diff options
Diffstat (limited to 'src/video_core')
28 files changed, 127 insertions, 47 deletions
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index 106991969..d7f7d336c 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -73,8 +73,6 @@ add_library(video_core STATIC macro/macro_hle.h macro/macro_interpreter.cpp macro/macro_interpreter.h - macro/macro_jit_x64.cpp - macro/macro_jit_x64.h fence_manager.h gpu.cpp gpu.h @@ -245,7 +243,7 @@ add_library(video_core STATIC create_target_directory_groups(video_core) target_link_libraries(video_core PUBLIC common core) -target_link_libraries(video_core PUBLIC glad shader_recompiler xbyak) +target_link_libraries(video_core PUBLIC glad shader_recompiler) if (YUZU_USE_BUNDLED_FFMPEG AND NOT WIN32) add_dependencies(video_core ffmpeg-build) @@ -282,8 +280,19 @@ else() -Wno-sign-conversion ) + + # xbyak + set_source_files_properties(macro/macro_jit_x64.cpp PROPERTIES COMPILE_OPTIONS "-Wno-conversion;-Wno-shadow") endif() if (ARCHITECTURE_x86_64) + target_sources(video_core PRIVATE + macro/macro_jit_x64.cpp + macro/macro_jit_x64.h + ) + target_link_libraries(video_core PUBLIC xbyak) +endif() + +if (ARCHITECTURE_x86_64 OR ARCHITECTURE_arm64) target_link_libraries(video_core PRIVATE dynarmic) endif() diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 4a2f2c1fd..5bb1427c1 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -232,7 +232,7 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume use_topology_override = true; return; case MAXWELL3D_REG_INDEX(clear_surface): - return ProcessClearBuffers(); + return ProcessClearBuffers(1); case MAXWELL3D_REG_INDEX(report_semaphore.query): return ProcessQueryGet(); case MAXWELL3D_REG_INDEX(render_enable.mode): @@ -249,6 +249,11 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume return; case MAXWELL3D_REG_INDEX(fragment_barrier): return rasterizer->FragmentBarrier(); + case MAXWELL3D_REG_INDEX(invalidate_texture_data_cache): + rasterizer->InvalidateGPUCache(); + return rasterizer->WaitForIdle(); + case MAXWELL3D_REG_INDEX(tiled_cache_barrier): + return rasterizer->TiledCacheBarrier(); } } @@ -591,8 +596,8 @@ u32 Maxwell3D::GetRegisterValue(u32 method) const { return regs.reg_array[method]; } -void Maxwell3D::ProcessClearBuffers() { - rasterizer->Clear(); +void Maxwell3D::ProcessClearBuffers(u32 layer_count) { + rasterizer->Clear(layer_count); } void Maxwell3D::ProcessDraw(u32 instance_count) { diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index a948fcb14..c3099f9a6 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -707,7 +707,7 @@ public: case Size::Size_A2_B10_G10_R10: return "2_10_10_10"; case Size::Size_B10_G11_R11: - return "10_11_12"; + return "10_11_11"; default: ASSERT(false); return {}; @@ -2639,7 +2639,7 @@ public: L2CacheControl l2_cache_control; ///< 0x0218 InvalidateShaderCache invalidate_shader_cache; ///< 0x021C INSERT_PADDING_BYTES_NOINIT(0xA8); - SyncInfo sync_info; ///< 0x02C8 + SyncInfo sync_info; ///< 0x02C8 INSERT_PADDING_BYTES_NOINIT(0x4); u32 prim_circular_buffer_throttle; ///< 0x02D0 u32 flush_invalidate_rop_mini_cache; ///< 0x02D4 @@ -2731,7 +2731,11 @@ public: s32 stencil_back_ref; ///< 0x0F54 u32 stencil_back_mask; ///< 0x0F58 u32 stencil_back_func_mask; ///< 0x0F5C - INSERT_PADDING_BYTES_NOINIT(0x24); + INSERT_PADDING_BYTES_NOINIT(0x14); + u32 invalidate_texture_data_cache; ///< 0x0F74 Assumed - Not in official docs. + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 tiled_cache_barrier; ///< 0x0F7C Assumed - Not in official docs. + INSERT_PADDING_BYTES_NOINIT(0x4); VertexStreamSubstitute vertex_stream_substitute; ///< 0x0F84 u32 line_mode_clip_generated_edge_do_not_draw; ///< 0x0F8C u32 color_mask_common; ///< 0x0F90 @@ -2791,7 +2795,8 @@ public: FillViaTriangleMode fill_via_triangle_mode; ///< 0x113C u32 blend_per_format_snorm8_unorm16_snorm16_enabled; ///< 0x1140 u32 flush_pending_writes_sm_gloal_store; ///< 0x1144 - INSERT_PADDING_BYTES_NOINIT(0x18); + u32 conservative_raster_enable; ///< 0x1148 Assumed - Not in official docs. + INSERT_PADDING_BYTES_NOINIT(0x14); std::array<VertexAttribute, NumVertexAttributes> vertex_attrib_format; ///< 0x1160 std::array<MsaaSampleLocation, 4> multisample_sample_locations; ///< 0x11E0 u32 offset_render_target_index_by_viewport_index; ///< 0x11F0 @@ -2970,7 +2975,7 @@ public: CullFace gl_cull_face; ///< 0x1920 Viewport::PixelCenter viewport_pixel_center; ///< 0x1924 INSERT_PADDING_BYTES_NOINIT(0x4); - u32 viewport_scale_offset_enbled; ///< 0x192C + u32 viewport_scale_offset_enabled; ///< 0x192C INSERT_PADDING_BYTES_NOINIT(0xC); ViewportClipControl viewport_clip_control; ///< 0x193C UserClip::Op user_clip_op; ///< 0x1940 @@ -3081,6 +3086,9 @@ public: std::vector<u8> inline_index_draw_indexes; + /// Handles a write to the CLEAR_BUFFERS register. + void ProcessClearBuffers(u32 layer_count); + private: void InitializeRegisterDefaults(); @@ -3115,9 +3123,6 @@ private: /// Handles firmware blob 4 void ProcessFirmwareCall4(); - /// Handles a write to the CLEAR_BUFFERS register. - void ProcessClearBuffers(); - /// Handles a write to the QUERY_GET register. void ProcessQueryGet(); @@ -3287,6 +3292,8 @@ ASSERT_REG_POSITION(const_color_rendering, 0x0F40); ASSERT_REG_POSITION(stencil_back_ref, 0x0F54); ASSERT_REG_POSITION(stencil_back_mask, 0x0F58); ASSERT_REG_POSITION(stencil_back_func_mask, 0x0F5C); +ASSERT_REG_POSITION(invalidate_texture_data_cache, 0x0F74); +ASSERT_REG_POSITION(tiled_cache_barrier, 0x0F7C); ASSERT_REG_POSITION(vertex_stream_substitute, 0x0F84); ASSERT_REG_POSITION(line_mode_clip_generated_edge_do_not_draw, 0x0F8C); ASSERT_REG_POSITION(color_mask_common, 0x0F90); @@ -3343,6 +3350,7 @@ ASSERT_REG_POSITION(post_ps_use_pre_ps_coverage, 0x1138); ASSERT_REG_POSITION(fill_via_triangle_mode, 0x113C); ASSERT_REG_POSITION(blend_per_format_snorm8_unorm16_snorm16_enabled, 0x1140); ASSERT_REG_POSITION(flush_pending_writes_sm_gloal_store, 0x1144); +ASSERT_REG_POSITION(conservative_raster_enable, 0x1148); ASSERT_REG_POSITION(vertex_attrib_format, 0x1160); ASSERT_REG_POSITION(multisample_sample_locations, 0x11E0); ASSERT_REG_POSITION(offset_render_target_index_by_viewport_index, 0x11F0); @@ -3482,7 +3490,7 @@ ASSERT_REG_POSITION(gl_cull_test_enabled, 0x1918); ASSERT_REG_POSITION(gl_front_face, 0x191C); ASSERT_REG_POSITION(gl_cull_face, 0x1920); ASSERT_REG_POSITION(viewport_pixel_center, 0x1924); -ASSERT_REG_POSITION(viewport_scale_offset_enbled, 0x192C); +ASSERT_REG_POSITION(viewport_scale_offset_enabled, 0x192C); ASSERT_REG_POSITION(viewport_clip_control, 0x193C); ASSERT_REG_POSITION(user_clip_op, 0x1940); ASSERT_REG_POSITION(render_enable_override, 0x1944); diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index 4eb7a100d..1bf6ca2dd 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -102,26 +102,29 @@ void MaxwellDMA::Launch() { const bool is_src_pitch = IsPitchKind(static_cast<PTEKind>(src_kind)); const bool is_dst_pitch = IsPitchKind(static_cast<PTEKind>(dst_kind)); if (!is_src_pitch && is_dst_pitch) { - std::vector<u8> tmp_buffer(regs.line_length_in); - std::vector<u8> dst_buffer(regs.line_length_in); - memory_manager.ReadBlockUnsafe(regs.offset_in, tmp_buffer.data(), - regs.line_length_in); - for (u32 offset = 0; offset < regs.line_length_in; ++offset) { - dst_buffer[offset] = - tmp_buffer[convert_linear_2_blocklinear_addr(regs.offset_in + offset) - - regs.offset_in]; + UNIMPLEMENTED_IF(regs.line_length_in % 16 != 0); + UNIMPLEMENTED_IF(regs.offset_in % 16 != 0); + UNIMPLEMENTED_IF(regs.offset_out % 16 != 0); + std::vector<u8> tmp_buffer(16); + for (u32 offset = 0; offset < regs.line_length_in; offset += 16) { + memory_manager.ReadBlockUnsafe( + convert_linear_2_blocklinear_addr(regs.offset_in + offset), + tmp_buffer.data(), tmp_buffer.size()); + memory_manager.WriteBlock(regs.offset_out + offset, tmp_buffer.data(), + tmp_buffer.size()); } - memory_manager.WriteBlock(regs.offset_out, dst_buffer.data(), regs.line_length_in); } else if (is_src_pitch && !is_dst_pitch) { - std::vector<u8> tmp_buffer(regs.line_length_in); - std::vector<u8> dst_buffer(regs.line_length_in); - memory_manager.ReadBlockUnsafe(regs.offset_in, tmp_buffer.data(), - regs.line_length_in); - for (u32 offset = 0; offset < regs.line_length_in; ++offset) { - dst_buffer[convert_linear_2_blocklinear_addr(regs.offset_out + offset) - - regs.offset_out] = tmp_buffer[offset]; + UNIMPLEMENTED_IF(regs.line_length_in % 16 != 0); + UNIMPLEMENTED_IF(regs.offset_in % 16 != 0); + UNIMPLEMENTED_IF(regs.offset_out % 16 != 0); + std::vector<u8> tmp_buffer(16); + for (u32 offset = 0; offset < regs.line_length_in; offset += 16) { + memory_manager.ReadBlockUnsafe(regs.offset_in + offset, tmp_buffer.data(), + tmp_buffer.size()); + memory_manager.WriteBlock( + convert_linear_2_blocklinear_addr(regs.offset_out + offset), + tmp_buffer.data(), tmp_buffer.size()); } - memory_manager.WriteBlock(regs.offset_out, dst_buffer.data(), regs.line_length_in); } else { if (!accelerate.BufferCopy(regs.offset_in, regs.offset_out, regs.line_length_in)) { std::vector<u8> tmp_buffer(regs.line_length_in); @@ -311,6 +314,7 @@ void MaxwellDMA::ReleaseSemaphore() { } default: ASSERT_MSG(false, "Unknown semaphore type: {}", static_cast<u32>(type.Value())); + break; } } diff --git a/src/video_core/engines/puller.cpp b/src/video_core/engines/puller.cpp index 3977bb0fb..4d2278811 100644 --- a/src/video_core/engines/puller.cpp +++ b/src/video_core/engines/puller.cpp @@ -50,6 +50,7 @@ void Puller::ProcessBindMethod(const MethodCall& method_call) { break; default: UNIMPLEMENTED_MSG("Unimplemented engine {:04X}", engine_id); + break; } } @@ -65,6 +66,7 @@ void Puller::ProcessFenceActionMethod() { break; default: UNIMPLEMENTED_MSG("Unimplemented operation {}", regs.fence_action.op.Value()); + break; } } @@ -228,6 +230,7 @@ void Puller::CallEngineMethod(const MethodCall& method_call) { break; default: UNIMPLEMENTED_MSG("Unimplemented engine"); + break; } } @@ -254,6 +257,7 @@ void Puller::CallEngineMultiMethod(u32 method, u32 subchannel, const u32* base_s break; default: UNIMPLEMENTED_MSG("Unimplemented engine"); + break; } } diff --git a/src/video_core/macro/macro.cpp b/src/video_core/macro/macro.cpp index f61d5998e..505d81c1e 100644 --- a/src/video_core/macro/macro.cpp +++ b/src/video_core/macro/macro.cpp @@ -16,7 +16,10 @@ #include "video_core/macro/macro.h" #include "video_core/macro/macro_hle.h" #include "video_core/macro/macro_interpreter.h" + +#ifdef ARCHITECTURE_x86_64 #include "video_core/macro/macro_jit_x64.h" +#endif namespace Tegra { diff --git a/src/video_core/macro/macro_hle.cpp b/src/video_core/macro/macro_hle.cpp index f896591bf..0f3262edb 100644 --- a/src/video_core/macro/macro_hle.cpp +++ b/src/video_core/macro/macro_hle.cpp @@ -126,11 +126,25 @@ void HLE_3F5E74B9C9A50164(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& } } -constexpr std::array<std::pair<u64, HLEFunction>, 4> hle_funcs{{ +// Multi-layer Clear +void HLE_EAD26C3E2109B06B(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) { + ASSERT(parameters.size() == 1); + + const Engines::Maxwell3D::Regs::ClearSurface clear_params{parameters[0]}; + const u32 rt_index = clear_params.RT; + const u32 num_layers = maxwell3d.regs.rt[rt_index].depth; + ASSERT(clear_params.layer == 0); + + maxwell3d.regs.clear_surface.raw = clear_params.raw; + maxwell3d.ProcessClearBuffers(num_layers); +} + +constexpr std::array<std::pair<u64, HLEFunction>, 5> hle_funcs{{ {0x771BB18C62444DA0, &HLE_771BB18C62444DA0}, {0x0D61FC9FAAC9FCAD, &HLE_0D61FC9FAAC9FCAD}, {0x0217920100488FF7, &HLE_0217920100488FF7}, {0x3F5E74B9C9A50164, &HLE_3F5E74B9C9A50164}, + {0xEAD26C3E2109B06B, &HLE_EAD26C3E2109B06B}, }}; class HLEMacroImpl final : public CachedMacro { diff --git a/src/video_core/macro/macro_interpreter.cpp b/src/video_core/macro/macro_interpreter.cpp index c0d32c112..0d63495a9 100644 --- a/src/video_core/macro/macro_interpreter.cpp +++ b/src/video_core/macro/macro_interpreter.cpp @@ -201,6 +201,7 @@ bool MacroInterpreterImpl::Step(bool is_delay_slot) { } default: UNIMPLEMENTED_MSG("Unimplemented macro operation {}", opcode.operation.Value()); + break; } // An instruction with the Exit flag will not actually @@ -297,6 +298,7 @@ void MacroInterpreterImpl::ProcessResult(Macro::ResultOperation operation, u32 r break; default: UNIMPLEMENTED_MSG("Unimplemented result operation {}", operation); + break; } } diff --git a/src/video_core/macro/macro_jit_x64.cpp b/src/video_core/macro/macro_jit_x64.cpp index 25c1ce798..7347cbd88 100644 --- a/src/video_core/macro/macro_jit_x64.cpp +++ b/src/video_core/macro/macro_jit_x64.cpp @@ -652,6 +652,7 @@ void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u3 break; default: UNIMPLEMENTED_MSG("Unimplemented macro operation {}", operation); + break; } } diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h index 1cbfef090..cfd872a40 100644 --- a/src/video_core/rasterizer_interface.h +++ b/src/video_core/rasterizer_interface.h @@ -43,7 +43,7 @@ public: virtual void Draw(bool is_indexed, u32 instance_count) = 0; /// Clear the current framebuffer - virtual void Clear() = 0; + virtual void Clear(u32 layer_count) = 0; /// Dispatches a compute shader invocation virtual void DispatchCompute() = 0; diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 0503196de..354c6e429 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -136,7 +136,7 @@ void RasterizerOpenGL::LoadDiskResources(u64 title_id, std::stop_token stop_load shader_cache.LoadDiskResources(title_id, stop_loading, callback); } -void RasterizerOpenGL::Clear() { +void RasterizerOpenGL::Clear(u32 layer_count) { MICROPROFILE_SCOPE(OpenGL_Clears); if (!maxwell3d->ShouldExecute()) { return; @@ -616,7 +616,7 @@ void RasterizerOpenGL::SyncViewport() { } flags[Dirty::Viewport0 + index] = false; - if (!regs.viewport_scale_offset_enbled) { + if (!regs.viewport_scale_offset_enabled) { const auto x = static_cast<GLfloat>(regs.surface_clip.x); const auto y = static_cast<GLfloat>(regs.surface_clip.y); const auto width = static_cast<GLfloat>(regs.surface_clip.width); @@ -768,7 +768,7 @@ void RasterizerOpenGL::SyncStencilTestState() { if (regs.stencil_two_side_enable) { glStencilFuncSeparate(GL_BACK, MaxwellToGL::ComparisonOp(regs.stencil_back_op.func), - regs.stencil_back_ref, regs.stencil_back_mask); + regs.stencil_back_ref, regs.stencil_back_func_mask); glStencilOpSeparate(GL_BACK, MaxwellToGL::StencilOp(regs.stencil_back_op.fail), MaxwellToGL::StencilOp(regs.stencil_back_op.zfail), MaxwellToGL::StencilOp(regs.stencil_back_op.zpass)); diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index c93ba3b42..fc183c3ca 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -69,7 +69,7 @@ public: ~RasterizerOpenGL() override; void Draw(bool is_indexed, u32 instance_count) override; - void Clear() override; + void Clear(u32 layer_count) override; void DispatchCompute() override; void ResetCounter(VideoCore::QueryType type) override; void Query(GPUVAddr gpu_addr, VideoCore::QueryType type, std::optional<u64> timestamp) override; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 977709518..3fe04a115 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -76,7 +76,8 @@ Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineKey& key, } break; case Shader::Stage::TessellationEval: - info.tess_clockwise = key.tessellation_clockwise != 0; + // Flip the face, as OpenGL's drawing is flipped. + info.tess_clockwise = key.tessellation_clockwise == 0; info.tess_primitive = [&key] { switch (key.tessellation_primitive) { case Maxwell::Tessellation::DomainType::Isolines: @@ -218,6 +219,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo .support_float16 = false, .support_int64 = device.HasShaderInt64(), .needs_demote_reorder = device.IsAmd(), + .support_snorm_render_buffer = false, } { if (use_asynchronous_shaders) { workers = CreateWorkers(); diff --git a/src/video_core/renderer_opengl/gl_state_tracker.cpp b/src/video_core/renderer_opengl/gl_state_tracker.cpp index a359f96f1..d53b422ca 100644 --- a/src/video_core/renderer_opengl/gl_state_tracker.cpp +++ b/src/video_core/renderer_opengl/gl_state_tracker.cpp @@ -70,8 +70,8 @@ void SetupDirtyViewports(Tables& tables) { FillBlock(tables[1], OFF(viewport_transform), NUM(viewport_transform), Viewports); FillBlock(tables[1], OFF(viewports), NUM(viewports), Viewports); - tables[0][OFF(viewport_scale_offset_enbled)] = ViewportTransform; - tables[1][OFF(viewport_scale_offset_enbled)] = Viewports; + tables[0][OFF(viewport_scale_offset_enabled)] = ViewportTransform; + tables[1][OFF(viewport_scale_offset_enabled)] = Viewports; } void SetupDirtyScissors(Tables& tables) { diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index 99cd11d1e..9f7ce7414 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -891,6 +891,7 @@ void Image::CopyBufferToImage(const VideoCommon::BufferImageCopy& copy, size_t b break; default: ASSERT(false); + break; } } @@ -927,6 +928,7 @@ void Image::CopyImageToBuffer(const VideoCommon::BufferImageCopy& copy, size_t b break; default: ASSERT(false); + break; } // Compressed formats don't have a pixel format or type const bool is_compressed = gl_format == GL_NONE; diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index 8bd5eba7e..f29462f7c 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -340,6 +340,7 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture, texture.gl_type = GL_UNSIGNED_INT_8_8_8_8_REV; // UNIMPLEMENTED_MSG("Unknown framebuffer pixel format: {}", // static_cast<u32>(framebuffer.pixel_format)); + break; } texture.resource.Release(); diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp index f85ed8e5b..98cc26679 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp @@ -90,6 +90,7 @@ void FixedPipelineState::Refresh(Tegra::Engines::Maxwell3D& maxwell3d, depth_format.Assign(static_cast<u32>(regs.zeta.format)); y_negate.Assign(regs.window_origin.mode != Maxwell::WindowOrigin::Mode::UpperLeft ? 1 : 0); provoking_vertex_last.Assign(regs.provoking_vertex == Maxwell::ProvokingVertex::Last ? 1 : 0); + conservative_raster_enable.Assign(regs.conservative_raster_enable != 0 ? 1 : 0); smooth_lines.Assign(regs.line_anti_alias_enable != 0 ? 1 : 0); for (size_t i = 0; i < regs.rt.size(); ++i) { diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.h b/src/video_core/renderer_vulkan/fixed_pipeline_state.h index 43441209c..1afdef329 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.h +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.h @@ -193,6 +193,7 @@ struct FixedPipelineState { BitField<6, 5, u32> depth_format; BitField<11, 1, u32> y_negate; BitField<12, 1, u32> provoking_vertex_last; + BitField<13, 1, u32> conservative_raster_enable; BitField<14, 1, u32> smooth_lines; }; std::array<u8, Maxwell::NumRenderTargets> color_formats; diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp index d8131232a..c2a95200b 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp +++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp @@ -172,6 +172,7 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { } void RendererVulkan::Report() const { + using namespace Common::Literals; const std::string vendor_name{device.GetVendorName()}; const std::string model_name{device.GetModelName()}; const std::string driver_version = GetDriverVersion(device); @@ -181,9 +182,12 @@ void RendererVulkan::Report() const { const std::string extensions = BuildCommaSeparatedExtensions(device.GetAvailableExtensions()); + const auto available_vram = static_cast<f64>(device.GetDeviceLocalMemory()) / f64{1_GiB}; + LOG_INFO(Render_Vulkan, "Driver: {}", driver_name); LOG_INFO(Render_Vulkan, "Device: {}", model_name); LOG_INFO(Render_Vulkan, "Vulkan: {}", api_version); + LOG_INFO(Render_Vulkan, "Available VRAM: {:.2f} GiB", available_vram); static constexpr auto field = Common::Telemetry::FieldType::UserSystem; telemetry_session.AddField(field, "GPU_Vendor", vendor_name); diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index 1aa116cea..ef75c126c 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -680,6 +680,15 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) { .lineStippleFactor = 0, .lineStipplePattern = 0, }; + VkPipelineRasterizationConservativeStateCreateInfoEXT conservative_raster{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_CONSERVATIVE_STATE_CREATE_INFO_EXT, + .pNext = nullptr, + .flags = 0, + .conservativeRasterizationMode = key.state.conservative_raster_enable != 0 + ? VK_CONSERVATIVE_RASTERIZATION_MODE_OVERESTIMATE_EXT + : VK_CONSERVATIVE_RASTERIZATION_MODE_DISABLED_EXT, + .extraPrimitiveOverestimationSize = 0.0f, + }; VkPipelineRasterizationProvokingVertexStateCreateInfoEXT provoking_vertex{ .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_PROVOKING_VERTEX_STATE_CREATE_INFO_EXT, .pNext = nullptr, @@ -690,6 +699,9 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) { if (IsLine(input_assembly_topology) && device.IsExtLineRasterizationSupported()) { line_state.pNext = std::exchange(rasterization_ci.pNext, &line_state); } + if (device.IsExtConservativeRasterizationSupported()) { + conservative_raster.pNext = std::exchange(rasterization_ci.pNext, &conservative_raster); + } if (device.IsExtProvokingVertexSupported()) { provoking_vertex.pNext = std::exchange(rasterization_ci.pNext, &provoking_vertex); } diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index b42e5be1e..d4b0a542a 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -166,6 +166,7 @@ Shader::RuntimeInfo MakeRuntimeInfo(std::span<const Shader::IR::Program> program } break; case Shader::Stage::TessellationEval: + info.tess_clockwise = key.state.tessellation_clockwise != 0; info.tess_primitive = [&key] { const u32 raw{key.state.tessellation_primitive.Value()}; switch (static_cast<Maxwell::Tessellation::DomainType>(raw)) { @@ -325,6 +326,7 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device .support_int64 = device.IsShaderInt64Supported(), .needs_demote_reorder = driver_id == VK_DRIVER_ID_AMD_PROPRIETARY_KHR || driver_id == VK_DRIVER_ID_AMD_OPEN_SOURCE_KHR, + .support_snorm_render_buffer = true, }; } diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index 6f94dcbc7..12b13cc59 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -211,7 +211,7 @@ void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) { EndTransformFeedback(); } -void RasterizerVulkan::Clear() { +void RasterizerVulkan::Clear(u32 layer_count) { MICROPROFILE_SCOPE(Vulkan_Clearing); if (!maxwell3d->ShouldExecute()) { @@ -254,7 +254,7 @@ void RasterizerVulkan::Clear() { .rect = regs.clear_control.use_scissor ? GetScissorState(regs, 0, up_scale, down_shift) : default_scissor, .baseArrayLayer = regs.clear_surface.layer, - .layerCount = 1, + .layerCount = layer_count, }; if (clear_rect.rect.extent.width == 0 || clear_rect.rect.extent.height == 0) { return; @@ -681,7 +681,7 @@ void RasterizerVulkan::UpdateViewportsState(Tegra::Engines::Maxwell3D::Regs& reg if (!state_tracker.TouchViewports()) { return; } - if (!regs.viewport_scale_offset_enbled) { + if (!regs.viewport_scale_offset_enabled) { const auto x = static_cast<float>(regs.surface_clip.x); const auto y = static_cast<float>(regs.surface_clip.y); const auto width = static_cast<float>(regs.surface_clip.width); diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h index fa2238c5f..ee483cfd9 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.h +++ b/src/video_core/renderer_vulkan/vk_rasterizer.h @@ -65,7 +65,7 @@ public: ~RasterizerVulkan() override; void Draw(bool is_indexed, u32 instance_count) override; - void Clear() override; + void Clear(u32 layer_count) override; void DispatchCompute() override; void ResetCounter(VideoCore::QueryType type) override; void Query(GPUVAddr gpu_addr, VideoCore::QueryType type, std::optional<u64> timestamp) override; diff --git a/src/video_core/renderer_vulkan/vk_scheduler.cpp b/src/video_core/renderer_vulkan/vk_scheduler.cpp index 7934f2a51..4a7b633b7 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.cpp +++ b/src/video_core/renderer_vulkan/vk_scheduler.cpp @@ -221,6 +221,7 @@ void Scheduler::SubmitExecution(VkSemaphore signal_semaphore, VkSemaphore wait_s [[fallthrough]]; default: vk::Check(result); + break; } }); chunk->MarkSubmit(); diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.cpp b/src/video_core/renderer_vulkan/vk_state_tracker.cpp index b87c3be66..edb41b171 100644 --- a/src/video_core/renderer_vulkan/vk_state_tracker.cpp +++ b/src/video_core/renderer_vulkan/vk_state_tracker.cpp @@ -51,7 +51,7 @@ Flags MakeInvalidationFlags() { void SetupDirtyViewports(Tables& tables) { FillBlock(tables[0], OFF(viewport_transform), NUM(viewport_transform), Viewports); FillBlock(tables[0], OFF(viewports), NUM(viewports), Viewports); - tables[0][OFF(viewport_scale_offset_enbled)] = Viewports; + tables[0][OFF(viewport_scale_offset_enabled)] = Viewports; tables[1][OFF(window_origin)] = Viewports; } diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index 853b80d8a..a65bbeb1c 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -108,6 +108,7 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) { break; default: ASSERT_MSG(false, "Invalid surface type"); + break; } } if (info.storage) { diff --git a/src/video_core/shader_environment.cpp b/src/video_core/shader_environment.cpp index 37bb76b72..f24f320b6 100644 --- a/src/video_core/shader_environment.cpp +++ b/src/video_core/shader_environment.cpp @@ -352,7 +352,7 @@ Shader::TexturePixelFormat GraphicsEnvironment::ReadTexturePixelFormat(u32 handl u32 GraphicsEnvironment::ReadViewportTransformState() { const auto& regs{maxwell3d->regs}; - viewport_transform_state = regs.viewport_scale_offset_enbled; + viewport_transform_state = regs.viewport_scale_offset_enabled; return viewport_transform_state; } diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index fd1a4b987..59120cd09 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -170,6 +170,7 @@ void Swizzle(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixe #undef BPP_CASE default: ASSERT_MSG(false, "Invalid bytes_per_pixel={}", bytes_per_pixel); + break; } } @@ -217,6 +218,7 @@ void SwizzleSubrect(std::span<u8> output, std::span<const u8> input, u32 bytes_p #undef BPP_CASE default: ASSERT_MSG(false, "Invalid bytes_per_pixel={}", bytes_per_pixel); + break; } } @@ -240,6 +242,7 @@ void UnswizzleSubrect(std::span<u8> output, std::span<const u8> input, u32 bytes #undef BPP_CASE default: ASSERT_MSG(false, "Invalid bytes_per_pixel={}", bytes_per_pixel); + break; } } |