reserve(static_cast<size_t>(input.size() / ratio + 4)); + + double& pos = state.position; + auto& h = state.history; + for (size_t i = 0; i < num_frames; ++i) { + std::rotate(h.begin(), h.end() - 1, h.end()); + h[0][0] = input[i * 2 + 0]; + h[0][1] = input[i * 2 + 1]; + + while (pos <= 1.0) { + double l = 0.0; + double r = 0.0; + for (size_t j = 0; j < h.size(); j++) { + l += Lanczos(taps, pos + j - taps + 1) * h[j][0]; + r += Lanczos(taps, pos + j - taps + 1) * h[j][1]; + } + output.emplace_back(static_cast<s16>(std::clamp(l, -32768.0, 32767.0))); + output.emplace_back(static_cast<s16>(std::clamp(r, -32768.0, 32767.0))); + + pos += ratio; + } + pos -= 1.0; + } + + return output; +} + +} // namespace AudioCore diff --git a/src/audio_core/algorithm/interpolate.h b/src/audio_core/algorithm/interpolate.h new file mode 100644 index 000000000..c79c2eef4 --- /dev/null +++ b/src/audio_core/algorithm/interpolate.h @@ -0,0 +1,43 @@ +// Copyright 2018 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/algorithm/filter.h" +#include "common/common_types.h" + +namespace AudioCore { + +struct InterpolationState { + static constexpr size_t lanczos_taps = 4; + static constexpr size_t history_size = lanczos_taps * 2 - 1; + + double current_ratio = 0.0; + CascadingFilter nyquist; + std::array<std::array<s16, 2>, history_size> history = {}; + double position = 0; +}; + +/// Interpolates input signal to produce output signal. +/// @param input The signal to interpolate. +/// @param ratio Interpolation ratio. +/// ratio > 1.0 results in fewer output samples. +/// ratio < 1.0 results in more output samples. +/// @returns Output signal. +std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, double ratio); + +/// Interpolates input signal to produce output signal. +/// @param input The signal to interpolate. +/// @param input_rate The sample rate of input. +/// @param output_rate The desired sample rate of the output. +/// @returns Output signal. +inline std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, + u32 input_rate, u32 output_rate) { + const double ratio = static_cast<double>(input_rate) / static_cast<double>(output_rate); + return Interpolate(state, std::move(input), ratio); +} + +} // namespace AudioCore diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp index 282f345c5..397b107f5 100644 --- a/src/audio_core/audio_renderer.cpp +++ b/src/audio_core/audio_renderer.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "audio_core/algorithm/interpolate.h" #include "audio_core/audio_renderer.h" #include "common/assert.h" #include "common/logging/log.h" @@ -26,6 +27,18 @@ AudioRenderer::AudioRenderer(AudioRendererParameter params, QueueMixedBuffer(2); } +u32 AudioRenderer::GetSampleRate() const { + return worker_params.sample_rate; +} + +u32 AudioRenderer::GetSampleCount() const { + return worker_params.sample_count; +} + +u32 AudioRenderer::GetMixBufferCount() const { + return worker_params.mix_buffer_count; +} + std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params) { // Copy UpdateDataHeader struct UpdateDataHeader config{}; @@ -187,6 +200,8 @@ void AudioRenderer::VoiceState::RefreshBuffer() { break; } + samples = Interpolate(interp_state, std::move(samples), Info().sample_rate, STREAM_SAMPLE_RATE); + is_refresh_pending = false; } @@ -212,7 +227,7 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) { break; } - samples_remaining -= samples.size(); + samples_remaining -= samples.size() / stream->GetNumChannels(); for (const auto& sample : samples) { const s32 buffer_sample{buffer[offset]}; diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h index 6950a4681..eba67f28e 100644 --- a/src/audio_core/audio_renderer.h +++ b/src/audio_core/audio_renderer.h @@ -8,6 +8,7 @@ #include <memory> #include <vector> +#include "audio_core/algorithm/interpolate.h" #include "audio_core/audio_out.h" #include "audio_core/codec.h" #include "audio_core/stream.h" @@ -26,7 +27,7 @@ enum class PlayState : u8 { struct AudioRendererParameter { u32_le sample_rate; u32_le sample_count; - u32_le unknown_8; + u32_le mix_buffer_count; u32_le unknown_c; u32_le voice_count; u32_le sink_count; @@ -160,6 +161,9 @@ public: std::vector<u8> UpdateAudioRenderer(const std::vector<u8>& input_params); void QueueMixedBuffer(Buffer::Tag tag); void ReleaseAndQueueBuffers(); + u32 GetSampleRate() const; + u32 GetSampleCount() const; + u32 GetMixBufferCount() const; private: class VoiceState { @@ -191,6 +195,7 @@ private: size_t wave_index{}; size_t offset{}; Codec::ADPCMState adpcm_state{}; + InterpolationState interp_state{}; std::vector<s16> samples; VoiceOutStatus out_status{}; VoiceInfo info{}; diff --git a/src/audio_core/cubeb_sink.cpp b/src/audio_core/cubeb_sink.cpp index 1501ef1f4..5a1177d0c 100644 --- a/src/audio_core/cubeb_sink.cpp +++ b/src/audio_core/cubeb_sink.cpp @@ -4,6 +4,7 @@ #include <algorithm> #include <cstring> +#include <mutex> #include "audio_core/cubeb_sink.h" #include "audio_core/stream.h" @@ -66,6 +67,8 @@ public: return; } + std::lock_guard lock{queue_mutex}; + queue.reserve(queue.size() + samples.size() * GetNumChannels()); if (is_6_channel) { @@ -94,6 +97,7 @@ private: u32 num_channels{}; bool is_6_channel{}; + std::mutex queue_mutex; std::vector<s16> queue; static long DataCallback(cubeb_stream* stream, void* user_data, const void* input_buffer, @@ -153,6 +157,8 @@ long SinkStreamImpl::DataCallback(cubeb_stream* stream, void* user_data, const v return {}; } + std::lock_guard lock{impl->queue_mutex}; + const size_t frames_to_write{ std::min(impl->queue.size() / impl->GetNumChannels(), static_cast<size_t>(num_frames))}; diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt index d5d4f6f82..d9424ea91 100644 --- a/src/common/CMakeLists.txt +++ b/src/common/CMakeLists.txt @@ -29,8 +29,6 @@ add_library(common STATIC assert.h bit_field.h bit_set.h - break_points.cpp - break_points.h cityhash.cpp cityhash.h color.h @@ -40,6 +38,8 @@ add_library(common STATIC file_util.cpp file_util.h hash.h + hex_util.cpp + hex_util.h logging/backend.cpp logging/backend.h logging/filter.cpp diff --git a/src/common/break_points.cpp b/src/common/break_points.cpp deleted file mode 100644 index fa367a4ca..000000000 --- a/src/common/break_points.cpp +++ /dev/null @@ -1,90 +0,0 @@ -// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <sstream> -#include "common/break_points.h" - -bool BreakPoints::IsAddressBreakPoint(u32 iAddress) const { - auto cond = [&iAddress](const TBreakPoint& bp) { return bp.iAddress == iAddress; }; - auto it = std::find_if(m_BreakPoints.begin(), m_BreakPoints.end(), cond); - return it != m_BreakPoints.end(); -} - -bool BreakPoints::IsTempBreakPoint(u32 iAddress) const { - auto cond = [&iAddress](const TBreakPoint& bp) { - return bp.iAddress == iAddress && bp.bTemporary; - }; - auto it = std::find_if(m_BreakPoints.begin(), m_BreakPoints.end(), cond); - return it != m_BreakPoints.end(); -} - -BreakPoints::TBreakPointsStr BreakPoints::GetStrings() const { - TBreakPointsStr bps; - for (auto breakpoint : m_BreakPoints) { - if (!breakpoint.bTemporary) { - std::stringstream bp; - bp << std::hex << breakpoint.iAddress << " " << (breakpoint.bOn ? "n" : ""); - bps.push_back(bp.str()); - } - } - - return bps; -} - -void BreakPoints::AddFromStrings(const TBreakPointsStr& bps) { - for (auto bps_item : bps) { - TBreakPoint bp; - std::stringstream bpstr; - bpstr << std::hex << bps_item; - bpstr >> bp.iAddress; - bp.bOn = bps_item.find("n") != bps_item.npos; - bp.bTemporary = false; - Add(bp); - } -} - -void BreakPoints::Add(const TBreakPoint& bp) { - if (!IsAddressBreakPoint(bp.iAddress)) { - m_BreakPoints.push_back(bp); - // if (jit) - // jit->GetBlockCache()->InvalidateICache(bp.iAddress, 4); - } -} - -void BreakPoints::Add(u32 em_address, bool temp) { - if (!IsAddressBreakPoint(em_address)) // only add new addresses - { - TBreakPoint pt; // breakpoint settings - pt.bOn = true; - pt.bTemporary = temp; - pt.iAddress = em_address; - - m_BreakPoints.push_back(pt); - - // if (jit) - // jit->GetBlockCache()->InvalidateICache(em_address, 4); - } -} - -void BreakPoints::Remove(u32 em_address) { - auto cond = [&em_address](const TBreakPoint& bp) { return bp.iAddress == em_address; }; - auto it = std::find_if(m_BreakPoints.begin(), m_BreakPoints.end(), cond); - if (it != m_BreakPoints.end()) - m_BreakPoints.erase(it); -} - -void BreakPoints::Clear() { - // if (jit) - //{ - // std::for_each(m_BreakPoints.begin(), m_BreakPoints.end(), - // [](const TBreakPoint& bp) - // { - // jit->GetBlockCache()->InvalidateICache(bp.iAddress, 4); - // } - // ); - //} - - m_BreakPoints.clear(); -} diff --git a/src/common/break_points.h b/src/common/break_points.h deleted file mode 100644 index e15b9f842..000000000 --- a/src/common/break_points.h +++ /dev/null @@ -1,49 +0,0 @@ -// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <string> -#include <vector> -#include "common/common_types.h" - -class DebugInterface; - -struct TBreakPoint { - u32 iAddress; - bool bOn; - bool bTemporary; -}; - -// Code breakpoints. -class BreakPoints { -public: - typedef std::vector<TBreakPoint> TBreakPoints; - typedef std::vector<std::string> TBreakPointsStr; - - const TBreakPoints& GetBreakPoints() { - return m_BreakPoints; - } - - TBreakPointsStr GetStrings() const; - void AddFromStrings(const TBreakPointsStr& bps); - - // is address breakpoint - bool IsAddressBreakPoint(u32 iAddress) const; - bool IsTempBreakPoint(u32 iAddress) const; - - // Add BreakPoint - void Add(u32 em_address, bool temp = false); - void Add(const TBreakPoint& bp); - - // Remove Breakpoint - void Remove(u32 iAddress); - void Clear(); - - void DeleteByAddress(u32 Address); - -private: - TBreakPoints m_BreakPoints; - u32 m_iBreakOnCount; -}; diff --git a/src/common/file_util.cpp b/src/common/file_util.cpp index 3ce590062..b30a67ff9 100644 --- a/src/common/file_util.cpp +++ b/src/common/file_util.cpp @@ -750,6 +750,12 @@ std::string GetHactoolConfigurationPath() { #endif } +std::string GetNANDRegistrationDir(bool system) { + if (system) + return GetUserPath(UserPath::NANDDir) + "system/Contents/registered/"; + return GetUserPath(UserPath::NANDDir) + "user/Contents/registered/"; +} + size_t WriteStringToFile(bool text_file, const std::string& str, const char* filename) { return FileUtil::IOFile(filename, text_file ? "w" : "wb").WriteBytes(str.data(), str.size()); } diff --git a/src/common/file_util.h b/src/common/file_util.h index 2711872ae..2f13d0b6b 100644 --- a/src/common/file_util.h +++ b/src/common/file_util.h @@ -129,6 +129,8 @@ const std::string& GetUserPath(UserPath path, const std::string& new_path = ""); std::string GetHactoolConfigurationPath(); +std::string GetNANDRegistrationDir(bool system = false); + // Returns the path to where the sys file are std::string GetSysDirectory(); diff --git a/src/common/hex_util.cpp b/src/common/hex_util.cpp new file mode 100644 index 000000000..609144def --- /dev/null +++ b/src/common/hex_util.cpp @@ -0,0 +1,31 @@ +// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/hex_util.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; + throw std::logic_error("Invalid hex digit"); +} + +std::array<u8, 16> operator""_array16(const char* str, size_t len) { + if (len != 32) + throw std::logic_error("Not of correct size."); + return HexStringToArray<16>(str); +} + +std::array<u8, 32> operator""_array32(const char* str, size_t len) { + if (len != 64) + throw std::logic_error("Not of correct size."); + return HexStringToArray<32>(str); +} + +} // namespace Common diff --git a/src/common/hex_util.h b/src/common/hex_util.h new file mode 100644 index 000000000..5fb79bb72 --- /dev/null +++ b/src/common/hex_util.h @@ -0,0 +1,41 @@ +// Copyright 2013 Dolphin Emulator Project / 2014 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <cstddef> +#include <string> +#include <fmt/format.h> +#include "common/common_types.h" + +namespace Common { + +u8 ToHexNibble(char c1); + +template <size_t Size, bool le = false> +std::array<u8, Size> HexStringToArray(std::string_view str) { + std::array<u8, Size> out{}; + if constexpr (le) { + for (size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2) + out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]); + } else { + for (size_t i = 0; i < 2 * Size; i += 2) + out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]); + } + return out; +} + +template <size_t Size> +std::string HexArrayToString(std::array<u8, Size> array, bool upper = true) { + std::string out; + for (u8 c : array) + out += fmt::format(upper ? "{:02X}" : "{:02x}", c); + return out; +} + +std::array<u8, 0x10> operator"" _array16(const char* str, size_t len); +std::array<u8, 0x20> operator"" _array32(const char* str, size_t len); + +} // namespace Common diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp index e80784c3c..1323f8d0f 100644 --- a/src/common/logging/backend.cpp +++ b/src/common/logging/backend.cpp @@ -302,13 +302,14 @@ Backend* GetBackend(std::string_view backend_name) { void FmtLogMessageImpl(Class log_class, Level log_level, const char* filename, unsigned int line_num, const char* function, const char* format, const fmt::format_args& args) { - auto filter = Impl::Instance().GetGlobalFilter(); + auto& instance = Impl::Instance(); + const auto& filter = instance.GetGlobalFilter(); if (!filter.CheckMessage(log_class, log_level)) return; Entry entry = CreateEntry(log_class, log_level, filename, line_num, function, fmt::vformat(format, args)); - Impl::Instance().PushEntry(std::move(entry)); + instance.PushEntry(std::move(entry)); } } // namespace Log diff --git a/src/common/misc.cpp b/src/common/misc.cpp index 217a87098..3fa8a3bc4 100644 --- a/src/common/misc.cpp +++ b/src/common/misc.cpp @@ -4,7 +4,7 @@ #include <cstddef> #ifdef _WIN32 -#include <Windows.h> +#include <windows.h> #else #include <cerrno> #include <cstring> diff --git a/src/common/thread_queue_list.h b/src/common/thread_queue_list.h index 38a450d69..133122c5f 100644 --- a/src/common/thread_queue_list.h +++ b/src/common/thread_queue_list.h @@ -16,7 +16,7 @@ struct ThreadQueueList { // (dynamically resizable) circular buffers to remove their overhead when // inserting and popping. - typedef unsigned int Priority; + using Priority = unsigned int; // Number of priority levels. (Valid levels are [0..NUM_QUEUES).) static const Priority NUM_QUEUES = N; @@ -26,9 +26,9 @@ struct ThreadQueueList { } // Only for debugging, returns priority level. - Priority contains(const T& uid) { + Priority contains(const T& uid) const { for (Priority i = 0; i < NUM_QUEUES; ++i) { - Queue& cur = queues[i]; + const Queue& cur = queues[i]; if (std::find(cur.data.cbegin(), cur.data.cend(), uid) != cur.data.cend()) { return i; } @@ -37,8 +37,8 @@ struct ThreadQueueList { return -1; } - T get_first() { - Queue* cur = first; + T get_first() const { + const Queue* cur = first; while (cur != nullptr) { if (!cur->data.empty()) { return cur->data.front(); diff --git a/src/common/x64/xbyak_abi.h b/src/common/x64/xbyak_abi.h index fd3fbdd4b..927da9187 100644 --- a/src/common/x64/xbyak_abi.h +++ b/src/common/x64/xbyak_abi.h @@ -9,10 +9,9 @@ #include "common/assert.h" #include "common/bit_set.h" -namespace Common { -namespace X64 { +namespace Common::X64 { -int RegToIndex(const Xbyak::Reg& reg) { +inline int 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."); @@ -152,8 +151,8 @@ constexpr size_t ABI_SHADOW_SPACE = 0; #endif -void ABI_CalculateFrameSize(BitSet32 regs, size_t rsp_alignment, size_t needed_frame_size, - s32* out_subtraction, s32* out_xmm_offset) { +inline void ABI_CalculateFrameSize(BitSet32 regs, size_t rsp_alignment, size_t needed_frame_size, + s32* out_subtraction, s32* out_xmm_offset) { int count = (regs & ABI_ALL_GPRS).Count(); rsp_alignment -= count * 8; size_t subtraction = 0; @@ -174,8 +173,8 @@ void ABI_CalculateFrameSize(BitSet32 regs, size_t rsp_alignment, size_t needed_f *out_xmm_offset = (s32)(subtraction - xmm_base_subtraction); } -size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs, - size_t rsp_alignment, size_t needed_frame_size = 0) { +inline size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs, + size_t rsp_alignment, size_t needed_frame_size = 0) { s32 subtraction, xmm_offset; ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size, &subtraction, &xmm_offset); @@ -195,8 +194,8 @@ size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs return ABI_SHADOW_SPACE; } -void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs, size_t rsp_alignment, - size_t needed_frame_size = 0) { +inline void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs, + size_t rsp_alignment, size_t needed_frame_size = 0) { s32 subtraction, xmm_offset; ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size, &subtraction, &xmm_offset); @@ -217,5 +216,4 @@ void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, BitSet32 regs, s } } -} // namespace X64 -} // namespace Common +} // namespace Common::X64 diff --git a/src/common/x64/xbyak_util.h b/src/common/x64/xbyak_util.h index ec76e0a47..02323a017 100644 --- a/src/common/x64/xbyak_util.h +++ b/src/common/x64/xbyak_util.h @@ -8,8 +8,7 @@ #include <xbyak.h> #include "common/x64/xbyak_abi.h" -namespace Common { -namespace X64 { +namespace Common::X64 { // Constants for use with cmpps/cmpss enum { @@ -45,5 +44,4 @@ inline void CallFarFunction(Xbyak::CodeGenerator& code, const T f) { } } -} // namespace X64 -} // namespace Common +} // namespace Common::X64 diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index 4d39ba409..31a7bf6fd 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -20,6 +20,8 @@ add_library(core STATIC crypto/key_manager.h crypto/ctr_encryption_layer.cpp crypto/ctr_encryption_layer.h + file_sys/bis_factory.cpp + file_sys/bis_factory.h file_sys/card_image.cpp file_sys/card_image.h file_sys/content_archive.cpp @@ -29,10 +31,14 @@ add_library(core STATIC file_sys/directory.h file_sys/errors.h file_sys/mode.h + file_sys/nca_metadata.cpp + file_sys/nca_metadata.h file_sys/partition_filesystem.cpp file_sys/partition_filesystem.h file_sys/program_metadata.cpp file_sys/program_metadata.h + file_sys/registered_cache.cpp + file_sys/registered_cache.h file_sys/romfs.cpp file_sys/romfs.h file_sys/romfs_factory.cpp @@ -43,6 +49,8 @@ add_library(core STATIC file_sys/sdmc_factory.h file_sys/vfs.cpp file_sys/vfs.h + file_sys/vfs_concat.cpp + file_sys/vfs_concat.h file_sys/vfs_offset.cpp file_sys/vfs_offset.h file_sys/vfs_real.cpp @@ -251,6 +259,10 @@ 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_nvjpg.cpp + hle/service/nvdrv/devices/nvhost_nvjpg.h + hle/service/nvdrv/devices/nvhost_vic.cpp + hle/service/nvdrv/devices/nvhost_vic.h hle/service/nvdrv/devices/nvmap.cpp hle/service/nvdrv/devices/nvmap.h hle/service/nvdrv/interface.cpp diff --git a/src/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp index ceb3f7683..2c817d7d1 100644 --- a/src/core/arm/dynarmic/arm_dynarmic.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic.cpp @@ -86,7 +86,16 @@ public: } void AddTicks(u64 ticks) override { - CoreTiming::AddTicks(ticks - num_interpreted_instructions); + // 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; + // Always execute at least one tick. + amortized_ticks = std::max<u64>(amortized_ticks, 1); + + CoreTiming::AddTicks(amortized_ticks); num_interpreted_instructions = 0; } u64 GetTicksRemaining() override { @@ -125,6 +134,9 @@ std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const { config.dczid_el0 = 4; config.ctr_el0 = 0x8444c004; + // Unpredictable instructions + config.define_unpredictable_behaviour = true; + return std::make_unique<Dynarmic::A64::Jit>(config); } @@ -234,9 +246,7 @@ void ARM_Dynarmic::LoadContext(const ThreadContext& ctx) { } void ARM_Dynarmic::PrepareReschedule() { - if (jit->IsExecuting()) { - jit->HaltExecution(); - } + jit->HaltExecution(); } void ARM_Dynarmic::ClearInstructionCache() { diff --git a/src/core/core.cpp b/src/core/core.cpp index 69c45c026..28038ff6f 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -5,6 +5,7 @@ #include <memory> #include <utility> #include "common/logging/log.h" +#include "common/string_util.h" #include "core/core.h" #include "core/core_timing.h" #include "core/gdbstub/gdbstub.h" @@ -17,6 +18,7 @@ #include "core/hle/service/sm/sm.h" #include "core/loader/loader.h" #include "core/settings.h" +#include "file_sys/vfs_concat.h" #include "file_sys/vfs_real.h" #include "video_core/renderer_base.h" #include "video_core/video_core.h" @@ -88,8 +90,39 @@ System::ResultStatus System::SingleStep() { return RunLoop(false); } -System::ResultStatus System::Load(EmuWindow& emu_window, const std::string& filepath) { - app_loader = Loader::GetLoader(virtual_filesystem->OpenFile(filepath, FileSys::Mode::Read)); +static FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs, + const std::string& path) { + // To account for split 00+01+etc files. + std::string dir_name; + std::string filename; + Common::SplitPath(path, &dir_name, &filename, nullptr); + if (filename == "00") { + const auto dir = vfs->OpenDirectory(dir_name, FileSys::Mode::Read); + std::vector<FileSys::VirtualFile> concat; + for (u8 i = 0; i < 0x10; ++i) { + auto next = dir->GetFile(fmt::format("{:02X}", i)); + if (next != nullptr) + concat.push_back(std::move(next)); + else { + next = dir->GetFile(fmt::format("{:02x}", i)); + if (next != nullptr) + concat.push_back(std::move(next)); + else + break; + } + } + + if (concat.empty()) + return nullptr; + + return FileSys::ConcatenateFiles(concat, dir->GetName()); + } + + return vfs->OpenFile(path, FileSys::Mode::Read); +} + +System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath) { + app_loader = Loader::GetLoader(GetGameFileFromPath(virtual_filesystem, filepath)); if (!app_loader) { LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath); @@ -102,18 +135,8 @@ System::ResultStatus System::Load(EmuWindow& emu_window, const std::string& file LOG_CRITICAL(Core, "Failed to determine system mode (Error {})!", static_cast<int>(system_mode.second)); - switch (system_mode.second) { - case Loader::ResultStatus::ErrorMissingKeys: - return ResultStatus::ErrorLoader_ErrorMissingKeys; - case Loader::ResultStatus::ErrorDecrypting: - return ResultStatus::ErrorLoader_ErrorDecrypting; - case Loader::ResultStatus::ErrorInvalidFormat: - return ResultStatus::ErrorLoader_ErrorInvalidFormat; - case Loader::ResultStatus::ErrorUnsupportedArch: - return ResultStatus::ErrorUnsupportedArch; - default: + if (system_mode.second != Loader::ResultStatus::Success) return ResultStatus::ErrorSystemMode; - } } ResultStatus init_result{Init(emu_window)}; @@ -129,17 +152,9 @@ System::ResultStatus System::Load(EmuWindow& emu_window, const std::string& file LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", static_cast<int>(load_result)); System::Shutdown(); - switch (load_result) { - case Loader::ResultStatus::ErrorMissingKeys: - return ResultStatus::ErrorLoader_ErrorMissingKeys; - case Loader::ResultStatus::ErrorDecrypting: - return ResultStatus::ErrorLoader_ErrorDecrypting; - case Loader::ResultStatus::ErrorInvalidFormat: - return ResultStatus::ErrorLoader_ErrorInvalidFormat; - case Loader::ResultStatus::ErrorUnsupportedArch: - return ResultStatus::ErrorUnsupportedArch; - default: - return ResultStatus::ErrorLoader; + if (load_result != Loader::ResultStatus::Success) { + return static_cast<ResultStatus>(static_cast<u32>(ResultStatus::ErrorLoader) + + static_cast<u32>(load_result)); } } status = ResultStatus::Success; @@ -169,7 +184,7 @@ Cpu& System::CpuCore(size_t core_index) { return *cpu_cores[core_index]; } -System::ResultStatus System::Init(EmuWindow& emu_window) { +System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) { LOG_DEBUG(HW_Memory, "initialized OK"); CoreTiming::Init(); diff --git a/src/core/core.h b/src/core/core.h index 7cf7ea4e1..790e23cae 100644 --- a/src/core/core.h +++ b/src/core/core.h @@ -22,9 +22,12 @@ #include "video_core/debug_utils/debug_utils.h" #include "video_core/gpu.h" -class EmuWindow; class ARM_Interface; +namespace Core::Frontend { +class EmuWindow; +} + namespace Service::SM { class ServiceManager; } @@ -37,6 +40,12 @@ namespace Core { class System { public: + System(const System&) = delete; + System& operator=(const System&) = delete; + + System(System&&) = delete; + System& operator=(System&&) = delete; + ~System(); /** @@ -49,21 +58,15 @@ public: /// Enumeration representing the return values of the System Initialize and Load process. enum class ResultStatus : u32 { - Success, ///< Succeeded - ErrorNotInitialized, ///< Error trying to use core prior to initialization - ErrorGetLoader, ///< Error finding the correct application loader - ErrorSystemMode, ///< Error determining the system mode - ErrorLoader, ///< Error loading the specified application - ErrorLoader_ErrorMissingKeys, ///< Error because the key/keys needed to run could not be - ///< found. - ErrorLoader_ErrorDecrypting, ///< Error loading the specified application due to encryption - ErrorLoader_ErrorInvalidFormat, ///< Error loading the specified application due to an - /// invalid format - ErrorSystemFiles, ///< Error in finding system files - ErrorSharedFont, ///< Error in finding shared font - ErrorVideoCore, ///< Error in the video core - ErrorUnsupportedArch, ///< Unsupported Architecture (32-Bit ROMs) - ErrorUnknown ///< Any other error + Success, ///< Succeeded + ErrorNotInitialized, ///< Error trying to use core prior to initialization + ErrorGetLoader, ///< Error finding the correct application loader + ErrorSystemMode, ///< Error determining the system mode + ErrorSystemFiles, ///< Error in finding system files + ErrorSharedFont, ///< Error in finding shared font + ErrorVideoCore, ///< Error in the video core + ErrorUnknown, ///< Any other error + ErrorLoader, ///< The base for loader errors (too many to repeat) }; /** @@ -105,7 +108,7 @@ 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(EmuWindow& emu_window, const std::string& filepath); + 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 @@ -233,7 +236,7 @@ private: * input. * @return ResultStatus code, indicating if the operation succeeded. */ - ResultStatus Init(EmuWindow& emu_window); + ResultStatus Init(Frontend::EmuWindow& emu_window); /// RealVfsFilesystem instance FileSys::VirtualFilesystem virtual_filesystem; diff --git a/src/core/core_cpu.cpp b/src/core/core_cpu.cpp index 46a522fcd..b042ee02b 100644 --- a/src/core/core_cpu.cpp +++ b/src/core/core_cpu.cpp @@ -14,6 +14,7 @@ #include "core/core_timing.h" #include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" +#include "core/hle/lock.h" #include "core/settings.h" namespace Core { @@ -90,6 +91,7 @@ void Cpu::RunLoop(bool tight_loop) { LOG_TRACE(Core, "Core-{} idling", core_index); if (IsMainCore()) { + // TODO(Subv): Only let CoreTiming idle if all 4 cores are idling. CoreTiming::Idle(); CoreTiming::Advance(); } @@ -125,6 +127,8 @@ void Cpu::Reschedule() { } reschedule_pending = false; + // Lock the global kernel mutex when we manipulate the HLE state + std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock); scheduler->Reschedule(); } diff --git a/src/core/core_cpu.h b/src/core/core_cpu.h index 976952903..56cdae194 100644 --- a/src/core/core_cpu.h +++ b/src/core/core_cpu.h @@ -79,7 +79,7 @@ private: std::shared_ptr<CpuBarrier> cpu_barrier; std::shared_ptr<Kernel::Scheduler> scheduler; - bool reschedule_pending{}; + std::atomic<bool> reschedule_pending = false; size_t core_index; }; diff --git a/src/core/core_timing.cpp b/src/core/core_timing.cpp index d3bb6f818..7953c8720 100644 --- a/src/core/core_timing.cpp +++ b/src/core/core_timing.cpp @@ -56,6 +56,9 @@ static u64 event_fifo_id; // to the event_queue by the emu thread static Common::MPSCQueue<Event, false> ts_queue; +// the queue for unscheduling the events from other threads threadsafe +static Common::MPSCQueue<std::pair<const EventType*, u64>, false> unschedule_queue; + constexpr int MAX_SLICE_LENGTH = 20000; static s64 idled_cycles; @@ -135,11 +138,9 @@ void ClearPendingEvents() { void ScheduleEvent(s64 cycles_into_future, const EventType* event_type, u64 userdata) { ASSERT(event_type != nullptr); s64 timeout = GetTicks() + cycles_into_future; - // If this event needs to be scheduled before the next advance(), force one early if (!is_global_timer_sane) ForceExceptionCheck(cycles_into_future); - event_queue.emplace_back(Event{timeout, event_fifo_id++, userdata, event_type}); std::push_heap(event_queue.begin(), event_queue.end(), std::greater<>()); } @@ -160,6 +161,10 @@ void UnscheduleEvent(const EventType* event_type, u64 userdata) { } } +void UnscheduleEventThreadsafe(const EventType* event_type, u64 userdata) { + unschedule_queue.Push(std::make_pair(event_type, userdata)); +} + void RemoveEvent(const EventType* event_type) { auto itr = std::remove_if(event_queue.begin(), event_queue.end(), [&](const Event& e) { return e.type == event_type; }); @@ -196,6 +201,9 @@ void MoveEvents() { void Advance() { MoveEvents(); + for (std::pair<const EventType*, u64> ev; unschedule_queue.Pop(ev);) { + UnscheduleEvent(ev.first, ev.second); + } int cycles_executed = slice_length - downcount; global_timer += cycles_executed; diff --git a/src/core/core_timing.h b/src/core/core_timing.h index dfa161c0d..9ed757bd7 100644 --- a/src/core/core_timing.h +++ b/src/core/core_timing.h @@ -65,6 +65,7 @@ void ScheduleEvent(s64 cycles_into_future, const EventType* event_type, u64 user void ScheduleEventThreadsafe(s64 cycles_into_future, const EventType* event_type, u64 userdata); void UnscheduleEvent(const EventType* event_type, u64 userdata); +void UnscheduleEventThreadsafe(const EventType* event_type, u64 userdata); /// We only permit one event of each type in the queue at a time. void RemoveEvent(const EventType* event_type); diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp index fc45e7ab5..db8b22c85 100644 --- a/src/core/crypto/key_manager.cpp +++ b/src/core/crypto/key_manager.cpp @@ -10,44 +10,13 @@ #include <string_view> #include "common/common_paths.h" #include "common/file_util.h" +#include "common/hex_util.h" +#include "common/logging/log.h" #include "core/crypto/key_manager.h" #include "core/settings.h" namespace Core::Crypto { -static 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; - throw std::logic_error("Invalid hex digit"); -} - -template <size_t Size> -static std::array<u8, Size> HexStringToArray(std::string_view str) { - std::array<u8, Size> out{}; - for (size_t i = 0; i < 2 * Size; i += 2) { - auto d1 = str[i]; - auto d2 = str[i + 1]; - out[i / 2] = (ToHexNibble(d1) << 4) | ToHexNibble(d2); - } - return out; -} - -std::array<u8, 16> operator""_array16(const char* str, size_t len) { - if (len != 32) - throw std::logic_error("Not of correct size."); - return HexStringToArray<16>(str); -} - -std::array<u8, 32> operator""_array32(const char* str, size_t len) { - if (len != 64) - throw std::logic_error("Not of correct size."); - return HexStringToArray<32>(str); -} - KeyManager::KeyManager() { // Initialize keys const std::string hactool_keys_dir = FileUtil::GetHactoolConfigurationPath(); @@ -83,20 +52,20 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) { out[1].erase(std::remove(out[1].begin(), out[1].end(), ' '), out[1].end()); if (is_title_keys) { - auto rights_id_raw = HexStringToArray<16>(out[0]); + auto rights_id_raw = Common::HexStringToArray<16>(out[0]); u128 rights_id{}; std::memcpy(rights_id.data(), rights_id_raw.data(), rights_id_raw.size()); - Key128 key = HexStringToArray<16>(out[1]); + Key128 key = Common::HexStringToArray<16>(out[1]); SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); } else { std::transform(out[0].begin(), out[0].end(), out[0].begin(), ::tolower); if (s128_file_id.find(out[0]) != s128_file_id.end()) { const auto index = s128_file_id.at(out[0]); - Key128 key = HexStringToArray<16>(out[1]); + Key128 key = Common::HexStringToArray<16>(out[1]); SetKey(index.type, key, index.field1, index.field2); } else if (s256_file_id.find(out[0]) != s256_file_id.end()) { const auto index = s256_file_id.at(out[0]); - Key256 key = HexStringToArray<32>(out[1]); + Key256 key = Common::HexStringToArray<32>(out[1]); SetKey(index.type, key, index.field1, index.field2); } } diff --git a/src/core/crypto/key_manager.h b/src/core/crypto/key_manager.h index c4c53cefc..0c62d4421 100644 --- a/src/core/crypto/key_manager.h +++ b/src/core/crypto/key_manager.h @@ -87,9 +87,6 @@ struct hash<Core::Crypto::KeyIndex<KeyType>> { namespace Core::Crypto { -std::array<u8, 0x10> operator"" _array16(const char* str, size_t len); -std::array<u8, 0x20> operator"" _array32(const char* str, size_t len); - class KeyManager { public: KeyManager(); diff --git a/src/core/file_sys/bis_factory.cpp b/src/core/file_sys/bis_factory.cpp new file mode 100644 index 000000000..ae4e33800 --- /dev/null +++ b/src/core/file_sys/bis_factory.cpp @@ -0,0 +1,31 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/file_sys/bis_factory.h" + +namespace FileSys { + +static VirtualDir GetOrCreateDirectory(const VirtualDir& dir, std::string_view path) { + const auto res = dir->GetDirectoryRelative(path); + if (res == nullptr) + return dir->CreateDirectoryRelative(path); + return res; +} + +BISFactory::BISFactory(VirtualDir nand_root_) + : nand_root(std::move(nand_root_)), + sysnand_cache(std::make_shared<RegisteredCache>( + GetOrCreateDirectory(nand_root, "/system/Contents/registered"))), + usrnand_cache(std::make_shared<RegisteredCache>( + GetOrCreateDirectory(nand_root, "/user/Contents/registered"))) {} + +std::shared_ptr<RegisteredCache> BISFactory::GetSystemNANDContents() const { + return sysnand_cache; +} + +std::shared_ptr<RegisteredCache> BISFactory::GetUserNANDContents() const { + return usrnand_cache; +} + +} // namespace FileSys diff --git a/src/core/file_sys/bis_factory.h b/src/core/file_sys/bis_factory.h new file mode 100644 index 000000000..a970a5e2e --- /dev/null +++ b/src/core/file_sys/bis_factory.h @@ -0,0 +1,30 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include "core/loader/loader.h" +#include "registered_cache.h" + +namespace FileSys { + +/// File system interface to the Built-In Storage +/// This is currently missing accessors to BIS partitions, but seemed like a good place for the NAND +/// registered caches. +class BISFactory { +public: + explicit BISFactory(VirtualDir nand_root); + + std::shared_ptr<RegisteredCache> GetSystemNANDContents() const; + std::shared_ptr<RegisteredCache> GetUserNANDContents() const; + +private: + VirtualDir nand_root; + + std::shared_ptr<RegisteredCache> sysnand_cache; + std::shared_ptr<RegisteredCache> usrnand_cache; +}; + +} // namespace FileSys diff --git a/src/core/file_sys/card_image.cpp b/src/core/file_sys/card_image.cpp index e897d9913..1d7c7fb10 100644 --- a/src/core/file_sys/card_image.cpp +++ b/src/core/file_sys/card_image.cpp @@ -4,22 +4,27 @@ #include <array> #include <string> -#include <core/loader/loader.h> + +#include <fmt/ostream.h> + #include "common/logging/log.h" #include "core/file_sys/card_image.h" #include "core/file_sys/partition_filesystem.h" #include "core/file_sys/vfs_offset.h" +#include "core/loader/loader.h" namespace FileSys { +constexpr std::array<const char*, 0x4> partition_names = {"update", "normal", "secure", "logo"}; + XCI::XCI(VirtualFile file_) : file(std::move(file_)), partitions(0x4) { if (file->ReadObject(&header) != sizeof(GamecardHeader)) { - status = Loader::ResultStatus::ErrorInvalidFormat; + status = Loader::ResultStatus::ErrorBadXCIHeader; return; } if (header.magic != Common::MakeMagic('H', 'E', 'A', 'D')) { - status = Loader::ResultStatus::ErrorInvalidFormat; + status = Loader::ResultStatus::ErrorBadXCIHeader; return; } @@ -31,9 +36,6 @@ XCI::XCI(VirtualFile file_) : file(std::move(file_)), partitions(0x4) { return; } - static constexpr std::array<const char*, 0x4> partition_names = {"update", "normal", "secure", - "logo"}; - for (XCIPartition partition : {XCIPartition::Update, XCIPartition::Normal, XCIPartition::Secure, XCIPartition::Logo}) { auto raw = main_hfs.GetFile(partition_names[static_cast<size_t>(partition)]); @@ -94,6 +96,10 @@ VirtualDir XCI::GetLogoPartition() const { return GetPartition(XCIPartition::Logo); } +const std::vector<std::shared_ptr<NCA>>& XCI::GetNCAs() const { + return ncas; +} + std::shared_ptr<NCA> XCI::GetNCAByType(NCAContentType type) const { const auto iter = std::find_if(ncas.begin(), ncas.end(), @@ -108,19 +114,19 @@ VirtualFile XCI::GetNCAFileByType(NCAContentType type) const { return nullptr; } -std::vector<std::shared_ptr<VfsFile>> XCI::GetFiles() const { +std::vector<VirtualFile> XCI::GetFiles() const { return {}; } -std::vector<std::shared_ptr<VfsDirectory>> XCI::GetSubdirectories() const { - return std::vector<std::shared_ptr<VfsDirectory>>(); +std::vector<VirtualDir> XCI::GetSubdirectories() const { + return {}; } std::string XCI::GetName() const { return file->GetName(); } -std::shared_ptr<VfsDirectory> XCI::GetParentDirectory() const { +VirtualDir XCI::GetParentDirectory() const { return file->GetContainingDirectory(); } @@ -130,15 +136,21 @@ bool XCI::ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) { Loader::ResultStatus XCI::AddNCAFromPartition(XCIPartition part) { if (partitions[static_cast<size_t>(part)] == nullptr) { - return Loader::ResultStatus::ErrorInvalidFormat; + return Loader::ResultStatus::ErrorXCIMissingPartition; } for (const VirtualFile& file : partitions[static_cast<size_t>(part)]->GetFiles()) { if (file->GetExtension() != "nca") continue; auto nca = std::make_shared<NCA>(file); - if (nca->GetStatus() == Loader::ResultStatus::Success) + if (nca->GetStatus() == Loader::ResultStatus::Success) { ncas.push_back(std::move(nca)); + } else { + const u16 error_id = static_cast<u16>(nca->GetStatus()); + LOG_CRITICAL(Loader, "Could not load NCA {}/{}, failed with error code {:04X} ({})", + partition_names[static_cast<size_t>(part)], nca->GetName(), error_id, + nca->GetStatus()); + } } return Loader::ResultStatus::Success; diff --git a/src/core/file_sys/card_image.h b/src/core/file_sys/card_image.h index e089d737c..a03d5264e 100644 --- a/src/core/file_sys/card_image.h +++ b/src/core/file_sys/card_image.h @@ -68,16 +68,17 @@ public: VirtualDir GetUpdatePartition() const; VirtualDir GetLogoPartition() const; + const std::vector<std::shared_ptr<NCA>>& GetNCAs() const; std::shared_ptr<NCA> GetNCAByType(NCAContentType type) const; VirtualFile GetNCAFileByType(NCAContentType type) const; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; + std::vector<VirtualFile> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualDir> GetSubdirectories() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; + VirtualDir GetParentDirectory() const override; protected: bool ReplaceFileWithSubdirectory(VirtualFile file, VirtualDir dir) override; diff --git a/src/core/file_sys/content_archive.cpp b/src/core/file_sys/content_archive.cpp index d3007d981..47afcad9b 100644 --- a/src/core/file_sys/content_archive.cpp +++ b/src/core/file_sys/content_archive.cpp @@ -113,17 +113,27 @@ boost::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType ty return out; } -boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() const { +boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() { const auto master_key_id = GetCryptoRevision(); u128 rights_id{}; memcpy(rights_id.data(), header.rights_id.data(), 16); - if (rights_id == u128{}) + if (rights_id == u128{}) { + status = Loader::ResultStatus::ErrorInvalidRightsID; return boost::none; + } auto titlekey = keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]); - if (titlekey == Core::Crypto::Key128{}) + if (titlekey == Core::Crypto::Key128{}) { + status = Loader::ResultStatus::ErrorMissingTitlekey; + return boost::none; + } + + if (!keys.HasKey(Core::Crypto::S128KeyType::Titlekek, master_key_id)) { + status = Loader::ResultStatus::ErrorMissingTitlekek; return boost::none; + } + Core::Crypto::AESCipher<Core::Crypto::Key128> cipher( keys.GetKey(Core::Crypto::S128KeyType::Titlekek, master_key_id), Core::Crypto::Mode::ECB); cipher.Transcode(titlekey.data(), titlekey.size(), titlekey.data(), Core::Crypto::Op::Decrypt); @@ -131,7 +141,7 @@ boost::optional<Core::Crypto::Key128> NCA::GetTitlekey() const { return titlekey; } -VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting_offset) const { +VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting_offset) { if (!encrypted) return in; @@ -143,15 +153,22 @@ VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting LOG_DEBUG(Crypto, "called with mode=CTR, starting_offset={:016X}", starting_offset); { boost::optional<Core::Crypto::Key128> key = boost::none; - if (std::find_if_not(header.rights_id.begin(), header.rights_id.end(), - [](char c) { return c == 0; }) == header.rights_id.end()) { - key = GetKeyAreaKey(NCASectionCryptoType::CTR); - } else { + if (has_rights_id) { + status = Loader::ResultStatus::Success; key = GetTitlekey(); + if (key == boost::none) { + if (status == Loader::ResultStatus::Success) + status = Loader::ResultStatus::ErrorMissingTitlekey; + return nullptr; + } + } else { + key = GetKeyAreaKey(NCASectionCryptoType::CTR); + if (key == boost::none) { + status = Loader::ResultStatus::ErrorMissingKeyAreaKey; + return nullptr; + } } - if (key == boost::none) - return nullptr; auto out = std::make_shared<Core::Crypto::CTREncryptionLayer>( std::move(in), key.value(), starting_offset); std::vector<u8> iv(16); @@ -170,16 +187,31 @@ VirtualFile NCA::Decrypt(NCASectionHeader s_header, VirtualFile in, u64 starting } NCA::NCA(VirtualFile file_) : file(std::move(file_)) { + status = Loader::ResultStatus::Success; + if (file == nullptr) { - status = Loader::ResultStatus::ErrorInvalidFormat; + status = Loader::ResultStatus::ErrorNullFile; return; } - if (sizeof(NCAHeader) != file->ReadObject(&header)) + + if (sizeof(NCAHeader) != file->ReadObject(&header)) { LOG_ERROR(Loader, "File reader errored out during header read."); + status = Loader::ResultStatus::ErrorBadNCAHeader; + return; + } encrypted = false; if (!IsValidNCA(header)) { + if (header.magic == Common::MakeMagic('N', 'C', 'A', '2')) { + status = Loader::ResultStatus::ErrorNCA2; + return; + } + if (header.magic == Common::MakeMagic('N', 'C', 'A', '0')) { + status = Loader::ResultStatus::ErrorNCA0; + return; + } + NCAHeader dec_header{}; Core::Crypto::AESCipher<Core::Crypto::Key256> cipher( keys.GetKey(Core::Crypto::S256KeyType::Header), Core::Crypto::Mode::XTS); @@ -189,14 +221,26 @@ NCA::NCA(VirtualFile file_) : file(std::move(file_)) { header = dec_header; encrypted = true; } else { + if (dec_header.magic == Common::MakeMagic('N', 'C', 'A', '2')) { + status = Loader::ResultStatus::ErrorNCA2; + return; + } + if (dec_header.magic == Common::MakeMagic('N', 'C', 'A', '0')) { + status = Loader::ResultStatus::ErrorNCA0; + return; + } + if (!keys.HasKey(Core::Crypto::S256KeyType::Header)) - status = Loader::ResultStatus::ErrorMissingKeys; + status = Loader::ResultStatus::ErrorMissingHeaderKey; else - status = Loader::ResultStatus::ErrorDecrypting; + status = Loader::ResultStatus::ErrorIncorrectHeaderKey; return; } } + has_rights_id = std::find_if_not(header.rights_id.begin(), header.rights_id.end(), + [](char c) { return c == '\0'; }) != header.rights_id.end(); + const std::ptrdiff_t number_sections = std::count_if(std::begin(header.section_tables), std::end(header.section_tables), [](NCASectionTableEntry entry) { return entry.media_offset > 0; }); @@ -229,7 +273,12 @@ NCA::NCA(VirtualFile file_) : file(std::move(file_)) { files.push_back(std::move(dec)); romfs = files.back(); } else { - status = Loader::ResultStatus::ErrorMissingKeys; + if (status != Loader::ResultStatus::Success) + return; + if (has_rights_id) + status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek; + else + status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey; return; } } else if (section.raw.header.filesystem_type == NCASectionFilesystemType::PFS0) { @@ -249,7 +298,12 @@ NCA::NCA(VirtualFile file_) : file(std::move(file_)) { exefs = dirs.back(); } } else { - status = Loader::ResultStatus::ErrorMissingKeys; + if (status != Loader::ResultStatus::Success) + return; + if (has_rights_id) + status = Loader::ResultStatus::ErrorIncorrectTitlekeyOrTitlekek; + else + status = Loader::ResultStatus::ErrorIncorrectKeyAreaKey; return; } } diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h index 5cfd5031a..b82e65ad5 100644 --- a/src/core/file_sys/content_archive.h +++ b/src/core/file_sys/content_archive.h @@ -98,8 +98,8 @@ protected: private: u8 GetCryptoRevision() const; boost::optional<Core::Crypto::Key128> GetKeyAreaKey(NCASectionCryptoType type) const; - boost::optional<Core::Crypto::Key128> GetTitlekey() const; - VirtualFile Decrypt(NCASectionHeader header, VirtualFile in, u64 starting_offset) const; + boost::optional<Core::Crypto::Key128> GetTitlekey(); + VirtualFile Decrypt(NCASectionHeader header, VirtualFile in, u64 starting_offset); std::vector<VirtualDir> dirs; std::vector<VirtualFile> files; @@ -109,6 +109,7 @@ private: VirtualFile file; NCAHeader header{}; + bool has_rights_id{}; Loader::ResultStatus status{}; diff --git a/src/core/file_sys/control_metadata.cpp b/src/core/file_sys/control_metadata.cpp index 3ddc9f162..ae21ad5b9 100644 --- a/src/core/file_sys/control_metadata.cpp +++ b/src/core/file_sys/control_metadata.cpp @@ -16,7 +16,7 @@ std::string LanguageEntry::GetDeveloperName() const { return Common::StringFromFixedZeroTerminatedBuffer(developer_name.data(), 0x100); } -NACP::NACP(VirtualFile file_) : file(std::move(file_)), raw(std::make_unique<RawNACP>()) { +NACP::NACP(VirtualFile file) : raw(std::make_unique<RawNACP>()) { file->ReadObject(raw.get()); } diff --git a/src/core/file_sys/control_metadata.h b/src/core/file_sys/control_metadata.h index 6582cc240..8c2cc1a2a 100644 --- a/src/core/file_sys/control_metadata.h +++ b/src/core/file_sys/control_metadata.h @@ -81,7 +81,6 @@ public: std::string GetVersionString() const; private: - VirtualFile file; std::unique_ptr<RawNACP> raw; }; diff --git a/src/core/file_sys/nca_metadata.cpp b/src/core/file_sys/nca_metadata.cpp new file mode 100644 index 000000000..449244444 --- /dev/null +++ b/src/core/file_sys/nca_metadata.cpp @@ -0,0 +1,131 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> +#include "common/common_funcs.h" +#include "common/logging/log.h" +#include "common/swap.h" +#include "content_archive.h" +#include "core/file_sys/nca_metadata.h" + +namespace FileSys { + +bool operator>=(TitleType lhs, TitleType rhs) { + return static_cast<size_t>(lhs) >= static_cast<size_t>(rhs); +} + +bool operator<=(TitleType lhs, TitleType rhs) { + return static_cast<size_t>(lhs) <= static_cast<size_t>(rhs); +} + +CNMT::CNMT(VirtualFile file) { + if (file->ReadObject(&header) != sizeof(CNMTHeader)) + return; + + // If type is {Application, Update, AOC} has opt-header. + if (header.type >= TitleType::Application && header.type <= TitleType::AOC) { + if (file->ReadObject(&opt_header, sizeof(CNMTHeader)) != sizeof(OptionalHeader)) { + LOG_WARNING(Loader, "Failed to read optional header."); + } + } + + for (u16 i = 0; i < header.number_content_entries; ++i) { + auto& next = content_records.emplace_back(ContentRecord{}); + if (file->ReadObject(&next, sizeof(CNMTHeader) + i * sizeof(ContentRecord) + + header.table_offset) != sizeof(ContentRecord)) { + content_records.erase(content_records.end() - 1); + } + } + + for (u16 i = 0; i < header.number_meta_entries; ++i) { + auto& next = meta_records.emplace_back(MetaRecord{}); + if (file->ReadObject(&next, sizeof(CNMTHeader) + i * sizeof(MetaRecord) + + header.table_offset) != sizeof(MetaRecord)) { + meta_records.erase(meta_records.end() - 1); + } + } +} + +CNMT::CNMT(CNMTHeader header, OptionalHeader opt_header, std::vector<ContentRecord> content_records, + std::vector<MetaRecord> meta_records) + : header(std::move(header)), opt_header(std::move(opt_header)), + content_records(std::move(content_records)), meta_records(std::move(meta_records)) {} + +u64 CNMT::GetTitleID() const { + return header.title_id; +} + +u32 CNMT::GetTitleVersion() const { + return header.title_version; +} + +TitleType CNMT::GetType() const { + return header.type; +} + +const std::vector<ContentRecord>& CNMT::GetContentRecords() const { + return content_records; +} + +const std::vector<MetaRecord>& CNMT::GetMetaRecords() const { + return meta_records; +} + +bool CNMT::UnionRecords(const CNMT& other) { + bool change = false; + for (const auto& rec : other.content_records) { + const auto iter = std::find_if(content_records.begin(), content_records.end(), + [&rec](const ContentRecord& r) { + return r.nca_id == rec.nca_id && r.type == rec.type; + }); + if (iter == content_records.end()) { + content_records.emplace_back(rec); + ++header.number_content_entries; + change = true; + } + } + for (const auto& rec : other.meta_records) { + const auto iter = + std::find_if(meta_records.begin(), meta_records.end(), [&rec](const MetaRecord& r) { + return r.title_id == rec.title_id && r.title_version == rec.title_version && + r.type == rec.type; + }); + if (iter == meta_records.end()) { + meta_records.emplace_back(rec); + ++header.number_meta_entries; + change = true; + } + } + return change; +} + +std::vector<u8> CNMT::Serialize() const { + const bool has_opt_header = + header.type >= TitleType::Application && header.type <= TitleType::AOC; + const auto dead_zone = header.table_offset + sizeof(CNMTHeader); + std::vector<u8> out( + std::max(sizeof(CNMTHeader) + (has_opt_header ? sizeof(OptionalHeader) : 0), dead_zone) + + content_records.size() * sizeof(ContentRecord) + meta_records.size() * sizeof(MetaRecord)); + memcpy(out.data(), &header, sizeof(CNMTHeader)); + + // Optional Header + if (has_opt_header) { + memcpy(out.data() + sizeof(CNMTHeader), &opt_header, sizeof(OptionalHeader)); + } + + auto offset = header.table_offset; + + for (const auto& rec : content_records) { + memcpy(out.data() + offset + sizeof(CNMTHeader), &rec, sizeof(ContentRecord)); + offset += sizeof(ContentRecord); + } + + for (const auto& rec : meta_records) { + memcpy(out.data() + offset + sizeof(CNMTHeader), &rec, sizeof(MetaRecord)); + offset += sizeof(MetaRecord); + } + + return out; +} +} // namespace FileSys diff --git a/src/core/file_sys/nca_metadata.h b/src/core/file_sys/nca_metadata.h new file mode 100644 index 000000000..88e66d4da --- /dev/null +++ b/src/core/file_sys/nca_metadata.h @@ -0,0 +1,111 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstring> +#include <memory> +#include <vector> +#include "common/common_types.h" +#include "common/swap.h" +#include "core/file_sys/vfs.h" + +namespace FileSys { +class CNMT; + +struct CNMTHeader; +struct OptionalHeader; + +enum class TitleType : u8 { + SystemProgram = 0x01, + SystemDataArchive = 0x02, + SystemUpdate = 0x03, + FirmwarePackageA = 0x04, + FirmwarePackageB = 0x05, + Application = 0x80, + Update = 0x81, + AOC = 0x82, + DeltaTitle = 0x83, +}; + +bool operator>=(TitleType lhs, TitleType rhs); +bool operator<=(TitleType lhs, TitleType rhs); + +enum class ContentRecordType : u8 { + Meta = 0, + Program = 1, + Data = 2, + Control = 3, + Manual = 4, + Legal = 5, + Patch = 6, +}; + +struct ContentRecord { + std::array<u8, 0x20> hash; + std::array<u8, 0x10> nca_id; + std::array<u8, 0x6> size; + ContentRecordType type; + INSERT_PADDING_BYTES(1); +}; +static_assert(sizeof(ContentRecord) == 0x38, "ContentRecord has incorrect size."); + +constexpr ContentRecord EMPTY_META_CONTENT_RECORD{{}, {}, {}, ContentRecordType::Meta, {}}; + +struct MetaRecord { + u64_le title_id; + u32_le title_version; + TitleType type; + u8 install_byte; + INSERT_PADDING_BYTES(2); +}; +static_assert(sizeof(MetaRecord) == 0x10, "MetaRecord has incorrect size."); + +struct OptionalHeader { + u64_le title_id; + u64_le minimum_version; +}; +static_assert(sizeof(OptionalHeader) == 0x10, "OptionalHeader has incorrect size."); + +struct CNMTHeader { + u64_le title_id; + u32_le title_version; + TitleType type; + INSERT_PADDING_BYTES(1); + u16_le table_offset; + u16_le number_content_entries; + u16_le number_meta_entries; + INSERT_PADDING_BYTES(12); +}; +static_assert(sizeof(CNMTHeader) == 0x20, "CNMTHeader has incorrect size."); + +// A class representing the format used by NCA metadata files, typically named {}.cnmt.nca or +// meta0.ncd. These describe which NCA's belong with which titles in the registered cache. +class CNMT { +public: + explicit CNMT(VirtualFile file); + CNMT(CNMTHeader header, OptionalHeader opt_header, std::vector<ContentRecord> content_records, + std::vector<MetaRecord> meta_records); + + u64 GetTitleID() const; + u32 GetTitleVersion() const; + TitleType GetType() const; + + const std::vector<ContentRecord>& GetContentRecords() const; + const std::vector<MetaRecord>& GetMetaRecords() const; + + bool UnionRecords(const CNMT& other); + std::vector<u8> Serialize() const; + +private: + CNMTHeader header; + OptionalHeader opt_header; + std::vector<ContentRecord> content_records; + std::vector<MetaRecord> meta_records; + + // TODO(DarkLordZach): According to switchbrew, for Patch-type there is additional data + // after the table. This is not documented, unfortunately. +}; + +} // namespace FileSys diff --git a/src/core/file_sys/partition_filesystem.cpp b/src/core/file_sys/partition_filesystem.cpp index 47e032b19..c377edc9c 100644 --- a/src/core/file_sys/partition_filesystem.cpp +++ b/src/core/file_sys/partition_filesystem.cpp @@ -24,19 +24,19 @@ bool PartitionFilesystem::Header::HasValidMagicValue() const { PartitionFilesystem::PartitionFilesystem(std::shared_ptr<VfsFile> file) { // At least be as large as the header if (file->GetSize() < sizeof(Header)) { - status = Loader::ResultStatus::Error; + status = Loader::ResultStatus::ErrorBadPFSHeader; return; } // For cartridges, HFSs can get very large, so we need to calculate the size up to // the actual content itself instead of just blindly reading in the entire file. if (sizeof(Header) != file->ReadObject(&pfs_header)) { - status = Loader::ResultStatus::Error; + status = Loader::ResultStatus::ErrorBadPFSHeader; return; } if (!pfs_header.HasValidMagicValue()) { - status = Loader::ResultStatus::ErrorInvalidFormat; + status = Loader::ResultStatus::ErrorBadPFSHeader; return; } @@ -51,7 +51,7 @@ PartitionFilesystem::PartitionFilesystem(std::shared_ptr<VfsFile> file) { const size_t total_size = file_data.size(); if (total_size != metadata_size) { - status = Loader::ResultStatus::Error; + status = Loader::ResultStatus::ErrorIncorrectPFSFileSize; return; } diff --git a/src/core/file_sys/partition_filesystem.h b/src/core/file_sys/partition_filesystem.h index 7c7a75816..be7bc32a8 100644 --- a/src/core/file_sys/partition_filesystem.h +++ b/src/core/file_sys/partition_filesystem.h @@ -13,7 +13,7 @@ #include "core/file_sys/vfs.h" namespace Loader { -enum class ResultStatus; +enum class ResultStatus : u16; } namespace FileSys { diff --git a/src/core/file_sys/program_metadata.cpp b/src/core/file_sys/program_metadata.cpp index 63d4b6e4f..279f987d4 100644 --- a/src/core/file_sys/program_metadata.cpp +++ b/src/core/file_sys/program_metadata.cpp @@ -12,26 +12,26 @@ namespace FileSys { Loader::ResultStatus ProgramMetadata::Load(VirtualFile file) { size_t total_size = static_cast<size_t>(file->GetSize()); if (total_size < sizeof(Header)) - return Loader::ResultStatus::Error; + return Loader::ResultStatus::ErrorBadNPDMHeader; // TODO(DarkLordZach): Use ReadObject when Header/AcidHeader becomes trivially copyable. std::vector<u8> npdm_header_data = file->ReadBytes(sizeof(Header)); if (sizeof(Header) != npdm_header_data.size()) - return Loader::ResultStatus::Error; + return Loader::ResultStatus::ErrorBadNPDMHeader; std::memcpy(&npdm_header, npdm_header_data.data(), sizeof(Header)); std::vector<u8> acid_header_data = file->ReadBytes(sizeof(AcidHeader), npdm_header.acid_offset); if (sizeof(AcidHeader) != acid_header_data.size()) - return Loader::ResultStatus::Error; + return Loader::ResultStatus::ErrorBadACIDHeader; std::memcpy(&acid_header, acid_header_data.data(), sizeof(AcidHeader)); if (sizeof(AciHeader) != file->ReadObject(&aci_header, npdm_header.aci_offset)) - return Loader::ResultStatus::Error; + return Loader::ResultStatus::ErrorBadACIHeader; if (sizeof(FileAccessControl) != file->ReadObject(&acid_file_access, acid_header.fac_offset)) - return Loader::ResultStatus::Error; + return Loader::ResultStatus::ErrorBadFileAccessControl; if (sizeof(FileAccessHeader) != file->ReadObject(&aci_file_access, aci_header.fah_offset)) - return Loader::ResultStatus::Error; + return Loader::ResultStatus::ErrorBadFileAccessHeader; return Loader::ResultStatus::Success; } diff --git a/src/core/file_sys/program_metadata.h b/src/core/file_sys/program_metadata.h index 06a7315db..74a91052b 100644 --- a/src/core/file_sys/program_metadata.h +++ b/src/core/file_sys/program_metadata.h @@ -13,7 +13,7 @@ #include "partition_filesystem.h" namespace Loader { -enum class ResultStatus; +enum class ResultStatus : u16; } namespace FileSys { diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp new file mode 100644 index 000000000..d25eeee34 --- /dev/null +++ b/src/core/file_sys/registered_cache.cpp @@ -0,0 +1,478 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <regex> +#include <mbedtls/sha256.h> +#include "common/assert.h" +#include "common/hex_util.h" +#include "common/logging/log.h" +#include "core/crypto/encryption_layer.h" +#include "core/file_sys/card_image.h" +#include "core/file_sys/nca_metadata.h" +#include "core/file_sys/registered_cache.h" +#include "core/file_sys/vfs_concat.h" + +namespace FileSys { +std::string RegisteredCacheEntry::DebugInfo() const { + return fmt::format("title_id={:016X}, content_type={:02X}", title_id, static_cast<u8>(type)); +} + +bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) { + return (lhs.title_id < rhs.title_id) || (lhs.title_id == rhs.title_id && lhs.type < rhs.type); +} + +static bool FollowsTwoDigitDirFormat(std::string_view name) { + static const std::regex two_digit_regex("000000[0-9A-F]{2}", std::regex_constants::ECMAScript | + std::regex_constants::icase); + return std::regex_match(name.begin(), name.end(), two_digit_regex); +} + +static bool FollowsNcaIdFormat(std::string_view name) { + static const std::regex nca_id_regex("[0-9A-F]{32}\\.nca", std::regex_constants::ECMAScript | + std::regex_constants::icase); + return name.size() == 36 && std::regex_match(name.begin(), name.end(), nca_id_regex); +} + +static std::string GetRelativePathFromNcaID(const std::array<u8, 16>& nca_id, bool second_hex_upper, + bool within_two_digit) { + if (!within_two_digit) + return fmt::format("/{}.nca", Common::HexArrayToString(nca_id, second_hex_upper)); + + Core::Crypto::SHA256Hash hash{}; + mbedtls_sha256(nca_id.data(), nca_id.size(), hash.data(), 0); + return fmt::format("/000000{:02X}/{}.nca", hash[0], + Common::HexArrayToString(nca_id, second_hex_upper)); +} + +static std::string GetCNMTName(TitleType type, u64 title_id) { + constexpr std::array<const char*, 9> TITLE_TYPE_NAMES{ + "SystemProgram", + "SystemData", + "SystemUpdate", + "BootImagePackage", + "BootImagePackageSafe", + "Application", + "Patch", + "AddOnContent", + "" ///< Currently unknown 'DeltaTitle' + }; + + auto index = static_cast<size_t>(type); + // If the index is after the jump in TitleType, subtract it out. + if (index >= static_cast<size_t>(TitleType::Application)) { + index -= static_cast<size_t>(TitleType::Application) - + static_cast<size_t>(TitleType::FirmwarePackageB); + } + return fmt::format("{}_{:016x}.cnmt", TITLE_TYPE_NAMES[index], title_id); +} + +static ContentRecordType GetCRTypeFromNCAType(NCAContentType type) { + switch (type) { + case NCAContentType::Program: + // TODO(DarkLordZach): Differentiate between Program and Patch + return ContentRecordType::Program; + case NCAContentType::Meta: + return ContentRecordType::Meta; + case NCAContentType::Control: + return ContentRecordType::Control; + case NCAContentType::Data: + return ContentRecordType::Data; + case NCAContentType::Manual: + // TODO(DarkLordZach): Peek at NCA contents to differentiate Manual and Legal. + return ContentRecordType::Manual; + default: + UNREACHABLE(); + } +} + +VirtualFile RegisteredCache::OpenFileOrDirectoryConcat(const VirtualDir& dir, + std::string_view path) const { + if (dir->GetFileRelative(path) != nullptr) + return dir->GetFileRelative(path); + if (dir->GetDirectoryRelative(path) != nullptr) { + const auto nca_dir = dir->GetDirectoryRelative(path); + VirtualFile file = nullptr; + + const auto files = nca_dir->GetFiles(); + if (files.size() == 1 && files[0]->GetName() == "00") { + file = files[0]; + } else { + std::vector<VirtualFile> concat; + // Since the files are a two-digit hex number, max is FF. + for (size_t i = 0; i < 0x100; ++i) { + auto next = nca_dir->GetFile(fmt::format("{:02X}", i)); + if (next != nullptr) { + concat.push_back(std::move(next)); + } else { + next = nca_dir->GetFile(fmt::format("{:02x}", i)); + if (next != nullptr) + concat.push_back(std::move(next)); + else + break; + } + } + + if (concat.empty()) + return nullptr; + + file = FileSys::ConcatenateFiles(concat); + } + + return file; + } + return nullptr; +} + +VirtualFile RegisteredCache::GetFileAtID(NcaID id) const { + VirtualFile file; + // Try all four modes of file storage: + // (bit 1 = uppercase/lower, bit 0 = within a two-digit dir) + // 00: /000000**/{:032X}.nca + // 01: /{:032X}.nca + // 10: /000000**/{:032x}.nca + // 11: /{:032x}.nca + for (u8 i = 0; i < 4; ++i) { + const auto path = GetRelativePathFromNcaID(id, (i & 0b10) == 0, (i & 0b01) == 0); + file = OpenFileOrDirectoryConcat(dir, path); + if (file != nullptr) + return file; + } + return file; +} + +static boost::optional<NcaID> CheckMapForContentRecord( + const boost::container::flat_map<u64, CNMT>& map, u64 title_id, ContentRecordType type) { + if (map.find(title_id) == map.end()) + return boost::none; + + const auto& cnmt = map.at(title_id); + + const auto iter = std::find_if(cnmt.GetContentRecords().begin(), cnmt.GetContentRecords().end(), + [type](const ContentRecord& rec) { return rec.type == type; }); + if (iter == cnmt.GetContentRecords().end()) + return boost::none; + + return boost::make_optional(iter->nca_id); +} + +boost::optional<NcaID> RegisteredCache::GetNcaIDFromMetadata(u64 title_id, + ContentRecordType type) const { + if (type == ContentRecordType::Meta && meta_id.find(title_id) != meta_id.end()) + return meta_id.at(title_id); + + const auto res1 = CheckMapForContentRecord(yuzu_meta, title_id, type); + if (res1 != boost::none) + return res1; + return CheckMapForContentRecord(meta, title_id, type); +} + +std::vector<NcaID> RegisteredCache::AccumulateFiles() const { + std::vector<NcaID> ids; + for (const auto& d2_dir : dir->GetSubdirectories()) { + if (FollowsNcaIdFormat(d2_dir->GetName())) { + ids.push_back(Common::HexStringToArray<0x10, true>(d2_dir->GetName().substr(0, 0x20))); + continue; + } + + if (!FollowsTwoDigitDirFormat(d2_dir->GetName())) + continue; + + for (const auto& nca_dir : d2_dir->GetSubdirectories()) { + if (!FollowsNcaIdFormat(nca_dir->GetName())) + continue; + + ids.push_back(Common::HexStringToArray<0x10, true>(nca_dir->GetName().substr(0, 0x20))); + } + + for (const auto& nca_file : d2_dir->GetFiles()) { + if (!FollowsNcaIdFormat(nca_file->GetName())) + continue; + + ids.push_back( + Common::HexStringToArray<0x10, true>(nca_file->GetName().substr(0, 0x20))); + } + } + + for (const auto& d2_file : dir->GetFiles()) { + if (FollowsNcaIdFormat(d2_file->GetName())) + ids.push_back(Common::HexStringToArray<0x10, true>(d2_file->GetName().substr(0, 0x20))); + } + return ids; +} + +void RegisteredCache::ProcessFiles(const std::vector<NcaID>& ids) { + for (const auto& id : ids) { + const auto file = GetFileAtID(id); + + if (file == nullptr) + continue; + const auto nca = std::make_shared<NCA>(parser(file, id)); + if (nca->GetStatus() != Loader::ResultStatus::Success || + nca->GetType() != NCAContentType::Meta) { + continue; + } + + const auto section0 = nca->GetSubdirectories()[0]; + + for (const auto& file : section0->GetFiles()) { + if (file->GetExtension() != "cnmt") + continue; + + meta.insert_or_assign(nca->GetTitleId(), CNMT(file)); + meta_id.insert_or_assign(nca->GetTitleId(), id); + break; + } + } +} + +void RegisteredCache::AccumulateYuzuMeta() { + const auto dir = this->dir->GetSubdirectory("yuzu_meta"); + if (dir == nullptr) + return; + + for (const auto& file : dir->GetFiles()) { + if (file->GetExtension() != "cnmt") + continue; + + CNMT cnmt(file); + yuzu_meta.insert_or_assign(cnmt.GetTitleID(), std::move(cnmt)); + } +} + +void RegisteredCache::Refresh() { + if (dir == nullptr) + return; + const auto ids = AccumulateFiles(); + ProcessFiles(ids); + AccumulateYuzuMeta(); +} + +RegisteredCache::RegisteredCache(VirtualDir dir_, RegisteredCacheParsingFunction parsing_function) + : dir(std::move(dir_)), parser(std::move(parsing_function)) { + Refresh(); +} + +bool RegisteredCache::HasEntry(u64 title_id, ContentRecordType type) const { + return GetEntryRaw(title_id, type) != nullptr; +} + +bool RegisteredCache::HasEntry(RegisteredCacheEntry entry) const { + return GetEntryRaw(entry) != nullptr; +} + +VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) const { + const auto id = GetNcaIDFromMetadata(title_id, type); + if (id == boost::none) + return nullptr; + + return parser(GetFileAtID(id.get()), id.get()); +} + +VirtualFile RegisteredCache::GetEntryRaw(RegisteredCacheEntry entry) const { + return GetEntryRaw(entry.title_id, entry.type); +} + +std::shared_ptr<NCA> RegisteredCache::GetEntry(u64 title_id, ContentRecordType type) const { + const auto raw = GetEntryRaw(title_id, type); + if (raw == nullptr) + return nullptr; + return std::make_shared<NCA>(raw); +} + +std::shared_ptr<NCA> RegisteredCache::GetEntry(RegisteredCacheEntry entry) const { + return GetEntry(entry.title_id, entry.type); +} + +template <typename T> +void RegisteredCache::IterateAllMetadata( + std::vector<T>& out, std::function<T(const CNMT&, const ContentRecord&)> proc, + std::function<bool(const CNMT&, const ContentRecord&)> filter) const { + for (const auto& kv : meta) { + const auto& cnmt = kv.second; + if (filter(cnmt, EMPTY_META_CONTENT_RECORD)) + out.push_back(proc(cnmt, EMPTY_META_CONTENT_RECORD)); + for (const auto& rec : cnmt.GetContentRecords()) { + if (GetFileAtID(rec.nca_id) != nullptr && filter(cnmt, rec)) { + out.push_back(proc(cnmt, rec)); + } + } + } + for (const auto& kv : yuzu_meta) { + const auto& cnmt = kv.second; + for (const auto& rec : cnmt.GetContentRecords()) { + if (GetFileAtID(rec.nca_id) != nullptr && filter(cnmt, rec)) { + out.push_back(proc(cnmt, rec)); + } + } + } +} + +std::vector<RegisteredCacheEntry> RegisteredCache::ListEntries() const { + std::vector<RegisteredCacheEntry> out; + IterateAllMetadata<RegisteredCacheEntry>( + out, + [](const CNMT& c, const ContentRecord& r) { + return RegisteredCacheEntry{c.GetTitleID(), r.type}; + }, + [](const CNMT& c, const ContentRecord& r) { return true; }); + return out; +} + +std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter( + boost::optional<TitleType> title_type, boost::optional<ContentRecordType> record_type, + boost::optional<u64> title_id) const { + std::vector<RegisteredCacheEntry> out; + IterateAllMetadata<RegisteredCacheEntry>( + out, + [](const CNMT& c, const ContentRecord& r) { + return RegisteredCacheEntry{c.GetTitleID(), r.type}; + }, + [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) { + if (title_type != boost::none && title_type.get() != c.GetType()) + return false; + if (record_type != boost::none && record_type.get() != r.type) + return false; + if (title_id != boost::none && title_id.get() != c.GetTitleID()) + return false; + return true; + }); + return out; +} + +static std::shared_ptr<NCA> GetNCAFromXCIForID(std::shared_ptr<XCI> xci, const NcaID& id) { + const auto filename = fmt::format("{}.nca", Common::HexArrayToString(id, false)); + const auto iter = + std::find_if(xci->GetNCAs().begin(), xci->GetNCAs().end(), + [&filename](std::shared_ptr<NCA> nca) { return nca->GetName() == filename; }); + return iter == xci->GetNCAs().end() ? nullptr : *iter; +} + +InstallResult RegisteredCache::InstallEntry(std::shared_ptr<XCI> xci, bool overwrite_if_exists, + const VfsCopyFunction& copy) { + const auto& ncas = xci->GetNCAs(); + const auto& meta_iter = std::find_if(ncas.begin(), ncas.end(), [](std::shared_ptr<NCA> nca) { + return nca->GetType() == NCAContentType::Meta; + }); + + if (meta_iter == ncas.end()) { + LOG_ERROR(Loader, "The XCI you are attempting to install does not have a metadata NCA and " + "is therefore malformed. Double check your encryption keys."); + 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; + + // Install all the other NCAs + const auto section0 = (*meta_iter)->GetSubdirectories()[0]; + const auto cnmt_file = section0->GetFiles()[0]; + const CNMT cnmt(cnmt_file); + for (const auto& record : cnmt.GetContentRecords()) { + const auto nca = GetNCAFromXCIForID(xci, record.nca_id); + if (nca == nullptr) + return InstallResult::ErrorCopyFailed; + const auto res2 = RawInstallNCA(nca, copy, overwrite_if_exists, record.nca_id); + if (res2 != InstallResult::Success) + return res2; + } + + Refresh(); + return InstallResult::Success; +} + +InstallResult RegisteredCache::InstallEntry(std::shared_ptr<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 + }; + OptionalHeader opt_header{0, 0}; + ContentRecord c_rec{{}, {}, {}, GetCRTypeFromNCAType(nca->GetType()), {}}; + const auto& data = nca->GetBaseFile()->ReadBytes(0x100000); + mbedtls_sha256(data.data(), data.size(), c_rec.hash.data(), 0); + memcpy(&c_rec.nca_id, &c_rec.hash, 16); + const CNMT new_cnmt(header, opt_header, {c_rec}, {}); + if (!RawInstallYuzuMeta(new_cnmt)) + return InstallResult::ErrorMetaFailed; + return RawInstallNCA(nca, copy, overwrite_if_exists, c_rec.nca_id); +} + +InstallResult RegisteredCache::RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy, + bool overwrite_if_exists, + boost::optional<NcaID> override_id) { + const auto in = nca->GetBaseFile(); + Core::Crypto::SHA256Hash hash{}; + + // Calculate NcaID + // NOTE: Because computing the SHA256 of an entire NCA is quite expensive (especially if the + // game is massive), we're going to cheat and only hash the first MB of the NCA. + // Also, for XCIs the NcaID matters, so if the override id isn't none, use that. + NcaID id{}; + if (override_id == boost::none) { + const auto& data = in->ReadBytes(0x100000); + mbedtls_sha256(data.data(), data.size(), hash.data(), 0); + memcpy(id.data(), hash.data(), 16); + } else { + id = override_id.get(); + } + + std::string path = GetRelativePathFromNcaID(id, false, true); + + if (GetFileAtID(id) != nullptr && !overwrite_if_exists) { + LOG_WARNING(Loader, "Attempting to overwrite existing NCA. Skipping..."); + return InstallResult::ErrorAlreadyExists; + } + + if (GetFileAtID(id) != nullptr) { + LOG_WARNING(Loader, "Overwriting existing NCA..."); + VirtualDir c_dir; + { c_dir = dir->GetFileRelative(path)->GetContainingDirectory(); } + c_dir->DeleteFile(FileUtil::GetFilename(path)); + } + + auto out = dir->CreateFileRelative(path); + if (out == nullptr) + return InstallResult::ErrorCopyFailed; + return copy(in, out) ? InstallResult::Success : InstallResult::ErrorCopyFailed; +} + +bool RegisteredCache::RawInstallYuzuMeta(const CNMT& cnmt) { + // Reasoning behind this method can be found in the comment for InstallEntry, NCA overload. + const auto dir = this->dir->CreateDirectoryRelative("yuzu_meta"); + const auto filename = GetCNMTName(cnmt.GetType(), cnmt.GetTitleID()); + if (dir->GetFile(filename) == nullptr) { + auto out = dir->CreateFile(filename); + const auto buffer = cnmt.Serialize(); + out->Resize(buffer.size()); + out->WriteBytes(buffer); + } else { + auto out = dir->GetFile(filename); + CNMT old_cnmt(out); + // Returns true on change + if (old_cnmt.UnionRecords(cnmt)) { + out->Resize(0); + const auto buffer = old_cnmt.Serialize(); + out->Resize(buffer.size()); + out->WriteBytes(buffer); + } + } + Refresh(); + return std::find_if(yuzu_meta.begin(), yuzu_meta.end(), + [&cnmt](const std::pair<u64, CNMT>& kv) { + return kv.second.GetType() == cnmt.GetType() && + kv.second.GetTitleID() == cnmt.GetTitleID(); + }) != yuzu_meta.end(); +} +} // namespace FileSys diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h new file mode 100644 index 000000000..a7c51a59c --- /dev/null +++ b/src/core/file_sys/registered_cache.h @@ -0,0 +1,124 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <functional> +#include <map> +#include <memory> +#include <string> +#include <vector> +#include <boost/container/flat_map.hpp> +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "content_archive.h" +#include "core/file_sys/nca_metadata.h" +#include "core/file_sys/vfs.h" + +namespace FileSys { +class XCI; +class CNMT; + +using NcaID = std::array<u8, 0x10>; +using RegisteredCacheParsingFunction = std::function<VirtualFile(const VirtualFile&, const NcaID&)>; +using VfsCopyFunction = std::function<bool(VirtualFile, VirtualFile)>; + +enum class InstallResult { + Success, + ErrorAlreadyExists, + ErrorCopyFailed, + ErrorMetaFailed, +}; + +struct RegisteredCacheEntry { + u64 title_id; + ContentRecordType type; + + std::string DebugInfo() const; +}; + +// boost flat_map requires operator< for O(log(n)) lookups. +bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs); + +/* + * A class that catalogues NCAs in the registered directory structure. + * Nintendo's registered format follows this structure: + * + * Root + * | 000000XX <- XX is the ____ two digits of the NcaID + * | <hash>.nca <- hash is the NcaID (first half of SHA256 over entire file) (folder) + * | 00 + * | 01 <- Actual content split along 4GB boundaries. (optional) + * + * (This impl also supports substituting the nca dir for an nca file, as that's more convenient when + * 4GB splitting can be ignored.) + */ +class RegisteredCache { +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, + RegisteredCacheParsingFunction parsing_function = + [](const VirtualFile& file, const NcaID& id) { return file; }); + + void Refresh(); + + bool HasEntry(u64 title_id, ContentRecordType type) const; + bool HasEntry(RegisteredCacheEntry entry) const; + + VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const; + VirtualFile GetEntryRaw(RegisteredCacheEntry entry) const; + + std::shared_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const; + std::shared_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const; + + std::vector<RegisteredCacheEntry> ListEntries() const; + // If a parameter is not boost::none, it will be filtered for from all entries. + std::vector<RegisteredCacheEntry> ListEntriesFilter( + boost::optional<TitleType> title_type = boost::none, + boost::optional<ContentRecordType> record_type = boost::none, + boost::optional<u64> title_id = boost::none) const; + + // Raw copies all the ncas from the xci to the csache. Does some quick checks to make sure there + // is a meta NCA and all of them are accessible. + InstallResult InstallEntry(std::shared_ptr<XCI> xci, bool overwrite_if_exists = false, + const VfsCopyFunction& copy = &VfsRawCopy); + + // Due to the fact that we must use Meta-type NCAs to determine the existance of files, this + // poses quite a challenge. Instead of creating a new meta NCA for this file, yuzu will create a + // dir inside the NAND called 'yuzu_meta' and store the raw CNMT there. + // TODO(DarkLordZach): Author real meta-type NCAs and install those. + InstallResult InstallEntry(std::shared_ptr<NCA> nca, TitleType type, + bool overwrite_if_exists = false, + const VfsCopyFunction& copy = &VfsRawCopy); + +private: + template <typename T> + void IterateAllMetadata(std::vector<T>& out, + std::function<T(const CNMT&, const ContentRecord&)> proc, + std::function<bool(const CNMT&, const ContentRecord&)> filter) const; + std::vector<NcaID> AccumulateFiles() const; + void ProcessFiles(const std::vector<NcaID>& ids); + void AccumulateYuzuMeta(); + boost::optional<NcaID> GetNcaIDFromMetadata(u64 title_id, ContentRecordType type) const; + VirtualFile GetFileAtID(NcaID id) const; + VirtualFile OpenFileOrDirectoryConcat(const VirtualDir& dir, std::string_view path) const; + InstallResult RawInstallNCA(std::shared_ptr<NCA> nca, const VfsCopyFunction& copy, + bool overwrite_if_exists, + boost::optional<NcaID> override_id = boost::none); + bool RawInstallYuzuMeta(const CNMT& cnmt); + + VirtualDir dir; + RegisteredCacheParsingFunction parser; + // maps tid -> NcaID of meta + boost::container::flat_map<u64, NcaID> meta_id; + // maps tid -> meta + boost::container::flat_map<u64, CNMT> meta; + // maps tid -> meta for CNMT in yuzu_meta + boost::container::flat_map<u64, CNMT> yuzu_meta; +}; + +} // namespace FileSys diff --git a/src/core/file_sys/romfs.cpp b/src/core/file_sys/romfs.cpp index ff3ddb29c..e490c8ace 100644 --- a/src/core/file_sys/romfs.cpp +++ b/src/core/file_sys/romfs.cpp @@ -65,7 +65,7 @@ void ProcessFile(VirtualFile file, size_t file_offset, size_t data_offset, u32 t auto entry = GetEntry<FileEntry>(file, file_offset + this_file_offset); parent->AddFile(std::make_shared<OffsetVfsFile>( - file, entry.first.size, entry.first.offset + data_offset, entry.second, parent)); + file, entry.first.size, entry.first.offset + data_offset, entry.second)); if (entry.first.sibling == ROMFS_ENTRY_EMPTY) break; @@ -79,7 +79,7 @@ void ProcessDirectory(VirtualFile file, size_t dir_offset, size_t file_offset, s while (true) { auto entry = GetEntry<DirectoryEntry>(file, dir_offset + this_dir_offset); auto current = std::make_shared<VectorVfsDirectory>( - std::vector<VirtualFile>{}, std::vector<VirtualDir>{}, parent, entry.second); + std::vector<VirtualFile>{}, std::vector<VirtualDir>{}, entry.second); if (entry.first.child_file != ROMFS_ENTRY_EMPTY) { ProcessFile(file, file_offset, data_offset, entry.first.child_file, current); @@ -108,9 +108,9 @@ VirtualDir ExtractRomFS(VirtualFile file) { const u64 file_offset = header.file_meta.offset; const u64 dir_offset = header.directory_meta.offset + 4; - const auto root = + auto root = std::make_shared<VectorVfsDirectory>(std::vector<VirtualFile>{}, std::vector<VirtualDir>{}, - file->GetContainingDirectory(), file->GetName()); + file->GetName(), file->GetContainingDirectory()); ProcessDirectory(file, dir_offset, file_offset, header.data_offset, 0, root); diff --git a/src/core/file_sys/vfs.cpp b/src/core/file_sys/vfs.cpp index 24e158962..a5ec50b1a 100644 --- a/src/core/file_sys/vfs.cpp +++ b/src/core/file_sys/vfs.cpp @@ -74,15 +74,15 @@ VirtualFile VfsFilesystem::CopyFile(std::string_view old_path_, std::string_view return new_file; } -VirtualFile VfsFilesystem::MoveFile(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_); +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); // Again, non-default impls are highly encouraged to provide a more optimized version of this. - auto out = CopyFile(old_path_, new_path_); + auto out = CopyFile(sanitized_old_path, sanitized_new_path); if (out == nullptr) return nullptr; - if (DeleteFile(old_path)) + if (DeleteFile(sanitized_old_path)) return out; return nullptr; } @@ -137,15 +137,15 @@ VirtualDir VfsFilesystem::CopyDirectory(std::string_view old_path_, std::string_ return new_dir; } -VirtualDir VfsFilesystem::MoveDirectory(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_); +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); // Non-default impls are highly encouraged to provide a more optimized version of this. - auto out = CopyDirectory(old_path_, new_path_); + auto out = CopyDirectory(sanitized_old_path, sanitized_new_path); if (out == nullptr) return nullptr; - if (DeleteDirectory(old_path)) + if (DeleteDirectory(sanitized_old_path)) return out; return nullptr; } diff --git a/src/core/file_sys/vfs.h b/src/core/file_sys/vfs.h index 141a053ce..78a63c59b 100644 --- a/src/core/file_sys/vfs.h +++ b/src/core/file_sys/vfs.h @@ -15,9 +15,9 @@ namespace FileSys { -struct VfsFilesystem; -struct VfsFile; -struct VfsDirectory; +class VfsDirectory; +class VfsFile; +class VfsFilesystem; // Convenience typedefs to use Vfs* interfaces using VirtualFilesystem = std::shared_ptr<VfsFilesystem>; @@ -34,8 +34,9 @@ enum class VfsEntryType { // A class representing an abstract filesystem. A default implementation given the root VirtualDir // is provided for convenience, but if the Vfs implementation has any additional state or // functionality, they will need to override. -struct VfsFilesystem : NonCopyable { - VfsFilesystem(VirtualDir root); +class VfsFilesystem : NonCopyable { +public: + explicit VfsFilesystem(VirtualDir root); virtual ~VfsFilesystem(); // Gets the friendly name for the filesystem. @@ -81,7 +82,8 @@ protected: }; // A class representing a file in an abstract filesystem. -struct VfsFile : NonCopyable { +class VfsFile : NonCopyable { +public: virtual ~VfsFile(); // Retrieves the file name. @@ -179,7 +181,8 @@ struct VfsFile : NonCopyable { }; // A class representing a directory in an abstract filesystem. -struct VfsDirectory : NonCopyable { +class VfsDirectory : NonCopyable { +public: virtual ~VfsDirectory(); // Retrives the file located at path as if the current directory was root. Returns nullptr if @@ -295,7 +298,8 @@ protected: // A convenience partial-implementation of VfsDirectory that stubs out methods that should only work // if writable. This is to avoid redundant empty methods everywhere. -struct ReadOnlyVfsDirectory : public VfsDirectory { +class ReadOnlyVfsDirectory : public VfsDirectory { +public: bool IsWritable() const override; bool IsReadable() const override; std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) override; diff --git a/src/core/file_sys/vfs_concat.cpp b/src/core/file_sys/vfs_concat.cpp new file mode 100644 index 000000000..e6bf586a3 --- /dev/null +++ b/src/core/file_sys/vfs_concat.cpp @@ -0,0 +1,94 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <utility> + +#include "core/file_sys/vfs_concat.h" + +namespace FileSys { + +VirtualFile ConcatenateFiles(std::vector<VirtualFile> files, std::string name) { + if (files.empty()) + return nullptr; + if (files.size() == 1) + return files[0]; + + return std::shared_ptr<VfsFile>(new ConcatenatedVfsFile(std::move(files), std::move(name))); +} + +ConcatenatedVfsFile::ConcatenatedVfsFile(std::vector<VirtualFile> files_, std::string name) + : name(std::move(name)) { + size_t next_offset = 0; + for (const auto& file : files_) { + files[next_offset] = file; + next_offset += file->GetSize(); + } +} + +std::string ConcatenatedVfsFile::GetName() const { + if (files.empty()) + return ""; + if (!name.empty()) + return name; + return files.begin()->second->GetName(); +} + +size_t ConcatenatedVfsFile::GetSize() const { + if (files.empty()) + return 0; + return files.rbegin()->first + files.rbegin()->second->GetSize(); +} + +bool ConcatenatedVfsFile::Resize(size_t new_size) { + return false; +} + +std::shared_ptr<VfsDirectory> ConcatenatedVfsFile::GetContainingDirectory() const { + if (files.empty()) + return nullptr; + return files.begin()->second->GetContainingDirectory(); +} + +bool ConcatenatedVfsFile::IsWritable() const { + return false; +} + +bool ConcatenatedVfsFile::IsReadable() const { + return true; +} + +size_t ConcatenatedVfsFile::Read(u8* data, size_t length, size_t offset) const { + auto entry = files.end(); + for (auto iter = files.begin(); iter != files.end(); ++iter) { + if (iter->first > offset) { + entry = --iter; + break; + } + } + + // Check if the entry should be the last one. The loop above will make it end(). + if (entry == files.end() && offset < files.rbegin()->first + files.rbegin()->second->GetSize()) + --entry; + + if (entry == files.end()) + return 0; + + const auto remaining = entry->second->GetSize() + offset - entry->first; + if (length > remaining) { + return entry->second->Read(data, remaining, offset - entry->first) + + Read(data + remaining, length - remaining, offset + remaining); + } + + return entry->second->Read(data, length, offset - entry->first); +} + +size_t ConcatenatedVfsFile::Write(const u8* data, size_t length, size_t offset) { + return 0; +} + +bool ConcatenatedVfsFile::Rename(std::string_view name) { + return false; +} +} // namespace FileSys diff --git a/src/core/file_sys/vfs_concat.h b/src/core/file_sys/vfs_concat.h new file mode 100644 index 000000000..686d32515 --- /dev/null +++ b/src/core/file_sys/vfs_concat.h @@ -0,0 +1,41 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <string_view> +#include <boost/container/flat_map.hpp> +#include "core/file_sys/vfs.h" + +namespace FileSys { + +// Wrapper function to allow for more efficient handling of files.size() == 0, 1 cases. +VirtualFile ConcatenateFiles(std::vector<VirtualFile> files, std::string name = ""); + +// Class that wraps multiple vfs files and concatenates them, making reads seamless. Currently +// read-only. +class ConcatenatedVfsFile : public VfsFile { + friend VirtualFile ConcatenateFiles(std::vector<VirtualFile> files, std::string name); + + ConcatenatedVfsFile(std::vector<VirtualFile> files, std::string name); + +public: + std::string GetName() const override; + size_t GetSize() const override; + bool Resize(size_t new_size) override; + std::shared_ptr<VfsDirectory> GetContainingDirectory() const override; + bool IsWritable() const override; + bool IsReadable() const override; + size_t Read(u8* data, size_t length, size_t offset) const override; + size_t Write(const u8* data, size_t length, size_t offset) override; + bool Rename(std::string_view name) override; + +private: + // Maps starting offset to file -- more efficient. + boost::container::flat_map<u64, VirtualFile> files; + std::string name; +}; + +} // namespace FileSys diff --git a/src/core/file_sys/vfs_offset.h b/src/core/file_sys/vfs_offset.h index 235970dc5..cb92d1570 100644 --- a/src/core/file_sys/vfs_offset.h +++ b/src/core/file_sys/vfs_offset.h @@ -15,7 +15,8 @@ namespace FileSys { // Similar to seeking to an offset. // If the file is writable, operations that would write past the end of the offset file will expand // the size of this wrapper. -struct OffsetVfsFile : public VfsFile { +class OffsetVfsFile : public VfsFile { +public: OffsetVfsFile(std::shared_ptr<VfsFile> file, size_t size, size_t offset = 0, std::string new_name = "", VirtualDir new_parent = nullptr); diff --git a/src/core/file_sys/vfs_real.cpp b/src/core/file_sys/vfs_real.cpp index 1b5919737..0afe515f0 100644 --- a/src/core/file_sys/vfs_real.cpp +++ b/src/core/file_sys/vfs_real.cpp @@ -83,8 +83,12 @@ VirtualFile RealVfsFilesystem::OpenFile(std::string_view path_, Mode perms) { VirtualFile RealVfsFilesystem::CreateFile(std::string_view path_, Mode perms) { const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - if (!FileUtil::Exists(path) && !FileUtil::CreateEmptyFile(path)) - return nullptr; + const auto path_fwd = FileUtil::SanitizePath(path, FileUtil::DirectorySeparator::ForwardSlash); + if (!FileUtil::Exists(path)) { + FileUtil::CreateFullPath(path_fwd); + if (!FileUtil::CreateEmptyFile(path)) + return nullptr; + } return OpenFile(path, perms); } @@ -140,8 +144,12 @@ VirtualDir RealVfsFilesystem::OpenDirectory(std::string_view path_, Mode perms) VirtualDir RealVfsFilesystem::CreateDirectory(std::string_view path_, Mode perms) { const auto path = FileUtil::SanitizePath(path_, FileUtil::DirectorySeparator::PlatformDefault); - if (!FileUtil::Exists(path) && !FileUtil::CreateDir(path)) - return nullptr; + const auto path_fwd = FileUtil::SanitizePath(path, FileUtil::DirectorySeparator::ForwardSlash); + if (!FileUtil::Exists(path)) { + FileUtil::CreateFullPath(path_fwd); + if (!FileUtil::CreateDir(path)) + return nullptr; + } // Cannot use make_shared as RealVfsDirectory constructor is private return std::shared_ptr<RealVfsDirectory>(new RealVfsDirectory(*this, path, perms)); } @@ -306,14 +314,14 @@ RealVfsDirectory::RealVfsDirectory(RealVfsFilesystem& base_, const std::string& 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)) + if (!FileUtil::Exists(full_path) || FileUtil::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)) + if (!FileUtil::Exists(full_path) || !FileUtil::IsDirectory(full_path)) return nullptr; return base.OpenDirectory(full_path, perms); } diff --git a/src/core/file_sys/vfs_real.h b/src/core/file_sys/vfs_real.h index 8a1e79ef6..989803d43 100644 --- a/src/core/file_sys/vfs_real.h +++ b/src/core/file_sys/vfs_real.h @@ -5,7 +5,6 @@ #pragma once #include <string_view> - #include <boost/container/flat_map.hpp> #include "common/file_util.h" #include "core/file_sys/mode.h" diff --git a/src/core/file_sys/vfs_vector.cpp b/src/core/file_sys/vfs_vector.cpp index fda603960..98e7c4598 100644 --- a/src/core/file_sys/vfs_vector.cpp +++ b/src/core/file_sys/vfs_vector.cpp @@ -8,8 +8,8 @@ namespace FileSys { VectorVfsDirectory::VectorVfsDirectory(std::vector<VirtualFile> files_, - std::vector<VirtualDir> dirs_, VirtualDir parent_, - std::string name_) + std::vector<VirtualDir> dirs_, std::string name_, + VirtualDir parent_) : files(std::move(files_)), dirs(std::move(dirs_)), parent(std::move(parent_)), name(std::move(name_)) {} diff --git a/src/core/file_sys/vfs_vector.h b/src/core/file_sys/vfs_vector.h index ba469647b..179f62e4b 100644 --- a/src/core/file_sys/vfs_vector.h +++ b/src/core/file_sys/vfs_vector.h @@ -10,10 +10,11 @@ namespace FileSys { // An implementation of VfsDirectory that maintains two vectors for subdirectories and files. // Vector data is supplied upon construction. -struct VectorVfsDirectory : public VfsDirectory { +class VectorVfsDirectory : public VfsDirectory { +public: explicit VectorVfsDirectory(std::vector<VirtualFile> files = {}, - std::vector<VirtualDir> dirs = {}, VirtualDir parent = nullptr, - std::string name = ""); + std::vector<VirtualDir> dirs = {}, std::string name = "", + VirtualDir parent = nullptr); std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; diff --git a/src/core/frontend/emu_window.cpp b/src/core/frontend/emu_window.cpp index 2d776c693..9dd493efb 100644 --- a/src/core/frontend/emu_window.cpp +++ b/src/core/frontend/emu_window.cpp @@ -8,6 +8,8 @@ #include "core/frontend/input.h" #include "core/settings.h" +namespace Core::Frontend { + class EmuWindow::TouchState : public Input::Factory<Input::TouchDevice>, public std::enable_shared_from_this<TouchState> { public: @@ -108,3 +110,5 @@ void EmuWindow::TouchMoved(unsigned framebuffer_x, unsigned framebuffer_y) { void EmuWindow::UpdateCurrentFramebufferLayout(unsigned width, unsigned height) { NotifyFramebufferLayoutChanged(Layout::DefaultFrameLayout(width, height)); } + +} // namespace Core::Frontend diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h index e8c29adfb..7006a37b3 100644 --- a/src/core/frontend/emu_window.h +++ b/src/core/frontend/emu_window.h @@ -10,6 +10,8 @@ #include "common/common_types.h" #include "core/frontend/framebuffer_layout.h" +namespace Core::Frontend { + /** * Abstraction class used to provide an interface between emulation code and the frontend * (e.g. SDL, QGLWidget, GLFW, etc...). @@ -32,9 +34,9 @@ class EmuWindow { public: /// Data structure to store emuwindow configuration struct WindowConfig { - bool fullscreen; - int res_width; - int res_height; + bool fullscreen = false; + int res_width = 0; + int res_height = 0; std::pair<unsigned, unsigned> min_client_area_size; }; @@ -166,3 +168,5 @@ private: */ std::tuple<unsigned, unsigned> ClipToTouchScreen(unsigned new_x, unsigned new_y); }; + +} // namespace Core::Frontend diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index 1b0cd0abf..8c19e86d3 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -11,7 +11,7 @@ namespace Kernel { -unsigned int Object::next_object_id; +std::atomic<u32> Object::next_object_id{0}; /// Initialize the kernel void Init() { diff --git a/src/core/hle/kernel/object.h b/src/core/hle/kernel/object.h index 83df68dfd..526ac9cc3 100644 --- a/src/core/hle/kernel/object.h +++ b/src/core/hle/kernel/object.h @@ -4,6 +4,7 @@ #pragma once +#include <atomic> #include <string> #include <utility> @@ -42,8 +43,8 @@ public: virtual ~Object(); /// Returns a unique identifier for the object. For debugging purposes only. - unsigned int GetObjectId() const { - return object_id; + u32 GetObjectId() const { + return object_id.load(std::memory_order_relaxed); } virtual std::string GetTypeName() const { @@ -61,23 +62,23 @@ public: bool IsWaitable() const; public: - static unsigned int next_object_id; + static std::atomic<u32> next_object_id; private: friend void intrusive_ptr_add_ref(Object*); friend void intrusive_ptr_release(Object*); - unsigned int ref_count = 0; - unsigned int object_id = next_object_id++; + std::atomic<u32> ref_count{0}; + std::atomic<u32> object_id{next_object_id++}; }; // Special functions used by boost::instrusive_ptr to do automatic ref-counting inline void intrusive_ptr_add_ref(Object* object) { - ++object->ref_count; + object->ref_count.fetch_add(1, std::memory_order_relaxed); } inline void intrusive_ptr_release(Object* object) { - if (--object->ref_count == 0) { + if (object->ref_count.fetch_sub(1, std::memory_order_acq_rel) == 1) { delete object; } } diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp index 94065c736..e770b9103 100644 --- a/src/core/hle/kernel/scheduler.cpp +++ b/src/core/hle/kernel/scheduler.cpp @@ -25,7 +25,7 @@ Scheduler::~Scheduler() { } } -bool Scheduler::HaveReadyThreads() { +bool Scheduler::HaveReadyThreads() const { std::lock_guard<std::mutex> lock(scheduler_mutex); return ready_queue.get_first() != nullptr; } diff --git a/src/core/hle/kernel/scheduler.h b/src/core/hle/kernel/scheduler.h index 1a4ee8f36..6a61ef64e 100644 --- a/src/core/hle/kernel/scheduler.h +++ b/src/core/hle/kernel/scheduler.h @@ -21,7 +21,7 @@ public: ~Scheduler(); /// Returns whether there are any threads that are ready to run. - bool HaveReadyThreads(); + bool HaveReadyThreads() const; /// Reschedules to the next available thread (call after current thread is suspended) void Reschedule(); diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp index 93560152f..51a1ec160 100644 --- a/src/core/hle/kernel/server_session.cpp +++ b/src/core/hle/kernel/server_session.cpp @@ -71,6 +71,14 @@ ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& con const u32 object_id{context.GetDomainMessageHeader()->object_id}; switch (domain_message_header->command) { case IPC::DomainMessageHeader::CommandType::SendMessage: + if (object_id > domain_request_handlers.size()) { + LOG_CRITICAL(IPC, + "object_id {} is too big! This probably means a recent service call " + "to {} needed to return a new interface!", + object_id, name); + UNREACHABLE(); + return RESULT_SUCCESS; // Ignore error if asserts are off + } return domain_request_handlers[object_id - 1]->HandleSyncRequest(context); case IPC::DomainMessageHeader::CommandType::CloseVirtualHandle: { @@ -144,7 +152,7 @@ ResultCode ServerSession::HandleSyncRequest(SharedPtr<Thread> thread) { // Handle scenario when ConvertToDomain command was issued, as we must do the conversion at the // end of the command such that only commands following this one are handled as domains if (convert_to_domain) { - ASSERT_MSG(domain_request_handlers.empty(), "already a domain"); + ASSERT_MSG(IsSession(), "ServerSession is already a domain instance."); domain_request_handlers = {hle_handler}; convert_to_domain = false; } diff --git a/src/core/hle/kernel/server_session.h b/src/core/hle/kernel/server_session.h index 2bce54fee..1a88e66b9 100644 --- a/src/core/hle/kernel/server_session.h +++ b/src/core/hle/kernel/server_session.h @@ -97,7 +97,12 @@ public: /// Returns true if the session has been converted to a domain, otherwise False bool IsDomain() const { - return !domain_request_handlers.empty(); + return !IsSession(); + } + + /// Returns true if this session has not been converted to a domain, otherwise false. + bool IsSession() const { + return domain_request_handlers.empty(); } /// Converts the session to a domain at the end of the current command diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index 5db2db687..6be5c474e 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -250,8 +250,11 @@ static ResultCode ArbitrateUnlock(VAddr mutex_addr) { } /// Break program execution -static void Break(u64 unk_0, u64 unk_1, u64 unk_2) { - LOG_CRITICAL(Debug_Emulated, "Emulated program broke execution!"); +static void Break(u64 reason, u64 info1, u64 info2) { + LOG_CRITICAL( + Debug_Emulated, + "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}", + reason, info1, info2); ASSERT(false); } @@ -532,7 +535,6 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V CASCADE_RESULT(thread->guest_handle, g_handle_table.Create(thread)); *out_handle = thread->guest_handle; - Core::System::GetInstance().PrepareReschedule(); Core::System::GetInstance().CpuCore(thread->processor_id).PrepareReschedule(); LOG_TRACE(Kernel_SVC, @@ -706,8 +708,7 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask); auto owner = g_handle_table.Get<Thread>(owner_handle); ASSERT(owner); - ASSERT(thread->status != ThreadStatus::Running); - thread->status = ThreadStatus::WaitMutex; + ASSERT(thread->status == ThreadStatus::WaitMutex); thread->wakeup_callback = nullptr; owner->AddMutexWaiter(thread); diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index b9022feae..cf4f94822 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -23,6 +23,7 @@ #include "core/hle/kernel/object.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/thread.h" +#include "core/hle/lock.h" #include "core/hle/result.h" #include "core/memory.h" @@ -104,6 +105,10 @@ void ExitCurrentThread() { */ static void ThreadWakeupCallback(u64 thread_handle, int cycles_late) { const auto proper_handle = static_cast<Handle>(thread_handle); + + // Lock the global kernel mutex when we enter the kernel HLE. + std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock); + SharedPtr<Thread> thread = wakeup_callback_handle_table.Get<Thread>(proper_handle); if (thread == nullptr) { LOG_CRITICAL(Kernel, "Callback fired for invalid thread {:08X}", proper_handle); @@ -155,12 +160,14 @@ void Thread::WakeAfterDelay(s64 nanoseconds) { if (nanoseconds == -1) return; - CoreTiming::ScheduleEvent(CoreTiming::nsToCycles(nanoseconds), ThreadWakeupEventType, - callback_handle); + // This function might be called from any thread so we have to be cautious and use the + // thread-safe version of ScheduleEvent. + CoreTiming::ScheduleEventThreadsafe(CoreTiming::nsToCycles(nanoseconds), ThreadWakeupEventType, + callback_handle); } void Thread::CancelWakeupTimer() { - CoreTiming::UnscheduleEvent(ThreadWakeupEventType, callback_handle); + CoreTiming::UnscheduleEventThreadsafe(ThreadWakeupEventType, callback_handle); } static boost::optional<s32> GetNextProcessorId(u64 mask) { @@ -419,12 +426,33 @@ VAddr Thread::GetCommandBufferAddress() const { } void Thread::AddMutexWaiter(SharedPtr<Thread> thread) { + if (thread->lock_owner == this) { + // If the thread is already waiting for this thread to release the mutex, ensure that the + // waiters list is consistent and return without doing anything. + auto itr = std::find(wait_mutex_threads.begin(), wait_mutex_threads.end(), thread); + ASSERT(itr != wait_mutex_threads.end()); + return; + } + + // A thread can't wait on two different mutexes at the same time. + ASSERT(thread->lock_owner == nullptr); + + // Ensure that the thread is not already in the list of mutex waiters + auto itr = std::find(wait_mutex_threads.begin(), wait_mutex_threads.end(), thread); + ASSERT(itr == wait_mutex_threads.end()); + thread->lock_owner = this; wait_mutex_threads.emplace_back(std::move(thread)); UpdatePriority(); } void Thread::RemoveMutexWaiter(SharedPtr<Thread> thread) { + ASSERT(thread->lock_owner == this); + + // Ensure that the thread is in the list of mutex waiters + auto itr = std::find(wait_mutex_threads.begin(), wait_mutex_threads.end(), thread); + ASSERT(itr != wait_mutex_threads.end()); + boost::remove_erase(wait_mutex_threads, thread); thread->lock_owner = nullptr; UpdatePriority(); diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index 762763463..c524e7a48 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -145,8 +145,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger {51, nullptr, "ApproveToDisplay"}, {60, nullptr, "OverrideAutoSleepTimeAndDimmingTime"}, {61, nullptr, "SetMediaPlaybackState"}, - {62, nullptr, "SetIdleTimeDetectionExtension"}, - {63, nullptr, "GetIdleTimeDetectionExtension"}, + {62, &ISelfController::SetIdleTimeDetectionExtension, "SetIdleTimeDetectionExtension"}, + {63, &ISelfController::GetIdleTimeDetectionExtension, "GetIdleTimeDetectionExtension"}, {64, nullptr, "SetInputDetectionSourceSet"}, {65, nullptr, "ReportUserIsActive"}, {66, nullptr, "GetCurrentIlluminance"}, @@ -281,6 +281,23 @@ void ISelfController::SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx) LOG_WARNING(Service_AM, "(STUBBED) called"); } +void ISelfController::SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + idle_time_detection_extension = rp.Pop<u32>(); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + + LOG_WARNING(Service_AM, "(STUBBED) called"); +} + +void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push<u32>(idle_time_detection_extension); + + LOG_WARNING(Service_AM, "(STUBBED) called"); +} + ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter") { static const FunctionInfo functions[] = { {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"}, @@ -306,7 +323,8 @@ ICommonStateGetter::ICommonStateGetter() : ServiceFramework("ICommonStateGetter" {52, nullptr, "SwitchLcdBacklight"}, {55, nullptr, "IsInControllerFirmwareUpdateSection"}, {60, nullptr, "GetDefaultDisplayResolution"}, - {61, nullptr, "GetDefaultDisplayResolutionChangeEvent"}, + {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent, + "GetDefaultDisplayResolutionChangeEvent"}, {62, nullptr, "GetHdcpAuthenticationState"}, {63, nullptr, "GetHdcpAuthenticationStateChangeEvent"}, }; @@ -341,6 +359,16 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_AM, "(STUBBED) called"); } +void ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent(Kernel::HLERequestContext& ctx) { + event->Signal(); + + IPC::ResponseBuilder rb{ctx, 2, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushCopyObjects(event); + + LOG_WARNING(Service_AM, "(STUBBED) called"); +} + void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) { const bool use_docked_mode{Settings::values.use_docked_mode}; IPC::ResponseBuilder rb{ctx, 3}; diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h index 862f338ac..b763aff6f 100644 --- a/src/core/hle/service/am/am.h +++ b/src/core/hle/service/am/am.h @@ -87,9 +87,12 @@ private: void CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx); void SetScreenShotPermission(Kernel::HLERequestContext& ctx); void SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx); + void SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx); + void GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx); std::shared_ptr<NVFlinger::NVFlinger> nvflinger; Kernel::SharedPtr<Kernel::Event> launchable_event; + u32 idle_time_detection_extension = 0; }; class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> { @@ -110,6 +113,7 @@ private: void GetEventHandle(Kernel::HLERequestContext& ctx); void ReceiveMessage(Kernel::HLERequestContext& ctx); void GetCurrentFocusState(Kernel::HLERequestContext& ctx); + void GetDefaultDisplayResolutionChangeEvent(Kernel::HLERequestContext& ctx); void GetOperationMode(Kernel::HLERequestContext& ctx); void GetPerformanceMode(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp index 108a7c6eb..ce709ccf4 100644 --- a/src/core/hle/service/audio/audout_u.cpp +++ b/src/core/hle/service/audio/audout_u.cpp @@ -28,7 +28,7 @@ constexpr int DefaultSampleRate{48000}; class IAudioOut final : public ServiceFramework<IAudioOut> { public: IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core) - : ServiceFramework("IAudioOut"), audio_params(audio_params), audio_core(audio_core) { + : ServiceFramework("IAudioOut"), audio_core(audio_core), audio_params(audio_params) { static const FunctionInfo functions[] = { {0, &IAudioOut::GetAudioOutState, "GetAudioOutState"}, diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index f99304de5..9e75eb3a6 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -20,9 +20,9 @@ public: explicit IAudioRenderer(AudioCore::AudioRendererParameter audren_params) : ServiceFramework("IAudioRenderer") { static const FunctionInfo functions[] = { - {0, nullptr, "GetAudioRendererSampleRate"}, - {1, nullptr, "GetAudioRendererSampleCount"}, - {2, nullptr, "GetAudioRendererMixBufferCount"}, + {0, &IAudioRenderer::GetAudioRendererSampleRate, "GetAudioRendererSampleRate"}, + {1, &IAudioRenderer::GetAudioRendererSampleCount, "GetAudioRendererSampleCount"}, + {2, &IAudioRenderer::GetAudioRendererMixBufferCount, "GetAudioRendererMixBufferCount"}, {3, nullptr, "GetAudioRendererState"}, {4, &IAudioRenderer::RequestUpdateAudioRenderer, "RequestUpdateAudioRenderer"}, {5, &IAudioRenderer::StartAudioRenderer, "StartAudioRenderer"}, @@ -45,6 +45,27 @@ private: system_event->Signal(); } + void GetAudioRendererSampleRate(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push<u32>(renderer->GetSampleRate()); + LOG_DEBUG(Service_Audio, "called"); + } + + void GetAudioRendererSampleCount(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push<u32>(renderer->GetSampleCount()); + LOG_DEBUG(Service_Audio, "called"); + } + + void GetAudioRendererMixBufferCount(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push<u32>(renderer->GetMixBufferCount()); + LOG_DEBUG(Service_Audio, "called"); + } + void RequestUpdateAudioRenderer(Kernel::HLERequestContext& ctx) { ctx.WriteBuffer(renderer->UpdateAudioRenderer(ctx.ReadBuffer())); IPC::ResponseBuilder rb{ctx, 2}; @@ -169,7 +190,8 @@ AudRenU::AudRenU() : ServiceFramework("audren:u") { {1, &AudRenU::GetAudioRendererWorkBufferSize, "GetAudioRendererWorkBufferSize"}, {2, &AudRenU::GetAudioDevice, "GetAudioDevice"}, {3, nullptr, "OpenAudioRendererAuto"}, - {4, nullptr, "GetAudioDeviceServiceWithRevisionInfo"}, + {4, &AudRenU::GetAudioDeviceServiceWithRevisionInfo, + "GetAudioDeviceServiceWithRevisionInfo"}, }; RegisterHandlers(functions); } @@ -189,7 +211,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; auto params = rp.PopRaw<AudioCore::AudioRendererParameter>(); - u64 buffer_sz = Common::AlignUp(4 * params.unknown_8, 0x40); + u64 buffer_sz = Common::AlignUp(4 * params.mix_buffer_count, 0x40); buffer_sz += params.unknown_c * 1024; buffer_sz += 0x940 * (params.unknown_c + 1); buffer_sz += 0x3F0 * params.voice_count; @@ -197,7 +219,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) { buffer_sz += Common::AlignUp(8 * params.voice_count, 0x10); buffer_sz += Common::AlignUp((0x3C0 * (params.sink_count + params.unknown_c) + 4 * params.sample_count) * - (params.unknown_8 + 6), + (params.mix_buffer_count + 6), 0x40); if (IsFeatureSupported(AudioFeatures::Splitter, params.revision)) { @@ -253,6 +275,16 @@ void AudRenU::GetAudioDevice(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Audio, "called"); } +void AudRenU::GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<Audio::IAudioDevice>(); + + LOG_WARNING(Service_Audio, "(STUBBED) called"); // TODO(ogniK): Figure out what is different + // based on the current revision +} + bool AudRenU::IsFeatureSupported(AudioFeatures feature, u32_le revision) const { u32_be version_num = (revision - Common::MakeMagic('R', 'E', 'V', '0')); // Byte swap switch (feature) { diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h index 14907f8ae..8600ac6e4 100644 --- a/src/core/hle/service/audio/audren_u.h +++ b/src/core/hle/service/audio/audren_u.h @@ -22,6 +22,7 @@ private: void OpenAudioRenderer(Kernel::HLERequestContext& ctx); void GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx); void GetAudioDevice(Kernel::HLERequestContext& ctx); + void GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& ctx); enum class AudioFeatures : u32 { Splitter, diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp index 5e416cde2..da658cbe6 100644 --- a/src/core/hle/service/filesystem/filesystem.cpp +++ b/src/core/hle/service/filesystem/filesystem.cpp @@ -226,6 +226,7 @@ ResultVal<FileSys::EntryType> VfsDirectoryServiceWrapper::GetEntryType( static std::unique_ptr<FileSys::RomFSFactory> romfs_factory; static std::unique_ptr<FileSys::SaveDataFactory> save_data_factory; static std::unique_ptr<FileSys::SDMCFactory> sdmc_factory; +static std::unique_ptr<FileSys::BISFactory> bis_factory; ResultCode RegisterRomFS(std::unique_ptr<FileSys::RomFSFactory>&& factory) { ASSERT_MSG(romfs_factory == nullptr, "Tried to register a second RomFS"); @@ -248,6 +249,13 @@ ResultCode RegisterSDMC(std::unique_ptr<FileSys::SDMCFactory>&& factory) { return RESULT_SUCCESS; } +ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory) { + ASSERT_MSG(bis_factory == nullptr, "Tried to register a second BIS"); + bis_factory = std::move(factory); + LOG_DEBUG(Service_FS, "Registred BIS"); + return RESULT_SUCCESS; +} + ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id) { LOG_TRACE(Service_FS, "Opening RomFS for title_id={:016X}", title_id); @@ -281,6 +289,14 @@ ResultVal<FileSys::VirtualDir> OpenSDMC() { return sdmc_factory->Open(); } +std::shared_ptr<FileSys::RegisteredCache> GetSystemNANDContents() { + return bis_factory->GetSystemNANDContents(); +} + +std::shared_ptr<FileSys::RegisteredCache> GetUserNANDContents() { + return bis_factory->GetUserNANDContents(); +} + void RegisterFileSystems(const FileSys::VirtualFilesystem& vfs) { romfs_factory = nullptr; save_data_factory = nullptr; @@ -291,6 +307,9 @@ void RegisterFileSystems(const FileSys::VirtualFilesystem& vfs) { auto sd_directory = vfs->OpenDirectory(FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir), FileSys::Mode::ReadWrite); + if (bis_factory == nullptr) + bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory); + auto savedata = std::make_unique<FileSys::SaveDataFactory>(std::move(nand_directory)); save_data_factory = std::move(savedata); diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h index 462c13f20..1d6f922dd 100644 --- a/src/core/hle/service/filesystem/filesystem.h +++ b/src/core/hle/service/filesystem/filesystem.h @@ -6,6 +6,7 @@ #include <memory> #include "common/common_types.h" +#include "core/file_sys/bis_factory.h" #include "core/file_sys/directory.h" #include "core/file_sys/mode.h" #include "core/file_sys/romfs_factory.h" @@ -24,16 +25,15 @@ namespace FileSystem { ResultCode RegisterRomFS(std::unique_ptr<FileSys::RomFSFactory>&& factory); ResultCode RegisterSaveData(std::unique_ptr<FileSys::SaveDataFactory>&& factory); ResultCode RegisterSDMC(std::unique_ptr<FileSys::SDMCFactory>&& factory); +ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory); -// TODO(DarkLordZach): BIS Filesystem -// ResultCode RegisterBIS(std::unique_ptr<FileSys::BISFactory>&& factory); ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id); ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space, FileSys::SaveDataDescriptor save_struct); ResultVal<FileSys::VirtualDir> OpenSDMC(); -// TODO(DarkLordZach): BIS Filesystem -// ResultVal<std::unique_ptr<FileSys::FileSystemBackend>> OpenBIS(); +std::shared_ptr<FileSys::RegisteredCache> GetSystemNANDContents(); +std::shared_ptr<FileSys::RegisteredCache> GetUserNANDContents(); /// Registers all Filesystem services with the specified service manager. void InstallInterfaces(SM::ServiceManager& service_manager, const FileSys::VirtualFilesystem& vfs); diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp index fb4d89068..f2b0e509a 100644 --- a/src/core/hle/service/friend/friend.cpp +++ b/src/core/hle/service/friend/friend.cpp @@ -9,10 +9,110 @@ namespace Service::Friend { +class IFriendService final : public ServiceFramework<IFriendService> { +public: + IFriendService() : ServiceFramework("IFriendService") { + static const FunctionInfo functions[] = { + {0, nullptr, "GetCompletionEvent"}, + {1, nullptr, "Cancel"}, + {10100, nullptr, "GetFriendListIds"}, + {10101, nullptr, "GetFriendList"}, + {10102, nullptr, "UpdateFriendInfo"}, + {10110, nullptr, "GetFriendProfileImage"}, + {10200, nullptr, "SendFriendRequestForApplication"}, + {10211, nullptr, "AddFacedFriendRequestForApplication"}, + {10400, nullptr, "GetBlockedUserListIds"}, + {10500, nullptr, "GetProfileList"}, + {10600, nullptr, "DeclareOpenOnlinePlaySession"}, + {10601, &IFriendService::DeclareCloseOnlinePlaySession, + "DeclareCloseOnlinePlaySession"}, + {10610, &IFriendService::UpdateUserPresence, "UpdateUserPresence"}, + {10700, nullptr, "GetPlayHistoryRegistrationKey"}, + {10701, nullptr, "GetPlayHistoryRegistrationKeyWithNetworkServiceAccountId"}, + {10702, nullptr, "AddPlayHistory"}, + {11000, nullptr, "GetProfileImageUrl"}, + {20100, nullptr, "GetFriendCount"}, + {20101, nullptr, "GetNewlyFriendCount"}, + {20102, nullptr, "GetFriendDetailedInfo"}, + {20103, nullptr, "SyncFriendList"}, + {20104, nullptr, "RequestSyncFriendList"}, + {20110, nullptr, "LoadFriendSetting"}, + {20200, nullptr, "GetReceivedFriendRequestCount"}, + {20201, nullptr, "GetFriendRequestList"}, + {20300, nullptr, "GetFriendCandidateList"}, + {20301, nullptr, "GetNintendoNetworkIdInfo"}, + {20302, nullptr, "GetSnsAccountLinkage"}, + {20303, nullptr, "GetSnsAccountProfile"}, + {20304, nullptr, "GetSnsAccountFriendList"}, + {20400, nullptr, "GetBlockedUserList"}, + {20401, nullptr, "SyncBlockedUserList"}, + {20500, nullptr, "GetProfileExtraList"}, + {20501, nullptr, "GetRelationship"}, + {20600, nullptr, "GetUserPresenceView"}, + {20700, nullptr, "GetPlayHistoryList"}, + {20701, nullptr, "GetPlayHistoryStatistics"}, + {20800, nullptr, "LoadUserSetting"}, + {20801, nullptr, "SyncUserSetting"}, + {20900, nullptr, "RequestListSummaryOverlayNotification"}, + {21000, nullptr, "GetExternalApplicationCatalog"}, + {30100, nullptr, "DropFriendNewlyFlags"}, + {30101, nullptr, "DeleteFriend"}, + {30110, nullptr, "DropFriendNewlyFlag"}, + {30120, nullptr, "ChangeFriendFavoriteFlag"}, + {30121, nullptr, "ChangeFriendOnlineNotificationFlag"}, + {30200, nullptr, "SendFriendRequest"}, + {30201, nullptr, "SendFriendRequestWithApplicationInfo"}, + {30202, nullptr, "CancelFriendRequest"}, + {30203, nullptr, "AcceptFriendRequest"}, + {30204, nullptr, "RejectFriendRequest"}, + {30205, nullptr, "ReadFriendRequest"}, + {30210, nullptr, "GetFacedFriendRequestRegistrationKey"}, + {30211, nullptr, "AddFacedFriendRequest"}, + {30212, nullptr, "CancelFacedFriendRequest"}, + {30213, nullptr, "GetFacedFriendRequestProfileImage"}, + {30214, nullptr, "GetFacedFriendRequestProfileImageFromPath"}, + {30215, nullptr, "SendFriendRequestWithExternalApplicationCatalogId"}, + {30216, nullptr, "ResendFacedFriendRequest"}, + {30217, nullptr, "SendFriendRequestWithNintendoNetworkIdInfo"}, + {30300, nullptr, "GetSnsAccountLinkPageUrl"}, + {30301, nullptr, "UnlinkSnsAccount"}, + {30400, nullptr, "BlockUser"}, + {30401, nullptr, "BlockUserWithApplicationInfo"}, + {30402, nullptr, "UnblockUser"}, + {30500, nullptr, "GetProfileExtraFromFriendCode"}, + {30700, nullptr, "DeletePlayHistory"}, + {30810, nullptr, "ChangePresencePermission"}, + {30811, nullptr, "ChangeFriendRequestReception"}, + {30812, nullptr, "ChangePlayLogPermission"}, + {30820, nullptr, "IssueFriendCode"}, + {30830, nullptr, "ClearPlayLog"}, + {49900, nullptr, "DeleteNetworkServiceAccountCache"}, + }; + + RegisterHandlers(functions); + } + +private: + void DeclareCloseOnlinePlaySession(Kernel::HLERequestContext& ctx) { + // Stub used by Splatoon 2 + LOG_WARNING(Service_ACC, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void UpdateUserPresence(Kernel::HLERequestContext& ctx) { + // Stub used by Retro City Rampage + LOG_WARNING(Service_ACC, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } +}; + void Module::Interface::CreateFriendService(Kernel::HLERequestContext& ctx) { - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - LOG_WARNING(Service_Friend, "(STUBBED) called"); + rb.PushIpcInterface<IFriendService>(); + LOG_DEBUG(Service_ACC, "called"); } Module::Interface::Interface(std::shared_ptr<Module> module, const char* name) diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index dcdfa0e19..970942d3f 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -291,6 +291,7 @@ private: class Hid final : public ServiceFramework<Hid> { public: Hid() : ServiceFramework("hid") { + // clang-format off static const FunctionInfo functions[] = { {0, &Hid::CreateAppletResource, "CreateAppletResource"}, {1, &Hid::ActivateDebugPad, "ActivateDebugPad"}, @@ -333,15 +334,13 @@ public: {102, &Hid::SetSupportedNpadIdType, "SetSupportedNpadIdType"}, {103, &Hid::ActivateNpad, "ActivateNpad"}, {104, nullptr, "DeactivateNpad"}, - {106, &Hid::AcquireNpadStyleSetUpdateEventHandle, - "AcquireNpadStyleSetUpdateEventHandle"}, - {107, nullptr, "DisconnectNpad"}, + {106, &Hid::AcquireNpadStyleSetUpdateEventHandle, "AcquireNpadStyleSetUpdateEventHandle"}, + {107, &Hid::DisconnectNpad, "DisconnectNpad"}, {108, &Hid::GetPlayerLedPattern, "GetPlayerLedPattern"}, {109, nullptr, "ActivateNpadWithRevision"}, {120, &Hid::SetNpadJoyHoldType, "SetNpadJoyHoldType"}, {121, &Hid::GetNpadJoyHoldType, "GetNpadJoyHoldType"}, - {122, &Hid::SetNpadJoyAssignmentModeSingleByDefault, - "SetNpadJoyAssignmentModeSingleByDefault"}, + {122, &Hid::SetNpadJoyAssignmentModeSingleByDefault, "SetNpadJoyAssignmentModeSingleByDefault"}, {123, nullptr, "SetNpadJoyAssignmentModeSingleByDefault"}, {124, &Hid::SetNpadJoyAssignmentModeDual, "SetNpadJoyAssignmentModeDual"}, {125, &Hid::MergeSingleJoyAsDualJoy, "MergeSingleJoyAsDualJoy"}, @@ -398,6 +397,8 @@ public: {1000, nullptr, "SetNpadCommunicationMode"}, {1001, nullptr, "GetNpadCommunicationMode"}, }; + // clang-format on + RegisterHandlers(functions); event = Kernel::Event::Create(Kernel::ResetType::OneShot, "hid:EventHandle"); @@ -496,6 +497,12 @@ private: LOG_WARNING(Service_HID, "(STUBBED) called"); } + void DisconnectNpad(Kernel::HLERequestContext& ctx) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + LOG_WARNING(Service_HID, "(STUBBED) called"); + } + void GetPlayerLedPattern(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); diff --git a/src/core/hle/service/lm/lm.cpp b/src/core/hle/service/lm/lm.cpp index 2e99ddf51..098da2a41 100644 --- a/src/core/hle/service/lm/lm.cpp +++ b/src/core/hle/service/lm/lm.cpp @@ -92,7 +92,11 @@ private: // Parse out log metadata u32 line{}; - std::string message, filename, function; + std::string module; + std::string message; + std::string filename; + std::string function; + std::string thread; while (addr < end_addr) { const Field field{static_cast<Field>(Memory::Read8(addr++))}; const size_t length{Memory::Read8(addr++)}; @@ -102,6 +106,8 @@ private: } switch (field) { + case Field::Skip: + break; case Field::Message: message = Memory::ReadCString(addr, length); break; @@ -114,6 +120,12 @@ private: case Field::Function: function = Memory::ReadCString(addr, length); break; + case Field::Module: + module = Memory::ReadCString(addr, length); + break; + case Field::Thread: + thread = Memory::ReadCString(addr, length); + break; } addr += length; @@ -128,12 +140,18 @@ private: if (!filename.empty()) { log_stream << filename << ':'; } + if (!module.empty()) { + log_stream << module << ':'; + } if (!function.empty()) { log_stream << function << ':'; } if (line) { log_stream << std::to_string(line) << ':'; } + if (!thread.empty()) { + log_stream << thread << ':'; + } if (log_stream.str().length() > 0 && log_stream.str().back() == ':') { log_stream << ' '; } @@ -142,7 +160,7 @@ private: if (header.IsTailLog()) { switch (header.severity) { case MessageHeader::Severity::Trace: - LOG_TRACE(Debug_Emulated, "{}", log_stream.str()); + LOG_DEBUG(Debug_Emulated, "{}", log_stream.str()); break; case MessageHeader::Severity::Info: LOG_INFO(Debug_Emulated, "{}", log_stream.str()); diff --git a/src/core/hle/service/mm/mm_u.cpp b/src/core/hle/service/mm/mm_u.cpp index 08f45b78a..7b91bb258 100644 --- a/src/core/hle/service/mm/mm_u.cpp +++ b/src/core/hle/service/mm/mm_u.cpp @@ -9,42 +9,63 @@ namespace Service::MM { -void InstallInterfaces(SM::ServiceManager& service_manager) { - std::make_shared<MM_U>()->InstallAsService(service_manager); -} +class MM_U final : public ServiceFramework<MM_U> { +public: + explicit MM_U() : ServiceFramework{"mm:u"} { + // clang-format off + static const FunctionInfo functions[] = { + {0, &MM_U::Initialize, "InitializeOld"}, + {1, &MM_U::Finalize, "FinalizeOld"}, + {2, &MM_U::SetAndWait, "SetAndWaitOld"}, + {3, &MM_U::Get, "GetOld"}, + {4, &MM_U::Initialize, "Initialize"}, + {5, &MM_U::Finalize, "Finalize"}, + {6, &MM_U::SetAndWait, "SetAndWait"}, + {7, &MM_U::Get, "Get"}, + }; + // clang-format on -void MM_U::Initialize(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_MM, "(STUBBED) called"); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); -} + RegisterHandlers(functions); + } -void MM_U::SetAndWait(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - min = rp.Pop<u32>(); - max = rp.Pop<u32>(); - current = min; +private: + void Initialize(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_MM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } - LOG_WARNING(Service_MM, "(STUBBED) called, min=0x{:X}, max=0x{:X}", min, max); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); -} + void Finalize(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_MM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } -void MM_U::Get(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_MM, "(STUBBED) called"); - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push(current); -} + void SetAndWait(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + min = rp.Pop<u32>(); + max = rp.Pop<u32>(); + current = min; -MM_U::MM_U() : ServiceFramework("mm:u") { - static const FunctionInfo functions[] = { - {0, nullptr, "InitializeOld"}, {1, nullptr, "FinalizeOld"}, - {2, nullptr, "SetAndWaitOld"}, {3, nullptr, "GetOld"}, - {4, &MM_U::Initialize, "Initialize"}, {5, nullptr, "Finalize"}, - {6, &MM_U::SetAndWait, "SetAndWait"}, {7, &MM_U::Get, "Get"}, - }; - RegisterHandlers(functions); + LOG_WARNING(Service_MM, "(STUBBED) called, min=0x{:X}, max=0x{:X}", min, max); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void Get(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_MM, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(current); + } + + u32 min{0}; + u32 max{0}; + u32 current{0}; +}; + +void InstallInterfaces(SM::ServiceManager& service_manager) { + std::make_shared<MM_U>()->InstallAsService(service_manager); } } // namespace Service::MM diff --git a/src/core/hle/service/mm/mm_u.h b/src/core/hle/service/mm/mm_u.h index 79eeedf9c..5439fa653 100644 --- a/src/core/hle/service/mm/mm_u.h +++ b/src/core/hle/service/mm/mm_u.h @@ -8,21 +8,6 @@ namespace Service::MM { -class MM_U final : public ServiceFramework<MM_U> { -public: - MM_U(); - ~MM_U() = default; - -private: - void Initialize(Kernel::HLERequestContext& ctx); - void SetAndWait(Kernel::HLERequestContext& ctx); - void Get(Kernel::HLERequestContext& ctx); - - u32 min{0}; - u32 max{0}; - u32 current{0}; -}; - /// Registers all MM services with the specified service manager. void InstallInterfaces(SM::ServiceManager& service_manager); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp new file mode 100644 index 000000000..51f01077b --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp @@ -0,0 +1,34 @@ +// Copyright 2018 yuzu emulator team +// 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/hle/service/nvdrv/devices/nvhost_nvjpg.h" + +namespace Service::Nvidia::Devices { + +u32 nvhost_nvjpg::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + 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); + } + + UNIMPLEMENTED_MSG("Unimplemented ioctl"); + return 0; +} + +u32 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; +} + +} // 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 new file mode 100644 index 000000000..2b0eb43ee --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h @@ -0,0 +1,36 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "common/common_types.h" +#include "common/swap.h" +#include "core/hle/service/nvdrv/devices/nvdevice.h" + +namespace Service::Nvidia::Devices { + +class nvhost_nvjpg final : public nvdevice { +public: + nvhost_nvjpg() = default; + ~nvhost_nvjpg() override = default; + + u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) 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); +}; + +} // 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 new file mode 100644 index 000000000..fcb488d50 --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp @@ -0,0 +1,34 @@ +// Copyright 2018 yuzu emulator team +// 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/hle/service/nvdrv/devices/nvhost_vic.h" + +namespace Service::Nvidia::Devices { + +u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + 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); + } + + UNIMPLEMENTED_MSG("Unimplemented ioctl"); + return 0; +} + +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); + nvmap_fd = params.nvmap_fd; + return 0; +} + +} // 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 new file mode 100644 index 000000000..c7d681e52 --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.h @@ -0,0 +1,36 @@ +// Copyright 2018 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "common/common_types.h" +#include "common/swap.h" +#include "core/hle/service/nvdrv/devices/nvdevice.h" + +namespace Service::Nvidia::Devices { + +class nvhost_vic final : public nvdevice { +public: + nvhost_vic() = default; + ~nvhost_vic() override = default; + + u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) 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); +}; + +} // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/nvdrv.cpp b/src/core/hle/service/nvdrv/nvdrv.cpp index 427f4b574..2de39822f 100644 --- a/src/core/hle/service/nvdrv/nvdrv.cpp +++ b/src/core/hle/service/nvdrv/nvdrv.cpp @@ -12,6 +12,8 @@ #include "core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h" #include "core/hle/service/nvdrv/devices/nvhost_gpu.h" #include "core/hle/service/nvdrv/devices/nvhost_nvdec.h" +#include "core/hle/service/nvdrv/devices/nvhost_nvjpg.h" +#include "core/hle/service/nvdrv/devices/nvhost_vic.h" #include "core/hle/service/nvdrv/devices/nvmap.h" #include "core/hle/service/nvdrv/interface.h" #include "core/hle/service/nvdrv/nvdrv.h" @@ -39,6 +41,8 @@ Module::Module() { devices["/dev/nvdisp_disp0"] = std::make_shared<Devices::nvdisp_disp0>(nvmap_dev); devices["/dev/nvhost-ctrl"] = std::make_shared<Devices::nvhost_ctrl>(); devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>(); + devices["/dev/nvhost-nvjpg"] = std::make_shared<Devices::nvhost_nvjpg>(); + devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>(); } u32 Module::Open(const std::string& device_name) { diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h index f86e1056c..db2e17c0c 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.h +++ b/src/core/hle/service/nvflinger/buffer_queue.h @@ -6,6 +6,7 @@ #include <vector> #include <boost/optional.hpp> +#include "common/common_funcs.h" #include "common/math_util.h" #include "common/swap.h" #include "core/hle/kernel/event.h" diff --git a/src/core/hle/service/service.h b/src/core/hle/service/service.h index 8a294c0f2..cd9c74f3d 100644 --- a/src/core/hle/service/service.h +++ b/src/core/hle/service/service.h @@ -23,7 +23,7 @@ class HLERequestContext; } // namespace Kernel namespace FileSys { -struct VfsFilesystem; +class VfsFilesystem; } namespace Service { diff --git a/src/core/hle/service/sm/controller.cpp b/src/core/hle/service/sm/controller.cpp index 518a0cc46..1cef73216 100644 --- a/src/core/hle/service/sm/controller.cpp +++ b/src/core/hle/service/sm/controller.cpp @@ -10,7 +10,7 @@ namespace Service::SM { void Controller::ConvertSessionToDomain(Kernel::HLERequestContext& ctx) { - ASSERT_MSG(!ctx.Session()->IsDomain(), "session is alread a domain"); + ASSERT_MSG(ctx.Session()->IsSession(), "Session is already a domain"); ctx.Session()->ConvertToDomain(); IPC::ResponseBuilder rb{ctx, 3}; @@ -41,7 +41,7 @@ void Controller::DuplicateSessionEx(Kernel::HLERequestContext& ctx) { void Controller::QueryPointerBufferSize(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0x500); + rb.Push<u16>(0x500); LOG_WARNING(Service, "(STUBBED) called"); } diff --git a/src/core/loader/deconstructed_rom_directory.cpp b/src/core/loader/deconstructed_rom_directory.cpp index 915d525b0..d575a9bea 100644 --- a/src/core/loader/deconstructed_rom_directory.cpp +++ b/src/core/loader/deconstructed_rom_directory.cpp @@ -83,13 +83,13 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load( if (dir == nullptr) { if (file == nullptr) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorNullFile; dir = file->GetContainingDirectory(); } const FileSys::VirtualFile npdm = dir->GetFile("main.npdm"); if (npdm == nullptr) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorMissingNPDM; ResultStatus result = metadata.Load(npdm); if (result != ResultStatus::Success) { @@ -99,7 +99,7 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load( const FileSys::ProgramAddressSpaceType arch_bits{metadata.GetAddressSpaceType()}; if (arch_bits == FileSys::ProgramAddressSpaceType::Is32Bit) { - return ResultStatus::ErrorUnsupportedArch; + return ResultStatus::Error32BitISA; } // Load NSO modules @@ -118,7 +118,6 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load( process->program_id = metadata.GetTitleID(); process->svc_access_mask.set(); - process->address_mappings = default_address_mappings; process->resource_limit = Kernel::ResourceLimit::GetForCategory(Kernel::ResourceLimitCategory::APPLICATION); process->Run(Memory::PROCESS_IMAGE_VADDR, metadata.GetMainThreadPriority(), @@ -143,28 +142,28 @@ ResultStatus AppLoader_DeconstructedRomDirectory::Load( ResultStatus AppLoader_DeconstructedRomDirectory::ReadRomFS(FileSys::VirtualFile& dir) { if (romfs == nullptr) - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoRomFS; dir = romfs; return ResultStatus::Success; } ResultStatus AppLoader_DeconstructedRomDirectory::ReadIcon(std::vector<u8>& buffer) { if (icon_data.empty()) - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoIcon; buffer = icon_data; return ResultStatus::Success; } ResultStatus AppLoader_DeconstructedRomDirectory::ReadProgramId(u64& out_program_id) { if (name.empty()) - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoControl; out_program_id = title_id; return ResultStatus::Success; } ResultStatus AppLoader_DeconstructedRomDirectory::ReadTitle(std::string& title) { if (name.empty()) - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoControl; title = name; return ResultStatus::Success; } diff --git a/src/core/loader/elf.cpp b/src/core/loader/elf.cpp index a7133f5a6..6420a7f11 100644 --- a/src/core/loader/elf.cpp +++ b/src/core/loader/elf.cpp @@ -390,7 +390,7 @@ ResultStatus AppLoader_ELF::Load(Kernel::SharedPtr<Kernel::Process>& process) { std::vector<u8> buffer = file->ReadAllBytes(); if (buffer.size() != file->GetSize()) - return ResultStatus::Error; + return ResultStatus::ErrorIncorrectELFFileSize; ElfReader elf_reader(&buffer[0]); SharedPtr<CodeSet> codeset = elf_reader.LoadInto(Memory::PROCESS_IMAGE_VADDR); @@ -398,7 +398,6 @@ ResultStatus AppLoader_ELF::Load(Kernel::SharedPtr<Kernel::Process>& process) { process->LoadModule(codeset, codeset->entrypoint); process->svc_access_mask.set(); - process->address_mappings = default_address_mappings; // Attach the default resource limit (APPLICATION) to the process process->resource_limit = diff --git a/src/core/loader/loader.cpp b/src/core/loader/loader.cpp index a288654df..70ef5d240 100644 --- a/src/core/loader/loader.cpp +++ b/src/core/loader/loader.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <memory> +#include <ostream> #include <string> #include "common/logging/log.h" #include "common/string_util.h" @@ -17,12 +18,6 @@ namespace Loader { -const std::initializer_list<Kernel::AddressMapping> default_address_mappings = { - {0x1FF50000, 0x8000, true}, // part of DSP RAM - {0x1FF70000, 0x8000, true}, // part of DSP RAM - {0x1F000000, 0x600000, false}, // entire VRAM -}; - FileType IdentifyFile(FileSys::VirtualFile file) { FileType type; @@ -46,6 +41,8 @@ FileType IdentifyFile(FileSys::VirtualFile file) { FileType GuessFromFilename(const std::string& name) { if (name == "main") return FileType::DeconstructedRomDirectory; + if (name == "00") + return FileType::NCA; const std::string extension = Common::ToLower(std::string(FileUtil::GetExtensionFromFilename(name))); @@ -86,6 +83,50 @@ std::string GetFileTypeString(FileType type) { return "unknown"; } +constexpr std::array<const char*, 36> RESULT_MESSAGES{ + "The operation completed successfully.", + "The loader requested to load is already loaded.", + "The operation is not implemented.", + "The loader is not initialized properly.", + "The NPDM file has a bad header.", + "The NPDM has a bad ACID header.", + "The NPDM has a bad ACI header,", + "The NPDM file has a bad file access control.", + "The NPDM has a bad file access header.", + "The PFS/HFS partition has a bad header.", + "The PFS/HFS partition has incorrect size as determined by the header.", + "The NCA file has a bad header.", + "The general keyfile could not be found.", + "The NCA Header key could not be found.", + "The NCA Header key is incorrect or the header is invalid.", + "Support for NCA2-type NCAs is not implemented.", + "Support for NCA0-type NCAs is not implemented.", + "The titlekey for this Rights ID could not be found.", + "The titlekek for this crypto revision could not be found.", + "The Rights ID in the header is invalid.", + "The key area key for this application type and crypto revision could not be found.", + "The key area key is incorrect or the section header is invalid.", + "The titlekey and/or titlekek is incorrect or the section header is invalid.", + "The XCI file is missing a Program-type NCA.", + "The NCA file is not an application.", + "The ExeFS partition could not be found.", + "The XCI file has a bad header.", + "The XCI file is missing a partition.", + "The file could not be found or does not exist.", + "The game is missing a program metadata file (main.npdm).", + "The game uses the currently-unimplemented 32-bit architecture.", + "The RomFS could not be found.", + "The ELF file has incorrect size as determined by the header.", + "There was a general error loading the NRO into emulated memory.", + "There is no icon available.", + "There is no control data available.", +}; + +std::ostream& operator<<(std::ostream& os, ResultStatus status) { + os << RESULT_MESSAGES.at(static_cast<size_t>(status)); + return os; +} + /** * Get a loader for a file with a specific type * @param file The file to load diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h index 6a9e5a68b..b74cfbf8a 100644 --- a/src/core/loader/loader.h +++ b/src/core/loader/loader.h @@ -5,7 +5,7 @@ #pragma once #include <algorithm> -#include <initializer_list> +#include <iosfwd> #include <memory> #include <string> #include <utility> @@ -56,20 +56,47 @@ FileType GuessFromFilename(const std::string& name); std::string GetFileTypeString(FileType type); /// Return type for functions in Loader namespace -enum class ResultStatus { +enum class ResultStatus : u16 { Success, - Error, - ErrorInvalidFormat, - ErrorNotImplemented, - ErrorNotLoaded, - ErrorNotUsed, ErrorAlreadyLoaded, - ErrorMemoryAllocationFailed, - ErrorMissingKeys, - ErrorDecrypting, - ErrorUnsupportedArch, + ErrorNotImplemented, + ErrorNotInitialized, + ErrorBadNPDMHeader, + ErrorBadACIDHeader, + ErrorBadACIHeader, + ErrorBadFileAccessControl, + ErrorBadFileAccessHeader, + ErrorBadPFSHeader, + ErrorIncorrectPFSFileSize, + ErrorBadNCAHeader, + ErrorMissingProductionKeyFile, + ErrorMissingHeaderKey, + ErrorIncorrectHeaderKey, + ErrorNCA2, + ErrorNCA0, + ErrorMissingTitlekey, + ErrorMissingTitlekek, + ErrorInvalidRightsID, + ErrorMissingKeyAreaKey, + ErrorIncorrectKeyAreaKey, + ErrorIncorrectTitlekeyOrTitlekek, + ErrorXCIMissingProgramNCA, + ErrorNCANotProgram, + ErrorNoExeFS, + ErrorBadXCIHeader, + ErrorXCIMissingPartition, + ErrorNullFile, + ErrorMissingNPDM, + Error32BitISA, + ErrorNoRomFS, + ErrorIncorrectELFFileSize, + ErrorLoadingNRO, + ErrorNoIcon, + ErrorNoControl, }; +std::ostream& operator<<(std::ostream& os, ResultStatus status); + /// Interface for loading an application class AppLoader : NonCopyable { public: @@ -180,12 +207,6 @@ protected: }; /** - * Common address mappings found in most games, used for binary formats that don't have this - * information. - */ -extern const std::initializer_list<Kernel::AddressMapping> default_address_mappings; - -/** * Identifies a bootable file and return a suitable loader * @param file The bootable file * @return the best loader for this file diff --git a/src/core/loader/nca.cpp b/src/core/loader/nca.cpp index 46f5cd393..9d50c7d42 100644 --- a/src/core/loader/nca.cpp +++ b/src/core/loader/nca.cpp @@ -3,28 +3,22 @@ // Refer to the license.txt file included. #include <utility> -#include <vector> #include "common/file_util.h" #include "common/logging/log.h" -#include "common/string_util.h" -#include "common/swap.h" -#include "core/core.h" #include "core/file_sys/content_archive.h" -#include "core/file_sys/program_metadata.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/resource_limit.h" #include "core/hle/service/filesystem/filesystem.h" +#include "core/loader/deconstructed_rom_directory.h" #include "core/loader/nca.h" -#include "core/loader/nso.h" -#include "core/memory.h" namespace Loader { AppLoader_NCA::AppLoader_NCA(FileSys::VirtualFile file_) : AppLoader(std::move(file_)), nca(std::make_unique<FileSys::NCA>(file)) {} +AppLoader_NCA::~AppLoader_NCA() = default; + FileType AppLoader_NCA::IdentifyType(const FileSys::VirtualFile& file) { FileSys::NCA nca(file); @@ -46,12 +40,12 @@ ResultStatus AppLoader_NCA::Load(Kernel::SharedPtr<Kernel::Process>& process) { } if (nca->GetType() != FileSys::NCAContentType::Program) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorNCANotProgram; const auto exefs = nca->GetExeFS(); if (exefs == nullptr) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorNoExeFS; directory_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(exefs); @@ -69,20 +63,18 @@ ResultStatus AppLoader_NCA::Load(Kernel::SharedPtr<Kernel::Process>& process) { ResultStatus AppLoader_NCA::ReadRomFS(FileSys::VirtualFile& dir) { if (nca == nullptr) - return ResultStatus::ErrorNotLoaded; + return ResultStatus::ErrorNotInitialized; if (nca->GetRomFS() == nullptr || nca->GetRomFS()->GetSize() == 0) - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoRomFS; dir = nca->GetRomFS(); return ResultStatus::Success; } ResultStatus AppLoader_NCA::ReadProgramId(u64& out_program_id) { if (nca == nullptr || nca->GetStatus() != ResultStatus::Success) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorNotInitialized; out_program_id = nca->GetTitleId(); return ResultStatus::Success; } -AppLoader_NCA::~AppLoader_NCA() = default; - } // namespace Loader diff --git a/src/core/loader/nca.h b/src/core/loader/nca.h index 7f7d8ea0b..326f84857 100644 --- a/src/core/loader/nca.h +++ b/src/core/loader/nca.h @@ -4,20 +4,24 @@ #pragma once -#include <string> #include "common/common_types.h" -#include "core/file_sys/content_archive.h" -#include "core/file_sys/program_metadata.h" +#include "core/file_sys/vfs.h" #include "core/hle/kernel/object.h" #include "core/loader/loader.h" -#include "deconstructed_rom_directory.h" + +namespace FileSys { +class NCA; +} namespace Loader { +class AppLoader_DeconstructedRomDirectory; + /// Loads an NCA file class AppLoader_NCA final : public AppLoader { public: explicit AppLoader_NCA(FileSys::VirtualFile file); + ~AppLoader_NCA() override; /** * Returns the type of the file @@ -35,12 +39,7 @@ public: ResultStatus ReadRomFS(FileSys::VirtualFile& dir) override; ResultStatus ReadProgramId(u64& out_program_id) override; - ~AppLoader_NCA(); - private: - FileSys::ProgramMetadata metadata; - - FileSys::NCAHeader header; std::unique_ptr<FileSys::NCA> nca; std::unique_ptr<AppLoader_DeconstructedRomDirectory> directory_loader; }; diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp index dc053cdad..2179cf2ea 100644 --- a/src/core/loader/nro.cpp +++ b/src/core/loader/nro.cpp @@ -182,11 +182,10 @@ ResultStatus AppLoader_NRO::Load(Kernel::SharedPtr<Kernel::Process>& process) { static constexpr VAddr base_addr{Memory::PROCESS_IMAGE_VADDR}; if (!LoadNro(file, base_addr)) { - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorLoadingNRO; } process->svc_access_mask.set(); - process->address_mappings = default_address_mappings; process->resource_limit = Kernel::ResourceLimit::GetForCategory(Kernel::ResourceLimitCategory::APPLICATION); process->Run(base_addr, THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE); @@ -197,7 +196,7 @@ ResultStatus AppLoader_NRO::Load(Kernel::SharedPtr<Kernel::Process>& process) { ResultStatus AppLoader_NRO::ReadIcon(std::vector<u8>& buffer) { if (icon_data.empty()) { - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoIcon; } buffer = icon_data; @@ -206,7 +205,7 @@ ResultStatus AppLoader_NRO::ReadIcon(std::vector<u8>& buffer) { ResultStatus AppLoader_NRO::ReadProgramId(u64& out_program_id) { if (nacp == nullptr) { - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoControl; } out_program_id = nacp->GetTitleId(); @@ -215,7 +214,7 @@ ResultStatus AppLoader_NRO::ReadProgramId(u64& out_program_id) { ResultStatus AppLoader_NRO::ReadRomFS(FileSys::VirtualFile& dir) { if (romfs == nullptr) { - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoRomFS; } dir = romfs; @@ -224,7 +223,7 @@ ResultStatus AppLoader_NRO::ReadRomFS(FileSys::VirtualFile& dir) { ResultStatus AppLoader_NRO::ReadTitle(std::string& title) { if (nacp == nullptr) { - return ResultStatus::ErrorNotUsed; + return ResultStatus::ErrorNoControl; } title = nacp->GetApplicationName(); diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp index fee7d58c6..a94558ac5 100644 --- a/src/core/loader/nso.cpp +++ b/src/core/loader/nso.cpp @@ -152,7 +152,6 @@ ResultStatus AppLoader_NSO::Load(Kernel::SharedPtr<Kernel::Process>& process) { LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", file->GetName(), Memory::PROCESS_IMAGE_VADDR); process->svc_access_mask.set(); - process->address_mappings = default_address_mappings; process->resource_limit = Kernel::ResourceLimit::GetForCategory(Kernel::ResourceLimitCategory::APPLICATION); process->Run(Memory::PROCESS_IMAGE_VADDR, THREADPRIO_DEFAULT, Memory::DEFAULT_STACK_SIZE); diff --git a/src/core/loader/xci.cpp b/src/core/loader/xci.cpp index d3fe24419..4c4979545 100644 --- a/src/core/loader/xci.cpp +++ b/src/core/loader/xci.cpp @@ -4,22 +4,14 @@ #include <vector> -#include "common/file_util.h" -#include "common/logging/log.h" -#include "common/string_util.h" -#include "common/swap.h" -#include "core/core.h" +#include "common/common_types.h" +#include "core/file_sys/card_image.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" -#include "core/file_sys/program_metadata.h" #include "core/file_sys/romfs.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/resource_limit.h" -#include "core/hle/service/filesystem/filesystem.h" -#include "core/loader/nso.h" +#include "core/loader/nca.h" #include "core/loader/xci.h" -#include "core/memory.h" namespace Loader { @@ -66,10 +58,13 @@ ResultStatus AppLoader_XCI::Load(Kernel::SharedPtr<Kernel::Process>& process) { return ResultStatus::ErrorAlreadyLoaded; } + if (xci->GetStatus() != ResultStatus::Success) + return xci->GetStatus(); + if (xci->GetNCAFileByType(FileSys::NCAContentType::Program) == nullptr) { if (!Core::Crypto::KeyManager::KeyFileExists(false)) - return ResultStatus::ErrorMissingKeys; - return ResultStatus::ErrorDecrypting; + return ResultStatus::ErrorMissingProductionKeyFile; + return ResultStatus::ErrorXCIMissingProgramNCA; } auto result = nca_loader->Load(process); @@ -91,14 +86,14 @@ ResultStatus AppLoader_XCI::ReadProgramId(u64& out_program_id) { ResultStatus AppLoader_XCI::ReadIcon(std::vector<u8>& buffer) { if (icon_file == nullptr) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorNoControl; buffer = icon_file->ReadAllBytes(); return ResultStatus::Success; } ResultStatus AppLoader_XCI::ReadTitle(std::string& title) { if (nacp_file == nullptr) - return ResultStatus::ErrorInvalidFormat; + return ResultStatus::ErrorNoControl; title = nacp_file->GetApplicationName(); return ResultStatus::Success; } diff --git a/src/core/loader/xci.h b/src/core/loader/xci.h index 973833050..cc4287e17 100644 --- a/src/core/loader/xci.h +++ b/src/core/loader/xci.h @@ -6,12 +6,18 @@ #include <memory> #include "common/common_types.h" -#include "core/file_sys/card_image.h" +#include "core/file_sys/vfs.h" #include "core/loader/loader.h" -#include "core/loader/nca.h" + +namespace FileSys { +class NACP; +class XCI; +} // namespace FileSys namespace Loader { +class AppLoader_NCA; + /// Loads an XCI file class AppLoader_XCI final : public AppLoader { public: @@ -37,8 +43,6 @@ public: ResultStatus ReadTitle(std::string& title) override; private: - FileSys::ProgramMetadata metadata; - std::unique_ptr<FileSys::XCI> xci; std::unique_ptr<AppLoader_NCA> nca_loader; diff --git a/src/core/settings.cpp b/src/core/settings.cpp index a4623223d..0da159559 100644 --- a/src/core/settings.cpp +++ b/src/core/settings.cpp @@ -7,22 +7,18 @@ #include "core/hle/service/hid/hid.h" #include "core/settings.h" #include "video_core/renderer_base.h" -#include "video_core/video_core.h" namespace Settings { Values values = {}; void Apply() { - GDBStub::SetServerPort(values.gdbstub_port); GDBStub::ToggleServer(values.use_gdbstub); - VideoCore::g_toggle_framelimit_enabled = values.toggle_framelimit; - auto& system_instance = Core::System::GetInstance(); if (system_instance.IsPoweredOn()) { - system_instance.Renderer().UpdateCurrentFramebufferLayout(); + system_instance.Renderer().RefreshBaseSettings(); } Service::HID::ReloadInputDevices(); diff --git a/src/video_core/command_processor.h b/src/video_core/command_processor.h index f7214ffec..a01153e0b 100644 --- a/src/video_core/command_processor.h +++ b/src/video_core/command_processor.h @@ -30,8 +30,7 @@ union CommandHeader { BitField<29, 3, SubmissionMode> mode; }; -static_assert(std::is_standard_layout<CommandHeader>::value == true, - "CommandHeader does not use standard layout"); +static_assert(std::is_standard_layout_v<CommandHeader>, "CommandHeader is not standard layout"); static_assert(sizeof(CommandHeader) == sizeof(u32), "CommandHeader has incorrect size!"); } // namespace Tegra diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index a46ed4bd7..68f91cc75 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -222,6 +222,18 @@ void Maxwell3D::DrawArrays() { debug_context->OnEvent(Tegra::DebugContext::Event::FinishedPrimitiveBatch, nullptr); } + // Both instance configuration registers can not be set at the same time. + ASSERT_MSG(!regs.draw.instance_next || !regs.draw.instance_cont, + "Illegal combination of instancing parameters"); + + if (regs.draw.instance_next) { + // Increment the current instance *before* drawing. + state.current_instance += 1; + } else if (!regs.draw.instance_cont) { + // Reset the current instance to 0. + state.current_instance = 0; + } + const bool is_indexed{regs.index_array.count && !regs.vertex_buffer.count}; rasterizer.AccelerateDrawBatch(is_indexed); diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 0506ac8fe..771eb5abc 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -93,6 +93,7 @@ public: struct VertexAttribute { enum class Size : u32 { + Invalid = 0x0, Size_32_32_32_32 = 0x01, Size_32_32_32 = 0x02, Size_16_16_16_16 = 0x03, @@ -257,6 +258,10 @@ public: bool IsNormalized() const { return (type == Type::SignedNorm) || (type == Type::UnsignedNorm); } + + bool IsValid() const { + return size != Size::Invalid; + } }; enum class PrimitiveTopology : u32 { @@ -352,6 +357,27 @@ public: OneMinusConstantColor = 0x62, ConstantAlpha = 0x63, OneMinusConstantAlpha = 0x64, + + // These values are used by Nouveau and some games. + ZeroGL = 0x4000, + OneGL = 0x4001, + SourceColorGL = 0x4300, + OneMinusSourceColorGL = 0x4301, + SourceAlphaGL = 0x4302, + OneMinusSourceAlphaGL = 0x4303, + DestAlphaGL = 0x4304, + OneMinusDestAlphaGL = 0x4305, + DestColorGL = 0x4306, + OneMinusDestColorGL = 0x4307, + SourceAlphaSaturateGL = 0x4308, + ConstantColorGL = 0xc001, + OneMinusConstantColorGL = 0xc002, + ConstantAlphaGL = 0xc003, + OneMinusConstantAlphaGL = 0xc004, + Source1ColorGL = 0xc900, + OneMinusSource1ColorGL = 0xc901, + Source1AlphaGL = 0xc902, + OneMinusSource1AlphaGL = 0xc903, }; u32 separate_alpha; @@ -612,6 +638,8 @@ public: union { u32 vertex_begin_gl; BitField<0, 16, PrimitiveTopology> topology; + BitField<26, 1, u32> instance_next; + BitField<27, 1, u32> instance_cont; }; } draw; @@ -804,6 +832,7 @@ public: }; std::array<ShaderStageInfo, Regs::MaxShaderStage> shader_stages; + u32 current_instance = 0; ///< Current instance to be used to simulate instanced rendering. }; State state{}; diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index 3d4557b7e..9413a81fb 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -12,6 +12,7 @@ #include <boost/optional.hpp> +#include "common/assert.h" #include "common/bit_field.h" #include "common/common_types.h" @@ -74,12 +75,14 @@ union Attribute { enum class Index : u64 { Position = 7, Attribute_0 = 8, + Attribute_31 = 39, // This attribute contains a tuple of (~, ~, InstanceId, VertexId) when inside a vertex // shader, and a tuple of (TessCoord.x, TessCoord.y, TessCoord.z, ~) when inside a Tess Eval // shader. TessCoordInstanceIDVertexID = 47, - // TODO(bunnei): Figure out what this is used for. Super Mario Odyssey uses this. - Unknown_63 = 63, + // This attribute contains a tuple of (Unk, Unk, Unk, gl_FrontFacing) when inside a fragment + // shader. It is unknown what the other values contain. + FrontFacing = 63, }; union { @@ -142,6 +145,7 @@ enum class PredCondition : u64 { NotEqual = 5, GreaterEqual = 6, LessThanWithNan = 9, + GreaterThanWithNan = 12, NotEqualWithNan = 13, // TODO(Subv): Other condition types }; @@ -201,11 +205,24 @@ enum class IMinMaxExchange : u64 { XHi = 3, }; +enum class XmadMode : u64 { + None = 0, + CLo = 1, + CHi = 2, + CSfu = 3, + CBcc = 4, +}; + enum class FlowCondition : u64 { Always = 0xF, Fcsm_Tr = 0x1C, // TODO(bunnei): What is this used for? }; +enum class PredicateResultMode : u64 { + None = 0x0, + NotZero = 0x3, +}; + union Instruction { Instruction& operator=(const Instruction& instr) { value = instr.value; @@ -246,7 +263,7 @@ union Instruction { BitField<39, 1, u64> invert_a; BitField<40, 1, u64> invert_b; BitField<41, 2, LogicOperation> operation; - BitField<44, 2, u64> unk44; + BitField<44, 2, PredicateResultMode> pred_result_mode; BitField<48, 3, Pred> pred48; } lop; @@ -276,6 +293,10 @@ union Instruction { } alu; union { + BitField<48, 1, u64> negate_b; + } fmul; + + union { BitField<48, 1, u64> is_signed; } shift; @@ -430,16 +451,20 @@ union Instruction { } bool IsComponentEnabled(size_t component) const { - static constexpr std::array<std::array<u32, 8>, 4> mask_lut{ - {{}, - {0x1, 0x2, 0x4, 0x8, 0x3}, - {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc}, - {0x7, 0xb, 0xd, 0xe, 0xf}}}; + static constexpr std::array<std::array<u32, 8>, 4> mask_lut{{ + {}, + {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc}, + {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc}, + {0x7, 0xb, 0xd, 0xe, 0xf}, + }}; size_t index{gpr0.Value() != Register::ZeroIndex ? 1U : 0U}; index |= gpr28.Value() != Register::ZeroIndex ? 2 : 0; - return ((1ull << component) & mask_lut[index][component_mask_selector]) != 0; + u32 mask = mask_lut[index][component_mask_selector]; + // A mask of 0 means this instruction uses an unimplemented mask. + ASSERT(mask != 0); + return ((1ull << component) & mask) != 0; } } texs; @@ -458,6 +483,18 @@ union Instruction { } bra; union { + BitField<20, 16, u64> imm20_16; + BitField<36, 1, u64> product_shift_left; + BitField<37, 1, u64> merge_37; + BitField<48, 1, u64> sign_a; + BitField<49, 1, u64> sign_b; + BitField<50, 3, XmadMode> mode; + BitField<52, 1, u64> high_b; + BitField<53, 1, u64> high_a; + BitField<56, 1, u64> merge_56; + } xmad; + + union { BitField<20, 14, u64> offset; BitField<34, 5, u64> index; } cbuf34; @@ -477,8 +514,7 @@ union Instruction { u64 value; }; static_assert(sizeof(Instruction) == 0x8, "Incorrect structure size"); -static_assert(std::is_standard_layout<Instruction>::value, - "Structure does not have standard layout"); +static_assert(std::is_standard_layout_v<Instruction>, "Instruction is not standard layout"); class OpCode { public: @@ -494,6 +530,8 @@ public: LD_A, LD_C, ST_A, + LDG, // Load from global memory + STG, // Store in global memory TEX, TEXQ, // Texture Query TEXS, // Texture Fetch with scalar/non-vec4 source/destinations @@ -595,9 +633,17 @@ public: IntegerSetPredicate, PredicateSetPredicate, Conversion, + Xmad, Unknown, }; + /// Returns whether an opcode has an execution predicate field or not (ie, whether it can be + /// conditionally executed). + static bool IsPredicatedInstruction(Id opcode) { + // TODO(Subv): Add the rest of unpredicated instructions. + return opcode != Id::SSY; + } + class Matcher { public: Matcher(const char* const name, u16 mask, u16 expected, OpCode::Id id, OpCode::Type type) @@ -697,6 +743,8 @@ private: INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"), INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"), INST("1110111111110---", Id::ST_A, Type::Memory, "ST_A"), + INST("1110111011010---", Id::LDG, Type::Memory, "LDG"), + INST("1110111011011---", Id::STG, Type::Memory, "STG"), INST("110000----111---", Id::TEX, Type::Memory, "TEX"), INST("1101111101001---", Id::TEXQ, Type::Memory, "TEXQ"), INST("1101100---------", Id::TEXS, Type::Memory, "TEXS"), @@ -777,10 +825,10 @@ private: INST("010010110101----", Id::ISET_C, Type::IntegerSet, "ISET_C"), INST("0011011-0101----", Id::ISET_IMM, Type::IntegerSet, "ISET_IMM"), INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"), - INST("0011011-00------", Id::XMAD_IMM, Type::Arithmetic, "XMAD_IMM"), - INST("0100111---------", Id::XMAD_CR, Type::Arithmetic, "XMAD_CR"), - INST("010100010-------", Id::XMAD_RC, Type::Arithmetic, "XMAD_RC"), - INST("0101101100------", Id::XMAD_RR, Type::Arithmetic, "XMAD_RR"), + INST("0011011-00------", Id::XMAD_IMM, Type::Xmad, "XMAD_IMM"), + INST("0100111---------", Id::XMAD_CR, Type::Xmad, "XMAD_CR"), + INST("010100010-------", Id::XMAD_RC, Type::Xmad, "XMAD_RC"), + INST("0101101100------", Id::XMAD_RR, Type::Xmad, "XMAD_RR"), }; #undef INST std::stable_sort(table.begin(), table.end(), [](const auto& a, const auto& b) { diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index 4ff4d71c5..9758adcfd 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/assert.h" #include "video_core/engines/fermi_2d.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/engines/maxwell_compute.h" @@ -11,6 +12,15 @@ namespace Tegra { +u32 FramebufferConfig::BytesPerPixel(PixelFormat format) { + switch (format) { + case PixelFormat::ABGR8: + return 4; + } + + UNREACHABLE(); +} + GPU::GPU(VideoCore::RasterizerInterface& rasterizer) { memory_manager = std::make_unique<MemoryManager>(); maxwell_3d = std::make_unique<Engines::Maxwell3D>(rasterizer, *memory_manager); @@ -34,19 +44,60 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format) { switch (format) { case RenderTargetFormat::RGBA32_FLOAT: + case RenderTargetFormat::RGBA32_UINT: return 16; + case RenderTargetFormat::RGBA16_UINT: + case RenderTargetFormat::RGBA16_UNORM: case RenderTargetFormat::RGBA16_FLOAT: case RenderTargetFormat::RG32_FLOAT: + case RenderTargetFormat::RG32_UINT: return 8; case RenderTargetFormat::RGBA8_UNORM: + case RenderTargetFormat::RGBA8_SNORM: + case RenderTargetFormat::RGBA8_SRGB: + case RenderTargetFormat::RGBA8_UINT: case RenderTargetFormat::RGB10_A2_UNORM: case RenderTargetFormat::BGRA8_UNORM: + case RenderTargetFormat::RG16_UNORM: + case RenderTargetFormat::RG16_SNORM: + case RenderTargetFormat::RG16_UINT: + case RenderTargetFormat::RG16_SINT: + case RenderTargetFormat::RG16_FLOAT: case RenderTargetFormat::R32_FLOAT: case RenderTargetFormat::R11G11B10_FLOAT: + case RenderTargetFormat::R32_UINT: return 4; + case RenderTargetFormat::R16_UNORM: + case RenderTargetFormat::R16_SNORM: + case RenderTargetFormat::R16_UINT: + case RenderTargetFormat::R16_SINT: + case RenderTargetFormat::R16_FLOAT: + case RenderTargetFormat::RG8_UNORM: + case RenderTargetFormat::RG8_SNORM: + return 2; + case RenderTargetFormat::R8_UNORM: + case RenderTargetFormat::R8_UINT: + return 1; default: UNIMPLEMENTED_MSG("Unimplemented render target format {}", static_cast<u32>(format)); } } +u32 DepthFormatBytesPerPixel(DepthFormat format) { + switch (format) { + case DepthFormat::Z32_S8_X24_FLOAT: + return 8; + case DepthFormat::Z32_FLOAT: + case DepthFormat::S8_Z24_UNORM: + case DepthFormat::Z24_X8_UNORM: + case DepthFormat::Z24_S8_UNORM: + case DepthFormat::Z24_C8_UNORM: + return 4; + case DepthFormat::Z16_UNORM: + return 2; + default: + UNIMPLEMENTED_MSG("Unimplemented Depth format {}", static_cast<u32>(format)); + } +} + } // namespace Tegra diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index 874eddd78..2697e1c27 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -6,7 +6,6 @@ #include <memory> #include <unordered_map> -#include <vector> #include "common/common_types.h" #include "core/hle/service/nvflinger/buffer_queue.h" #include "video_core/memory_manager.h" @@ -21,22 +20,35 @@ enum class RenderTargetFormat : u32 { NONE = 0x0, RGBA32_FLOAT = 0xC0, RGBA32_UINT = 0xC2, + RGBA16_UNORM = 0xC6, + RGBA16_UINT = 0xC9, RGBA16_FLOAT = 0xCA, RG32_FLOAT = 0xCB, + RG32_UINT = 0xCD, BGRA8_UNORM = 0xCF, 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, + R32_UINT = 0xE4, R32_FLOAT = 0xE5, B5G6R5_UNORM = 0xE8, + RG8_UNORM = 0xEA, + RG8_SNORM = 0xEB, + R16_UNORM = 0xEE, + R16_SNORM = 0xEF, + R16_SINT = 0xF0, + R16_UINT = 0xF1, R16_FLOAT = 0xF2, R8_UNORM = 0xF3, + R8_UINT = 0xF6, }; enum class DepthFormat : u32 { @@ -52,6 +64,9 @@ enum class DepthFormat : u32 { /// Returns the number of bytes per pixel of each rendertarget format. u32 RenderTargetBytesPerPixel(RenderTargetFormat format); +/// Returns the number of bytes per pixel of each depth format. +u32 DepthFormatBytesPerPixel(DepthFormat format); + class DebugContext; /** @@ -65,14 +80,7 @@ struct FramebufferConfig { /** * Returns the number of bytes per pixel. */ - static u32 BytesPerPixel(PixelFormat format) { - switch (format) { - case PixelFormat::ABGR8: - return 4; - } - - UNREACHABLE(); - } + static u32 BytesPerPixel(PixelFormat format); VAddr address; u32 offset; diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp index 3ca350243..afd86a83a 100644 --- a/src/video_core/renderer_base.cpp +++ b/src/video_core/renderer_base.cpp @@ -4,18 +4,23 @@ #include <memory> #include "core/frontend/emu_window.h" +#include "core/settings.h" #include "video_core/renderer_base.h" #include "video_core/renderer_opengl/gl_rasterizer.h" namespace VideoCore { -RendererBase::RendererBase(EmuWindow& window) : render_window{window} {} +RendererBase::RendererBase(Core::Frontend::EmuWindow& window) : render_window{window} { + RefreshBaseSettings(); +} + RendererBase::~RendererBase() = default; -void RendererBase::UpdateCurrentFramebufferLayout() { - const Layout::FramebufferLayout& layout = render_window.GetFramebufferLayout(); +void RendererBase::RefreshBaseSettings() { + RefreshRasterizerSetting(); + UpdateCurrentFramebufferLayout(); - render_window.UpdateCurrentFramebufferLayout(layout.width, layout.height); + renderer_settings.use_framelimiter = Settings::values.toggle_framelimit; } void RendererBase::RefreshRasterizerSetting() { @@ -24,4 +29,10 @@ void RendererBase::RefreshRasterizerSetting() { } } +void RendererBase::UpdateCurrentFramebufferLayout() { + const Layout::FramebufferLayout& layout = render_window.GetFramebufferLayout(); + + render_window.UpdateCurrentFramebufferLayout(layout.width, layout.height); +} + } // namespace VideoCore diff --git a/src/video_core/renderer_base.h b/src/video_core/renderer_base.h index 235de23a1..d9f16b8e6 100644 --- a/src/video_core/renderer_base.h +++ b/src/video_core/renderer_base.h @@ -4,23 +4,26 @@ #pragma once +#include <atomic> #include <memory> #include <boost/optional.hpp> -#include "common/assert.h" #include "common/common_types.h" #include "video_core/gpu.h" #include "video_core/rasterizer_interface.h" +namespace Core::Frontend { class EmuWindow; +} namespace VideoCore { +struct RendererSettings { + std::atomic_bool use_framelimiter{false}; +}; + class RendererBase : NonCopyable { public: - /// Used to reference a framebuffer - enum kFramebuffer { kFramebuffer_VirtualXFB = 0, kFramebuffer_EFB, kFramebuffer_Texture }; - - explicit RendererBase(EmuWindow& window); + explicit RendererBase(Core::Frontend::EmuWindow& window); virtual ~RendererBase(); /// Swap buffers (render frame) @@ -32,9 +35,6 @@ public: /// Shutdown the renderer virtual void ShutDown() = 0; - /// Updates the framebuffer layout of the contained render window handle. - void UpdateCurrentFramebufferLayout(); - // Getter/setter functions: // ------------------------ @@ -54,13 +54,23 @@ public: return *rasterizer; } - void RefreshRasterizerSetting(); + /// Refreshes the settings common to all renderers + void RefreshBaseSettings(); protected: - EmuWindow& render_window; ///< Reference to the render window handle. + /// Refreshes settings specific to the rasterizer. + void RefreshRasterizerSetting(); + + Core::Frontend::EmuWindow& render_window; ///< Reference to the render window handle. std::unique_ptr<RasterizerInterface> rasterizer; 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 + + RendererSettings renderer_settings; + +private: + /// Updates the framebuffer layout of the contained render window handle. + void UpdateCurrentFramebufferLayout(); }; } // namespace VideoCore diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 8360feb5d..93eadde7a 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -36,30 +36,21 @@ MICROPROFILE_DEFINE(OpenGL_Drawing, "OpenGL", "Drawing", MP_RGB(128, 128, 192)); MICROPROFILE_DEFINE(OpenGL_Blits, "OpenGL", "Blits", MP_RGB(100, 100, 255)); MICROPROFILE_DEFINE(OpenGL_CacheManagement, "OpenGL", "Cache Mgmt", MP_RGB(100, 255, 100)); -RasterizerOpenGL::RasterizerOpenGL(EmuWindow& window) : emu_window{window} { +RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window) + : emu_window{window}, stream_buffer(GL_ARRAY_BUFFER, STREAM_BUFFER_SIZE) { // Create sampler objects for (size_t i = 0; i < texture_samplers.size(); ++i) { texture_samplers[i].Create(); state.texture_units[i].sampler = texture_samplers[i].sampler.handle; } - // Create SSBOs - for (size_t stage = 0; stage < ssbos.size(); ++stage) { - for (size_t buffer = 0; buffer < ssbos[stage].size(); ++buffer) { - ssbos[stage][buffer].Create(); - state.draw.const_buffers[stage][buffer].ssbo = ssbos[stage][buffer].handle; - } - } - GLint ext_num; glGetIntegerv(GL_NUM_EXTENSIONS, &ext_num); for (GLint i = 0; i < ext_num; i++) { const std::string_view extension{ reinterpret_cast<const char*>(glGetStringi(GL_EXTENSIONS, i))}; - if (extension == "GL_ARB_buffer_storage") { - has_ARB_buffer_storage = true; - } else if (extension == "GL_ARB_direct_state_access") { + if (extension == "GL_ARB_direct_state_access") { has_ARB_direct_state_access = true; } else if (extension == "GL_ARB_separate_shader_objects") { has_ARB_separate_shader_objects = true; @@ -86,47 +77,31 @@ RasterizerOpenGL::RasterizerOpenGL(EmuWindow& window) : emu_window{window} { hw_vao.Create(); - stream_buffer = OGLStreamBuffer::MakeBuffer(has_ARB_buffer_storage, GL_ARRAY_BUFFER); - stream_buffer->Create(STREAM_BUFFER_SIZE, STREAM_BUFFER_SIZE / 2); - state.draw.vertex_buffer = stream_buffer->GetHandle(); + state.draw.vertex_buffer = stream_buffer.GetHandle(); shader_program_manager = std::make_unique<GLShader::ProgramManager>(); state.draw.shader_program = 0; state.draw.vertex_array = hw_vao.handle; state.Apply(); - glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, stream_buffer->GetHandle()); - - for (unsigned index = 0; index < uniform_buffers.size(); ++index) { - auto& buffer = uniform_buffers[index]; - buffer.Create(); - glBindBuffer(GL_UNIFORM_BUFFER, buffer.handle); - glBufferData(GL_UNIFORM_BUFFER, sizeof(GLShader::MaxwellUniformData), nullptr, - GL_STREAM_COPY); - glBindBufferBase(GL_UNIFORM_BUFFER, index, buffer.handle); - } + glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, stream_buffer.GetHandle()); glEnable(GL_BLEND); + glGetIntegerv(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT, &uniform_buffer_alignment); + LOG_CRITICAL(Render_OpenGL, "Sync fixed function OpenGL state here!"); } -RasterizerOpenGL::~RasterizerOpenGL() { - if (stream_buffer != nullptr) { - state.draw.vertex_buffer = stream_buffer->GetHandle(); - state.Apply(); - stream_buffer->Release(); - } -} +RasterizerOpenGL::~RasterizerOpenGL() {} std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr, GLintptr buffer_offset) { MICROPROFILE_SCOPE(OpenGL_VAO); const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; - const auto& memory_manager = Core::System::GetInstance().GPU().memory_manager; state.draw.vertex_array = hw_vao.handle; - state.draw.vertex_buffer = stream_buffer->GetHandle(); + state.draw.vertex_buffer = stream_buffer.GetHandle(); state.Apply(); // Upload all guest vertex arrays sequentially to our buffer @@ -141,16 +116,15 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr, ASSERT(end > start); u64 size = end - start + 1; - // Copy vertex array data - Memory::ReadBlock(*memory_manager->GpuToCpuAddress(start), array_ptr, size); + GLintptr vertex_buffer_offset; + std::tie(array_ptr, buffer_offset, vertex_buffer_offset) = + UploadMemory(array_ptr, buffer_offset, start, size); // Bind the vertex array to the buffer at the current offset. - glBindVertexBuffer(index, stream_buffer->GetHandle(), buffer_offset, vertex_array.stride); + glBindVertexBuffer(index, stream_buffer.GetHandle(), vertex_buffer_offset, + vertex_array.stride); - ASSERT_MSG(vertex_array.divisor == 0, "Vertex buffer divisor unimplemented"); - - array_ptr += size; - buffer_offset += size; + ASSERT_MSG(vertex_array.divisor == 0, "Instanced vertex arrays are not supported"); } // Use the vertex array as-is, assumes that the data is formatted correctly for OpenGL. @@ -161,11 +135,16 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr, // assume every shader uses them all. for (unsigned index = 0; index < 16; ++index) { auto& attrib = regs.vertex_attrib_format[index]; + + // Ignore invalid attributes. + if (!attrib.IsValid()) + continue; + + auto& buffer = regs.vertex_array[attrib.buffer]; LOG_TRACE(HW_GPU, "vertex attrib {}, count={}, size={}, type={}, offset={}, normalize={}", index, attrib.ComponentCount(), attrib.SizeString(), attrib.TypeString(), attrib.offset.Value(), attrib.IsNormalized()); - auto& buffer = regs.vertex_array[attrib.buffer]; ASSERT(buffer.IsEnabled()); glEnableVertexAttribArray(index); @@ -196,22 +175,12 @@ static GLShader::ProgramCode GetShaderProgramCode(Maxwell::ShaderProgram program return program_code; } -void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) { - // Helper function for uploading uniform data - const auto copy_buffer = [&](GLuint handle, GLintptr offset, GLsizeiptr size) { - if (has_ARB_direct_state_access) { - glCopyNamedBufferSubData(stream_buffer->GetHandle(), handle, offset, 0, size); - } else { - glBindBuffer(GL_COPY_WRITE_BUFFER, handle); - glCopyBufferSubData(GL_ARRAY_BUFFER, GL_COPY_WRITE_BUFFER, offset, 0, size); - } - }; - +std::pair<u8*, GLintptr> RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) { auto& gpu = Core::System::GetInstance().GPU().Maxwell3D(); // Next available bindpoints to use when uploading the const buffers and textures to the GLSL // shaders. The constbuffer bindpoint starts after the shader stage configuration bind points. - u32 current_constbuffer_bindpoint = uniform_buffers.size(); + u32 current_constbuffer_bindpoint = Tegra::Engines::Maxwell3D::Regs::MaxShaderStage; u32 current_texture_bindpoint = 0; for (size_t index = 0; index < Maxwell::MaxShaderProgram; ++index) { @@ -223,22 +192,21 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) { continue; } + std::tie(buffer_ptr, buffer_offset) = + AlignBuffer(buffer_ptr, buffer_offset, static_cast<size_t>(uniform_buffer_alignment)); + const size_t stage{index == 0 ? 0 : index - 1}; // Stage indices are 0 - 5 GLShader::MaxwellUniformData ubo{}; ubo.SetFromRegs(gpu.state.shader_stages[stage]); std::memcpy(buffer_ptr, &ubo, sizeof(ubo)); - // Flush the buffer so that the GPU can see the data we just wrote. - glFlushMappedBufferRange(GL_ARRAY_BUFFER, buffer_offset, sizeof(ubo)); - - // Upload uniform data as one UBO per stage - const GLintptr ubo_offset = buffer_offset; - copy_buffer(uniform_buffers[stage].handle, ubo_offset, - sizeof(GLShader::MaxwellUniformData)); + // Bind the buffer + glBindBufferRange(GL_UNIFORM_BUFFER, stage, stream_buffer.GetHandle(), buffer_offset, + sizeof(ubo)); - buffer_ptr += sizeof(GLShader::MaxwellUniformData); - buffer_offset += sizeof(GLShader::MaxwellUniformData); + buffer_ptr += sizeof(ubo); + buffer_offset += sizeof(ubo); GLShader::ShaderSetup setup{GetShaderProgramCode(program)}; GLShader::ShaderEntries shader_resources; @@ -277,9 +245,9 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) { static_cast<Maxwell::ShaderStage>(stage)); // Configure the const buffers for this shader stage. - current_constbuffer_bindpoint = - SetupConstBuffers(static_cast<Maxwell::ShaderStage>(stage), gl_stage_program, - current_constbuffer_bindpoint, shader_resources.const_buffer_entries); + std::tie(buffer_ptr, buffer_offset, current_constbuffer_bindpoint) = SetupConstBuffers( + buffer_ptr, buffer_offset, static_cast<Maxwell::ShaderStage>(stage), gl_stage_program, + current_constbuffer_bindpoint, shader_resources.const_buffer_entries); // Configure the textures for this shader stage. current_texture_bindpoint = @@ -294,6 +262,8 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) { } shader_program_manager->UseTrivialGeometryShader(); + + return {buffer_ptr, buffer_offset}; } size_t RasterizerOpenGL::CalculateVertexArraysSize() const { @@ -427,6 +397,31 @@ void RasterizerOpenGL::Clear() { } } +std::pair<u8*, GLintptr> RasterizerOpenGL::AlignBuffer(u8* buffer_ptr, GLintptr buffer_offset, + size_t alignment) { + // Align the offset, not the mapped pointer + GLintptr offset_aligned = + static_cast<GLintptr>(Common::AlignUp(static_cast<size_t>(buffer_offset), alignment)); + return {buffer_ptr + (offset_aligned - buffer_offset), offset_aligned}; +} + +std::tuple<u8*, GLintptr, GLintptr> RasterizerOpenGL::UploadMemory(u8* buffer_ptr, + GLintptr buffer_offset, + Tegra::GPUVAddr gpu_addr, + size_t size, size_t alignment) { + std::tie(buffer_ptr, buffer_offset) = AlignBuffer(buffer_ptr, buffer_offset, alignment); + GLintptr uploaded_offset = buffer_offset; + + const auto& memory_manager = Core::System::GetInstance().GPU().memory_manager; + const boost::optional<VAddr> cpu_addr{memory_manager->GpuToCpuAddress(gpu_addr)}; + Memory::ReadBlock(*cpu_addr, buffer_ptr, size); + + buffer_ptr += size; + buffer_offset += size; + + return {buffer_ptr, buffer_offset, uploaded_offset}; +} + void RasterizerOpenGL::DrawArrays() { if (accelerate_draw == AccelDraw::Disabled) return; @@ -451,7 +446,7 @@ void RasterizerOpenGL::DrawArrays() { const u64 index_buffer_size{regs.index_array.count * regs.index_array.FormatSizeInBytes()}; const unsigned vertex_num{is_indexed ? regs.index_array.count : regs.vertex_buffer.count}; - state.draw.vertex_buffer = stream_buffer->GetHandle(); + state.draw.vertex_buffer = stream_buffer.GetHandle(); state.Apply(); size_t buffer_size = CalculateVertexArraysSize(); @@ -461,41 +456,31 @@ void RasterizerOpenGL::DrawArrays() { } // Uniform space for the 5 shader stages - buffer_size = Common::AlignUp<size_t>(buffer_size, 4) + - sizeof(GLShader::MaxwellUniformData) * Maxwell::MaxShaderStage; + buffer_size = + Common::AlignUp<size_t>(buffer_size, 4) + + (sizeof(GLShader::MaxwellUniformData) + uniform_buffer_alignment) * Maxwell::MaxShaderStage; + + // Add space for at least 18 constant buffers + buffer_size += Maxwell::MaxConstBuffers * (MaxConstbufferSize + uniform_buffer_alignment); u8* buffer_ptr; GLintptr buffer_offset; - std::tie(buffer_ptr, buffer_offset) = - stream_buffer->Map(static_cast<GLsizeiptr>(buffer_size), 4); + std::tie(buffer_ptr, buffer_offset, std::ignore) = + stream_buffer.Map(static_cast<GLsizeiptr>(buffer_size), 4); + u8* buffer_ptr_base = buffer_ptr; - u8* offseted_buffer; - std::tie(offseted_buffer, buffer_offset) = SetupVertexArrays(buffer_ptr, buffer_offset); - - offseted_buffer = - reinterpret_cast<u8*>(Common::AlignUp(reinterpret_cast<size_t>(offseted_buffer), 4)); - buffer_offset = Common::AlignUp<size_t>(buffer_offset, 4); + std::tie(buffer_ptr, buffer_offset) = SetupVertexArrays(buffer_ptr, buffer_offset); // If indexed mode, copy the index buffer GLintptr index_buffer_offset = 0; if (is_indexed) { - const auto& memory_manager = Core::System::GetInstance().GPU().memory_manager; - const boost::optional<VAddr> index_data_addr{ - memory_manager->GpuToCpuAddress(regs.index_array.StartAddress())}; - Memory::ReadBlock(*index_data_addr, offseted_buffer, index_buffer_size); - - index_buffer_offset = buffer_offset; - offseted_buffer += index_buffer_size; - buffer_offset += index_buffer_size; + std::tie(buffer_ptr, buffer_offset, index_buffer_offset) = UploadMemory( + buffer_ptr, buffer_offset, regs.index_array.StartAddress(), index_buffer_size); } - offseted_buffer = - reinterpret_cast<u8*>(Common::AlignUp(reinterpret_cast<size_t>(offseted_buffer), 4)); - buffer_offset = Common::AlignUp<size_t>(buffer_offset, 4); - - SetupShaders(offseted_buffer, buffer_offset); + std::tie(buffer_ptr, buffer_offset) = SetupShaders(buffer_ptr, buffer_offset); - stream_buffer->Unmap(); + stream_buffer.Unmap(buffer_ptr - buffer_ptr_base); shader_program_manager->ApplyTo(state); state.Apply(); @@ -642,45 +627,32 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr } } -u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, GLuint program, - u32 current_bindpoint, - const std::vector<GLShader::ConstBufferEntry>& entries) { +std::tuple<u8*, GLintptr, u32> RasterizerOpenGL::SetupConstBuffers( + u8* buffer_ptr, GLintptr buffer_offset, Maxwell::ShaderStage stage, GLuint program, + u32 current_bindpoint, const std::vector<GLShader::ConstBufferEntry>& entries) { const auto& gpu = Core::System::GetInstance().GPU(); const auto& maxwell3d = gpu.Maxwell3D(); - // Reset all buffer draw state for this stage. - for (auto& buffer : state.draw.const_buffers[static_cast<size_t>(stage)]) { - buffer.bindpoint = 0; - buffer.enabled = false; - } - // Upload only the enabled buffers from the 16 constbuffers of each shader stage const auto& shader_stage = maxwell3d.state.shader_stages[static_cast<size_t>(stage)]; for (u32 bindpoint = 0; bindpoint < entries.size(); ++bindpoint) { const auto& used_buffer = entries[bindpoint]; const auto& buffer = shader_stage.const_buffers[used_buffer.GetIndex()]; - auto& buffer_draw_state = - state.draw.const_buffers[static_cast<size_t>(stage)][used_buffer.GetIndex()]; if (!buffer.enabled) { continue; } - buffer_draw_state.enabled = true; - buffer_draw_state.bindpoint = current_bindpoint + bindpoint; - - boost::optional<VAddr> addr = gpu.memory_manager->GpuToCpuAddress(buffer.address); - size_t size = 0; if (used_buffer.IsIndirect()) { // Buffer is accessed indirectly, so upload the entire thing - size = buffer.size * sizeof(float); + size = buffer.size; if (size > MaxConstbufferSize) { - LOG_ERROR(HW_GPU, "indirect constbuffer size {} exceeds maximum {}", size, - MaxConstbufferSize); + LOG_CRITICAL(HW_GPU, "indirect constbuffer size {} exceeds maximum {}", size, + MaxConstbufferSize); size = MaxConstbufferSize; } } else { @@ -693,25 +665,26 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, GLuint progr size = Common::AlignUp(size, sizeof(GLvec4)); ASSERT_MSG(size <= MaxConstbufferSize, "Constbuffer too big"); - std::vector<u8> data(size); - Memory::ReadBlock(*addr, data.data(), data.size()); + GLintptr const_buffer_offset; + std::tie(buffer_ptr, buffer_offset, const_buffer_offset) = + UploadMemory(buffer_ptr, buffer_offset, buffer.address, size, + static_cast<size_t>(uniform_buffer_alignment)); - glBindBuffer(GL_UNIFORM_BUFFER, buffer_draw_state.ssbo); - glBufferData(GL_UNIFORM_BUFFER, data.size(), data.data(), GL_DYNAMIC_DRAW); - glBindBuffer(GL_UNIFORM_BUFFER, 0); + glBindBufferRange(GL_UNIFORM_BUFFER, current_bindpoint + bindpoint, + stream_buffer.GetHandle(), const_buffer_offset, size); // Now configure the bindpoint of the buffer inside the shader const std::string buffer_name = used_buffer.GetName(); const GLuint index = glGetProgramResourceIndex(program, GL_UNIFORM_BLOCK, buffer_name.c_str()); if (index != GL_INVALID_INDEX) { - glUniformBlockBinding(program, index, buffer_draw_state.bindpoint); + glUniformBlockBinding(program, index, current_bindpoint + bindpoint); } } state.Apply(); - return current_bindpoint + static_cast<u32>(entries.size()); + return {buffer_ptr, buffer_offset, current_bindpoint + static_cast<u32>(entries.size())}; } u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, GLuint program, u32 current_unit, diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 6d6d85cc1..74307f626 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -7,6 +7,7 @@ #include <array> #include <cstddef> #include <memory> +#include <tuple> #include <utility> #include <vector> #include <glad/glad.h> @@ -21,12 +22,15 @@ #include "video_core/renderer_opengl/gl_state.h" #include "video_core/renderer_opengl/gl_stream_buffer.h" -class EmuWindow; struct ScreenInfo; +namespace Core::Frontend { +class EmuWindow; +} + class RasterizerOpenGL : public VideoCore::RasterizerInterface { public: - explicit RasterizerOpenGL(EmuWindow& renderer); + explicit RasterizerOpenGL(Core::Frontend::EmuWindow& renderer); ~RasterizerOpenGL() override; void DrawArrays() override; @@ -97,9 +101,10 @@ private: * @param entries Vector describing the buffers that are actually used in the guest shader. * @returns The next available bindpoint for use in the next shader stage. */ - u32 SetupConstBuffers(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, GLuint program, - u32 current_bindpoint, - const std::vector<GLShader::ConstBufferEntry>& entries); + std::tuple<u8*, GLintptr, u32> SetupConstBuffers( + u8* buffer_ptr, GLintptr buffer_offset, Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, + GLuint program, u32 current_bindpoint, + const std::vector<GLShader::ConstBufferEntry>& entries); /* * Configures the current textures to use for the draw command. @@ -136,7 +141,6 @@ private: /// Syncs the blend state to match the guest state void SyncBlendState(); - bool has_ARB_buffer_storage = false; bool has_ARB_direct_state_access = false; bool has_ARB_separate_shader_objects = false; bool has_ARB_vertex_attrib_binding = false; @@ -145,29 +149,31 @@ private: RasterizerCacheOpenGL res_cache; - EmuWindow& emu_window; + Core::Frontend::EmuWindow& emu_window; std::unique_ptr<GLShader::ProgramManager> shader_program_manager; OGLVertexArray sw_vao; OGLVertexArray hw_vao; std::array<SamplerInfo, GLShader::NumTextureSamplers> texture_samplers; - std::array<std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers>, - Tegra::Engines::Maxwell3D::Regs::MaxShaderStage> - ssbos; static constexpr size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024; - std::unique_ptr<OGLStreamBuffer> stream_buffer; + OGLStreamBuffer stream_buffer; OGLBuffer uniform_buffer; OGLFramebuffer framebuffer; + GLint uniform_buffer_alignment; size_t CalculateVertexArraysSize() const; std::pair<u8*, GLintptr> SetupVertexArrays(u8* array_ptr, GLintptr buffer_offset); - std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::MaxShaderStage> uniform_buffers; + std::pair<u8*, GLintptr> SetupShaders(u8* buffer_ptr, GLintptr buffer_offset); + + std::pair<u8*, GLintptr> AlignBuffer(u8* buffer_ptr, GLintptr buffer_offset, size_t alignment); - void SetupShaders(u8* buffer_ptr, GLintptr buffer_offset); + std::tuple<u8*, GLintptr, GLintptr> UploadMemory(u8* buffer_ptr, GLintptr buffer_offset, + Tegra::GPUVAddr gpu_addr, size_t size, + size_t alignment = 4); enum class AccelDraw { Disabled, Arrays, Indexed }; AccelDraw accelerate_draw = AccelDraw::Disabled; diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index 9fb734b77..fb7476fb8 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -92,13 +92,18 @@ struct FormatTuple { } static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{ - {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8 - {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, ComponentType::UNorm, false}, // B5G6R5 + {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8U + {GL_RGBA8, GL_RGBA, GL_BYTE, ComponentType::SNorm, false}, // ABGR8S + {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false}, // ABGR8UI + {GL_RGB, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, ComponentType::UNorm, false}, // B5G6R5U {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV, ComponentType::UNorm, - false}, // A2B10G10R10 - {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV, ComponentType::UNorm, false}, // A1B5G5R5 - {GL_R8, GL_RED, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // R8 + false}, // A2B10G10R10U + {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV, ComponentType::UNorm, false}, // A1B5G5R5U + {GL_R8, GL_RED, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // R8U + {GL_R8UI, GL_RED_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false}, // R8UI {GL_RGBA16F, GL_RGBA, GL_HALF_FLOAT, ComponentType::Float, false}, // RGBA16F + {GL_RGBA16, GL_RGBA, GL_UNSIGNED_SHORT, ComponentType::UNorm, false}, // RGBA16U + {GL_RGBA16UI, GL_RGBA, GL_UNSIGNED_SHORT, ComponentType::UInt, false}, // RGBA16UI {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV, ComponentType::Float, false}, // R11FG11FB10F {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false}, // RGBA32UI @@ -112,16 +117,20 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form {GL_COMPRESSED_RG_RGTC2, GL_RG, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, true}, // DXN2UNORM {GL_COMPRESSED_SIGNED_RG_RGTC2, GL_RG, GL_INT, ComponentType::SNorm, true}, // DXN2SNORM - {GL_COMPRESSED_RGBA_BPTC_UNORM_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, + {GL_COMPRESSED_RGBA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm, true}, // BC7U {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_4X4 - {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // G8R8 + {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // G8R8U + {GL_RG8, GL_RG, GL_BYTE, ComponentType::SNorm, false}, // G8R8S {GL_RGBA8, GL_BGRA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // BGRA8 {GL_RGBA32F, GL_RGBA, GL_FLOAT, ComponentType::Float, false}, // RGBA32F {GL_RG32F, GL_RG, GL_FLOAT, ComponentType::Float, false}, // RG32F {GL_R32F, GL_RED, GL_FLOAT, ComponentType::Float, false}, // R32F {GL_R16F, GL_RED, GL_HALF_FLOAT, ComponentType::Float, false}, // R16F - {GL_R16, GL_RED, GL_UNSIGNED_SHORT, ComponentType::UNorm, false}, // R16UNORM + {GL_R16, GL_RED, GL_UNSIGNED_SHORT, ComponentType::UNorm, false}, // R16U + {GL_R16_SNORM, GL_RED, GL_SHORT, ComponentType::SNorm, false}, // R16S + {GL_R16UI, GL_RED_INTEGER, GL_UNSIGNED_SHORT, ComponentType::UInt, false}, // R16UI + {GL_R16I, GL_RED_INTEGER, GL_SHORT, ComponentType::SInt, false}, // R16I {GL_RG16, GL_RG, GL_UNSIGNED_SHORT, ComponentType::UNorm, false}, // RG16 {GL_RG16F, GL_RG, GL_HALF_FLOAT, ComponentType::Float, false}, // RG16F {GL_RG16UI, GL_RG_INTEGER, GL_UNSIGNED_SHORT, ComponentType::UInt, false}, // RG16UI @@ -129,15 +138,21 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form {GL_RG16_SNORM, GL_RG, GL_SHORT, ComponentType::SNorm, false}, // RG16S {GL_RGB32F, GL_RGB, GL_FLOAT, ComponentType::Float, false}, // RGB32F {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // SRGBA8 + {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // RG8U + {GL_RG8, GL_RG, GL_BYTE, ComponentType::SNorm, false}, // RG8S + {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false}, // RG32UI + {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT, ComponentType::UInt, false}, // R32UI + + // Depth formats + {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F + {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT, ComponentType::UNorm, + false}, // Z16 // DepthStencil formats {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8, ComponentType::UNorm, false}, // Z24S8 {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8, ComponentType::UNorm, - false}, // S8Z24 - {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F - {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT, ComponentType::UNorm, - false}, // Z16 + false}, // S8Z24 {GL_DEPTH32F_STENCIL8, GL_DEPTH_STENCIL, GL_FLOAT_32_UNSIGNED_INT_24_8_REV, ComponentType::Float, false}, // Z32FS8 }}; @@ -228,35 +243,73 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, std::vector<u8>& gl_bu static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPUVAddr), SurfaceParams::MaxPixelFormat> morton_to_gl_fns = { - MortonCopy<true, PixelFormat::ABGR8>, MortonCopy<true, PixelFormat::B5G6R5>, - MortonCopy<true, PixelFormat::A2B10G10R10>, MortonCopy<true, PixelFormat::A1B5G5R5>, - MortonCopy<true, PixelFormat::R8>, MortonCopy<true, PixelFormat::RGBA16F>, - 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::ASTC_2D_4X4>, - MortonCopy<true, PixelFormat::G8R8>, MortonCopy<true, PixelFormat::BGRA8>, - MortonCopy<true, PixelFormat::RGBA32F>, MortonCopy<true, PixelFormat::RG32F>, - MortonCopy<true, PixelFormat::R32F>, MortonCopy<true, PixelFormat::R16F>, - MortonCopy<true, PixelFormat::R16UNORM>, 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::SRGBA8>, - MortonCopy<true, PixelFormat::Z24S8>, MortonCopy<true, PixelFormat::S8Z24>, - MortonCopy<true, PixelFormat::Z32F>, MortonCopy<true, PixelFormat::Z16>, + // clang-format off + 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::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::ASTC_2D_4X4>, + MortonCopy<true, PixelFormat::G8R8U>, + MortonCopy<true, PixelFormat::G8R8S>, + 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::SRGBA8>, + MortonCopy<true, PixelFormat::RG8U>, + MortonCopy<true, PixelFormat::RG8S>, + MortonCopy<true, PixelFormat::RG32UI>, + MortonCopy<true, PixelFormat::R32UI>, + MortonCopy<true, PixelFormat::Z32F>, + MortonCopy<true, PixelFormat::Z16>, + MortonCopy<true, PixelFormat::Z24S8>, + MortonCopy<true, PixelFormat::S8Z24>, MortonCopy<true, PixelFormat::Z32FS8>, + // clang-format on }; static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPUVAddr), SurfaceParams::MaxPixelFormat> gl_to_morton_fns = { - MortonCopy<false, PixelFormat::ABGR8>, - MortonCopy<false, PixelFormat::B5G6R5>, - MortonCopy<false, PixelFormat::A2B10G10R10>, - MortonCopy<false, PixelFormat::A1B5G5R5>, - MortonCopy<false, PixelFormat::R8>, + // clang-format off + 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::RGBA16U>, + MortonCopy<false, PixelFormat::RGBA16UI>, MortonCopy<false, PixelFormat::R11FG11FB10F>, MortonCopy<false, PixelFormat::RGBA32UI>, // TODO(Subv): Swizzling DXT1/DXT23/DXT45/DXN1/DXN2/BC7U/ASTC_2D_4X4 formats is not @@ -269,13 +322,17 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU nullptr, nullptr, nullptr, - MortonCopy<false, PixelFormat::G8R8>, + MortonCopy<false, PixelFormat::G8R8U>, + MortonCopy<false, PixelFormat::G8R8S>, MortonCopy<false, PixelFormat::BGRA8>, MortonCopy<false, PixelFormat::RGBA32F>, MortonCopy<false, PixelFormat::RG32F>, MortonCopy<false, PixelFormat::R32F>, MortonCopy<false, PixelFormat::R16F>, - MortonCopy<false, PixelFormat::R16UNORM>, + 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>, @@ -283,11 +340,16 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU MortonCopy<false, PixelFormat::RG16S>, MortonCopy<false, PixelFormat::RGB32F>, MortonCopy<false, PixelFormat::SRGBA8>, - MortonCopy<false, PixelFormat::Z24S8>, - MortonCopy<false, PixelFormat::S8Z24>, + MortonCopy<false, PixelFormat::RG8U>, + MortonCopy<false, PixelFormat::RG8S>, + MortonCopy<false, PixelFormat::RG32UI>, + MortonCopy<false, PixelFormat::R32UI>, MortonCopy<false, PixelFormat::Z32F>, MortonCopy<false, PixelFormat::Z16>, + MortonCopy<false, PixelFormat::Z24S8>, + MortonCopy<false, PixelFormat::S8Z24>, MortonCopy<false, PixelFormat::Z32FS8>, + // clang-format on }; // Allocate an uninitialized texture of appropriate size and format for the surface @@ -407,7 +469,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) { } static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) { - const auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8)}; + const auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8U)}; for (size_t y = 0; y < height; ++y) { for (size_t x = 0; x < width; ++x) { const size_t offset{bpp * (y * width + x)}; @@ -439,7 +501,8 @@ static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelForma ConvertS8Z24ToZ24S8(data, width, height); break; - case PixelFormat::G8R8: + case PixelFormat::G8R8U: + case PixelFormat::G8R8S: // Convert the G8R8 color format to R8G8, as OpenGL does not support G8R8. ConvertG8R8ToR8G8(data, width, height); break; @@ -730,8 +793,6 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& surface, // Verify surface is compatible for blitting const auto& params{surface->GetSurfaceParams()}; ASSERT(params.type == new_params.type); - ASSERT(params.pixel_format == new_params.pixel_format); - ASSERT(params.component_type == new_params.component_type); // Create a new surface with the new parameters, and blit the previous surface to it Surface new_surface{std::make_shared<CachedSurface>(new_params)}; diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h index 829a76dfe..fc8b44219 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -23,45 +23,62 @@ using PageMap = boost::icl::interval_map<u64, int>; struct SurfaceParams { enum class PixelFormat { - ABGR8 = 0, - B5G6R5 = 1, - A2B10G10R10 = 2, - A1B5G5R5 = 3, - R8 = 4, - RGBA16F = 5, - R11FG11FB10F = 6, - RGBA32UI = 7, - DXT1 = 8, - DXT23 = 9, - DXT45 = 10, - DXN1 = 11, // This is also known as BC4 - DXN2UNORM = 12, - DXN2SNORM = 13, - BC7U = 14, - ASTC_2D_4X4 = 15, - G8R8 = 16, - BGRA8 = 17, - RGBA32F = 18, - RG32F = 19, - R32F = 20, - R16F = 21, - R16UNORM = 22, - RG16 = 23, - RG16F = 24, - RG16UI = 25, - RG16I = 26, - RG16S = 27, - RGB32F = 28, - SRGBA8 = 29, + ABGR8U = 0, + ABGR8S = 1, + ABGR8UI = 2, + B5G6R5U = 3, + A2B10G10R10U = 4, + A1B5G5R5U = 5, + R8U = 6, + R8UI = 7, + RGBA16F = 8, + RGBA16U = 9, + RGBA16UI = 10, + R11FG11FB10F = 11, + RGBA32UI = 12, + DXT1 = 13, + DXT23 = 14, + DXT45 = 15, + DXN1 = 16, // This is also known as BC4 + DXN2UNORM = 17, + DXN2SNORM = 18, + BC7U = 19, + ASTC_2D_4X4 = 20, + G8R8U = 21, + G8R8S = 22, + BGRA8 = 23, + RGBA32F = 24, + RG32F = 25, + R32F = 26, + R16F = 27, + R16U = 28, + R16S = 29, + R16UI = 30, + R16I = 31, + RG16 = 32, + RG16F = 33, + RG16UI = 34, + RG16I = 35, + RG16S = 36, + RGB32F = 37, + SRGBA8 = 38, + RG8U = 39, + RG8S = 40, + RG32UI = 41, + R32UI = 42, MaxColorFormat, + // Depth formats + Z32F = 43, + Z16 = 44, + + MaxDepthFormat, + // DepthStencil formats - Z24S8 = 30, - S8Z24 = 31, - Z32F = 32, - Z16 = 33, - Z32FS8 = 34, + Z24S8 = 45, + S8Z24 = 46, + Z32FS8 = 47, MaxDepthStencilFormat, @@ -99,12 +116,17 @@ struct SurfaceParams { return 0; constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{ - 1, // ABGR8 - 1, // B5G6R5 - 1, // A2B10G10R10 - 1, // A1B5G5R5 - 1, // R8 + 1, // ABGR8U + 1, // ABGR8S + 1, // ABGR8UI + 1, // B5G6R5U + 1, // A2B10G10R10U + 1, // A1B5G5R5U + 1, // R8U + 1, // R8UI 1, // RGBA16F + 1, // RGBA16U + 1, // RGBA16UI 1, // R11FG11FB10F 1, // RGBA32UI 4, // DXT1 @@ -115,13 +137,17 @@ struct SurfaceParams { 4, // DXN2SNORM 4, // BC7U 4, // ASTC_2D_4X4 - 1, // G8R8 + 1, // G8R8U + 1, // G8R8S 1, // BGRA8 1, // RGBA32F 1, // RG32F 1, // R32F 1, // R16F - 1, // R16UNORM + 1, // R16U + 1, // R16S + 1, // R16UI + 1, // R16I 1, // RG16 1, // RG16F 1, // RG16UI @@ -129,10 +155,14 @@ struct SurfaceParams { 1, // RG16S 1, // RGB32F 1, // SRGBA8 - 1, // Z24S8 - 1, // S8Z24 + 1, // RG8U + 1, // RG8S + 1, // RG32UI + 1, // R32UI 1, // Z32F 1, // Z16 + 1, // Z24S8 + 1, // S8Z24 1, // Z32FS8 }}; @@ -145,12 +175,17 @@ struct SurfaceParams { return 0; constexpr std::array<u32, MaxPixelFormat> bpp_table = {{ - 32, // ABGR8 - 16, // B5G6R5 - 32, // A2B10G10R10 - 16, // A1B5G5R5 - 8, // R8 + 32, // ABGR8U + 32, // ABGR8S + 32, // ABGR8UI + 16, // B5G6R5U + 32, // A2B10G10R10U + 16, // A1B5G5R5U + 8, // R8U + 8, // R8UI 64, // RGBA16F + 64, // RGBA16U + 64, // RGBA16UI 32, // R11FG11FB10F 128, // RGBA32UI 64, // DXT1 @@ -161,13 +196,17 @@ struct SurfaceParams { 128, // DXN2SNORM 128, // BC7U 32, // ASTC_2D_4X4 - 16, // G8R8 + 16, // G8R8U + 16, // G8R8S 32, // BGRA8 128, // RGBA32F 64, // RG32F 32, // R32F 16, // R16F - 16, // R16UNORM + 16, // R16U + 16, // R16S + 16, // R16UI + 16, // R16I 32, // RG16 32, // RG16F 32, // RG16UI @@ -175,10 +214,14 @@ struct SurfaceParams { 32, // RG16S 96, // RGB32F 32, // SRGBA8 - 32, // Z24S8 - 32, // S8Z24 + 16, // RG8U + 16, // RG8S + 64, // RG32UI + 32, // R32UI 32, // Z32F 16, // Z16 + 32, // Z24S8 + 32, // S8Z24 64, // Z32FS8 }}; @@ -214,13 +257,21 @@ struct SurfaceParams { // gamma. case Tegra::RenderTargetFormat::RGBA8_SRGB: case Tegra::RenderTargetFormat::RGBA8_UNORM: - return PixelFormat::ABGR8; + return PixelFormat::ABGR8U; + case Tegra::RenderTargetFormat::RGBA8_SNORM: + return PixelFormat::ABGR8S; + case Tegra::RenderTargetFormat::RGBA8_UINT: + return PixelFormat::ABGR8UI; case Tegra::RenderTargetFormat::BGRA8_UNORM: return PixelFormat::BGRA8; case Tegra::RenderTargetFormat::RGB10_A2_UNORM: - return PixelFormat::A2B10G10R10; + return PixelFormat::A2B10G10R10U; case Tegra::RenderTargetFormat::RGBA16_FLOAT: return PixelFormat::RGBA16F; + case Tegra::RenderTargetFormat::RGBA16_UNORM: + return PixelFormat::RGBA16U; + case Tegra::RenderTargetFormat::RGBA16_UINT: + return PixelFormat::RGBA16UI; case Tegra::RenderTargetFormat::RGBA32_FLOAT: return PixelFormat::RGBA32F; case Tegra::RenderTargetFormat::RG32_FLOAT: @@ -228,11 +279,13 @@ struct SurfaceParams { case Tegra::RenderTargetFormat::R11G11B10_FLOAT: return PixelFormat::R11FG11FB10F; case Tegra::RenderTargetFormat::B5G6R5_UNORM: - return PixelFormat::B5G6R5; + return PixelFormat::B5G6R5U; case Tegra::RenderTargetFormat::RGBA32_UINT: return PixelFormat::RGBA32UI; case Tegra::RenderTargetFormat::R8_UNORM: - return PixelFormat::R8; + return PixelFormat::R8U; + case Tegra::RenderTargetFormat::R8_UINT: + return PixelFormat::R8UI; case Tegra::RenderTargetFormat::RG16_FLOAT: return PixelFormat::RG16F; case Tegra::RenderTargetFormat::RG16_UINT: @@ -243,10 +296,26 @@ struct SurfaceParams { 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::R16_UNORM: + return PixelFormat::R16U; + case Tegra::RenderTargetFormat::R16_SNORM: + return PixelFormat::R16S; + case Tegra::RenderTargetFormat::R16_UINT: + return PixelFormat::R16UI; + case Tegra::RenderTargetFormat::R16_SINT: + return PixelFormat::R16I; case Tegra::RenderTargetFormat::R32_FLOAT: return PixelFormat::R32F; + case Tegra::RenderTargetFormat::R32_UINT: + return PixelFormat::R32UI; + case Tegra::RenderTargetFormat::RG32_UINT: + return PixelFormat::RG32UI; default: LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format)); UNREACHABLE(); @@ -258,21 +327,79 @@ struct SurfaceParams { // TODO(Subv): Properly implement this switch (format) { case Tegra::Texture::TextureFormat::A8R8G8B8: - return PixelFormat::ABGR8; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::ABGR8U; + case Tegra::Texture::ComponentType::SNORM: + return PixelFormat::ABGR8S; + case Tegra::Texture::ComponentType::UINT: + return PixelFormat::ABGR8UI; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::B5G6R5: - return PixelFormat::B5G6R5; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::B5G6R5U; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::A2B10G10R10: - return PixelFormat::A2B10G10R10; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::A2B10G10R10U; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::A1B5G5R5: - return PixelFormat::A1B5G5R5; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::A1B5G5R5U; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::R8: - return PixelFormat::R8; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::R8U; + case Tegra::Texture::ComponentType::UINT: + return PixelFormat::R8UI; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::G8R8: - return PixelFormat::G8R8; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::G8R8U; + case Tegra::Texture::ComponentType::SNORM: + return PixelFormat::G8R8S; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::R16_G16_B16_A16: - return PixelFormat::RGBA16F; + switch (component_type) { + case Tegra::Texture::ComponentType::UNORM: + return PixelFormat::RGBA16U; + case Tegra::Texture::ComponentType::FLOAT: + return PixelFormat::RGBA16F; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::BF10GF11RF11: - return PixelFormat::R11FG11FB10F; + switch (component_type) { + case Tegra::Texture::ComponentType::FLOAT: + return PixelFormat::R11FG11FB10F; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::R32_G32_B32_A32: switch (component_type) { case Tegra::Texture::ComponentType::FLOAT: @@ -284,23 +411,53 @@ struct SurfaceParams { static_cast<u32>(component_type)); UNREACHABLE(); case Tegra::Texture::TextureFormat::R32_G32: - return PixelFormat::RG32F; + switch (component_type) { + case Tegra::Texture::ComponentType::FLOAT: + return PixelFormat::RG32F; + case Tegra::Texture::ComponentType::UINT: + return PixelFormat::RG32UI; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::R32_G32_B32: - return PixelFormat::RGB32F; + switch (component_type) { + case Tegra::Texture::ComponentType::FLOAT: + return PixelFormat::RGB32F; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::R16: switch (component_type) { case Tegra::Texture::ComponentType::FLOAT: return PixelFormat::R16F; case Tegra::Texture::ComponentType::UNORM: - return PixelFormat::R16UNORM; + return PixelFormat::R16U; + case Tegra::Texture::ComponentType::SNORM: + return PixelFormat::R16S; + case Tegra::Texture::ComponentType::UINT: + return PixelFormat::R16UI; + case Tegra::Texture::ComponentType::SINT: + return PixelFormat::R16I; } LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", static_cast<u32>(component_type)); UNREACHABLE(); case Tegra::Texture::TextureFormat::R32: - return PixelFormat::R32F; + switch (component_type) { + case Tegra::Texture::ComponentType::FLOAT: + return PixelFormat::R32F; + case Tegra::Texture::ComponentType::UINT: + return PixelFormat::R32UI; + } + LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}", + static_cast<u32>(component_type)); + UNREACHABLE(); case Tegra::Texture::TextureFormat::ZF32: return PixelFormat::Z32F; + case Tegra::Texture::TextureFormat::Z16: + return PixelFormat::Z16; case Tegra::Texture::TextureFormat::Z24S8: return PixelFormat::Z24S8; case Tegra::Texture::TextureFormat::DXT1: @@ -376,9 +533,15 @@ struct SurfaceParams { case Tegra::RenderTargetFormat::RGB10_A2_UNORM: case Tegra::RenderTargetFormat::R8_UNORM: case Tegra::RenderTargetFormat::RG16_UNORM: + case Tegra::RenderTargetFormat::R16_UNORM: case Tegra::RenderTargetFormat::B5G6R5_UNORM: + case Tegra::RenderTargetFormat::RG8_UNORM: + case Tegra::RenderTargetFormat::RGBA16_UNORM: return ComponentType::UNorm; + case Tegra::RenderTargetFormat::RGBA8_SNORM: case Tegra::RenderTargetFormat::RG16_SNORM: + case Tegra::RenderTargetFormat::R16_SNORM: + case Tegra::RenderTargetFormat::RG8_SNORM: return ComponentType::SNorm; case Tegra::RenderTargetFormat::RGBA16_FLOAT: case Tegra::RenderTargetFormat::R11G11B10_FLOAT: @@ -389,9 +552,16 @@ struct SurfaceParams { case Tegra::RenderTargetFormat::R32_FLOAT: return ComponentType::Float; case Tegra::RenderTargetFormat::RGBA32_UINT: + case Tegra::RenderTargetFormat::RGBA16_UINT: case Tegra::RenderTargetFormat::RG16_UINT: + case Tegra::RenderTargetFormat::R8_UINT: + case Tegra::RenderTargetFormat::R16_UINT: + case Tegra::RenderTargetFormat::RG32_UINT: + case Tegra::RenderTargetFormat::R32_UINT: + case Tegra::RenderTargetFormat::RGBA8_UINT: return ComponentType::UInt; case Tegra::RenderTargetFormat::RG16_SINT: + case Tegra::RenderTargetFormat::R16_SINT: return ComponentType::SInt; default: LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format)); @@ -402,7 +572,7 @@ struct SurfaceParams { static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) { switch (format) { case Tegra::FramebufferConfig::PixelFormat::ABGR8: - return PixelFormat::ABGR8; + return PixelFormat::ABGR8U; default: LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format)); UNREACHABLE(); @@ -429,6 +599,10 @@ struct SurfaceParams { return SurfaceType::ColorTexture; } + if (static_cast<size_t>(pixel_format) < static_cast<size_t>(PixelFormat::MaxDepthFormat)) { + return SurfaceType::Depth; + } + if (static_cast<size_t>(pixel_format) < static_cast<size_t>(PixelFormat::MaxDepthStencilFormat)) { return SurfaceType::DepthStencil; diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index 32f06f409..57cf9f213 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -141,6 +141,15 @@ private: ExitMethod jmp = Scan(target, end, labels); return exit_method = ParallelExit(no_jmp, jmp); } + case OpCode::Id::SSY: { + // The SSY instruction uses a similar encoding as the BRA instruction. + ASSERT_MSG(instr.bra.constant_buffer == 0, + "Constant buffer SSY is not supported"); + u32 target = offset + instr.bra.GetBranchTarget(); + labels.insert(target); + // Continue scanning for an exit method. + break; + } } } } @@ -347,35 +356,43 @@ public: * @param reg The register to use as the source value. */ void SetOutputAttributeToRegister(Attribute::Index attribute, u64 elem, const Register& reg) { - std::string dest = GetOutputAttribute(attribute) + GetSwizzle(elem); + std::string dest = GetOutputAttribute(attribute); std::string src = GetRegisterAsFloat(reg); - shader.AddLine(dest + " = " + src + ';'); + + if (!dest.empty()) { + // Can happen with unknown/unimplemented output attributes, in which case we ignore the + // instruction for now. + shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';'); + } } /// Generates code representing a uniform (C buffer) register, interpreted as the input type. - std::string GetUniform(u64 index, u64 offset, GLSLRegister::Type type) { + std::string GetUniform(u64 index, u64 offset, GLSLRegister::Type type, + Register::Size size = Register::Size::Word) { declr_const_buffers[index].MarkAsUsed(index, offset, stage); std::string value = 'c' + std::to_string(index) + '[' + std::to_string(offset / 4) + "][" + std::to_string(offset % 4) + ']'; if (type == GLSLRegister::Type::Float) { - return value; + // Do nothing, default } else if (type == GLSLRegister::Type::Integer) { - return "floatBitsToInt(" + value + ')'; + value = "floatBitsToInt(" + value + ')'; + } else if (type == GLSLRegister::Type::UnsignedInteger) { + value = "floatBitsToUint(" + value + ')'; } else { UNREACHABLE(); } + + return ConvertIntegerSize(value, size); } - std::string GetUniformIndirect(u64 index, s64 offset, const Register& index_reg, + std::string GetUniformIndirect(u64 cbuf_index, s64 offset, const std::string& index_str, GLSLRegister::Type type) { - declr_const_buffers[index].MarkAsUsedIndirect(index, stage); - - std::string final_offset = "((floatBitsToInt(" + GetRegister(index_reg, 0) + ") + " + - std::to_string(offset) + ") / 4)"; + declr_const_buffers[cbuf_index].MarkAsUsedIndirect(cbuf_index, stage); - std::string value = - 'c' + std::to_string(index) + '[' + final_offset + " / 4][" + final_offset + " % 4]"; + std::string final_offset = fmt::format("({} + {})", index_str, offset / 4); + std::string value = 'c' + std::to_string(cbuf_index) + '[' + final_offset + " / 4][" + + final_offset + " % 4]"; if (type == GLSLRegister::Type::Float) { return value; @@ -524,21 +541,21 @@ private: // vertex shader, and what's the value of the fourth element when inside a Tess Eval // shader. ASSERT(stage == Maxwell3D::Regs::ShaderStage::Vertex); - return "vec4(0, 0, uintBitsToFloat(gl_InstanceID), uintBitsToFloat(gl_VertexID))"; - case Attribute::Index::Unknown_63: - // TODO(bunnei): Figure out what this is used for. Super Mario Odyssey uses this. - LOG_CRITICAL(HW_GPU, "Unhandled input attribute Unknown_63"); - UNREACHABLE(); - break; + return "vec4(0, 0, uintBitsToFloat(instance_id.x), uintBitsToFloat(gl_VertexID))"; + case Attribute::Index::FrontFacing: + // TODO(Subv): Find out what the values are for the other elements. + ASSERT(stage == Maxwell3D::Regs::ShaderStage::Fragment); + return "vec4(0, 0, 0, uintBitsToFloat(gl_FrontFacing ? 1 : 0))"; default: const u32 index{static_cast<u32>(attribute) - static_cast<u32>(Attribute::Index::Attribute_0)}; - if (attribute >= Attribute::Index::Attribute_0) { + if (attribute >= Attribute::Index::Attribute_0 && + attribute <= Attribute::Index::Attribute_31) { declr_input_attribute.insert(attribute); return "input_attribute_" + std::to_string(index); } - LOG_CRITICAL(HW_GPU, "Unhandled input attribute: {}", index); + LOG_CRITICAL(HW_GPU, "Unhandled input attribute: {}", static_cast<u32>(attribute)); UNREACHABLE(); } @@ -560,6 +577,7 @@ private: LOG_CRITICAL(HW_GPU, "Unhandled output attribute: {}", index); UNREACHABLE(); + return {}; } } @@ -689,10 +707,11 @@ private: const std::string& op_a, const std::string& op_b) const { using Tegra::Shader::PredCondition; static const std::unordered_map<PredCondition, const char*> PredicateComparisonStrings = { - {PredCondition::LessThan, "<"}, {PredCondition::Equal, "=="}, - {PredCondition::LessEqual, "<="}, {PredCondition::GreaterThan, ">"}, - {PredCondition::NotEqual, "!="}, {PredCondition::GreaterEqual, ">="}, - {PredCondition::LessThanWithNan, "<"}, {PredCondition::NotEqualWithNan, "!="}, + {PredCondition::LessThan, "<"}, {PredCondition::Equal, "=="}, + {PredCondition::LessEqual, "<="}, {PredCondition::GreaterThan, ">"}, + {PredCondition::NotEqual, "!="}, {PredCondition::GreaterEqual, ">="}, + {PredCondition::LessThanWithNan, "<"}, {PredCondition::NotEqualWithNan, "!="}, + {PredCondition::GreaterThanWithNan, ">"}, }; const auto& comparison{PredicateComparisonStrings.find(condition)}; @@ -701,7 +720,8 @@ private: std::string predicate{'(' + op_a + ") " + comparison->second + " (" + op_b + ')'}; if (condition == PredCondition::LessThanWithNan || - condition == PredCondition::NotEqualWithNan) { + condition == PredCondition::NotEqualWithNan || + condition == PredCondition::GreaterThanWithNan) { predicate += " || isnan(" + op_a + ") || isnan(" + op_b + ')'; } @@ -727,6 +747,30 @@ private: return op->second; } + /** + * Transforms the input string GLSL operand into one that applies the abs() function and negates + * the output if necessary. When both abs and neg are true, the negation will be applied after + * taking the absolute value. + * @param operand The input operand to take the abs() of, negate, or both. + * @param abs Whether to apply the abs() function to the input operand. + * @param neg Whether to negate the input operand. + * @returns String corresponding to the operand after being transformed by the abs() and + * negation operations. + */ + static std::string GetOperandAbsNeg(const std::string& operand, bool abs, bool neg) { + std::string result = operand; + + if (abs) { + result = "abs(" + result + ')'; + } + + if (neg) { + result = "-(" + result + ')'; + } + + return result; + } + /* * Returns whether the instruction at the specified offset is a 'sched' instruction. * Sched instructions always appear before a sequence of 3 instructions. @@ -740,28 +784,51 @@ private: } void WriteLogicOperation(Register dest, LogicOperation logic_op, const std::string& op_a, - const std::string& op_b) { + const std::string& op_b, + Tegra::Shader::PredicateResultMode predicate_mode, + Tegra::Shader::Pred predicate) { + std::string result{}; switch (logic_op) { case LogicOperation::And: { - regs.SetRegisterToInteger(dest, true, 0, '(' + op_a + " & " + op_b + ')', 1, 1); + result = '(' + op_a + " & " + op_b + ')'; break; } case LogicOperation::Or: { - regs.SetRegisterToInteger(dest, true, 0, '(' + op_a + " | " + op_b + ')', 1, 1); + result = '(' + op_a + " | " + op_b + ')'; break; } case LogicOperation::Xor: { - regs.SetRegisterToInteger(dest, true, 0, '(' + op_a + " ^ " + op_b + ')', 1, 1); + result = '(' + op_a + " ^ " + op_b + ')'; break; } case LogicOperation::PassB: { - regs.SetRegisterToInteger(dest, true, 0, op_b, 1, 1); + result = op_b; break; } default: LOG_CRITICAL(HW_GPU, "Unimplemented logic operation: {}", static_cast<u32>(logic_op)); UNREACHABLE(); } + + if (dest != Tegra::Shader::Register::ZeroIndex) { + regs.SetRegisterToInteger(dest, true, 0, result, 1, 1); + } + + using Tegra::Shader::PredicateResultMode; + // Write the predicate value depending on the predicate mode. + switch (predicate_mode) { + case PredicateResultMode::None: + // Do nothing. + return; + case PredicateResultMode::NotZero: + // Set the predicate to true if the result is not zero. + SetPredicate(static_cast<u64>(predicate), '(' + result + ") != 0"); + break; + default: + LOG_CRITICAL(HW_GPU, "Unimplemented predicate result mode: {}", + static_cast<u32>(predicate_mode)); + UNREACHABLE(); + } } void WriteTexsInstruction(const Instruction& instr, const std::string& coord, @@ -799,6 +866,33 @@ private: shader.AddLine('}'); } + /* + * Emits code to push the input target address to the SSY address stack, incrementing the stack + * top. + */ + void EmitPushToSSYStack(u32 target) { + shader.AddLine('{'); + ++shader.scope; + shader.AddLine("ssy_stack[ssy_stack_top] = " + std::to_string(target) + "u;"); + shader.AddLine("ssy_stack_top++;"); + --shader.scope; + shader.AddLine('}'); + } + + /* + * Emits code to pop an address from the SSY address stack, setting the jump address to the + * popped address and decrementing the stack top. + */ + void EmitPopFromSSYStack() { + shader.AddLine('{'); + ++shader.scope; + shader.AddLine("ssy_stack_top--;"); + shader.AddLine("jmp_to = ssy_stack[ssy_stack_top];"); + shader.AddLine("break;"); + --shader.scope; + shader.AddLine('}'); + } + /** * Compiles a single instruction from Tegra to GLSL. * @param offset the offset of the Tegra shader instruction. @@ -828,7 +922,11 @@ private: ASSERT_MSG(instr.pred.full_pred != Pred::NeverExecute, "NeverExecute predicate not implemented"); - if (instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) { + // Some instructions (like SSY) don't have a predicate field, they are always + // unconditionally executed. + bool can_be_predicated = OpCode::IsPredicatedInstruction(opcode->GetId()); + + if (can_be_predicated && instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) { shader.AddLine("if (" + GetPredicateCondition(instr.pred.pred_index, instr.negate_pred != 0) + ')'); @@ -839,13 +937,6 @@ private: switch (opcode->GetType()) { case OpCode::Type::Arithmetic: { std::string op_a = regs.GetRegisterAsFloat(instr.gpr8); - if (instr.alu.abs_a) { - op_a = "abs(" + op_a + ')'; - } - - if (instr.alu.negate_a) { - op_a = "-(" + op_a + ')'; - } std::string op_b; @@ -860,17 +951,10 @@ private: } } - if (instr.alu.abs_b) { - op_b = "abs(" + op_b + ')'; - } - - if (instr.alu.negate_b) { - op_b = "-(" + op_b + ')'; - } - switch (opcode->GetId()) { case OpCode::Id::MOV_C: case OpCode::Id::MOV_R: { + // MOV does not have neither 'abs' nor 'neg' bits. regs.SetRegisterToFloat(instr.gpr0, 0, op_b, 1, 1); break; } @@ -878,6 +962,8 @@ private: case OpCode::Id::FMUL_C: case OpCode::Id::FMUL_R: case OpCode::Id::FMUL_IMM: { + // FMUL does not have 'abs' bits and only the second operand has a 'neg' bit. + op_b = GetOperandAbsNeg(op_b, false, instr.fmul.negate_b); regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1, instr.alu.saturate_d); break; @@ -885,11 +971,14 @@ private: case OpCode::Id::FADD_C: case OpCode::Id::FADD_R: case OpCode::Id::FADD_IMM: { + op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a); + op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b); regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1, instr.alu.saturate_d); break; } case OpCode::Id::MUFU: { + op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a); switch (instr.sub_op) { case SubOp::Cos: regs.SetRegisterToFloat(instr.gpr0, 0, "cos(" + op_a + ')', 1, 1, @@ -929,6 +1018,9 @@ private: case OpCode::Id::FMNMX_C: case OpCode::Id::FMNMX_R: case OpCode::Id::FMNMX_IMM: { + op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a); + op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b); + std::string condition = GetPredicateCondition(instr.alu.fmnmx.pred, instr.alu.fmnmx.negate_pred != 0); std::string parameters = op_a + ',' + op_b; @@ -942,7 +1034,7 @@ private: case OpCode::Id::RRO_R: case OpCode::Id::RRO_IMM: { // Currently RRO is only implemented as a register move. - // Usage of `abs_b` and `negate_b` here should also be correct. + op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b); regs.SetRegisterToFloat(instr.gpr0, 0, op_b, 1, 1); LOG_WARNING(HW_GPU, "RRO instruction is incomplete"); break; @@ -1079,7 +1171,9 @@ private: if (instr.alu.lop32i.invert_b) op_b = "~(" + op_b + ')'; - WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b); + WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b, + Tegra::Shader::PredicateResultMode::None, + Tegra::Shader::Pred::UnusedIndex); break; } default: { @@ -1145,16 +1239,14 @@ private: case OpCode::Id::LOP_C: case OpCode::Id::LOP_R: case OpCode::Id::LOP_IMM: { - ASSERT_MSG(!instr.alu.lop.unk44, "Unimplemented"); - ASSERT_MSG(instr.alu.lop.pred48 == Pred::UnusedIndex, "Unimplemented"); - if (instr.alu.lop.invert_a) op_a = "~(" + op_a + ')'; if (instr.alu.lop.invert_b) op_b = "~(" + op_b + ')'; - WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b); + WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b, + instr.alu.lop.pred_result_mode, instr.alu.lop.pred48); break; } case OpCode::Id::IMNMX_C: @@ -1219,8 +1311,6 @@ private: break; } case OpCode::Type::Conversion: { - ASSERT_MSG(!instr.conversion.negate_a, "Unimplemented"); - switch (opcode->GetId()) { case OpCode::Id::I2I_R: { ASSERT_MSG(!instr.conversion.selector, "Unimplemented"); @@ -1232,20 +1322,41 @@ private: op_a = "abs(" + op_a + ')'; } + if (instr.conversion.negate_a) { + op_a = "-(" + op_a + ')'; + } + regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_output_signed, 0, op_a, 1, 1, instr.alu.saturate_d, 0, instr.conversion.dest_size); break; } - case OpCode::Id::I2F_R: { + case OpCode::Id::I2F_R: + case OpCode::Id::I2F_C: { ASSERT_MSG(instr.conversion.dest_size == Register::Size::Word, "Unimplemented"); ASSERT_MSG(!instr.conversion.selector, "Unimplemented"); - std::string op_a = regs.GetRegisterAsInteger( - instr.gpr20, 0, instr.conversion.is_input_signed, instr.conversion.src_size); + + std::string op_a{}; + + if (instr.is_b_gpr) { + op_a = + regs.GetRegisterAsInteger(instr.gpr20, 0, instr.conversion.is_input_signed, + instr.conversion.src_size); + } else { + op_a = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, + instr.conversion.is_input_signed + ? GLSLRegister::Type::Integer + : GLSLRegister::Type::UnsignedInteger, + instr.conversion.src_size); + } if (instr.conversion.abs_a) { op_a = "abs(" + op_a + ')'; } + if (instr.conversion.negate_a) { + op_a = "-(" + op_a + ')'; + } + regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1); break; } @@ -1254,6 +1365,14 @@ private: ASSERT_MSG(instr.conversion.src_size == Register::Size::Word, "Unimplemented"); std::string op_a = regs.GetRegisterAsFloat(instr.gpr20); + if (instr.conversion.abs_a) { + op_a = "abs(" + op_a + ')'; + } + + if (instr.conversion.negate_a) { + op_a = "-(" + op_a + ')'; + } + switch (instr.conversion.f2f.rounding) { case Tegra::Shader::F2fRoundingOp::None: break; @@ -1276,21 +1395,29 @@ private: break; } - if (instr.conversion.abs_a) { - op_a = "abs(" + op_a + ')'; - } - regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1, instr.alu.saturate_d); break; } - case OpCode::Id::F2I_R: { + case OpCode::Id::F2I_R: + case OpCode::Id::F2I_C: { ASSERT_MSG(instr.conversion.src_size == Register::Size::Word, "Unimplemented"); - std::string op_a = regs.GetRegisterAsFloat(instr.gpr20); + std::string op_a{}; + + if (instr.is_b_gpr) { + op_a = regs.GetRegisterAsFloat(instr.gpr20); + } else { + op_a = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, + GLSLRegister::Type::Float); + } if (instr.conversion.abs_a) { op_a = "abs(" + op_a + ')'; } + if (instr.conversion.negate_a) { + op_a = "-(" + op_a + ')'; + } + switch (instr.conversion.f2i.rounding) { case Tegra::Shader::F2iRoundingOp::None: break; @@ -1338,11 +1465,16 @@ private: case OpCode::Id::LD_C: { ASSERT_MSG(instr.ld_c.unknown == 0, "Unimplemented"); + // Add an extra scope and declare the index register inside to prevent + // overwriting it in case it is used as an output of the LD instruction. + shader.AddLine("{"); + ++shader.scope; + + shader.AddLine("uint index = (" + regs.GetRegisterAsInteger(instr.gpr8, 0, false) + + " / 4) & (MAX_CONSTBUFFER_ELEMENTS - 1);"); + std::string op_a = - regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 0, instr.gpr8, - GLSLRegister::Type::Float); - std::string op_b = - regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 4, instr.gpr8, + regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 0, "index", GLSLRegister::Type::Float); switch (instr.ld_c.type.Value()) { @@ -1350,16 +1482,22 @@ private: regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1); break; - case Tegra::Shader::UniformType::Double: + case Tegra::Shader::UniformType::Double: { + std::string op_b = + regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 4, + "index", GLSLRegister::Type::Float); regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1); regs.SetRegisterToFloat(instr.gpr0.Value() + 1, 0, op_b, 1, 1); break; - + } default: LOG_CRITICAL(HW_GPU, "Unhandled type: {}", static_cast<unsigned>(instr.ld_c.type.Value())); UNREACHABLE(); } + + --shader.scope; + shader.AddLine("}"); break; } case OpCode::Id::ST_A: { @@ -1615,6 +1753,99 @@ private: } break; } + case OpCode::Type::Xmad: { + ASSERT_MSG(!instr.xmad.sign_a, "Unimplemented"); + ASSERT_MSG(!instr.xmad.sign_b, "Unimplemented"); + + std::string op_a{regs.GetRegisterAsInteger(instr.gpr8, 0, instr.xmad.sign_a)}; + std::string op_b; + std::string op_c; + + // TODO(bunnei): Needs to be fixed once op_a or op_b is signed + ASSERT_MSG(instr.xmad.sign_a == instr.xmad.sign_b, "Unimplemented"); + const bool is_signed{instr.xmad.sign_a == 1}; + + bool is_merge{}; + switch (opcode->GetId()) { + case OpCode::Id::XMAD_CR: { + is_merge = instr.xmad.merge_56; + op_b += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, + instr.xmad.sign_b ? GLSLRegister::Type::Integer + : GLSLRegister::Type::UnsignedInteger); + op_c += regs.GetRegisterAsInteger(instr.gpr39, 0, is_signed); + break; + } + case OpCode::Id::XMAD_RR: { + is_merge = instr.xmad.merge_37; + op_b += regs.GetRegisterAsInteger(instr.gpr20, 0, instr.xmad.sign_b); + op_c += regs.GetRegisterAsInteger(instr.gpr39, 0, is_signed); + break; + } + case OpCode::Id::XMAD_RC: { + op_b += regs.GetRegisterAsInteger(instr.gpr39, 0, instr.xmad.sign_b); + op_c += regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset, + is_signed ? GLSLRegister::Type::Integer + : GLSLRegister::Type::UnsignedInteger); + break; + } + case OpCode::Id::XMAD_IMM: { + is_merge = instr.xmad.merge_37; + op_b += std::to_string(instr.xmad.imm20_16); + op_c += regs.GetRegisterAsInteger(instr.gpr39, 0, is_signed); + break; + } + default: { + LOG_CRITICAL(HW_GPU, "Unhandled XMAD instruction: {}", opcode->GetName()); + UNREACHABLE(); + } + } + + // TODO(bunnei): Ensure this is right with signed operands + if (instr.xmad.high_a) { + op_a = "((" + op_a + ") >> 16)"; + } else { + op_a = "((" + op_a + ") & 0xFFFF)"; + } + + std::string src2 = '(' + op_b + ')'; // Preserve original source 2 + if (instr.xmad.high_b) { + op_b = '(' + src2 + " >> 16)"; + } else { + op_b = '(' + src2 + " & 0xFFFF)"; + } + + std::string product = '(' + op_a + " * " + op_b + ')'; + if (instr.xmad.product_shift_left) { + product = '(' + product + " << 16)"; + } + + switch (instr.xmad.mode) { + case Tegra::Shader::XmadMode::None: + break; + case Tegra::Shader::XmadMode::CLo: + op_c = "((" + op_c + ") & 0xFFFF)"; + break; + case Tegra::Shader::XmadMode::CHi: + op_c = "((" + op_c + ") >> 16)"; + break; + case Tegra::Shader::XmadMode::CBcc: + op_c = "((" + op_c + ") + (" + src2 + "<< 16))"; + break; + default: { + LOG_CRITICAL(HW_GPU, "Unhandled XMAD mode: {}", + static_cast<u32>(instr.xmad.mode.Value())); + UNREACHABLE(); + } + } + + std::string sum{'(' + product + " + " + op_c + ')'}; + if (is_merge) { + sum = "((" + sum + " & 0xFFFF) | (" + src2 + "<< 16))"; + } + + regs.SetRegisterToInteger(instr.gpr0, is_signed, 0, sum, 1, 1); + break; + } default: { switch (opcode->GetId()) { case OpCode::Id::EXIT: { @@ -1652,7 +1883,15 @@ private: } case OpCode::Id::KIL: { ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always); + + // Enclose "discard" in a conditional, so that GLSL compilation does not complain + // about unexecuted instructions that may follow this. + shader.AddLine("if (true) {"); + ++shader.scope; shader.AddLine("discard;"); + --shader.scope; + shader.AddLine("}"); + break; } case OpCode::Id::BRA: { @@ -1668,16 +1907,25 @@ private: break; } case OpCode::Id::SSY: { - // The SSY opcode tells the GPU where to re-converge divergent execution paths, we - // can ignore this when generating GLSL code. + // The SSY opcode tells the GPU where to re-converge divergent execution paths, it + // sets the target of the jump that the SYNC instruction will make. The SSY opcode + // has a similar structure to the BRA opcode. + ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported"); + + u32 target = offset + instr.bra.GetBranchTarget(); + EmitPushToSSYStack(target); break; } - case OpCode::Id::SYNC: + case OpCode::Id::SYNC: { + // The SYNC opcode jumps to the address previously set by the SSY opcode ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always); + EmitPopFromSSYStack(); + break; + } case OpCode::Id::DEPBAR: { - // TODO(Subv): Find out if we actually have to care about these instructions or if + // TODO(Subv): Find out if we actually have to care about this instruction or if // the GLSL compiler takes care of that for us. - LOG_WARNING(HW_GPU, "DEPBAR/SYNC instruction is stubbed"); + LOG_WARNING(HW_GPU, "DEPBAR instruction is stubbed"); break; } default: { @@ -1691,7 +1939,7 @@ private: } // Close the predicate condition scope. - if (instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) { + if (can_be_predicated && instr.pred.pred_index != static_cast<u64>(Pred::UnusedIndex)) { --shader.scope; shader.AddLine('}'); } @@ -1742,6 +1990,13 @@ private: } else { labels.insert(subroutine.begin); shader.AddLine("uint jmp_to = " + std::to_string(subroutine.begin) + "u;"); + + // TODO(Subv): Figure out the actual depth of the SSY stack, for now it seems + // unlikely that shaders will use 20 nested SSYs. + constexpr u32 SSY_STACK_SIZE = 20; + shader.AddLine("uint ssy_stack[" + std::to_string(SSY_STACK_SIZE) + "];"); + shader.AddLine("uint ssy_stack_top = 0u;"); + shader.AddLine("while (true) {"); ++shader.scope; @@ -1757,7 +2012,7 @@ private: u32 compile_end = CompileRange(label, next_label); if (compile_end > next_label && compile_end != PROGRAM_END) { // This happens only when there is a label inside a IF/LOOP block - shader.AddLine("{ jmp_to = " + std::to_string(compile_end) + "u; break; }"); + shader.AddLine(" jmp_to = " + std::to_string(compile_end) + "u; break; }"); labels.emplace(compile_end); } diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index 129c777d1..57e0e1726 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -38,6 +38,7 @@ out vec4 position; layout (std140) uniform vs_config { vec4 viewport_flip; + uvec4 instance_id; }; void main() { @@ -90,6 +91,7 @@ out vec4 color; layout (std140) uniform fs_config { vec4 viewport_flip; + uvec4 instance_id; }; void main() { diff --git a/src/video_core/renderer_opengl/gl_shader_manager.cpp b/src/video_core/renderer_opengl/gl_shader_manager.cpp index 415d42fda..f0886caac 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.cpp +++ b/src/video_core/renderer_opengl/gl_shader_manager.cpp @@ -37,11 +37,16 @@ void SetShaderUniformBlockBindings(GLuint shader) { } // namespace Impl void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage) { - const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; + const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D(); + const auto& regs = gpu.regs; + const auto& state = gpu.state; // TODO(bunnei): Support more than one viewport viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0f : 1.0f; viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0f : 1.0f; + + // We only assign the instance to the first component of the vector, the rest is just padding. + instance_id[0] = state.current_instance; } } // namespace GLShader diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h index 716933a0b..75fa73605 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.h +++ b/src/video_core/renderer_opengl/gl_shader_manager.h @@ -24,14 +24,15 @@ void SetShaderUniformBlockBindings(GLuint shader); } // namespace Impl /// 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 wether the alignment at +// NOTE: Always keep a vec4 at the end. The GL spec is not clear whether the alignment at // the end of a uniform block is included in UNIFORM_BLOCK_DATA_SIZE or not. // Not following that rule will cause problems on some AMD drivers. struct MaxwellUniformData { void SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage); alignas(16) GLvec4 viewport_flip; + alignas(16) GLuvec4 instance_id; }; -static_assert(sizeof(MaxwellUniformData) == 16, "MaxwellUniformData structure size is incorrect"); +static_assert(sizeof(MaxwellUniformData) == 32, "MaxwellUniformData structure size is incorrect"); static_assert(sizeof(MaxwellUniformData) < 16384, "MaxwellUniformData structure must be less than 16kb as per the OpenGL spec"); diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp index 68bacd4c5..1d1975179 100644 --- a/src/video_core/renderer_opengl/gl_state.cpp +++ b/src/video_core/renderer_opengl/gl_state.cpp @@ -203,21 +203,6 @@ void OpenGLState::Apply() const { } } - // Constbuffers - for (std::size_t stage = 0; stage < draw.const_buffers.size(); ++stage) { - for (std::size_t buffer_id = 0; buffer_id < draw.const_buffers[stage].size(); ++buffer_id) { - const auto& current = cur_state.draw.const_buffers[stage][buffer_id]; - const auto& new_state = draw.const_buffers[stage][buffer_id]; - - if (current.enabled != new_state.enabled || current.bindpoint != new_state.bindpoint || - current.ssbo != new_state.ssbo) { - if (new_state.enabled) { - glBindBufferBase(GL_UNIFORM_BUFFER, new_state.bindpoint, new_state.ssbo); - } - } - } - } - // Framebuffer if (draw.read_framebuffer != cur_state.draw.read_framebuffer) { glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer); diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h index 5c7b636e4..bdb02ba25 100644 --- a/src/video_core/renderer_opengl/gl_state.h +++ b/src/video_core/renderer_opengl/gl_state.h @@ -119,12 +119,6 @@ public: GLuint uniform_buffer; // GL_UNIFORM_BUFFER_BINDING GLuint shader_program; // GL_CURRENT_PROGRAM GLuint program_pipeline; // GL_PROGRAM_PIPELINE_BINDING - struct ConstBufferConfig { - bool enabled = false; - GLuint bindpoint; - GLuint ssbo; - }; - std::array<std::array<ConstBufferConfig, Regs::MaxConstBuffers>, 5> const_buffers; } draw; struct { diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp index a2713e9f0..03a8ed8b7 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp +++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp @@ -9,174 +9,91 @@ #include "video_core/renderer_opengl/gl_state.h" #include "video_core/renderer_opengl/gl_stream_buffer.h" -class OrphanBuffer : public OGLStreamBuffer { -public: - explicit OrphanBuffer(GLenum target) : OGLStreamBuffer(target) {} - ~OrphanBuffer() override; - -private: - void Create(size_t size, size_t sync_subdivide) override; - void Release() override; - - std::pair<u8*, GLintptr> Map(size_t size, size_t alignment) override; - void Unmap() override; - - std::vector<u8> data; -}; - -class StorageBuffer : public OGLStreamBuffer { -public: - explicit StorageBuffer(GLenum target) : OGLStreamBuffer(target) {} - ~StorageBuffer() override; - -private: - void Create(size_t size, size_t sync_subdivide) override; - void Release() override; - - std::pair<u8*, GLintptr> Map(size_t size, size_t alignment) override; - void Unmap() override; - - struct Fence { - OGLSync sync; - size_t offset; - }; - std::deque<Fence> head; - std::deque<Fence> tail; - - u8* mapped_ptr; -}; - -OGLStreamBuffer::OGLStreamBuffer(GLenum target) { - gl_target = target; -} - -GLuint OGLStreamBuffer::GetHandle() const { - return gl_buffer.handle; -} +OGLStreamBuffer::OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coherent) + : gl_target(target), buffer_size(size) { + gl_buffer.Create(); + glBindBuffer(gl_target, gl_buffer.handle); -std::unique_ptr<OGLStreamBuffer> OGLStreamBuffer::MakeBuffer(bool storage_buffer, GLenum target) { - if (storage_buffer) { - return std::make_unique<StorageBuffer>(target); + GLsizeiptr allocate_size = size; + if (target == GL_ARRAY_BUFFER) { + // On AMD GPU there is a strange crash in indexed drawing. The crash happens when the buffer + // read position is near the end and is an out-of-bound access to the vertex buffer. This is + // probably a bug in the driver and is related to the usage of vec3<byte> attributes in the + // vertex array. Doubling the allocation size for the vertex buffer seems to avoid the + // crash. + allocate_size *= 2; } - return std::make_unique<OrphanBuffer>(target); -} -OrphanBuffer::~OrphanBuffer() { - Release(); + if (GLAD_GL_ARB_buffer_storage) { + persistent = true; + coherent = prefer_coherent; + GLbitfield flags = + GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | (coherent ? GL_MAP_COHERENT_BIT : 0); + glBufferStorage(gl_target, allocate_size, nullptr, flags); + mapped_ptr = static_cast<u8*>(glMapBufferRange( + gl_target, 0, buffer_size, flags | (coherent ? 0 : GL_MAP_FLUSH_EXPLICIT_BIT))); + } else { + glBufferData(gl_target, allocate_size, nullptr, GL_STREAM_DRAW); + } } -void OrphanBuffer::Create(size_t size, size_t /*sync_subdivide*/) { - buffer_pos = 0; - buffer_size = size; - data.resize(buffer_size); - - if (gl_buffer.handle == 0) { - gl_buffer.Create(); +OGLStreamBuffer::~OGLStreamBuffer() { + if (persistent) { glBindBuffer(gl_target, gl_buffer.handle); + glUnmapBuffer(gl_target); } - - glBufferData(gl_target, static_cast<GLsizeiptr>(buffer_size), nullptr, GL_STREAM_DRAW); -} - -void OrphanBuffer::Release() { gl_buffer.Release(); } -std::pair<u8*, GLintptr> OrphanBuffer::Map(size_t size, size_t alignment) { - buffer_pos = Common::AlignUp(buffer_pos, alignment); - - if (buffer_pos + size > buffer_size) { - Create(std::max(buffer_size, size), 0); - } - - mapped_size = size; - return std::make_pair(&data[buffer_pos], static_cast<GLintptr>(buffer_pos)); -} - -void OrphanBuffer::Unmap() { - glBufferSubData(gl_target, static_cast<GLintptr>(buffer_pos), - static_cast<GLsizeiptr>(mapped_size), &data[buffer_pos]); - buffer_pos += mapped_size; -} - -StorageBuffer::~StorageBuffer() { - Release(); +GLuint OGLStreamBuffer::GetHandle() const { + return gl_buffer.handle; } -void StorageBuffer::Create(size_t size, size_t sync_subdivide) { - if (gl_buffer.handle != 0) - return; - - buffer_pos = 0; - buffer_size = size; - buffer_sync_subdivide = std::max<size_t>(sync_subdivide, 1); - - gl_buffer.Create(); - glBindBuffer(gl_target, gl_buffer.handle); - - glBufferStorage(gl_target, static_cast<GLsizeiptr>(buffer_size), nullptr, - GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT); - mapped_ptr = reinterpret_cast<u8*>( - glMapBufferRange(gl_target, 0, static_cast<GLsizeiptr>(buffer_size), - GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | GL_MAP_FLUSH_EXPLICIT_BIT)); +GLsizeiptr OGLStreamBuffer::GetSize() const { + return buffer_size; } -void StorageBuffer::Release() { - if (gl_buffer.handle == 0) - return; - - glUnmapBuffer(gl_target); - - gl_buffer.Release(); - head.clear(); - tail.clear(); -} - -std::pair<u8*, GLintptr> StorageBuffer::Map(size_t size, size_t alignment) { +std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr alignment) { ASSERT(size <= buffer_size); + ASSERT(alignment <= buffer_size); + mapped_size = size; - OGLSync sync; - - buffer_pos = Common::AlignUp(buffer_pos, alignment); - size_t effective_offset = Common::AlignDown(buffer_pos, buffer_sync_subdivide); - - if (!head.empty() && - (effective_offset > head.back().offset || buffer_pos + size > buffer_size)) { - ASSERT(head.back().sync.handle == 0); - head.back().sync.Create(); + if (alignment > 0) { + buffer_pos = Common::AlignUp<size_t>(buffer_pos, alignment); } + bool invalidate = false; if (buffer_pos + size > buffer_size) { - if (!tail.empty()) { - std::swap(sync, tail.back().sync); - tail.clear(); - } - std::swap(tail, head); buffer_pos = 0; - effective_offset = 0; - } + invalidate = true; - while (!tail.empty() && buffer_pos + size > tail.front().offset) { - std::swap(sync, tail.front().sync); - tail.pop_front(); + if (persistent) { + glUnmapBuffer(gl_target); + } } - if (sync.handle != 0) { - glClientWaitSync(sync.handle, GL_SYNC_FLUSH_COMMANDS_BIT, GL_TIMEOUT_IGNORED); - sync.Release(); + if (invalidate | !persistent) { + 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*>( + glMapBufferRange(gl_target, buffer_pos, buffer_size - buffer_pos, flags)); + mapped_offset = buffer_pos; } - if (head.empty() || effective_offset > head.back().offset) { - head.emplace_back(); - head.back().offset = effective_offset; + return std::make_tuple(mapped_ptr + buffer_pos - mapped_offset, buffer_pos, invalidate); +} + +void OGLStreamBuffer::Unmap(GLsizeiptr size) { + ASSERT(size <= mapped_size); + + if (!coherent && size > 0) { + glFlushMappedBufferRange(gl_target, buffer_pos - mapped_offset, size); } - mapped_size = size; - return std::make_pair(&mapped_ptr[buffer_pos], static_cast<GLintptr>(buffer_pos)); -} + if (!persistent) { + glUnmapBuffer(gl_target); + } -void StorageBuffer::Unmap() { - glFlushMappedBufferRange(gl_target, static_cast<GLintptr>(buffer_pos), - static_cast<GLsizeiptr>(mapped_size)); - buffer_pos += mapped_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 e78dc5784..45592daaf 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.h +++ b/src/video_core/renderer_opengl/gl_stream_buffer.h @@ -2,35 +2,41 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#pragma once - -#include <memory> +#include <tuple> #include <glad/glad.h> #include "common/common_types.h" #include "video_core/renderer_opengl/gl_resource_manager.h" class OGLStreamBuffer : private NonCopyable { public: - explicit OGLStreamBuffer(GLenum target); - virtual ~OGLStreamBuffer() = default; - -public: - static std::unique_ptr<OGLStreamBuffer> MakeBuffer(bool storage_buffer, GLenum target); - - virtual void Create(size_t size, size_t sync_subdivide) = 0; - virtual void Release() {} + explicit OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coherent = false); + ~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. + * If the buffer is full, the whole buffer is reallocated which invalidates old chunks. + * The return values are the pointer to the new chunk, the offset within the buffer, + * and the invalidation flag for previous chunks. + * The actual used size must be specified on unmapping the chunk. + */ + std::tuple<u8*, GLintptr, bool> Map(GLsizeiptr size, GLintptr alignment = 0); - virtual std::pair<u8*, GLintptr> Map(size_t size, size_t alignment) = 0; - virtual void Unmap() = 0; + void Unmap(GLsizeiptr size); -protected: +private: OGLBuffer gl_buffer; GLenum gl_target; - size_t buffer_pos = 0; - size_t buffer_size = 0; - size_t buffer_sync_subdivide = 0; - size_t mapped_size = 0; + bool coherent = false; + bool persistent = false; + + 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/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index c439446b1..5d91a0c2f 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -24,15 +24,25 @@ using Maxwell = Tegra::Engines::Maxwell3D::Regs; inline GLenum VertexType(Maxwell::VertexAttribute attrib) { switch (attrib.type) { + case Maxwell::VertexAttribute::Type::UnsignedInt: case Maxwell::VertexAttribute::Type::UnsignedNorm: { 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; + case Maxwell::VertexAttribute::Size::Size_32: + case Maxwell::VertexAttribute::Size::Size_32_32: + case Maxwell::VertexAttribute::Size::Size_32_32_32: + case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + return GL_UNSIGNED_INT; case Maxwell::VertexAttribute::Size::Size_10_10_10_2: return GL_UNSIGNED_INT_2_10_10_10_REV; } @@ -42,16 +52,25 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) { return {}; } + case Maxwell::VertexAttribute::Type::SignedInt: case Maxwell::VertexAttribute::Type::SignedNorm: { switch (attrib.size) { - case Maxwell::VertexAttribute::Size::Size_32_32_32: - return GL_INT; + 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; + case Maxwell::VertexAttribute::Size::Size_32: + case Maxwell::VertexAttribute::Size::Size_32_32: + case Maxwell::VertexAttribute::Size::Size_32_32_32: + case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + return GL_INT; case Maxwell::VertexAttribute::Size::Size_10_10_10_2: return GL_INT_2_10_10_10_REV; } @@ -61,9 +80,6 @@ inline GLenum VertexType(Maxwell::VertexAttribute attrib) { return {}; } - case Maxwell::VertexAttribute::Type::UnsignedInt: - return GL_UNSIGNED_INT; - case Maxwell::VertexAttribute::Type::Float: return GL_FLOAT; } @@ -91,6 +107,8 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) { switch (topology) { case Maxwell::PrimitiveTopology::Points: return GL_POINTS; + case Maxwell::PrimitiveTopology::LineStrip: + return GL_LINE_STRIP; case Maxwell::PrimitiveTopology::Triangles: return GL_TRIANGLES; case Maxwell::PrimitiveTopology::TriangleStrip: @@ -129,6 +147,8 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) { // GL_CLAMP_TO_BORDER to get the border color of the texture, and then sample the edge to // manually mix them. However the shader part of this is not yet implemented. return GL_CLAMP_TO_BORDER; + case Tegra::Texture::WrapMode::MirrorOnceClampToEdge: + return GL_MIRROR_CLAMP_TO_EDGE; } LOG_CRITICAL(Render_OpenGL, "Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode)); UNREACHABLE(); @@ -156,42 +176,61 @@ inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { switch (factor) { case Maxwell::Blend::Factor::Zero: + case Maxwell::Blend::Factor::ZeroGL: return GL_ZERO; case Maxwell::Blend::Factor::One: + case Maxwell::Blend::Factor::OneGL: return GL_ONE; case Maxwell::Blend::Factor::SourceColor: + case Maxwell::Blend::Factor::SourceColorGL: return GL_SRC_COLOR; case Maxwell::Blend::Factor::OneMinusSourceColor: + case Maxwell::Blend::Factor::OneMinusSourceColorGL: return GL_ONE_MINUS_SRC_COLOR; case Maxwell::Blend::Factor::SourceAlpha: + case Maxwell::Blend::Factor::SourceAlphaGL: return GL_SRC_ALPHA; case Maxwell::Blend::Factor::OneMinusSourceAlpha: + case Maxwell::Blend::Factor::OneMinusSourceAlphaGL: return GL_ONE_MINUS_SRC_ALPHA; case Maxwell::Blend::Factor::DestAlpha: + case Maxwell::Blend::Factor::DestAlphaGL: return GL_DST_ALPHA; case Maxwell::Blend::Factor::OneMinusDestAlpha: + case Maxwell::Blend::Factor::OneMinusDestAlphaGL: return GL_ONE_MINUS_DST_ALPHA; case Maxwell::Blend::Factor::DestColor: + case Maxwell::Blend::Factor::DestColorGL: return GL_DST_COLOR; case Maxwell::Blend::Factor::OneMinusDestColor: + case Maxwell::Blend::Factor::OneMinusDestColorGL: return GL_ONE_MINUS_DST_COLOR; case Maxwell::Blend::Factor::SourceAlphaSaturate: + case Maxwell::Blend::Factor::SourceAlphaSaturateGL: return GL_SRC_ALPHA_SATURATE; case Maxwell::Blend::Factor::Source1Color: + case Maxwell::Blend::Factor::Source1ColorGL: return GL_SRC1_COLOR; case Maxwell::Blend::Factor::OneMinusSource1Color: + case Maxwell::Blend::Factor::OneMinusSource1ColorGL: return GL_ONE_MINUS_SRC1_COLOR; case Maxwell::Blend::Factor::Source1Alpha: + case Maxwell::Blend::Factor::Source1AlphaGL: return GL_SRC1_ALPHA; case Maxwell::Blend::Factor::OneMinusSource1Alpha: + case Maxwell::Blend::Factor::OneMinusSource1AlphaGL: return GL_ONE_MINUS_SRC1_ALPHA; case Maxwell::Blend::Factor::ConstantColor: + case Maxwell::Blend::Factor::ConstantColorGL: return GL_CONSTANT_COLOR; case Maxwell::Blend::Factor::OneMinusConstantColor: + case Maxwell::Blend::Factor::OneMinusConstantColorGL: return GL_ONE_MINUS_CONSTANT_COLOR; case Maxwell::Blend::Factor::ConstantAlpha: + case Maxwell::Blend::Factor::ConstantAlphaGL: return GL_CONSTANT_ALPHA; case Maxwell::Blend::Factor::OneMinusConstantAlpha: + case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: return GL_ONE_MINUS_CONSTANT_ALPHA; } LOG_CRITICAL(Render_OpenGL, "Unimplemented blend factor={}", static_cast<u32>(factor)); diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index 899865e3b..bf30eda6d 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -18,7 +18,6 @@ #include "core/tracer/recorder.h" #include "video_core/renderer_opengl/renderer_opengl.h" #include "video_core/utils.h" -#include "video_core/video_core.h" static const char vertex_shader[] = R"( #version 150 core @@ -92,7 +91,8 @@ static std::array<GLfloat, 3 * 2> MakeOrthographicMatrix(const float width, cons return matrix; } -ScopeAcquireGLContext::ScopeAcquireGLContext(EmuWindow& emu_window_) : emu_window{emu_window_} { +ScopeAcquireGLContext::ScopeAcquireGLContext(Core::Frontend::EmuWindow& emu_window_) + : emu_window{emu_window_} { if (Settings::values.use_multi_core) { emu_window.MakeCurrent(); } @@ -103,7 +103,9 @@ ScopeAcquireGLContext::~ScopeAcquireGLContext() { } } -RendererOpenGL::RendererOpenGL(EmuWindow& window) : VideoCore::RendererBase{window} {} +RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& window) + : VideoCore::RendererBase{window} {} + RendererOpenGL::~RendererOpenGL() = default; /// Swap buffers (render frame) @@ -423,7 +425,7 @@ static void APIENTRY DebugHandler(GLenum source, GLenum type, GLuint id, GLenum switch (severity) { case GL_DEBUG_SEVERITY_HIGH: - LOG_ERROR(Render_OpenGL, format, str_source, str_type, id, message); + LOG_CRITICAL(Render_OpenGL, format, str_source, str_type, id, message); break; case GL_DEBUG_SEVERITY_MEDIUM: LOG_WARNING(Render_OpenGL, format, str_source, str_type, id, message); diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h index 428afa3b7..a5eab6997 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.h +++ b/src/video_core/renderer_opengl/renderer_opengl.h @@ -12,7 +12,9 @@ #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_state.h" +namespace Core::Frontend { class EmuWindow; +} /// Structure used for storing information about the textures for the Switch screen struct TextureInfo { @@ -34,16 +36,16 @@ struct ScreenInfo { /// Helper class to acquire/release OpenGL context within a given scope class ScopeAcquireGLContext : NonCopyable { public: - explicit ScopeAcquireGLContext(EmuWindow& window); + explicit ScopeAcquireGLContext(Core::Frontend::EmuWindow& window); ~ScopeAcquireGLContext(); private: - EmuWindow& emu_window; + Core::Frontend::EmuWindow& emu_window; }; class RendererOpenGL : public VideoCore::RendererBase { public: - explicit RendererOpenGL(EmuWindow& window); + explicit RendererOpenGL(Core::Frontend::EmuWindow& window); ~RendererOpenGL() override; /// Swap buffers (render frame) diff --git a/src/video_core/video_core.cpp b/src/video_core/video_core.cpp index 5085ef96b..6780d1c16 100644 --- a/src/video_core/video_core.cpp +++ b/src/video_core/video_core.cpp @@ -9,9 +9,7 @@ namespace VideoCore { -std::atomic<bool> g_toggle_framelimit_enabled; - -std::unique_ptr<RendererBase> CreateRenderer(EmuWindow& emu_window) { +std::unique_ptr<RendererBase> CreateRenderer(Core::Frontend::EmuWindow& emu_window) { return std::make_unique<RendererOpenGL>(emu_window); } diff --git a/src/video_core/video_core.h b/src/video_core/video_core.h index 7c01c0b8d..f79f85dfe 100644 --- a/src/video_core/video_core.h +++ b/src/video_core/video_core.h @@ -4,27 +4,22 @@ #pragma once -#include <atomic> #include <memory> +namespace Core::Frontend { class EmuWindow; +} namespace VideoCore { class RendererBase; -enum class Renderer { Software, OpenGL }; - -// TODO: Wrap these in a user settings struct along with any other graphics settings (often set from -// qt ui) -extern std::atomic<bool> g_toggle_framelimit_enabled; - /** * Creates a renderer instance. * * @note The returned renderer instance is simply allocated. Its Init() * function still needs to be called to fully complete its setup. */ -std::unique_ptr<RendererBase> CreateRenderer(EmuWindow& emu_window); +std::unique_ptr<RendererBase> CreateRenderer(Core::Frontend::EmuWindow& emu_window); } // namespace VideoCore diff --git a/src/yuzu/bootmanager.h b/src/yuzu/bootmanager.h index d0f990c64..f133bfadf 100644 --- a/src/yuzu/bootmanager.h +++ b/src/yuzu/bootmanager.h @@ -101,7 +101,7 @@ signals: void ErrorThrown(Core::System::ResultStatus, std::string); }; -class GRenderWindow : public QWidget, public EmuWindow { +class GRenderWindow : public QWidget, public Core::Frontend::EmuWindow { Q_OBJECT public: diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp index 1c738d2a4..bc4b93033 100644 --- a/src/yuzu/game_list.cpp +++ b/src/yuzu/game_list.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <regex> #include <QApplication> #include <QDir> #include <QFileInfo> @@ -402,12 +403,72 @@ void GameList::RefreshGameDirectory() { } } -void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsigned int recursion) { - boost::container::flat_map<u64, std::shared_ptr<FileSys::NCA>> nca_control_map; +static void GetMetadataFromControlNCA(const std::shared_ptr<FileSys::NCA>& nca, + std::vector<u8>& icon, std::string& name) { + const auto control_dir = FileSys::ExtractRomFS(nca->GetRomFS()); + if (control_dir == nullptr) + return; + + const auto nacp_file = control_dir->GetFile("control.nacp"); + if (nacp_file == nullptr) + return; + FileSys::NACP nacp(nacp_file); + name = nacp.GetApplicationName(); + + FileSys::VirtualFile icon_file = nullptr; + for (const auto& language : FileSys::LANGUAGE_NAMES) { + icon_file = control_dir->GetFile("icon_" + std::string(language) + ".dat"); + if (icon_file != nullptr) { + icon = icon_file->ReadAllBytes(); + break; + } + } +} + +void GameListWorker::AddInstalledTitlesToGameList() { + const auto usernand = Service::FileSystem::GetUserNANDContents(); + const auto installed_games = usernand->ListEntriesFilter(FileSys::TitleType::Application, + FileSys::ContentRecordType::Program); + + for (const auto& game : installed_games) { + const auto& file = usernand->GetEntryRaw(game); + std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(file); + if (!loader) + continue; + + std::vector<u8> icon; + std::string name; + u64 program_id = 0; + loader->ReadProgramId(program_id); + + const auto& control = + usernand->GetEntry(game.title_id, FileSys::ContentRecordType::Control); + if (control != nullptr) + GetMetadataFromControlNCA(control, icon, name); + emit EntryReady({ + new GameListItemPath( + FormatGameName(file->GetFullPath()), icon, QString::fromStdString(name), + QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType())), + program_id), + new GameListItem( + QString::fromStdString(Loader::GetFileTypeString(loader->GetFileType()))), + new GameListItemSize(file->GetSize()), + }); + } + + const auto control_data = usernand->ListEntriesFilter(FileSys::TitleType::Application, + FileSys::ContentRecordType::Control); + + for (const auto& entry : control_data) { + const auto nca = usernand->GetEntry(entry); + if (nca != nullptr) + nca_control_map.insert_or_assign(entry.title_id, nca); + } +} - const auto nca_control_callback = - [this, &nca_control_map](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { +void GameListWorker::FillControlMap(const std::string& dir_path) { + const auto nca_control_callback = [this](u64* num_entries_out, const std::string& directory, + const std::string& virtual_name) -> bool { std::string physical_name = directory + DIR_SEP + virtual_name; if (stop_processing) @@ -425,10 +486,11 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign }; FileUtil::ForeachDirectoryEntry(nullptr, dir_path, nca_control_callback); +} - const auto callback = [this, recursion, - &nca_control_map](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { +void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsigned int recursion) { + const auto callback = [this, recursion](u64* num_entries_out, const std::string& directory, + const std::string& virtual_name) -> bool { std::string physical_name = directory + DIR_SEP + virtual_name; if (stop_processing) @@ -447,34 +509,18 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign std::vector<u8> icon; const auto res1 = loader->ReadIcon(icon); - u64 program_id; + u64 program_id = 0; const auto res2 = loader->ReadProgramId(program_id); std::string name = " "; const auto res3 = loader->ReadTitle(name); - if ((res1 == Loader::ResultStatus::ErrorNotUsed || - res1 == Loader::ResultStatus::ErrorNotImplemented) && - (res3 == Loader::ResultStatus::ErrorNotUsed || - res3 == Loader::ResultStatus::ErrorNotImplemented) && + if (res1 != Loader::ResultStatus::Success && res3 != Loader::ResultStatus::Success && res2 == Loader::ResultStatus::Success) { // Use from metadata pool. if (nca_control_map.find(program_id) != nca_control_map.end()) { const auto nca = nca_control_map[program_id]; - const auto control_dir = FileSys::ExtractRomFS(nca->GetRomFS()); - - const auto nacp_file = control_dir->GetFile("control.nacp"); - FileSys::NACP nacp(nacp_file); - name = nacp.GetApplicationName(); - - FileSys::VirtualFile icon_file = nullptr; - for (const auto& language : FileSys::LANGUAGE_NAMES) { - icon_file = control_dir->GetFile("icon_" + std::string(language) + ".dat"); - if (icon_file != nullptr) { - icon = icon_file->ReadAllBytes(); - break; - } - } + GetMetadataFromControlNCA(nca, icon, name); } } @@ -501,7 +547,10 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign void GameListWorker::run() { stop_processing = false; watch_list.append(dir_path); + FillControlMap(dir_path.toStdString()); + AddInstalledTitlesToGameList(); AddFstEntriesToGameList(dir_path.toStdString(), deep_scan ? 256 : 0); + nca_control_map.clear(); emit Finished(watch_list); } diff --git a/src/yuzu/game_list_p.h b/src/yuzu/game_list_p.h index 114a0fc7f..10c2ef075 100644 --- a/src/yuzu/game_list_p.h +++ b/src/yuzu/game_list_p.h @@ -4,6 +4,7 @@ #pragma once +#include <array> #include <atomic> #include <utility> #include <QImage> @@ -39,7 +40,6 @@ public: * If this class receives valid SMDH data, it will also display game icons and titles. */ class GameListItemPath : public GameListItem { - public: static const int FullPathRole = Qt::UserRole + 1; static const int TitleRole = Qt::UserRole + 2; @@ -48,18 +48,18 @@ public: GameListItemPath() = default; GameListItemPath(const QString& game_path, const std::vector<u8>& picture_data, - const QString& game_name, const QString& game_type, u64 program_id) - : GameListItem() { + const QString& game_name, const QString& game_type, u64 program_id) { setData(game_path, FullPathRole); setData(game_name, TitleRole); setData(qulonglong(program_id), ProgramIdRole); setData(game_type, FileTypeRole); + const u32 size = UISettings::values.icon_size; + QPixmap picture; - u32 size = UISettings::values.icon_size; - if (!picture.loadFromData(picture_data.data(), picture_data.size())) + if (!picture.loadFromData(picture_data.data(), static_cast<u32>(picture_data.size()))) { picture = GetDefaultIcon(size); - + } picture = picture.scaled(size, size); setData(picture, Qt::DecorationRole); @@ -70,17 +70,16 @@ public: std::string filename; Common::SplitPath(data(FullPathRole).toString().toStdString(), nullptr, &filename, nullptr); - QString title = data(TitleRole).toString(); - std::vector<QString> row_data{ + const std::array<QString, 4> row_data{{ QString::fromStdString(filename), data(FileTypeRole).toString(), QString::fromStdString(fmt::format("0x{:016X}", data(ProgramIdRole).toULongLong())), data(TitleRole).toString(), - }; + }}; - auto row1 = row_data.at(UISettings::values.row_1_text_id); - auto row2 = row_data.at(UISettings::values.row_2_text_id); + const auto& row1 = row_data.at(UISettings::values.row_1_text_id); + const auto& row2 = row_data.at(UISettings::values.row_2_text_id); if (row1.isEmpty() || row1 == row2) return row2; @@ -88,9 +87,9 @@ public: return row1; return row1 + "\n " + row2; - } else { - return GameListItem::data(role); } + + return GameListItem::data(role); } }; @@ -164,10 +163,13 @@ signals: private: FileSys::VirtualFilesystem vfs; + std::map<u64, std::shared_ptr<FileSys::NCA>> nca_control_map; QStringList watch_list; QString dir_path; bool deep_scan; std::atomic_bool stop_processing; + void AddInstalledTitlesToGameList(); + void FillControlMap(const std::string& dir_path); void AddFstEntriesToGameList(const std::string& dir_path, unsigned int recursion = 0); }; diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index 67e3c6549..11d2331df 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -6,7 +6,10 @@ #include <clocale> #include <memory> #include <thread> + +#include <fmt/ostream.h> #include <glad/glad.h> + #define QT_NO_OPENGL #include <QDesktopWidget> #include <QFileDialog> @@ -24,6 +27,7 @@ #include "common/string_util.h" #include "core/core.h" #include "core/crypto/key_manager.h" +#include "core/file_sys/card_image.h" #include "core/file_sys/vfs_real.h" #include "core/gdbstub/gdbstub.h" #include "core/loader/loader.h" @@ -114,6 +118,9 @@ GMainWindow::GMainWindow() .arg(Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc)); show(); + // Necessary to load titles from nand in gamelist. + Service::FileSystem::RegisterBIS(std::make_unique<FileSys::BISFactory>(vfs->OpenDirectory( + FileUtil::GetUserPath(FileUtil::UserPath::NANDDir), FileSys::Mode::ReadWrite))); game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan); // Show one-time "callout" messages to the user @@ -309,6 +316,8 @@ void GMainWindow::ConnectMenuEvents() { // File connect(ui.action_Load_File, &QAction::triggered, this, &GMainWindow::OnMenuLoadFile); 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_Game_List_Root, &QAction::triggered, this, &GMainWindow::OnMenuSelectGameListRoot); connect(ui.action_Exit, &QAction::triggered, this, &QMainWindow::close); @@ -368,6 +377,8 @@ bool GMainWindow::SupportsRequiredGLExtensions() { unsupported_ext.append("ARB_vertex_attrib_binding"); if (!GLAD_GL_ARB_vertex_type_10f_11f_11f_rev) unsupported_ext.append("ARB_vertex_type_10f_11f_11f_rev"); + if (!GLAD_GL_ARB_texture_mirror_clamp_to_edge) + unsupported_ext.append("ARB_texture_mirror_clamp_to_edge"); // Extensions required to support some texture formats. if (!GLAD_GL_EXT_texture_compression_s3tc) @@ -424,67 +435,11 @@ bool GMainWindow::LoadROM(const QString& filename) { QMessageBox::critical(this, tr("Error while loading ROM!"), tr("The ROM format is not supported.")); break; - case Core::System::ResultStatus::ErrorUnsupportedArch: - LOG_CRITICAL(Frontend, "Unsupported architecture detected!", filename.toStdString()); - QMessageBox::critical(this, tr("Error while loading ROM!"), - tr("The ROM uses currently unusable 32-bit architecture")); - break; case Core::System::ResultStatus::ErrorSystemMode: LOG_CRITICAL(Frontend, "Failed to load ROM!"); QMessageBox::critical(this, tr("Error while loading ROM!"), tr("Could not determine the system mode.")); break; - - case Core::System::ResultStatus::ErrorLoader_ErrorMissingKeys: { - const auto reg_found = Core::Crypto::KeyManager::KeyFileExists(false); - const auto title_found = Core::Crypto::KeyManager::KeyFileExists(true); - - std::string file_text; - - if (!reg_found && !title_found) { - file_text = "A proper key file (prod.keys, dev.keys, or title.keys) could not be " - "found. You will need to dump your keys from your switch to continue."; - } else if (reg_found && title_found) { - file_text = - "Both key files were found in your config directory, but the correct key could" - "not be found. You may be missing a titlekey or general key, depending on " - "the game."; - } else if (reg_found) { - file_text = - "The regular keys file (prod.keys/dev.keys) was found in your config, but the " - "titlekeys file (title.keys) was not. You are either missing the correct " - "titlekey or missing a general key required to decrypt the game."; - } else { - file_text = "The title keys file (title.keys) was found in your config, but " - "the regular keys file (prod.keys/dev.keys) was not. Unfortunately, " - "having the titlekey is not enough, you need additional general keys " - "to properly decrypt the game. You should double-check to make sure " - "your keys are correct."; - } - - QMessageBox::critical( - this, tr("Error while loading ROM!"), - tr(("The game you are trying to load is encrypted and the required keys to load " - "the game could not be found in your configuration. " + - file_text + " Please refer to the yuzu wiki for help.") - .c_str())); - break; - } - case Core::System::ResultStatus::ErrorLoader_ErrorDecrypting: { - QMessageBox::critical( - this, tr("Error while loading ROM!"), - tr("There was a general error while decrypting the game. This means that the keys " - "necessary were found, but were either incorrect, the game itself was not a " - "valid game or the game uses an unhandled cryptographic scheme. Please double " - "check that you have the correct " - "keys.")); - break; - } - case Core::System::ResultStatus::ErrorLoader_ErrorInvalidFormat: - QMessageBox::critical(this, tr("Error while loading ROM!"), - tr("The ROM format is not supported.")); - break; - case Core::System::ResultStatus::ErrorVideoCore: QMessageBox::critical( this, tr("An error occurred initializing the video core."), @@ -499,9 +454,23 @@ bool GMainWindow::LoadROM(const QString& filename) { break; default: - QMessageBox::critical( - this, tr("Error while loading ROM!"), - tr("An unknown error occurred. Please see the log for more details.")); + 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; + QMessageBox::critical( + this, tr("Error while loading ROM!"), + 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)))); + } else { + QMessageBox::critical( + this, tr("Error while loading ROM!"), + tr("An unknown error occurred. Please see the log for more details.")); + } break; } return false; @@ -654,6 +623,148 @@ void GMainWindow::OnMenuLoadFolder() { } } +void GMainWindow::OnMenuInstallToNAND() { + const QString file_filter = + tr("Installable Switch File (*.nca *.xci);;Nintendo Content Archive (*.nca);;NX Cartridge " + "Image (*.xci)"); + QString filename = QFileDialog::getOpenFileName(this, tr("Install File"), + UISettings::values.roms_path, file_filter); + + if (filename.isEmpty()) { + return; + } + + const auto qt_raw_copy = [this](FileSys::VirtualFile src, FileSys::VirtualFile dest) { + if (src == nullptr || dest == nullptr) + return false; + 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 (size_t i = 0; i < src->GetSize(); i += buffer.size()) { + if (progress.wasCanceled()) { + dest->Resize(0); + return false; + } + + const int progress_value = static_cast<int>(i / buffer.size()); + progress.setValue(progress_value); + + 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.gamedir, UISettings::values.gamedir_deepscan); + }; + + 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("xci", Qt::CaseInsensitive)) { + const auto xci = std::make_shared<FileSys::XCI>( + vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); + if (xci->GetStatus() != Loader::ResultStatus::Success) { + failed(); + return; + } + const auto res = + Service::FileSystem::GetUserNANDContents()->InstallEntry(xci, false, qt_raw_copy); + if (res == FileSys::InstallResult::Success) { + success(); + } else { + if (res == FileSys::InstallResult::ErrorAlreadyExists) { + if (overwrite()) { + const auto res2 = Service::FileSystem::GetUserNANDContents()->InstallEntry( + xci, true, qt_raw_copy); + if (res2 == FileSys::InstallResult::Success) { + success(); + } else { + failed(); + } + } + } else { + failed(); + } + } + } else { + const auto nca = std::make_shared<FileSys::NCA>( + vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); + if (nca->GetStatus() != Loader::ResultStatus::Success) { + failed(); + return; + } + + 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; + } + + if (index >= 5) + index += 0x7B; + + const auto res = Service::FileSystem::GetUserNANDContents()->InstallEntry( + nca, static_cast<FileSys::TitleType>(index), false, qt_raw_copy); + if (res == FileSys::InstallResult::Success) { + success(); + } else if (res == FileSys::InstallResult::ErrorAlreadyExists) { + if (overwrite()) { + const auto res2 = Service::FileSystem::GetUserNANDContents()->InstallEntry( + nca, static_cast<FileSys::TitleType>(index), true, qt_raw_copy); + if (res2 == FileSys::InstallResult::Success) { + success(); + } else { + failed(); + } + } + } else { + failed(); + } + } +} + void GMainWindow::OnMenuSelectGameListRoot() { QString dir_path = QFileDialog::getExistingDirectory(this, tr("Select Directory")); if (!dir_path.isEmpty()) { diff --git a/src/yuzu/main.h b/src/yuzu/main.h index 74487c58c..5f4d2ab9a 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -125,6 +125,7 @@ private slots: void OnGameListOpenSaveFolder(u64 program_id); void OnMenuLoadFile(); void OnMenuLoadFolder(); + void OnMenuInstallToNAND(); /// Called whenever a user selects the "File->Select Game List Root" menu item void OnMenuSelectGameListRoot(); void OnMenuRecentFile(); diff --git a/src/yuzu/main.ui b/src/yuzu/main.ui index 22c4cad08..a3bfb2af3 100644 --- a/src/yuzu/main.ui +++ b/src/yuzu/main.ui @@ -57,6 +57,8 @@ <string>Recent Files</string> </property> </widget> + <addaction name="action_Install_File_NAND" /> + <addaction name="separator"/> <addaction name="action_Load_File"/> <addaction name="action_Load_Folder"/> <addaction name="separator"/> @@ -102,6 +104,11 @@ <addaction name="menu_View"/> <addaction name="menu_Help"/> </widget> + <action name="action_Install_File_NAND"> + <property name="text"> + <string>Install File to NAND...</string> + </property> + </action> <action name="action_Load_File"> <property name="text"> <string>Load File...</string> diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp index e2945b6cf..351dd9225 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp @@ -89,6 +89,8 @@ bool EmuWindow_SDL2::SupportsRequiredGLExtensions() { unsupported_ext.push_back("ARB_vertex_attrib_binding"); if (!GLAD_GL_ARB_vertex_type_10f_11f_11f_rev) unsupported_ext.push_back("ARB_vertex_type_10f_11f_11f_rev"); + if (!GLAD_GL_ARB_texture_mirror_clamp_to_edge) + unsupported_ext.push_back("ARB_texture_mirror_clamp_to_edge"); // Extensions required to support some texture formats. if (!GLAD_GL_EXT_texture_compression_s3tc) diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.h b/src/yuzu_cmd/emu_window/emu_window_sdl2.h index 1d835c3c6..d34902109 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.h +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.h @@ -10,7 +10,7 @@ struct SDL_Window; -class EmuWindow_SDL2 : public EmuWindow { +class EmuWindow_SDL2 : public Core::Frontend::EmuWindow { public: explicit EmuWindow_SDL2(bool fullscreen); ~EmuWindow_SDL2(); diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp index 0605c92e3..9095cf27d 100644 --- a/src/yuzu_cmd/yuzu.cpp +++ b/src/yuzu_cmd/yuzu.cpp @@ -7,6 +7,8 @@ #include <string> #include <thread> +#include <fmt/ostream.h> + #include "common/common_paths.h" #include "common/logging/backend.h" #include "common/logging/filter.h" @@ -174,19 +176,6 @@ int main(int argc, char** argv) { case Core::System::ResultStatus::ErrorLoader: LOG_CRITICAL(Frontend, "Failed to load ROM!"); return -1; - case Core::System::ResultStatus::ErrorLoader_ErrorMissingKeys: - LOG_CRITICAL(Frontend, "The game you are trying to load is encrypted and the keys required " - "could not be found. Please refer to the yuzu wiki for help"); - return -1; - case Core::System::ResultStatus::ErrorLoader_ErrorDecrypting: - LOG_CRITICAL(Frontend, "The game you are trying to load is encrypted and there was a " - "general error while decrypting. This could mean that the keys are " - "incorrect, game is invalid or game uses an unsupported method of " - "crypto. Please double-check your keys"); - return -1; - case Core::System::ResultStatus::ErrorLoader_ErrorInvalidFormat: - LOG_CRITICAL(Frontend, "Error while loading ROM: The ROM format is not supported."); - return -1; case Core::System::ResultStatus::ErrorNotInitialized: LOG_CRITICAL(Frontend, "CPUCore not initialized"); return -1; @@ -198,6 +187,17 @@ int main(int argc, char** argv) { return -1; case Core::System::ResultStatus::Success: break; // Expected case + default: + if (static_cast<u32>(load_result) > + static_cast<u32>(Core::System::ResultStatus::ErrorLoader)) { + const u16 loader_id = static_cast<u16>(Core::System::ResultStatus::ErrorLoader); + const u16 error_id = static_cast<u16>(load_result) - loader_id; + LOG_CRITICAL(Frontend, + "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)); + } } Core::Telemetry().AddField(Telemetry::FieldType::App, "Frontend", "SDL"); |