diff options
Diffstat (limited to 'src/video_core')
19 files changed, 729 insertions, 341 deletions
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index ddb1a1d69..0b1cc1290 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -33,6 +33,7 @@ add_library(video_core STATIC renderer_opengl/gl_rasterizer.h renderer_opengl/gl_rasterizer_cache.cpp renderer_opengl/gl_rasterizer_cache.h + renderer_opengl/gl_resource_manager.cpp renderer_opengl/gl_resource_manager.h renderer_opengl/gl_shader_cache.cpp renderer_opengl/gl_shader_cache.h diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index d79c50919..2cd595f26 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -37,6 +37,22 @@ void Maxwell3D::InitializeRegisterDefaults() { regs.viewport[viewport].depth_range_near = 0.0f; regs.viewport[viewport].depth_range_far = 1.0f; } + // Doom and Bomberman seems to use the uninitialized registers and just enable blend + // so initialize blend registers with sane values + regs.blend.equation_rgb = Regs::Blend::Equation::Add; + regs.blend.factor_source_rgb = Regs::Blend::Factor::One; + regs.blend.factor_dest_rgb = Regs::Blend::Factor::Zero; + regs.blend.equation_a = Regs::Blend::Equation::Add; + regs.blend.factor_source_a = Regs::Blend::Factor::One; + regs.blend.factor_dest_a = Regs::Blend::Factor::Zero; + for (std::size_t blend_index = 0; blend_index < Regs::NumRenderTargets; blend_index++) { + regs.independent_blend[blend_index].equation_rgb = Regs::Blend::Equation::Add; + regs.independent_blend[blend_index].factor_source_rgb = Regs::Blend::Factor::One; + regs.independent_blend[blend_index].factor_dest_rgb = Regs::Blend::Factor::Zero; + regs.independent_blend[blend_index].equation_a = Regs::Blend::Equation::Add; + regs.independent_blend[blend_index].factor_source_a = Regs::Blend::Factor::One; + regs.independent_blend[blend_index].factor_dest_a = Regs::Blend::Factor::Zero; + } } void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) { diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 50873813e..0509ba3a2 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -462,6 +462,16 @@ public: } }; + struct ColorMask { + union { + u32 raw; + BitField<0, 4, u32> R; + BitField<4, 4, u32> G; + BitField<8, 4, u32> B; + BitField<12, 4, u32> A; + }; + }; + bool IsShaderConfigEnabled(std::size_t index) const { // The VertexB is always enabled. if (index == static_cast<std::size_t>(Regs::ShaderProgram::VertexB)) { @@ -571,7 +581,11 @@ public: u32 stencil_back_mask; u32 stencil_back_func_mask; - INSERT_PADDING_WORDS(0x13); + INSERT_PADDING_WORDS(0xC); + + u32 color_mask_common; + + INSERT_PADDING_WORDS(0x6); u32 rt_separate_frag_data; @@ -646,8 +660,14 @@ public: ComparisonOp depth_test_func; float alpha_test_ref; ComparisonOp alpha_test_func; - - INSERT_PADDING_WORDS(0x9); + u32 draw_tfb_stride; + struct { + float r; + float g; + float b; + float a; + } blend_color; + INSERT_PADDING_WORDS(0x4); struct { u32 separate_alpha; @@ -841,8 +861,9 @@ public: BitField<6, 4, u32> RT; BitField<10, 11, u32> layer; } clear_buffers; - - INSERT_PADDING_WORDS(0x4B); + INSERT_PADDING_WORDS(0xB); + std::array<ColorMask, NumRenderTargets> color_mask; + INSERT_PADDING_WORDS(0x38); struct { u32 query_address_high; @@ -1075,6 +1096,7 @@ ASSERT_REG_POSITION(scissor_test, 0x380); ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5); ASSERT_REG_POSITION(stencil_back_mask, 0x3D6); ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7); +ASSERT_REG_POSITION(color_mask_common, 0x3E4); ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB); ASSERT_REG_POSITION(zeta, 0x3F8); ASSERT_REG_POSITION(vertex_attrib_format, 0x458); @@ -1087,6 +1109,10 @@ ASSERT_REG_POSITION(depth_write_enabled, 0x4BA); ASSERT_REG_POSITION(alpha_test_enabled, 0x4BB); ASSERT_REG_POSITION(d3d_cull_mode, 0x4C2); ASSERT_REG_POSITION(depth_test_func, 0x4C3); +ASSERT_REG_POSITION(alpha_test_ref, 0x4C4); +ASSERT_REG_POSITION(alpha_test_func, 0x4C5); +ASSERT_REG_POSITION(draw_tfb_stride, 0x4C6); +ASSERT_REG_POSITION(blend_color, 0x4C7); ASSERT_REG_POSITION(blend, 0x4CF); ASSERT_REG_POSITION(stencil_enable, 0x4E0); ASSERT_REG_POSITION(stencil_front_op_fail, 0x4E1); @@ -1117,6 +1143,7 @@ ASSERT_REG_POSITION(instanced_arrays, 0x620); ASSERT_REG_POSITION(cull, 0x646); ASSERT_REG_POSITION(logic_op, 0x671); ASSERT_REG_POSITION(clear_buffers, 0x674); +ASSERT_REG_POSITION(color_mask, 0x680); ASSERT_REG_POSITION(query, 0x6C0); ASSERT_REG_POSITION(vertex_array[0], 0x700); ASSERT_REG_POSITION(independent_blend, 0x780); diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index a0527fe57..bb263b6aa 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -511,10 +511,10 @@ void RasterizerOpenGL::Clear() { OpenGLState clear_state; clear_state.draw.draw_framebuffer = framebuffer.handle; - clear_state.color_mask.red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE; - clear_state.color_mask.green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE; - clear_state.color_mask.blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE; - clear_state.color_mask.alpha_enabled = regs.clear_buffers.A ? GL_TRUE : GL_FALSE; + clear_state.color_mask[0].red_enabled = regs.clear_buffers.R ? GL_TRUE : GL_FALSE; + clear_state.color_mask[0].green_enabled = regs.clear_buffers.G ? GL_TRUE : GL_FALSE; + clear_state.color_mask[0].blue_enabled = regs.clear_buffers.B ? GL_TRUE : GL_FALSE; + clear_state.color_mask[0].alpha_enabled = regs.clear_buffers.A ? GL_TRUE : GL_FALSE; if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B || regs.clear_buffers.A) { @@ -573,14 +573,13 @@ void RasterizerOpenGL::DrawArrays() { ScopeAcquireGLContext acquire_context{emu_window}; ConfigureFramebuffers(); - + SyncColorMask(); SyncDepthTestState(); SyncStencilTestState(); SyncBlendState(); SyncLogicOpState(); SyncCullMode(); SyncPrimitiveRestart(); - SyncDepthRange(); SyncScissorTest(); // Alpha Testing is synced on shaders. SyncTransformFeedback(); @@ -899,12 +898,16 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, Shader& shader, void RasterizerOpenGL::SyncViewport() { const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; - const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()}; - - state.viewport.x = viewport_rect.left; - state.viewport.y = viewport_rect.bottom; - state.viewport.width = static_cast<GLsizei>(viewport_rect.GetWidth()); - state.viewport.height = static_cast<GLsizei>(viewport_rect.GetHeight()); + for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { + const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()}; + auto& viewport = state.viewports[i]; + viewport.x = viewport_rect.left; + viewport.y = viewport_rect.bottom; + viewport.width = static_cast<GLsizei>(viewport_rect.GetWidth()); + viewport.height = static_cast<GLsizei>(viewport_rect.GetHeight()); + viewport.depth_range_far = regs.viewport[i].depth_range_far; + viewport.depth_range_near = regs.viewport[i].depth_range_near; + } } void RasterizerOpenGL::SyncClipEnabled() { @@ -946,13 +949,6 @@ void RasterizerOpenGL::SyncPrimitiveRestart() { state.primitive_restart.index = regs.primitive_restart.index; } -void RasterizerOpenGL::SyncDepthRange() { - const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; - - state.depth.depth_range_near = regs.viewport->depth_range_near; - state.depth.depth_range_far = regs.viewport->depth_range_far; -} - void RasterizerOpenGL::SyncDepthTestState() { const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; @@ -993,26 +989,60 @@ void RasterizerOpenGL::SyncStencilTestState() { state.stencil.back.write_mask = regs.stencil_back_mask; } -void RasterizerOpenGL::SyncBlendState() { +void RasterizerOpenGL::SyncColorMask() { const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; + for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { + const auto& source = regs.color_mask[regs.color_mask_common ? 0 : i]; + auto& dest = state.color_mask[i]; + dest.red_enabled = (source.R == 0) ? GL_FALSE : GL_TRUE; + dest.green_enabled = (source.G == 0) ? GL_FALSE : GL_TRUE; + dest.blue_enabled = (source.B == 0) ? GL_FALSE : GL_TRUE; + dest.alpha_enabled = (source.A == 0) ? GL_FALSE : GL_TRUE; + } +} - // TODO(Subv): Support more than just render target 0. - state.blend.enabled = regs.blend.enable[0] != 0; +void RasterizerOpenGL::SyncBlendState() { + const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; - if (!state.blend.enabled) + state.blend_color.red = regs.blend_color.r; + state.blend_color.green = regs.blend_color.g; + state.blend_color.blue = regs.blend_color.b; + state.blend_color.alpha = regs.blend_color.a; + + state.independant_blend.enabled = regs.independent_blend_enable; + if (!state.independant_blend.enabled) { + auto& blend = state.blend[0]; + blend.enabled = regs.blend.enable[0] != 0; + blend.separate_alpha = regs.blend.separate_alpha; + blend.rgb_equation = MaxwellToGL::BlendEquation(regs.blend.equation_rgb); + blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb); + blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb); + if (blend.separate_alpha) { + blend.a_equation = MaxwellToGL::BlendEquation(regs.blend.equation_a); + blend.src_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_a); + blend.dst_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_a); + } + for (size_t i = 1; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { + state.blend[i].enabled = false; + } return; + } - ASSERT_MSG(regs.logic_op.enable == 0, - "Blending and logic op can't be enabled at the same time."); - - ASSERT_MSG(regs.independent_blend_enable == 1, "Only independent blending is implemented"); - ASSERT_MSG(!regs.independent_blend[0].separate_alpha, "Unimplemented"); - state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_rgb); - state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_rgb); - state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_rgb); - state.blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_a); - state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_a); - state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_a); + for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { + auto& blend = state.blend[i]; + blend.enabled = regs.blend.enable[i] != 0; + if (!blend.enabled) + continue; + blend.separate_alpha = regs.independent_blend[i].separate_alpha; + blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[i].equation_rgb); + blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_source_rgb); + blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_dest_rgb); + if (blend.separate_alpha) { + blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[i].equation_a); + blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_source_a); + blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[i].factor_dest_a); + } + } } void RasterizerOpenGL::SyncLogicOpState() { @@ -1031,19 +1061,19 @@ void RasterizerOpenGL::SyncLogicOpState() { } void RasterizerOpenGL::SyncScissorTest() { + // TODO: what is the correct behavior here, a single scissor for all targets + // or scissor disabled for the rest of the targets? const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs; - state.scissor.enabled = (regs.scissor_test.enable != 0); - // TODO(Blinkhawk): Figure if the hardware supports scissor testing per viewport and how it's - // implemented. - if (regs.scissor_test.enable != 0) { - const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x; - const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y; - state.scissor.x = regs.scissor_test.min_x; - state.scissor.y = regs.scissor_test.min_y; - state.scissor.width = width; - state.scissor.height = height; + if (regs.scissor_test.enable == 0) { + return; } + const u32 width = regs.scissor_test.max_x - regs.scissor_test.min_x; + const u32 height = regs.scissor_test.max_y - regs.scissor_test.min_y; + state.scissor.x = regs.scissor_test.min_x; + state.scissor.y = regs.scissor_test.min_y; + state.scissor.width = width; + state.scissor.height = height; } void RasterizerOpenGL::SyncTransformFeedback() { diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 47097c569..60e783803 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -133,7 +133,7 @@ private: u32 SetupTextures(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, Shader& shader, GLenum primitive_mode, u32 current_unit); - /// Syncs the viewport to match the guest state + /// Syncs the viewport and depth range to match the guest state void SyncViewport(); /// Syncs the clip enabled status to match the guest state @@ -148,9 +148,6 @@ private: /// Syncs the primitve restart to match the guest state void SyncPrimitiveRestart(); - /// Syncs the depth range to match the guest state - void SyncDepthRange(); - /// Syncs the depth test state to match the guest state void SyncDepthTestState(); @@ -172,6 +169,9 @@ private: /// Syncs the point state to match the guest state void SyncPointState(); + /// Syncs Color Mask + void SyncColorMask(); + /// Check asserts for alpha testing. void CheckAlphaTests(); diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index f194a7687..49d63e6f3 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -16,6 +16,7 @@ #include "core/settings.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_opengl/gl_rasterizer_cache.h" +#include "video_core/renderer_opengl/gl_state.h" #include "video_core/renderer_opengl/utils.h" #include "video_core/surface.h" #include "video_core/textures/astc.h" @@ -58,16 +59,14 @@ void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) { std::size_t SurfaceParams::InnerMipmapMemorySize(u32 mip_level, bool force_gl, bool layer_only, bool uncompressed) const { - const u32 compression_factor{GetCompressionFactor(pixel_format)}; + const u32 tile_x{GetDefaultBlockWidth(pixel_format)}; + const u32 tile_y{GetDefaultBlockHeight(pixel_format)}; const u32 bytes_per_pixel{GetBytesPerPixel(pixel_format)}; u32 m_depth = (layer_only ? 1U : depth); u32 m_width = MipWidth(mip_level); u32 m_height = MipHeight(mip_level); - m_width = uncompressed ? m_width - : std::max(1U, (m_width + compression_factor - 1) / compression_factor); - m_height = uncompressed - ? m_height - : std::max(1U, (m_height + compression_factor - 1) / compression_factor); + m_width = uncompressed ? m_width : std::max(1U, (m_width + tile_x - 1) / tile_x); + m_height = uncompressed ? m_height : std::max(1U, (m_height + tile_y - 1) / tile_y); m_depth = std::max(1U, m_depth >> mip_level); u32 m_block_height = MipBlockHeight(mip_level); u32 m_block_depth = MipBlockDepth(mip_level); @@ -128,6 +127,13 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only, params.target = SurfaceTarget::Texture2D; } break; + case SurfaceTarget::TextureCubeArray: + params.depth = config.tic.Depth() * 6; + if (!entry.IsArray()) { + ASSERT(params.depth == 6); + params.target = SurfaceTarget::TextureCubemap; + } + break; default: LOG_CRITICAL(HW_GPU, "Unknown depth for target={}", static_cast<u32>(params.target)); UNREACHABLE(); @@ -305,6 +311,8 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8_SRGB {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5_SRGB {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X4_SRGB + {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X5 + {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_5X5_SRGB // Depth formats {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT, ComponentType::Float, false}, // Z32F @@ -334,6 +342,8 @@ static GLenum SurfaceTargetToGL(SurfaceTarget target) { return GL_TEXTURE_2D_ARRAY; case SurfaceTarget::TextureCubemap: return GL_TEXTURE_CUBE_MAP; + case SurfaceTarget::TextureCubeArray: + return GL_TEXTURE_CUBE_MAP_ARRAY_ARB; } LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target)); UNREACHABLE(); @@ -364,15 +374,18 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 d // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual // pixel values. - const u32 tile_size{IsFormatBCn(format) ? 4U : 1U}; + const u32 tile_size_x{GetDefaultBlockWidth(format)}; + const u32 tile_size_y{GetDefaultBlockHeight(format)}; if (morton_to_gl) { - const std::vector<u8> data = Tegra::Texture::UnswizzleTexture( - addr, tile_size, bytes_per_pixel, stride, height, depth, block_height, block_depth); + const std::vector<u8> data = + Tegra::Texture::UnswizzleTexture(addr, tile_size_x, tile_size_y, bytes_per_pixel, + stride, height, depth, block_height, block_depth); const std::size_t size_to_copy{std::min(gl_buffer_size, data.size())}; memcpy(gl_buffer, data.data(), size_to_copy); } else { - Tegra::Texture::CopySwizzledData(stride / tile_size, height / tile_size, depth, + Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x, + (height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr), gl_buffer, false, block_height, block_depth); } @@ -440,6 +453,8 @@ static constexpr GLConversionArray morton_to_gl_fns = { MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>, MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>, MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>, + MortonCopy<true, PixelFormat::ASTC_2D_5X5>, + MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>, MortonCopy<true, PixelFormat::Z32F>, MortonCopy<true, PixelFormat::Z16>, MortonCopy<true, PixelFormat::Z24S8>, @@ -508,6 +523,8 @@ static constexpr GLConversionArray gl_to_morton_fns = { nullptr, nullptr, nullptr, + nullptr, + nullptr, MortonCopy<false, PixelFormat::Z32F>, MortonCopy<false, PixelFormat::Z16>, MortonCopy<false, PixelFormat::Z24S8>, @@ -545,9 +562,11 @@ void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params } } +MICROPROFILE_DEFINE(OpenGL_BlitSurface, "OpenGL", "BlitSurface", MP_RGB(128, 192, 64)); static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface, GLuint read_fb_handle, GLuint draw_fb_handle, GLenum src_attachment = 0, GLenum dst_attachment = 0, std::size_t cubemap_face = 0) { + MICROPROFILE_SCOPE(OpenGL_BlitSurface); const auto& src_params{src_surface->GetSurfaceParams()}; const auto& dst_params{dst_surface->GetSurfaceParams()}; @@ -687,9 +706,11 @@ static void FastCopySurface(const Surface& src_surface, const Surface& dst_surfa 0, 0, width, height, 1); } +MICROPROFILE_DEFINE(OpenGL_CopySurface, "OpenGL", "CopySurface", MP_RGB(128, 192, 64)); static void CopySurface(const Surface& src_surface, const Surface& dst_surface, GLuint copy_pbo_handle, GLenum src_attachment = 0, GLenum dst_attachment = 0, std::size_t cubemap_face = 0) { + MICROPROFILE_SCOPE(OpenGL_CopySurface); ASSERT_MSG(dst_attachment == 0, "Unimplemented"); const auto& src_params{src_surface->GetSurfaceParams()}; @@ -754,6 +775,7 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface, break; case SurfaceTarget::Texture3D: case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0, 0, width, height, static_cast<GLsizei>(dst_params.depth), dest_format.format, dest_format.type, nullptr); @@ -806,6 +828,7 @@ CachedSurface::CachedSurface(const SurfaceParams& params) break; case SurfaceTarget::Texture3D: case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: glTexStorage3D(SurfaceTargetToGL(params.target), params.max_mip_level, format_tuple.internal_format, rect.GetWidth(), rect.GetHeight(), params.depth); @@ -897,21 +920,24 @@ static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) { * typical desktop GPUs. */ static void ConvertFormatAsNeeded_LoadGLBuffer(std::vector<u8>& data, PixelFormat pixel_format, - u32 width, u32 height) { + u32 width, u32 height, u32 depth) { switch (pixel_format) { case PixelFormat::ASTC_2D_4X4: case PixelFormat::ASTC_2D_8X8: case PixelFormat::ASTC_2D_8X5: case PixelFormat::ASTC_2D_5X4: + case PixelFormat::ASTC_2D_5X5: case PixelFormat::ASTC_2D_4X4_SRGB: case PixelFormat::ASTC_2D_8X8_SRGB: case PixelFormat::ASTC_2D_8X5_SRGB: - case PixelFormat::ASTC_2D_5X4_SRGB: { + case PixelFormat::ASTC_2D_5X4_SRGB: + case PixelFormat::ASTC_2D_5X5_SRGB: { // Convert ASTC pixel formats to RGBA8, as most desktop GPUs do not support ASTC. u32 block_width{}; u32 block_height{}; std::tie(block_width, block_height) = GetASTCBlockSize(pixel_format); - data = Tegra::Texture::ASTC::Decompress(data, width, height, block_width, block_height); + data = + Tegra::Texture::ASTC::Decompress(data, width, height, depth, block_width, block_height); break; } case PixelFormat::S8Z24: @@ -953,7 +979,7 @@ static void ConvertFormatAsNeeded_FlushGLBuffer(std::vector<u8>& data, PixelForm } } -MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 64, 192)); +MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 192, 64)); void CachedSurface::LoadGLBuffer() { MICROPROFILE_SCOPE(OpenGL_SurfaceLoad); gl_buffer.resize(params.max_mip_level); @@ -971,7 +997,7 @@ void CachedSurface::LoadGLBuffer() { } for (u32 i = 0; i < params.max_mip_level; i++) ConvertFormatAsNeeded_LoadGLBuffer(gl_buffer[i], params.pixel_format, params.MipWidth(i), - params.MipHeight(i)); + params.MipHeight(i), params.MipDepth(i)); } MICROPROFILE_DEFINE(OpenGL_SurfaceFlush, "OpenGL", "Surface Flush", MP_RGB(128, 192, 64)); @@ -1055,6 +1081,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, &gl_buffer[mip_map][buffer_offset]); break; case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format, static_cast<GLsizei>(params.MipWidth(mip_map)), static_cast<GLsizei>(params.MipHeight(mip_map)), @@ -1104,6 +1131,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]); break; case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0, static_cast<GLsizei>(rect.GetWidth()), static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format, @@ -1133,7 +1161,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, glPixelStorei(GL_UNPACK_ROW_LENGTH, 0); } -MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 64, 192)); +MICROPROFILE_DEFINE(OpenGL_TextureUL, "OpenGL", "Texture Upload", MP_RGB(128, 192, 64)); void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle) { if (params.type == SurfaceType::Fill) return; @@ -1306,6 +1334,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface, break; case SurfaceTarget::TextureCubemap: case SurfaceTarget::Texture3D: + case SurfaceTarget::TextureCubeArray: AccurateCopySurface(old_surface, new_surface); break; default: diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h index f255f4419..c0b6bc4e6 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -49,6 +49,8 @@ struct SurfaceParams { return "Texture2DArray"; case SurfaceTarget::TextureCubemap: return "TextureCubemap"; + case SurfaceTarget::TextureCubeArray: + return "TextureCubeArray"; default: LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target)); UNREACHABLE(); @@ -139,7 +141,7 @@ struct SurfaceParams { } u32 MipDepth(u32 mip_level) const { - return std::max(1U, depth >> mip_level); + return is_layered ? depth : std::max(1U, depth >> mip_level); } // Auto block resizing algorithm from: diff --git a/src/video_core/renderer_opengl/gl_resource_manager.cpp b/src/video_core/renderer_opengl/gl_resource_manager.cpp new file mode 100644 index 000000000..161318c5f --- /dev/null +++ b/src/video_core/renderer_opengl/gl_resource_manager.cpp @@ -0,0 +1,188 @@ +// Copyright 2015 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <utility> +#include <glad/glad.h> +#include "common/common_types.h" +#include "common/microprofile.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/renderer_opengl/gl_shader_util.h" +#include "video_core/renderer_opengl/gl_state.h" + +MICROPROFILE_DEFINE(OpenGL_ResourceCreation, "OpenGL", "Resource Creation", + MP_RGB(128, 128, 192)); +MICROPROFILE_DEFINE(OpenGL_ResourceDeletion, "OpenGL", "Resource Deletion", + MP_RGB(128, 128, 192)); + +namespace OpenGL { + +void OGLTexture::Create() { + if (handle != 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + glGenTextures(1, &handle); +} + +void OGLTexture::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteTextures(1, &handle); + OpenGLState::GetCurState().UnbindTexture(handle).Apply(); + handle = 0; +} + +void OGLSampler::Create() { + if (handle != 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + glGenSamplers(1, &handle); +} + +void OGLSampler::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteSamplers(1, &handle); + OpenGLState::GetCurState().ResetSampler(handle).Apply(); + handle = 0; +} + +void OGLShader::Create(const char* source, GLenum type) { + if (handle != 0) + return; + if (source == nullptr) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + handle = GLShader::LoadShader(source, type); +} + +void OGLShader::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteShader(handle); + handle = 0; +} + +void OGLProgram::CreateFromSource(const char* vert_shader, const char* geo_shader, + const char* frag_shader, bool separable_program) { + OGLShader vert, geo, frag; + if (vert_shader) + vert.Create(vert_shader, GL_VERTEX_SHADER); + if (geo_shader) + geo.Create(geo_shader, GL_GEOMETRY_SHADER); + if (frag_shader) + frag.Create(frag_shader, GL_FRAGMENT_SHADER); + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + Create(separable_program, vert.handle, geo.handle, frag.handle); +} + +void OGLProgram::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteProgram(handle); + OpenGLState::GetCurState().ResetProgram(handle).Apply(); + handle = 0; +} + +void OGLPipeline::Create() { + if (handle != 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + glGenProgramPipelines(1, &handle); +} + +void OGLPipeline::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteProgramPipelines(1, &handle); + OpenGLState::GetCurState().ResetPipeline(handle).Apply(); + handle = 0; +} + +void OGLBuffer::Create() { + if (handle != 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + glGenBuffers(1, &handle); +} + +void OGLBuffer::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteBuffers(1, &handle); + OpenGLState::GetCurState().ResetBuffer(handle).Apply(); + handle = 0; +} + +void OGLSync::Create() { + if (handle != 0) + return; + + // Don't profile here, this one is expected to happen ingame. + handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0); +} + +void OGLSync::Release() { + if (handle == 0) + return; + + // Don't profile here, this one is expected to happen ingame. + glDeleteSync(handle); + handle = 0; +} + +void OGLVertexArray::Create() { + if (handle != 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + glGenVertexArrays(1, &handle); +} + +void OGLVertexArray::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteVertexArrays(1, &handle); + OpenGLState::GetCurState().ResetVertexArray(handle).Apply(); + handle = 0; +} + +void OGLFramebuffer::Create() { + if (handle != 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceCreation); + glGenFramebuffers(1, &handle); +} + +void OGLFramebuffer::Release() { + if (handle == 0) + return; + + MICROPROFILE_SCOPE(OpenGL_ResourceDeletion); + glDeleteFramebuffers(1, &handle); + OpenGLState::GetCurState().ResetFramebuffer(handle).Apply(); + handle = 0; +} + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_resource_manager.h b/src/video_core/renderer_opengl/gl_resource_manager.h index 3bc1b83b5..e33f1e973 100644 --- a/src/video_core/renderer_opengl/gl_resource_manager.h +++ b/src/video_core/renderer_opengl/gl_resource_manager.h @@ -8,7 +8,6 @@ #include <glad/glad.h> #include "common/common_types.h" #include "video_core/renderer_opengl/gl_shader_util.h" -#include "video_core/renderer_opengl/gl_state.h" namespace OpenGL { @@ -29,20 +28,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenTextures(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteTextures(1, &handle); - OpenGLState::GetCurState().UnbindTexture(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -64,20 +53,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenSamplers(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteSamplers(1, &handle); - OpenGLState::GetCurState().ResetSampler(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -98,20 +77,9 @@ public: return *this; } - void Create(const char* source, GLenum type) { - if (handle != 0) - return; - if (source == nullptr) - return; - handle = GLShader::LoadShader(source, type); - } + void Create(const char* source, GLenum type); - void Release() { - if (handle == 0) - return; - glDeleteShader(handle); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -141,25 +109,10 @@ public: /// Creates a new internal OpenGL resource and stores the handle void CreateFromSource(const char* vert_shader, const char* geo_shader, const char* frag_shader, - bool separable_program = false) { - OGLShader vert, geo, frag; - if (vert_shader) - vert.Create(vert_shader, GL_VERTEX_SHADER); - if (geo_shader) - geo.Create(geo_shader, GL_GEOMETRY_SHADER); - if (frag_shader) - frag.Create(frag_shader, GL_FRAGMENT_SHADER); - Create(separable_program, vert.handle, geo.handle, frag.handle); - } + bool separable_program = false); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteProgram(handle); - OpenGLState::GetCurState().ResetProgram(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -178,20 +131,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenProgramPipelines(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteProgramPipelines(1, &handle); - OpenGLState::GetCurState().ResetPipeline(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -213,20 +156,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenBuffers(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteBuffers(1, &handle); - OpenGLState::GetCurState().ResetBuffer(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -247,19 +180,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - handle = glFenceSync(GL_SYNC_GPU_COMMANDS_COMPLETE, 0); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteSync(handle); - handle = 0; - } + void Release(); GLsync handle = 0; }; @@ -281,20 +205,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenVertexArrays(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteVertexArrays(1, &handle); - OpenGLState::GetCurState().ResetVertexArray(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; @@ -316,20 +230,10 @@ public: } /// Creates a new internal OpenGL resource and stores the handle - void Create() { - if (handle != 0) - return; - glGenFramebuffers(1, &handle); - } + void Create(); /// Deletes the internal OpenGL resource - void Release() { - if (handle == 0) - return; - glDeleteFramebuffers(1, &handle); - OpenGLState::GetCurState().ResetFramebuffer(handle).Apply(); - handle = 0; - } + void Release(); GLuint handle = 0; }; diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h index 36fe1f04c..2a069cdd8 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.h +++ b/src/video_core/renderer_opengl/gl_shader_manager.h @@ -7,6 +7,7 @@ #include <glad/glad.h> #include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/renderer_opengl/gl_state.h" #include "video_core/renderer_opengl/maxwell_to_gl.h" namespace OpenGL::GLShader { diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp index b6b426f34..9517285e5 100644 --- a/src/video_core/renderer_opengl/gl_state.cpp +++ b/src/video_core/renderer_opengl/gl_state.cpp @@ -22,17 +22,15 @@ OpenGLState::OpenGLState() { depth.test_enabled = false; depth.test_func = GL_LESS; depth.write_mask = GL_TRUE; - depth.depth_range_near = 0.0f; - depth.depth_range_far = 1.0f; primitive_restart.enabled = false; primitive_restart.index = 0; - - color_mask.red_enabled = GL_TRUE; - color_mask.green_enabled = GL_TRUE; - color_mask.blue_enabled = GL_TRUE; - color_mask.alpha_enabled = GL_TRUE; - + for (auto& item : color_mask) { + item.red_enabled = GL_TRUE; + item.green_enabled = GL_TRUE; + item.blue_enabled = GL_TRUE; + item.alpha_enabled = GL_TRUE; + } stencil.test_enabled = false; auto reset_stencil = [](auto& config) { config.test_func = GL_ALWAYS; @@ -45,19 +43,33 @@ OpenGLState::OpenGLState() { }; reset_stencil(stencil.front); reset_stencil(stencil.back); - - blend.enabled = true; - blend.rgb_equation = GL_FUNC_ADD; - blend.a_equation = GL_FUNC_ADD; - blend.src_rgb_func = GL_ONE; - blend.dst_rgb_func = GL_ZERO; - blend.src_a_func = GL_ONE; - blend.dst_a_func = GL_ZERO; - blend.color.red = 0.0f; - blend.color.green = 0.0f; - blend.color.blue = 0.0f; - blend.color.alpha = 0.0f; - + for (auto& item : viewports) { + item.x = 0; + item.y = 0; + item.width = 0; + item.height = 0; + item.depth_range_near = 0.0f; + item.depth_range_far = 1.0f; + } + scissor.enabled = false; + scissor.x = 0; + scissor.y = 0; + scissor.width = 0; + scissor.height = 0; + for (auto& item : blend) { + item.enabled = true; + item.rgb_equation = GL_FUNC_ADD; + item.a_equation = GL_FUNC_ADD; + item.src_rgb_func = GL_ONE; + item.dst_rgb_func = GL_ZERO; + item.src_a_func = GL_ONE; + item.dst_a_func = GL_ZERO; + } + independant_blend.enabled = false; + blend_color.red = 0.0f; + blend_color.green = 0.0f; + blend_color.blue = 0.0f; + blend_color.alpha = 0.0f; logic_op.enabled = false; logic_op.operation = GL_COPY; @@ -73,17 +85,6 @@ OpenGLState::OpenGLState() { draw.shader_program = 0; draw.program_pipeline = 0; - scissor.enabled = false; - scissor.x = 0; - scissor.y = 0; - scissor.width = 0; - scissor.height = 0; - - viewport.x = 0; - viewport.y = 0; - viewport.width = 0; - viewport.height = 0; - clip_distance = {}; point.size = 1; @@ -134,6 +135,32 @@ void OpenGLState::ApplyCulling() const { } } +void OpenGLState::ApplyColorMask() const { + if (GLAD_GL_ARB_viewport_array) { + for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { + const auto& updated = color_mask[i]; + const auto& current = cur_state.color_mask[i]; + if (updated.red_enabled != current.red_enabled || + updated.green_enabled != current.green_enabled || + updated.blue_enabled != current.blue_enabled || + updated.alpha_enabled != current.alpha_enabled) { + glColorMaski(static_cast<GLuint>(i), updated.red_enabled, updated.green_enabled, + updated.blue_enabled, updated.alpha_enabled); + } + } + } else { + const auto& updated = color_mask[0]; + const auto& current = cur_state.color_mask[0]; + if (updated.red_enabled != current.red_enabled || + updated.green_enabled != current.green_enabled || + updated.blue_enabled != current.blue_enabled || + updated.alpha_enabled != current.alpha_enabled) { + glColorMask(updated.red_enabled, updated.green_enabled, updated.blue_enabled, + updated.alpha_enabled); + } + } +} + void OpenGLState::ApplyDepth() const { // Depth test const bool depth_test_changed = depth.test_enabled != cur_state.depth.test_enabled; @@ -152,11 +179,6 @@ void OpenGLState::ApplyDepth() const { if (depth.write_mask != cur_state.depth.write_mask) { glDepthMask(depth.write_mask); } - // Depth range - if (depth.depth_range_near != cur_state.depth.depth_range_near || - depth.depth_range_far != cur_state.depth.depth_range_far) { - glDepthRange(depth.depth_range_near, depth.depth_range_far); - } } void OpenGLState::ApplyPrimitiveRestart() const { @@ -208,7 +230,7 @@ void OpenGLState::ApplyStencilTest() const { } } -void OpenGLState::ApplyScissorTest() const { +void OpenGLState::ApplyScissor() const { const bool scissor_changed = scissor.enabled != cur_state.scissor.enabled; if (scissor_changed) { if (scissor.enabled) { @@ -217,51 +239,141 @@ void OpenGLState::ApplyScissorTest() const { glDisable(GL_SCISSOR_TEST); } } - if (scissor_changed || scissor_changed || scissor.x != cur_state.scissor.x || - scissor.y != cur_state.scissor.y || scissor.width != cur_state.scissor.width || - scissor.height != cur_state.scissor.height) { + if (scissor.enabled && + (scissor_changed || scissor.x != cur_state.scissor.x || scissor.y != cur_state.scissor.y || + scissor.width != cur_state.scissor.width || scissor.height != cur_state.scissor.height)) { glScissor(scissor.x, scissor.y, scissor.width, scissor.height); } } -void OpenGLState::ApplyBlending() const { - const bool blend_changed = blend.enabled != cur_state.blend.enabled; +void OpenGLState::ApplyViewport() const { + if (GLAD_GL_ARB_viewport_array) { + for (GLuint i = 0; + i < static_cast<GLuint>(Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); i++) { + const auto& current = cur_state.viewports[i]; + const auto& updated = viewports[i]; + if (updated.x != current.x || updated.y != current.y || + updated.width != current.width || updated.height != current.height) { + glViewportIndexedf(i, updated.x, updated.y, updated.width, updated.height); + } + if (updated.depth_range_near != current.depth_range_near || + updated.depth_range_far != current.depth_range_far) { + glDepthRangeIndexed(i, updated.depth_range_near, updated.depth_range_far); + } + } + } else { + const auto& current = cur_state.viewports[0]; + const auto& updated = viewports[0]; + if (updated.x != current.x || updated.y != current.y || updated.width != current.width || + updated.height != current.height) { + glViewport(updated.x, updated.y, updated.width, updated.height); + } + if (updated.depth_range_near != current.depth_range_near || + updated.depth_range_far != current.depth_range_far) { + glDepthRange(updated.depth_range_near, updated.depth_range_far); + } + } +} + +void OpenGLState::ApplyGlobalBlending() const { + const Blend& current = cur_state.blend[0]; + const Blend& updated = blend[0]; + const bool blend_changed = updated.enabled != current.enabled; if (blend_changed) { - if (blend.enabled) { - ASSERT(!logic_op.enabled); + if (updated.enabled) { glEnable(GL_BLEND); } else { glDisable(GL_BLEND); } } - if (blend.enabled) { - if (blend_changed || blend.color.red != cur_state.blend.color.red || - blend.color.green != cur_state.blend.color.green || - blend.color.blue != cur_state.blend.color.blue || - blend.color.alpha != cur_state.blend.color.alpha) { - glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha); + if (!updated.enabled) { + return; + } + if (updated.separate_alpha) { + if (blend_changed || updated.src_rgb_func != current.src_rgb_func || + updated.dst_rgb_func != current.dst_rgb_func || + updated.src_a_func != current.src_a_func || updated.dst_a_func != current.dst_a_func) { + glBlendFuncSeparate(updated.src_rgb_func, updated.dst_rgb_func, updated.src_a_func, + updated.dst_a_func); } - if (blend_changed || blend.src_rgb_func != cur_state.blend.src_rgb_func || - blend.dst_rgb_func != cur_state.blend.dst_rgb_func || - blend.src_a_func != cur_state.blend.src_a_func || - blend.dst_a_func != cur_state.blend.dst_a_func) { - glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func, - blend.dst_a_func); + if (blend_changed || updated.rgb_equation != current.rgb_equation || + updated.a_equation != current.a_equation) { + glBlendEquationSeparate(updated.rgb_equation, updated.a_equation); + } + } else { + if (blend_changed || updated.src_rgb_func != current.src_rgb_func || + updated.dst_rgb_func != current.dst_rgb_func) { + glBlendFunc(updated.src_rgb_func, updated.dst_rgb_func); } - if (blend_changed || blend.rgb_equation != cur_state.blend.rgb_equation || - blend.a_equation != cur_state.blend.a_equation) { - glBlendEquationSeparate(blend.rgb_equation, blend.a_equation); + if (blend_changed || updated.rgb_equation != current.rgb_equation) { + glBlendEquation(updated.rgb_equation); } } } +void OpenGLState::ApplyTargetBlending(int target, bool force) const { + const Blend& updated = blend[target]; + const Blend& current = cur_state.blend[target]; + const bool blend_changed = updated.enabled != current.enabled || force; + if (blend_changed) { + if (updated.enabled) { + glEnablei(GL_BLEND, static_cast<GLuint>(target)); + } else { + glDisablei(GL_BLEND, static_cast<GLuint>(target)); + } + } + if (!updated.enabled) { + return; + } + if (updated.separate_alpha) { + if (blend_changed || updated.src_rgb_func != current.src_rgb_func || + updated.dst_rgb_func != current.dst_rgb_func || + updated.src_a_func != current.src_a_func || updated.dst_a_func != current.dst_a_func) { + glBlendFuncSeparateiARB(static_cast<GLuint>(target), updated.src_rgb_func, + updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func); + } + + if (blend_changed || updated.rgb_equation != current.rgb_equation || + updated.a_equation != current.a_equation) { + glBlendEquationSeparateiARB(static_cast<GLuint>(target), updated.rgb_equation, + updated.a_equation); + } + } else { + if (blend_changed || updated.src_rgb_func != current.src_rgb_func || + updated.dst_rgb_func != current.dst_rgb_func) { + glBlendFunciARB(static_cast<GLuint>(target), updated.src_rgb_func, + updated.dst_rgb_func); + } + + if (blend_changed || updated.rgb_equation != current.rgb_equation) { + glBlendEquationiARB(static_cast<GLuint>(target), updated.rgb_equation); + } + } +} + +void OpenGLState::ApplyBlending() const { + if (independant_blend.enabled) { + for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { + ApplyTargetBlending(i, + independant_blend.enabled != cur_state.independant_blend.enabled); + } + } else { + ApplyGlobalBlending(); + } + if (blend_color.red != cur_state.blend_color.red || + blend_color.green != cur_state.blend_color.green || + blend_color.blue != cur_state.blend_color.blue || + blend_color.alpha != cur_state.blend_color.alpha) { + glBlendColor(blend_color.red, blend_color.green, blend_color.blue, blend_color.alpha); + } +} + void OpenGLState::ApplyLogicOp() const { const bool logic_op_changed = logic_op.enabled != cur_state.logic_op.enabled; if (logic_op_changed) { if (logic_op.enabled) { - ASSERT(!blend.enabled); glEnable(GL_COLOR_LOGIC_OP); } else { glDisable(GL_COLOR_LOGIC_OP); @@ -348,12 +460,6 @@ void OpenGLState::Apply() const { if (draw.program_pipeline != cur_state.draw.program_pipeline) { glBindProgramPipeline(draw.program_pipeline); } - // Viewport - if (viewport.x != cur_state.viewport.x || viewport.y != cur_state.viewport.y || - viewport.width != cur_state.viewport.width || - viewport.height != cur_state.viewport.height) { - glViewport(viewport.x, viewport.y, viewport.width, viewport.height); - } // Clip distance for (std::size_t i = 0; i < clip_distance.size(); ++i) { if (clip_distance[i] != cur_state.clip_distance[i]) { @@ -364,19 +470,13 @@ void OpenGLState::Apply() const { } } } - // Color mask - if (color_mask.red_enabled != cur_state.color_mask.red_enabled || - color_mask.green_enabled != cur_state.color_mask.green_enabled || - color_mask.blue_enabled != cur_state.color_mask.blue_enabled || - color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) { - glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled, - color_mask.alpha_enabled); - } // Point if (point.size != cur_state.point.size) { glPointSize(point.size); } - ApplyScissorTest(); + ApplyColorMask(); + ApplyViewport(); + ApplyScissor(); ApplyStencilTest(); ApplySRgb(); ApplyCulling(); diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h index fe648aff6..b8cf1f637 100644 --- a/src/video_core/renderer_opengl/gl_state.h +++ b/src/video_core/renderer_opengl/gl_state.h @@ -46,11 +46,9 @@ public: } cull; struct { - bool test_enabled; // GL_DEPTH_TEST - GLenum test_func; // GL_DEPTH_FUNC - GLboolean write_mask; // GL_DEPTH_WRITEMASK - GLfloat depth_range_near; // GL_DEPTH_RANGE - GLfloat depth_range_far; // GL_DEPTH_RANGE + bool test_enabled; // GL_DEPTH_TEST + GLenum test_func; // GL_DEPTH_FUNC + GLboolean write_mask; // GL_DEPTH_WRITEMASK } depth; struct { @@ -58,13 +56,14 @@ public: GLuint index; } primitive_restart; // GL_PRIMITIVE_RESTART - struct { + struct ColorMask { GLboolean red_enabled; GLboolean green_enabled; GLboolean blue_enabled; GLboolean alpha_enabled; - } color_mask; // GL_COLOR_WRITEMASK - + }; + std::array<ColorMask, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> + color_mask; // GL_COLOR_WRITEMASK struct { bool test_enabled; // GL_STENCIL_TEST struct { @@ -78,22 +77,28 @@ public: } front, back; } stencil; - struct { + struct Blend { bool enabled; // GL_BLEND + bool separate_alpha; // Independent blend enabled GLenum rgb_equation; // GL_BLEND_EQUATION_RGB GLenum a_equation; // GL_BLEND_EQUATION_ALPHA GLenum src_rgb_func; // GL_BLEND_SRC_RGB GLenum dst_rgb_func; // GL_BLEND_DST_RGB GLenum src_a_func; // GL_BLEND_SRC_ALPHA GLenum dst_a_func; // GL_BLEND_DST_ALPHA + }; + std::array<Blend, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> blend; - struct { - GLclampf red; - GLclampf green; - GLclampf blue; - GLclampf alpha; - } color; // GL_BLEND_COLOR - } blend; + struct { + bool enabled; + } independant_blend; + + struct { + GLclampf red; + GLclampf green; + GLclampf blue; + GLclampf alpha; + } blend_color; // GL_BLEND_COLOR struct { bool enabled; // GL_LOGIC_OP_MODE @@ -138,6 +143,16 @@ public: GLuint program_pipeline; // GL_PROGRAM_PIPELINE_BINDING } draw; + struct viewport { + GLfloat x; + GLfloat y; + GLfloat width; + GLfloat height; + GLfloat depth_range_near; // GL_DEPTH_RANGE + GLfloat depth_range_far; // GL_DEPTH_RANGE + }; + std::array<viewport, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> viewports; + struct { bool enabled; // GL_SCISSOR_TEST GLint x; @@ -147,13 +162,6 @@ public: } scissor; struct { - GLint x; - GLint y; - GLsizei width; - GLsizei height; - } viewport; - - struct { float size; // GL_POINT_SIZE } point; @@ -191,14 +199,18 @@ private: static bool s_rgb_used; void ApplySRgb() const; void ApplyCulling() const; + void ApplyColorMask() const; void ApplyDepth() const; void ApplyPrimitiveRestart() const; void ApplyStencilTest() const; - void ApplyScissorTest() const; + void ApplyViewport() const; + void ApplyTargetBlending(int target, bool force) const; + void ApplyGlobalBlending() const; void ApplyBlending() const; void ApplyLogicOp() const; void ApplyTextures() const; void ApplySamplers() const; + void ApplyScissor() const; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp index e409228cc..b97b895a4 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp +++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp @@ -6,9 +6,13 @@ #include <vector> #include "common/alignment.h" #include "common/assert.h" +#include "common/microprofile.h" #include "video_core/renderer_opengl/gl_state.h" #include "video_core/renderer_opengl/gl_stream_buffer.h" +MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning", + MP_RGB(128, 128, 192)); + namespace OpenGL { OGLStreamBuffer::OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coherent) @@ -75,6 +79,7 @@ std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr a } if (invalidate || !persistent) { + MICROPROFILE_SCOPE(OpenGL_StreamBuffer); GLbitfield flags = GL_MAP_WRITE_BIT | (persistent ? GL_MAP_PERSISTENT_BIT : 0) | (coherent ? GL_MAP_COHERENT_BIT : GL_MAP_FLUSH_EXPLICIT_BIT) | (invalidate ? GL_MAP_INVALIDATE_BUFFER_BIT : GL_MAP_UNSYNCHRONIZED_BIT); diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index d9a97e30b..051ad3964 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -19,6 +19,8 @@ SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_t return SurfaceTarget::Texture3D; case Tegra::Texture::TextureType::TextureCubemap: return SurfaceTarget::TextureCubemap; + case Tegra::Texture::TextureType::TextureCubeArray: + return SurfaceTarget::TextureCubeArray; case Tegra::Texture::TextureType::Texture1DArray: return SurfaceTarget::Texture1DArray; case Tegra::Texture::TextureType::Texture2DArray: @@ -39,6 +41,7 @@ bool SurfaceTargetIsLayered(SurfaceTarget target) { case SurfaceTarget::Texture1DArray: case SurfaceTarget::Texture2DArray: case SurfaceTarget::TextureCubemap: + case SurfaceTarget::TextureCubeArray: return true; default: LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target)); @@ -297,6 +300,8 @@ PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format, return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4; case Tegra::Texture::TextureFormat::ASTC_2D_5X4: return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4; + case Tegra::Texture::TextureFormat::ASTC_2D_5X5: + return is_srgb ? PixelFormat::ASTC_2D_5X5_SRGB : PixelFormat::ASTC_2D_5X5; case Tegra::Texture::TextureFormat::ASTC_2D_8X8: return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8; case Tegra::Texture::TextureFormat::ASTC_2D_8X5: @@ -440,10 +445,12 @@ bool IsPixelFormatASTC(PixelFormat format) { switch (format) { case PixelFormat::ASTC_2D_4X4: case PixelFormat::ASTC_2D_5X4: + case PixelFormat::ASTC_2D_5X5: case PixelFormat::ASTC_2D_8X8: case PixelFormat::ASTC_2D_8X5: case PixelFormat::ASTC_2D_4X4_SRGB: case PixelFormat::ASTC_2D_5X4_SRGB: + case PixelFormat::ASTC_2D_5X5_SRGB: case PixelFormat::ASTC_2D_8X8_SRGB: case PixelFormat::ASTC_2D_8X5_SRGB: return true; @@ -453,27 +460,7 @@ bool IsPixelFormatASTC(PixelFormat format) { } std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) { - switch (format) { - case PixelFormat::ASTC_2D_4X4: - return {4, 4}; - case PixelFormat::ASTC_2D_5X4: - return {5, 4}; - case PixelFormat::ASTC_2D_8X8: - return {8, 8}; - case PixelFormat::ASTC_2D_8X5: - return {8, 5}; - case PixelFormat::ASTC_2D_4X4_SRGB: - return {4, 4}; - case PixelFormat::ASTC_2D_5X4_SRGB: - return {5, 4}; - case PixelFormat::ASTC_2D_8X8_SRGB: - return {8, 8}; - case PixelFormat::ASTC_2D_8X5_SRGB: - return {8, 5}; - default: - LOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format)); - UNREACHABLE(); - } + return {GetDefaultBlockWidth(format), GetDefaultBlockHeight(format)}; } bool IsFormatBCn(PixelFormat format) { diff --git a/src/video_core/surface.h b/src/video_core/surface.h index 3232e437f..dfdb8d122 100644 --- a/src/video_core/surface.h +++ b/src/video_core/surface.h @@ -72,19 +72,21 @@ enum class PixelFormat { ASTC_2D_8X8_SRGB = 54, ASTC_2D_8X5_SRGB = 55, ASTC_2D_5X4_SRGB = 56, + ASTC_2D_5X5 = 57, + ASTC_2D_5X5_SRGB = 58, MaxColorFormat, // Depth formats - Z32F = 57, - Z16 = 58, + Z32F = 59, + Z16 = 60, MaxDepthFormat, // DepthStencil formats - Z24S8 = 59, - S8Z24 = 60, - Z32FS8 = 61, + Z24S8 = 61, + S8Z24 = 62, + Z32FS8 = 63, MaxDepthStencilFormat, @@ -118,6 +120,7 @@ enum class SurfaceTarget { Texture1DArray, Texture2DArray, TextureCubemap, + TextureCubeArray, }; /** @@ -188,6 +191,8 @@ static constexpr u32 GetCompressionFactor(PixelFormat format) { 4, // ASTC_2D_8X8_SRGB 4, // ASTC_2D_8X5_SRGB 4, // ASTC_2D_5X4_SRGB + 4, // ASTC_2D_5X5 + 4, // ASTC_2D_5X5_SRGB 1, // Z32F 1, // Z16 1, // Z24S8 @@ -199,6 +204,79 @@ static constexpr u32 GetCompressionFactor(PixelFormat format) { return compression_factor_table[static_cast<std::size_t>(format)]; } +static constexpr u32 GetDefaultBlockWidth(PixelFormat format) { + if (format == PixelFormat::Invalid) + return 0; + constexpr std::array<u32, MaxPixelFormat> block_width_table = {{ + 1, // ABGR8U + 1, // ABGR8S + 1, // ABGR8UI + 1, // B5G6R5U + 1, // A2B10G10R10U + 1, // A1B5G5R5U + 1, // R8U + 1, // R8UI + 1, // RGBA16F + 1, // RGBA16U + 1, // RGBA16UI + 1, // R11FG11FB10F + 1, // RGBA32UI + 4, // DXT1 + 4, // DXT23 + 4, // DXT45 + 4, // DXN1 + 4, // DXN2UNORM + 4, // DXN2SNORM + 4, // BC7U + 4, // BC6H_UF16 + 4, // BC6H_SF16 + 4, // ASTC_2D_4X4 + 1, // G8R8U + 1, // G8R8S + 1, // BGRA8 + 1, // RGBA32F + 1, // RG32F + 1, // R32F + 1, // R16F + 1, // R16U + 1, // R16S + 1, // R16UI + 1, // R16I + 1, // RG16 + 1, // RG16F + 1, // RG16UI + 1, // RG16I + 1, // RG16S + 1, // RGB32F + 1, // RGBA8_SRGB + 1, // RG8U + 1, // RG8S + 1, // RG32UI + 1, // R32UI + 8, // ASTC_2D_8X8 + 8, // ASTC_2D_8X5 + 5, // ASTC_2D_5X4 + 1, // BGRA8_SRGB + 4, // DXT1_SRGB + 4, // DXT23_SRGB + 4, // DXT45_SRGB + 4, // BC7U_SRGB + 4, // ASTC_2D_4X4_SRGB + 8, // ASTC_2D_8X8_SRGB + 8, // ASTC_2D_8X5_SRGB + 5, // ASTC_2D_5X4_SRGB + 5, // ASTC_2D_5X5 + 5, // ASTC_2D_5X5_SRGB + 1, // Z32F + 1, // Z16 + 1, // Z24S8 + 1, // S8Z24 + 1, // Z32FS8 + }}; + ASSERT(static_cast<std::size_t>(format) < block_width_table.size()); + return block_width_table[static_cast<std::size_t>(format)]; +} + static constexpr u32 GetDefaultBlockHeight(PixelFormat format) { if (format == PixelFormat::Invalid) return 0; @@ -261,6 +339,8 @@ static constexpr u32 GetDefaultBlockHeight(PixelFormat format) { 8, // ASTC_2D_8X8_SRGB 5, // ASTC_2D_8X5_SRGB 4, // ASTC_2D_5X4_SRGB + 5, // ASTC_2D_5X5 + 5, // ASTC_2D_5X5_SRGB 1, // Z32F 1, // Z16 1, // Z24S8 @@ -299,7 +379,7 @@ static constexpr u32 GetFormatBpp(PixelFormat format) { 128, // BC7U 128, // BC6H_UF16 128, // BC6H_SF16 - 32, // ASTC_2D_4X4 + 128, // ASTC_2D_4X4 16, // G8R8U 16, // G8R8S 32, // BGRA8 @@ -322,18 +402,20 @@ static constexpr u32 GetFormatBpp(PixelFormat format) { 16, // RG8S 64, // RG32UI 32, // R32UI - 16, // ASTC_2D_8X8 - 16, // ASTC_2D_8X5 - 32, // ASTC_2D_5X4 + 128, // ASTC_2D_8X8 + 128, // ASTC_2D_8X5 + 128, // ASTC_2D_5X4 32, // BGRA8_SRGB 64, // DXT1_SRGB 128, // DXT23_SRGB 128, // DXT45_SRGB 128, // BC7U - 32, // ASTC_2D_4X4_SRGB - 16, // ASTC_2D_8X8_SRGB - 16, // ASTC_2D_8X5_SRGB - 32, // ASTC_2D_5X4_SRGB + 128, // ASTC_2D_4X4_SRGB + 128, // ASTC_2D_8X8_SRGB + 128, // ASTC_2D_8X5_SRGB + 128, // ASTC_2D_5X4_SRGB + 128, // ASTC_2D_5X5 + 128, // ASTC_2D_5X5_SRGB 32, // Z32F 16, // Z16 32, // Z24S8 diff --git a/src/video_core/textures/astc.cpp b/src/video_core/textures/astc.cpp index b1feacae9..bc50a4876 100644 --- a/src/video_core/textures/astc.cpp +++ b/src/video_core/textures/astc.cpp @@ -1598,27 +1598,29 @@ static void DecompressBlock(uint8_t inBuf[16], const uint32_t blockWidth, namespace Tegra::Texture::ASTC { std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height, - uint32_t block_width, uint32_t block_height) { + uint32_t depth, uint32_t block_width, uint32_t block_height) { uint32_t blockIdx = 0; - std::vector<uint8_t> outData(height * width * 4); - for (uint32_t j = 0; j < height; j += block_height) { - for (uint32_t i = 0; i < width; i += block_width) { + std::vector<uint8_t> outData(height * width * depth * 4); + for (uint32_t k = 0; k < depth; k++) { + for (uint32_t j = 0; j < height; j += block_height) { + for (uint32_t i = 0; i < width; i += block_width) { - uint8_t* blockPtr = data.data() + blockIdx * 16; + uint8_t* blockPtr = data.data() + blockIdx * 16; - // Blocks can be at most 12x12 - uint32_t uncompData[144]; - ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData); + // Blocks can be at most 12x12 + uint32_t uncompData[144]; + ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData); - uint32_t decompWidth = std::min(block_width, width - i); - uint32_t decompHeight = std::min(block_height, height - j); + uint32_t decompWidth = std::min(block_width, width - i); + uint32_t decompHeight = std::min(block_height, height - j); - uint8_t* outRow = outData.data() + (j * width + i) * 4; - for (uint32_t jj = 0; jj < decompHeight; jj++) { - memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4); - } + uint8_t* outRow = outData.data() + (j * width + i) * 4; + for (uint32_t jj = 0; jj < decompHeight; jj++) { + memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4); + } - blockIdx++; + blockIdx++; + } } } diff --git a/src/video_core/textures/astc.h b/src/video_core/textures/astc.h index f0d7c0e56..d419dd025 100644 --- a/src/video_core/textures/astc.h +++ b/src/video_core/textures/astc.h @@ -10,6 +10,6 @@ namespace Tegra::Texture::ASTC { std::vector<uint8_t> Decompress(std::vector<uint8_t>& data, uint32_t width, uint32_t height, - uint32_t block_width, uint32_t block_height); + uint32_t depth, uint32_t block_width, uint32_t block_height); } // namespace Tegra::Texture::ASTC diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index 550ca856c..3066abf61 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -227,12 +227,14 @@ u32 BytesPerPixel(TextureFormat format) { } } -std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width, - u32 height, u32 depth, u32 block_height, u32 block_depth) { +std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y, + u32 bytes_per_pixel, u32 width, u32 height, u32 depth, + u32 block_height, u32 block_depth) { std::vector<u8> unswizzled_data(width * height * depth * bytes_per_pixel); - CopySwizzledData(width / tile_size, height / tile_size, depth, bytes_per_pixel, bytes_per_pixel, - Memory::GetPointer(address), unswizzled_data.data(), true, block_height, - block_depth); + CopySwizzledData((width + tile_size_x - 1) / tile_size_x, + (height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel, + bytes_per_pixel, Memory::GetPointer(address), unswizzled_data.data(), true, + block_height, block_depth); return unswizzled_data; } diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index b390219e4..ba065510b 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -19,8 +19,8 @@ inline std::size_t GetGOBSize() { /** * Unswizzles a swizzled texture without changing its format. */ -std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size, u32 bytes_per_pixel, u32 width, - u32 height, u32 depth, +std::vector<u8> UnswizzleTexture(VAddr address, u32 tile_size_x, u32 tile_size_y, + u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height = TICEntry::DefaultBlockHeight, u32 block_depth = TICEntry::DefaultBlockHeight); |