diff options
Diffstat (limited to '')
-rw-r--r-- | src/video_core/renderer_opengl/gl_rasterizer.cpp | 1 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/gl_shader_cache.cpp | 11 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/gl_shader_decompiler.cpp | 12 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/gl_state.cpp | 1 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/gl_state.h | 3 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/gl_texture_cache.cpp | 16 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/utils.cpp | 17 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/utils.h | 14 |
8 files changed, 43 insertions, 32 deletions
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 672051102..926bccd42 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -1272,6 +1272,7 @@ void RasterizerOpenGL::SyncPointState() { const auto& regs = system.GPU().Maxwell3D().regs; // Limit the point size to 1 since nouveau sometimes sets a point size of 0 (and that's invalid // in OpenGL). + state.point.program_control = regs.vp_point_size.enable ? GL_TRUE : GL_FALSE; state.point.size = std::max(1.0f, regs.point_size); } diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index de742d11c..a4acb3796 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -34,9 +34,6 @@ using VideoCommon::Shader::ShaderIR; namespace { -// One UBO is always reserved for emulation values on staged shaders -constexpr u32 STAGE_RESERVED_UBOS = 1; - constexpr u32 STAGE_MAIN_OFFSET = 10; constexpr u32 KERNEL_MAIN_OFFSET = 0; @@ -243,7 +240,6 @@ CachedProgram BuildShader(const Device& device, u64 unique_identifier, ShaderTyp if (!code_b.empty()) { ir_b.emplace(code_b, main_offset, COMPILER_SETTINGS, locker); } - const auto entries = GLShader::GetEntries(ir); std::string source = fmt::format(R"(// {} #version 430 core @@ -314,9 +310,10 @@ std::unordered_set<GLenum> GetSupportedFormats() { CachedShader::CachedShader(const ShaderParameters& params, ShaderType shader_type, GLShader::ShaderEntries entries, ProgramCode code, ProgramCode code_b) - : RasterizerCacheObject{params.host_ptr}, system{params.system}, disk_cache{params.disk_cache}, - device{params.device}, cpu_addr{params.cpu_addr}, unique_identifier{params.unique_identifier}, - shader_type{shader_type}, entries{entries}, code{std::move(code)}, code_b{std::move(code_b)} { + : RasterizerCacheObject{params.host_ptr}, system{params.system}, + disk_cache{params.disk_cache}, device{params.device}, cpu_addr{params.cpu_addr}, + unique_identifier{params.unique_identifier}, shader_type{shader_type}, + entries{std::move(entries)}, code{std::move(code)}, code_b{std::move(code_b)} { if (!params.precompiled_variants) { return; } diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index f9f7a97b5..19751939a 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -1856,6 +1856,16 @@ private: Type::Uint}; } + template <const std::string_view& opname, Type type> + Expression Atomic(Operation operation) { + ASSERT(stage == ShaderType::Compute); + auto& smem = std::get<SmemNode>(*operation[0]); + + return {fmt::format("atomic{}(smem[{} >> 2], {})", opname, Visit(smem.GetAddress()).AsInt(), + Visit(operation[1]).As(type)), + type}; + } + Expression Branch(Operation operation) { const auto target = std::get_if<ImmediateNode>(&*operation[0]); UNIMPLEMENTED_IF(!target); @@ -2194,6 +2204,8 @@ private: &GLSLDecompiler::AtomicImage<Func::Xor>, &GLSLDecompiler::AtomicImage<Func::Exchange>, + &GLSLDecompiler::Atomic<Func::Add, Type::Uint>, + &GLSLDecompiler::Branch, &GLSLDecompiler::BranchIndirect, &GLSLDecompiler::PushFlowStack, diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp index df2e2395a..cc185e9e1 100644 --- a/src/video_core/renderer_opengl/gl_state.cpp +++ b/src/video_core/renderer_opengl/gl_state.cpp @@ -127,6 +127,7 @@ void OpenGLState::ApplyClipDistances() { } void OpenGLState::ApplyPointSize() { + Enable(GL_PROGRAM_POINT_SIZE, cur_state.point.program_control, point.program_control); if (UpdateValue(cur_state.point.size, point.size)) { glPointSize(point.size); } diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h index fb180f302..71d418776 100644 --- a/src/video_core/renderer_opengl/gl_state.h +++ b/src/video_core/renderer_opengl/gl_state.h @@ -131,7 +131,8 @@ public: std::array<Viewport, Tegra::Engines::Maxwell3D::Regs::NumViewports> viewports; struct { - float size = 1.0f; // GL_POINT_SIZE + GLboolean program_control = GL_FALSE; // GL_PROGRAM_POINT_SIZE + GLfloat size = 1.0f; // GL_POINT_SIZE } point; struct { diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index b790b0ef4..e95eb069e 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -44,7 +44,7 @@ struct FormatTuple { constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{ {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, false}, // ABGR8U - {GL_RGBA8, GL_RGBA, GL_BYTE, false}, // ABGR8S + {GL_RGBA8_SNORM, GL_RGBA, GL_BYTE, false}, // ABGR8S {GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, false}, // ABGR8UI {GL_RGB565, GL_RGB, GL_UNSIGNED_SHORT_5_6_5_REV, false}, // B5G6R5U {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV, false}, // A2B10G10R10U @@ -83,9 +83,9 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format {GL_RGB32F, GL_RGB, GL_FLOAT, false}, // RGB32F {GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, false}, // RGBA8_SRGB {GL_RG8, GL_RG, GL_UNSIGNED_BYTE, false}, // RG8U - {GL_RG8, GL_RG, GL_BYTE, false}, // RG8S + {GL_RG8_SNORM, GL_RG, GL_BYTE, false}, // RG8S {GL_RG32UI, GL_RG_INTEGER, GL_UNSIGNED_INT, false}, // RG32UI - {GL_RGB16F, GL_RGBA16, GL_HALF_FLOAT, false}, // RGBX16F + {GL_RGB16F, GL_RGBA, GL_HALF_FLOAT, false}, // RGBX16F {GL_R32UI, GL_RED_INTEGER, GL_UNSIGNED_INT, false}, // R32UI {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, false}, // ASTC_2D_8X8 {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, false}, // ASTC_2D_8X5 @@ -253,14 +253,12 @@ void CachedSurface::DownloadTexture(std::vector<u8>& staging_buffer) { glPixelStorei(GL_PACK_ALIGNMENT, std::min(8U, params.GetRowAlignment(level))); glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.GetMipWidth(level))); const std::size_t mip_offset = params.GetHostMipmapLevelOffset(level); + u8* const mip_data = staging_buffer.data() + mip_offset; + const GLsizei size = static_cast<GLsizei>(params.GetHostMipmapSize(level)); if (is_compressed) { - glGetCompressedTextureImage(texture.handle, level, - static_cast<GLsizei>(params.GetHostMipmapSize(level)), - staging_buffer.data() + mip_offset); + glGetCompressedTextureImage(texture.handle, level, size, mip_data); } else { - glGetTextureImage(texture.handle, level, format, type, - static_cast<GLsizei>(params.GetHostMipmapSize(level)), - staging_buffer.data() + mip_offset); + glGetTextureImage(texture.handle, level, format, type, size, mip_data); } } } diff --git a/src/video_core/renderer_opengl/utils.cpp b/src/video_core/renderer_opengl/utils.cpp index 9770dda1c..ac99e6385 100644 --- a/src/video_core/renderer_opengl/utils.cpp +++ b/src/video_core/renderer_opengl/utils.cpp @@ -6,16 +6,20 @@ #include <vector> #include <fmt/format.h> - #include <glad/glad.h> -#include "common/assert.h" #include "common/common_types.h" -#include "common/scope_exit.h" #include "video_core/renderer_opengl/utils.h" namespace OpenGL { +struct VertexArrayPushBuffer::Entry { + GLuint binding_index{}; + const GLuint* buffer{}; + GLintptr offset{}; + GLsizei stride{}; +}; + VertexArrayPushBuffer::VertexArrayPushBuffer() = default; VertexArrayPushBuffer::~VertexArrayPushBuffer() = default; @@ -47,6 +51,13 @@ void VertexArrayPushBuffer::Bind() { } } +struct BindBuffersRangePushBuffer::Entry { + GLuint binding; + const GLuint* buffer; + GLintptr offset; + GLsizeiptr size; +}; + BindBuffersRangePushBuffer::BindBuffersRangePushBuffer(GLenum target) : target{target} {} BindBuffersRangePushBuffer::~BindBuffersRangePushBuffer() = default; diff --git a/src/video_core/renderer_opengl/utils.h b/src/video_core/renderer_opengl/utils.h index d56153fe7..3ad7c02d4 100644 --- a/src/video_core/renderer_opengl/utils.h +++ b/src/video_core/renderer_opengl/utils.h @@ -26,12 +26,7 @@ public: void Bind(); private: - struct Entry { - GLuint binding_index{}; - const GLuint* buffer{}; - GLintptr offset{}; - GLsizei stride{}; - }; + struct Entry; GLuint vao{}; const GLuint* index_buffer{}; @@ -50,12 +45,7 @@ public: void Bind(); private: - struct Entry { - GLuint binding; - const GLuint* buffer; - GLintptr offset; - GLsizeiptr size; - }; + struct Entry; GLenum target; std::vector<Entry> entries; |