diff options
24 files changed, 316 insertions, 145 deletions
diff --git a/appveyor.yml b/appveyor.yml index a6f12b267..436d98fa1 100644 --- a/appveyor.yml +++ b/appveyor.yml @@ -53,7 +53,7 @@ build_script: # https://www.appveyor.com/docs/build-phase msbuild msvc_build/yuzu.sln /maxcpucount /logger:"C:\Program Files\AppVeyor\BuildAgent\Appveyor.MSBuildLogger.dll" } else { - C:\msys64\usr\bin\bash.exe -lc 'mingw32-make -C mingw_build/ 2>&1' + C:\msys64\usr\bin\bash.exe -lc 'mingw32-make -j4 -C mingw_build/ 2>&1' } after_build: diff --git a/externals/dynarmic b/externals/dynarmic -Subproject 0118ee04f90faaff951989f3c2494bc6ffb70cf +Subproject 550d662f0ece868209f0f760dddc8b2ad4698b1 diff --git a/src/common/hex_util.cpp b/src/common/hex_util.cpp index ae17c89d4..609144def 100644 --- a/src/common/hex_util.cpp +++ b/src/common/hex_util.cpp @@ -4,6 +4,8 @@ #include "common/hex_util.h" +namespace Common { + u8 ToHexNibble(char c1) { if (c1 >= 65 && c1 <= 70) return c1 - 55; @@ -25,3 +27,5 @@ std::array<u8, 32> operator""_array32(const char* str, size_t len) { 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 index 13d586015..5fb79bb72 100644 --- a/src/common/hex_util.h +++ b/src/common/hex_util.h @@ -10,6 +10,8 @@ #include <fmt/format.h> #include "common/common_types.h" +namespace Common { + u8 ToHexNibble(char c1); template <size_t Size, bool le = false> @@ -35,3 +37,5 @@ std::string HexArrayToString(std::array<u8, Size> array, bool upper = true) { 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/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp index 20e5200a8..2c817d7d1 100644 --- a/src/core/arm/dynarmic/arm_dynarmic.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic.cpp @@ -134,6 +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); } diff --git a/src/core/core.h b/src/core/core.h index d98b15a71..790e23cae 100644 --- a/src/core/core.h +++ b/src/core/core.h @@ -40,6 +40,12 @@ namespace Core { class System { public: + System(const System&) = delete; + System& operator=(const System&) = delete; + + System(System&&) = delete; + System& operator=(System&&) = delete; + ~System(); /** diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp index 94d92579f..db8b22c85 100644 --- a/src/core/crypto/key_manager.cpp +++ b/src/core/crypto/key_manager.cpp @@ -52,20 +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/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp index a5e935f64..d25eeee34 100644 --- a/src/core/file_sys/registered_cache.cpp +++ b/src/core/file_sys/registered_cache.cpp @@ -37,11 +37,12 @@ static bool FollowsNcaIdFormat(std::string_view name) { 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", HexArrayToString(nca_id, second_hex_upper)); + 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], HexArrayToString(nca_id, second_hex_upper)); + return fmt::format("/000000{:02X}/{}.nca", hash[0], + Common::HexArrayToString(nca_id, second_hex_upper)); } static std::string GetCNMTName(TitleType type, u64 title_id) { @@ -170,7 +171,7 @@ 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(HexStringToArray<0x10, true>(d2_dir->GetName().substr(0, 0x20))); + ids.push_back(Common::HexStringToArray<0x10, true>(d2_dir->GetName().substr(0, 0x20))); continue; } @@ -181,20 +182,21 @@ std::vector<NcaID> RegisteredCache::AccumulateFiles() const { if (!FollowsNcaIdFormat(nca_dir->GetName())) continue; - ids.push_back(HexStringToArray<0x10, true>(nca_dir->GetName().substr(0, 0x20))); + 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(HexStringToArray<0x10, true>(nca_file->GetName().substr(0, 0x20))); + 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(HexStringToArray<0x10, true>(d2_file->GetName().substr(0, 0x20))); + ids.push_back(Common::HexStringToArray<0x10, true>(d2_file->GetName().substr(0, 0x20))); } return ids; } @@ -339,7 +341,7 @@ std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter( } static std::shared_ptr<NCA> GetNCAFromXCIForID(std::shared_ptr<XCI> xci, const NcaID& id) { - const auto filename = fmt::format("{}.nca", HexArrayToString(id, false)); + 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; }); @@ -361,7 +363,7 @@ InstallResult RegisteredCache::InstallEntry(std::shared_ptr<XCI> xci, bool overw // Install Metadata File const auto meta_id_raw = (*meta_iter)->GetName().substr(0, 32); - const auto meta_id = HexStringToArray<16>(meta_id_raw); + 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) 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/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 1b30ce018..771eb5abc 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -638,6 +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; @@ -830,6 +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 f438fa809..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" @@ -79,6 +80,9 @@ union Attribute { // shader, and a tuple of (TessCoord.x, TessCoord.y, TessCoord.z, ~) when inside a Tess Eval // shader. TessCoordInstanceIDVertexID = 47, + // 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 { @@ -141,6 +145,7 @@ enum class PredCondition : u64 { NotEqual = 5, GreaterEqual = 6, LessThanWithNan = 9, + GreaterThanWithNan = 12, NotEqualWithNan = 13, // TODO(Subv): Other condition types }; @@ -213,6 +218,11 @@ enum class FlowCondition : u64 { 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; @@ -253,7 +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; @@ -441,16 +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; @@ -516,6 +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 @@ -727,6 +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"), diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 9d1549fe9..93eadde7a 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -124,7 +124,7 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr, glBindVertexBuffer(index, stream_buffer.GetHandle(), vertex_buffer_offset, vertex_array.stride); - ASSERT_MSG(vertex_array.divisor == 0, "Vertex buffer divisor unimplemented"); + 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. diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index b6947b97b..38aa067b6 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -142,14 +142,16 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form {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 }}; @@ -283,10 +285,10 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU MortonCopy<true, PixelFormat::RG8S>, MortonCopy<true, PixelFormat::RG32UI>, MortonCopy<true, PixelFormat::R32UI>, - MortonCopy<true, PixelFormat::Z24S8>, - MortonCopy<true, PixelFormat::S8Z24>, MortonCopy<true, PixelFormat::Z32F>, MortonCopy<true, PixelFormat::Z16>, + MortonCopy<true, PixelFormat::Z24S8>, + MortonCopy<true, PixelFormat::S8Z24>, MortonCopy<true, PixelFormat::Z32FS8>, // clang-format on }; @@ -339,10 +341,10 @@ static constexpr std::array<void (*)(u32, u32, u32, std::vector<u8>&, Tegra::GPU MortonCopy<false, PixelFormat::RG8S>, MortonCopy<false, PixelFormat::RG32UI>, MortonCopy<false, PixelFormat::R32UI>, - MortonCopy<false, PixelFormat::Z24S8>, - MortonCopy<false, PixelFormat::S8Z24>, MortonCopy<false, PixelFormat::Z32F>, MortonCopy<false, PixelFormat::Z16>, + MortonCopy<false, PixelFormat::Z24S8>, + MortonCopy<false, PixelFormat::S8Z24>, MortonCopy<false, PixelFormat::Z32FS8>, // clang-format on }; @@ -788,8 +790,6 @@ 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 55cf3782c..beec01746 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -68,11 +68,15 @@ struct SurfaceParams { MaxColorFormat, + // Depth formats + Z32F = 42, + Z16 = 43, + + MaxDepthFormat, + // DepthStencil formats - Z24S8 = 42, - S8Z24 = 43, - Z32F = 44, - Z16 = 45, + Z24S8 = 44, + S8Z24 = 45, Z32FS8 = 46, MaxDepthStencilFormat, @@ -153,10 +157,10 @@ struct SurfaceParams { 1, // RG8S 1, // RG32UI 1, // R32UI - 1, // Z24S8 - 1, // S8Z24 1, // Z32F 1, // Z16 + 1, // Z24S8 + 1, // S8Z24 1, // Z32FS8 }}; @@ -211,10 +215,10 @@ struct SurfaceParams { 16, // RG8S 64, // RG32UI 32, // R32UI - 32, // Z24S8 - 32, // S8Z24 32, // Z32F 16, // Z16 + 32, // Z24S8 + 32, // S8Z24 64, // Z32FS8 }}; @@ -587,6 +591,10 @@ 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 e899237e5..57cf9f213 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -541,7 +541,11 @@ 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))"; + 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)}; @@ -703,10 +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)}; @@ -715,7 +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 + ')'; } @@ -778,28 +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, @@ -837,6 +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. @@ -1115,7 +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: { @@ -1181,16 +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: @@ -1255,8 +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"); @@ -1859,13 +1913,13 @@ private: ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported"); u32 target = offset + instr.bra.GetBranchTarget(); - shader.AddLine("ssy_target = " + std::to_string(target) + "u;"); + EmitPushToSSYStack(target); break; } 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); - shader.AddLine("{ jmp_to = ssy_target; break; }"); + EmitPopFromSSYStack(); break; } case OpCode::Id::DEPBAR: { @@ -1936,7 +1990,13 @@ private: } else { labels.insert(subroutine.begin); shader.AddLine("uint jmp_to = " + std::to_string(subroutine.begin) + "u;"); - shader.AddLine("uint ssy_target = 0u;"); + + // 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; 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/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 8f719fdd8..5d91a0c2f 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -147,6 +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(); diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index 95f1aa0fe..bf30eda6d 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -425,7 +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/yuzu/main.cpp b/src/yuzu/main.cpp index f7eee7769..11d2331df 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -377,6 +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) @@ -628,24 +630,33 @@ void GMainWindow::OnMenuInstallToNAND() { 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; - QProgressDialog progress(fmt::format("Installing file \"{}\"...", src->GetName()).c_str(), - "Cancel", 0, src->GetSize() / 0x1000, this); + 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); - std::array<u8, 0x1000> buffer{}; - for (size_t i = 0; i < src->GetSize(); i += 0x1000) { + for (size_t i = 0; i < src->GetSize(); i += buffer.size()) { if (progress.wasCanceled()) { dest->Resize(0); return false; } - progress.setValue(i / 0x1000); + 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); } @@ -668,92 +679,88 @@ void GMainWindow::OnMenuInstallToNAND() { }; const auto overwrite = [this]() { - return QMessageBox::question(this, "Failed to Install", - "The file you are attempting to install already exists " - "in the cache. Would you like to overwrite it?") == + 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.isEmpty()) { - 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(); - } + 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) { + } else { failed(); - return; - } - - static const QStringList tt_options{"System Application", - "System Archive", - "System Application Update", - "Firmware Package (Type A)", - "Firmware Package (Type B)", - "Game", - "Game Update", - "Game DLC", - "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; } + } + } else { + const auto nca = std::make_shared<FileSys::NCA>( + vfs->OpenFile(filename.toStdString(), FileSys::Mode::Read)); + if (nca->GetStatus() != Loader::ResultStatus::Success) { + failed(); + return; + } - if (index >= 5) - index += 0x7B; + 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; + } - 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(); - } - } + 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(); } } } 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) |