Improve directory creation in WindowsCopyFiles.cmake

externals/cmake-modules/WindowsCopyFiles.cmake

@@ -1,4 +1,4 @@
-# Copyright 2016 Citra Emulator Project
+# Copyright 2018 Yuzu Emulator Project
 # Licensed under GPLv2 or any later version
 # Refer to the license.txt file included.
 
@@ -22,7 +22,7 @@ function(windows_copy_files TARGET SOURCE_DIR DEST_DIR)
     # cmake adds an extra check for command success which doesn't work too well with robocopy
     # so trick it into thinking the command was successful with the || cmd /c "exit /b 0"
     add_custom_command(TARGET ${TARGET} POST_BUILD
-        COMMAND if not exist ${DEST_DIR} mkdir ${DEST_DIR} 2> nul
+        COMMAND ${CMAKE_COMMAND} -E make_directory ${DEST_DIR}
         COMMAND robocopy ${SOURCE_DIR} ${DEST_DIR} ${ARGN} /NJH /NJS /NDL /NFL /NC /NS /NP || cmd /c "exit /b 0"
     )
-endfunction()
+endfunction()

src/core/hle/service/audio/audren_u.cpp

@@ -150,7 +150,7 @@ private:
         UpdateDataHeader() {}
 
         UpdateDataHeader(const AudioRendererParameter& config) {
-            revision = Common::MakeMagic('R', 'E', 'V', '4'); // 5.1.0 Revision
+            revision = config.revision;
             behavior_size = 0xb0;
             memory_pools_size = (config.effect_count + (config.voice_count * 4)) * 0x10;
             voices_size = config.voice_count * 0x10;

src/core/settings.h

@@ -129,7 +129,6 @@ struct Values {
     // Renderer
     float resolution_factor;
     bool toggle_framelimit;
-    bool use_accurate_framebuffers;
 
     float bg_red;
     float bg_green;

src/core/telemetry_session.cpp

@@ -161,8 +161,6 @@ TelemetrySession::TelemetrySession() {
              Settings::values.resolution_factor);
     AddField(Telemetry::FieldType::UserConfig, "Renderer_ToggleFramelimit",
              Settings::values.toggle_framelimit);
-    AddField(Telemetry::FieldType::UserConfig, "Renderer_UseAccurateFramebuffers",
-             Settings::values.use_accurate_framebuffers);
     AddField(Telemetry::FieldType::UserConfig, "System_UseDockedMode",
              Settings::values.use_docked_mode);
 }

src/video_core/debug_utils/debug_utils.h

@@ -55,10 +55,8 @@ public:
         virtual ~BreakPointObserver() {
             auto context = context_weak.lock();
             if (context) {
-                {
-                    std::unique_lock<std::mutex> lock(context->breakpoint_mutex);
-                    context->breakpoint_observers.remove(this);
-                }
+                std::unique_lock<std::mutex> lock(context->breakpoint_mutex);
+                context->breakpoint_observers.remove(this);
 
                 // If we are the last observer to be destroyed, tell the debugger context that
                 // it is free to continue. In particular, this is required for a proper yuzu

src/video_core/engines/maxwell_3d.h

@@ -321,24 +321,6 @@ public:
             INSERT_PADDING_WORDS(1);
         };
 
-        struct RenderTargetConfig {
-            u32 address_high;
-            u32 address_low;
-            u32 width;
-            u32 height;
-            Tegra::RenderTargetFormat format;
-            u32 block_dimensions;
-            u32 array_mode;
-            u32 layer_stride;
-            u32 base_layer;
-            INSERT_PADDING_WORDS(7);
-
-            GPUVAddr Address() const {
-                return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) |
-                                             address_low);
-            }
-        };
-
         union {
             struct {
                 INSERT_PADDING_WORDS(0x45);
@@ -351,7 +333,23 @@ public:
 
                 INSERT_PADDING_WORDS(0x1B8);
 
-                RenderTargetConfig rt[NumRenderTargets];
+                struct {
+                    u32 address_high;
+                    u32 address_low;
+                    u32 width;
+                    u32 height;
+                    Tegra::RenderTargetFormat format;
+                    u32 block_dimensions;
+                    u32 array_mode;
+                    u32 layer_stride;
+                    u32 base_layer;
+                    INSERT_PADDING_WORDS(7);
+
+                    GPUVAddr Address() const {
+                        return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) |
+                                                     address_low);
+                    }
+                } rt[NumRenderTargets];
 
                 struct {
                     f32 scale_x;

src/video_core/rasterizer_interface.h

@@ -51,8 +51,9 @@ public:
     }
 
     /// Attempt to use a faster method to display the framebuffer to screen
-    virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
-                                   u32 pixel_stride, ScreenInfo& screen_info) {
+    virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& framebuffer,
+                                   VAddr framebuffer_addr, u32 pixel_stride,
+                                   ScreenInfo& screen_info) {
         return false;
     }
 

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -146,6 +146,7 @@ std::pair<u8*, GLintptr> RasterizerOpenGL::SetupVertexArrays(u8* array_ptr,
         u64 size = end - start + 1;
 
         // Copy vertex array data
+        res_cache.FlushRegion(start, size, nullptr);
         Memory::ReadBlock(*memory_manager->GpuToCpuAddress(start), array_ptr, size);
 
         // Bind the vertex array to the buffer at the current offset.
@@ -196,8 +197,8 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
     ASSERT_MSG(!gpu.regs.shader_config[0].enable, "VertexA is unsupported!");
 
     // Next available bindpoints to use when uploading the const buffers and textures to the GLSL
-    // shaders. The constbuffer bindpoint starts after the shader stage configuration bind points.
-    u32 current_constbuffer_bindpoint = uniform_buffers.size();
+    // shaders.
+    u32 current_constbuffer_bindpoint = 0;
     u32 current_texture_bindpoint = 0;
 
     for (unsigned index = 1; index < Maxwell::MaxShaderProgram; ++index) {
@@ -324,22 +325,29 @@ void RasterizerOpenGL::DrawArrays() {
     std::tie(color_surface, depth_surface, surfaces_rect) =
         res_cache.GetFramebufferSurfaces(using_color_fb, using_depth_fb, viewport_rect);
 
+    const u16 res_scale = color_surface != nullptr
+                              ? color_surface->res_scale
+                              : (depth_surface == nullptr ? 1u : depth_surface->res_scale);
+
     MathUtil::Rectangle<u32> draw_rect{
-        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.left) + viewport_rect.left,
-                                         surfaces_rect.left, surfaces_rect.right)), // Left
-        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.bottom) + viewport_rect.top,
-                                         surfaces_rect.bottom, surfaces_rect.top)), // Top
-        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.left) + viewport_rect.right,
-                                         surfaces_rect.left, surfaces_rect.right)), // Right
         static_cast<u32>(
-            std::clamp<s32>(static_cast<s32>(surfaces_rect.bottom) + viewport_rect.bottom,
-                            surfaces_rect.bottom, surfaces_rect.top))}; // Bottom
+            std::clamp<s32>(static_cast<s32>(surfaces_rect.left) + viewport_rect.left * res_scale,
+                            surfaces_rect.left, surfaces_rect.right)), // Left
+        static_cast<u32>(
+            std::clamp<s32>(static_cast<s32>(surfaces_rect.bottom) + viewport_rect.top * res_scale,
+                            surfaces_rect.bottom, surfaces_rect.top)), // Top
+        static_cast<u32>(
+            std::clamp<s32>(static_cast<s32>(surfaces_rect.left) + viewport_rect.right * res_scale,
+                            surfaces_rect.left, surfaces_rect.right)), // Right
+        static_cast<u32>(std::clamp<s32>(static_cast<s32>(surfaces_rect.bottom) +
+                                             viewport_rect.bottom * res_scale,
+                                         surfaces_rect.bottom, surfaces_rect.top))}; // Bottom
 
     // Bind the framebuffer surfaces
     BindFramebufferSurfaces(color_surface, depth_surface, has_stencil);
 
     // Sync the viewport
-    SyncViewport(surfaces_rect);
+    SyncViewport(surfaces_rect, res_scale);
 
     // Sync the blend state registers
     SyncBlendState();
@@ -429,16 +437,24 @@ void RasterizerOpenGL::DrawArrays() {
 
     // Unbind textures for potential future use as framebuffer attachments
     for (auto& texture_unit : state.texture_units) {
-        texture_unit.Unbind();
+        texture_unit.texture_2d = 0;
     }
     state.Apply();
 
     // Mark framebuffer surfaces as dirty
+    MathUtil::Rectangle<u32> draw_rect_unscaled{
+        draw_rect.left / res_scale, draw_rect.top / res_scale, draw_rect.right / res_scale,
+        draw_rect.bottom / res_scale};
+
     if (color_surface != nullptr && write_color_fb) {
-        res_cache.MarkSurfaceAsDirty(color_surface);
+        auto interval = color_surface->GetSubRectInterval(draw_rect_unscaled);
+        res_cache.InvalidateRegion(boost::icl::first(interval), boost::icl::length(interval),
+                                   color_surface);
     }
     if (depth_surface != nullptr && write_depth_fb) {
-        res_cache.MarkSurfaceAsDirty(depth_surface);
+        auto interval = depth_surface->GetSubRectInterval(draw_rect_unscaled);
+        res_cache.InvalidateRegion(boost::icl::first(interval), boost::icl::length(interval),
+                                   depth_surface);
     }
 }
 
@@ -446,7 +462,7 @@ void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {}
 
 void RasterizerOpenGL::FlushAll() {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
-    res_cache.FlushRegion(0, Kernel::VMManager::MAX_ADDRESS);
+    res_cache.FlushAll();
 }
 
 void RasterizerOpenGL::FlushRegion(Tegra::GPUVAddr addr, u64 size) {
@@ -456,13 +472,13 @@ void RasterizerOpenGL::FlushRegion(Tegra::GPUVAddr addr, u64 size) {
 
 void RasterizerOpenGL::InvalidateRegion(Tegra::GPUVAddr addr, u64 size) {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
-    res_cache.InvalidateRegion(addr, size);
+    res_cache.InvalidateRegion(addr, size, nullptr);
 }
 
 void RasterizerOpenGL::FlushAndInvalidateRegion(Tegra::GPUVAddr addr, u64 size) {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
     res_cache.FlushRegion(addr, size);
-    res_cache.InvalidateRegion(addr, size);
+    res_cache.InvalidateRegion(addr, size, nullptr);
 }
 
 bool RasterizerOpenGL::AccelerateDisplayTransfer(const void* config) {
@@ -481,28 +497,45 @@ bool RasterizerOpenGL::AccelerateFill(const void* config) {
     return true;
 }
 
-bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
+bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& framebuffer,
                                          VAddr framebuffer_addr, u32 pixel_stride,
                                          ScreenInfo& screen_info) {
-    if (!framebuffer_addr) {
-        return {};
+    if (framebuffer_addr == 0) {
+        return false;
     }
-
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
 
-    const auto& surface{res_cache.TryFindFramebufferSurface(framebuffer_addr)};
-    if (!surface) {
-        return {};
+    SurfaceParams src_params;
+    src_params.cpu_addr = framebuffer_addr;
+    src_params.addr = res_cache.TryFindFramebufferGpuAddress(framebuffer_addr).get_value_or(0);
+    src_params.width = std::min(framebuffer.width, pixel_stride);
+    src_params.height = framebuffer.height;
+    src_params.stride = pixel_stride;
+    src_params.is_tiled = true;
+    src_params.block_height = Tegra::Texture::TICEntry::DefaultBlockHeight;
+    src_params.pixel_format =
+        SurfaceParams::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format);
+    src_params.component_type =
+        SurfaceParams::ComponentTypeFromGPUPixelFormat(framebuffer.pixel_format);
+    src_params.UpdateParams();
+
+    MathUtil::Rectangle<u32> src_rect;
+    Surface src_surface;
+    std::tie(src_surface, src_rect) =
+        res_cache.GetSurfaceSubRect(src_params, ScaleMatch::Ignore, true);
+
+    if (src_surface == nullptr) {
+        return false;
     }
 
-    // Verify that the cached surface is the same size and format as the requested framebuffer
-    const auto& params{surface->GetSurfaceParams()};
-    const auto& pixel_format{SurfaceParams::PixelFormatFromGPUPixelFormat(config.pixel_format)};
-    ASSERT_MSG(params.width == config.width, "Framebuffer width is different");
-    ASSERT_MSG(params.height == config.height, "Framebuffer height is different");
-    ASSERT_MSG(params.pixel_format == pixel_format, "Framebuffer pixel_format is different");
+    u32 scaled_width = src_surface->GetScaledWidth();
+    u32 scaled_height = src_surface->GetScaledHeight();
 
-    screen_info.display_texture = surface->Texture().handle;
+    screen_info.display_texcoords = MathUtil::Rectangle<float>(
+        (float)src_rect.bottom / (float)scaled_height, (float)src_rect.left / (float)scaled_width,
+        (float)src_rect.top / (float)scaled_height, (float)src_rect.right / (float)scaled_width);
+
+    screen_info.display_texture = src_surface->texture.handle;
 
     return true;
 }
@@ -575,39 +608,27 @@ u32 RasterizerOpenGL::SetupConstBuffers(Maxwell::ShaderStage stage, GLuint progr
 
         boost::optional<VAddr> addr = gpu.memory_manager->GpuToCpuAddress(buffer.address);
 
-        size_t size = 0;
-
+        std::vector<u8> data;
         if (used_buffer.IsIndirect()) {
             // Buffer is accessed indirectly, so upload the entire thing
-            size = buffer.size * sizeof(float);
-
-            if (size > MaxConstbufferSize) {
-                NGLOG_ERROR(HW_GPU, "indirect constbuffer size {} exceeds maximum {}", size,
-                            MaxConstbufferSize);
-                size = MaxConstbufferSize;
-            }
+            data.resize(buffer.size * sizeof(float));
         } else {
             // Buffer is accessed directly, upload just what we use
-            size = used_buffer.GetSize() * sizeof(float);
+            data.resize(used_buffer.GetSize() * sizeof(float));
         }
 
-        // Align the actual size so it ends up being a multiple of vec4 to meet the OpenGL std140
-        // UBO alignment requirements.
-        size = Common::AlignUp(size, sizeof(GLvec4));
-        ASSERT_MSG(size <= MaxConstbufferSize, "Constbuffer too big");
-
-        std::vector<u8> data(size);
         Memory::ReadBlock(*addr, data.data(), data.size());
 
-        glBindBuffer(GL_UNIFORM_BUFFER, buffer_draw_state.ssbo);
-        glBufferData(GL_UNIFORM_BUFFER, data.size(), data.data(), GL_DYNAMIC_DRAW);
-        glBindBuffer(GL_UNIFORM_BUFFER, 0);
+        glBindBuffer(GL_SHADER_STORAGE_BUFFER, buffer_draw_state.ssbo);
+        glBufferData(GL_SHADER_STORAGE_BUFFER, data.size(), data.data(), GL_DYNAMIC_DRAW);
+        glBindBuffer(GL_SHADER_STORAGE_BUFFER, 0);
 
         // Now configure the bindpoint of the buffer inside the shader
         std::string buffer_name = used_buffer.GetName();
-        GLuint index = glGetProgramResourceIndex(program, GL_UNIFORM_BLOCK, buffer_name.c_str());
+        GLuint index =
+            glGetProgramResourceIndex(program, GL_SHADER_STORAGE_BLOCK, buffer_name.c_str());
         if (index != -1)
-            glUniformBlockBinding(program, index, buffer_draw_state.bindpoint);
+            glShaderStorageBlockBinding(program, index, buffer_draw_state.bindpoint);
     }
 
     state.Apply();
@@ -641,7 +662,7 @@ u32 RasterizerOpenGL::SetupTextures(Maxwell::ShaderStage stage, GLuint program,
         texture_samplers[current_bindpoint].SyncWithConfig(texture.tsc);
         Surface surface = res_cache.GetTextureSurface(texture);
         if (surface != nullptr) {
-            state.texture_units[current_bindpoint].texture_2d = surface->Texture().handle;
+            state.texture_units[current_bindpoint].texture_2d = surface->texture.handle;
             state.texture_units[current_bindpoint].swizzle.r =
                 MaxwellToGL::SwizzleSource(texture.tic.x_source);
             state.texture_units[current_bindpoint].swizzle.g =
@@ -667,16 +688,16 @@ void RasterizerOpenGL::BindFramebufferSurfaces(const Surface& color_surface,
     state.Apply();
 
     glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D,
-                           color_surface != nullptr ? color_surface->Texture().handle : 0, 0);
+                           color_surface != nullptr ? color_surface->texture.handle : 0, 0);
     if (depth_surface != nullptr) {
         if (has_stencil) {
             // attach both depth and stencil
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
-                                   depth_surface->Texture().handle, 0);
+                                   depth_surface->texture.handle, 0);
         } else {
             // attach depth
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D,
-                                   depth_surface->Texture().handle, 0);
+                                   depth_surface->texture.handle, 0);
             // clear stencil attachment
             glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0);
         }
@@ -687,14 +708,14 @@ void RasterizerOpenGL::BindFramebufferSurfaces(const Surface& color_surface,
     }
 }
 
-void RasterizerOpenGL::SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect) {
+void RasterizerOpenGL::SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect, u16 res_scale) {
     const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
     const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[0].GetRect()};
 
-    state.viewport.x = static_cast<GLint>(surfaces_rect.left) + viewport_rect.left;
-    state.viewport.y = static_cast<GLint>(surfaces_rect.bottom) + viewport_rect.bottom;
-    state.viewport.width = static_cast<GLsizei>(viewport_rect.GetWidth());
-    state.viewport.height = static_cast<GLsizei>(viewport_rect.GetHeight());
+    state.viewport.x = static_cast<GLint>(surfaces_rect.left) + viewport_rect.left * res_scale;
+    state.viewport.y = static_cast<GLint>(surfaces_rect.bottom) + viewport_rect.bottom * res_scale;
+    state.viewport.width = static_cast<GLsizei>(viewport_rect.GetWidth() * res_scale);
+    state.viewport.height = static_cast<GLsizei>(viewport_rect.GetHeight() * res_scale);
 }
 
 void RasterizerOpenGL::SyncClipEnabled() {

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -54,11 +54,6 @@ public:
         OGLShader shader;
     };
 
-    /// Maximum supported size that a constbuffer can have in bytes.
-    static constexpr size_t MaxConstbufferSize = 0x10000;
-    static_assert(MaxConstbufferSize % sizeof(GLvec4) == 0,
-                  "The maximum size of a constbuffer must be a multiple of the size of GLvec4");
-
 private:
     class SamplerInfo {
     public:
@@ -109,7 +104,7 @@ private:
                       u32 current_unit, const std::vector<GLShader::SamplerEntry>& entries);
 
     /// Syncs the viewport to match the guest state
-    void SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect);
+    void SyncViewport(const MathUtil::Rectangle<u32>& surfaces_rect, u16 res_scale);
 
     /// Syncs the clip enabled status to match the guest state
     void SyncClipEnabled();

src/video_core/renderer_opengl/gl_rasterizer_cache.h

@@ -1,26 +1,57 @@
-// Copyright 2018 yuzu Emulator Project
+// Copyright 2015 Citra Emulator Project
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
 #pragma once
 
 #include <array>
-#include <map>
 #include <memory>
-#include <vector>
+#include <set>
+#include <tuple>
+#ifdef __GNUC__
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wunused-local-typedefs"
+#endif
 #include <boost/icl/interval_map.hpp>
+#include <boost/icl/interval_set.hpp>
+#ifdef __GNUC__
+#pragma GCC diagnostic pop
+#endif
+#include <boost/optional.hpp>
+#include <glad/glad.h>
+#include "common/assert.h"
+#include "common/common_funcs.h"
 #include "common/common_types.h"
-#include "common/hash.h"
 #include "common/math_util.h"
-#include "video_core/engines/maxwell_3d.h"
+#include "video_core/gpu.h"
+#include "video_core/memory_manager.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/textures/texture.h"
 
-class CachedSurface;
+struct CachedSurface;
 using Surface = std::shared_ptr<CachedSurface>;
+using SurfaceSet = std::set<Surface>;
+
+using SurfaceRegions = boost::icl::interval_set<Tegra::GPUVAddr>;
+using SurfaceMap = boost::icl::interval_map<Tegra::GPUVAddr, Surface>;
+using SurfaceCache = boost::icl::interval_map<Tegra::GPUVAddr, SurfaceSet>;
+
+using SurfaceInterval = SurfaceCache::interval_type;
+static_assert(std::is_same<SurfaceRegions::interval_type, SurfaceCache::interval_type>() &&
+                  std::is_same<SurfaceMap::interval_type, SurfaceCache::interval_type>(),
+              "incorrect interval types");
+
+using SurfaceRect_Tuple = std::tuple<Surface, MathUtil::Rectangle<u32>>;
 using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
+
 using PageMap = boost::icl::interval_map<u64, int>;
 
+enum class ScaleMatch {
+    Exact,   // only accept same res scale
+    Upscale, // only allow higher scale than params
+    Ignore   // accept every scaled res
+};
+
 struct SurfaceParams {
     enum class PixelFormat {
         ABGR8 = 0,
@@ -62,10 +93,10 @@ struct SurfaceParams {
     /**
      * Gets the compression factor for the specified PixelFormat. This applies to just the
      * "compressed width" and "compressed height", not the overall compression factor of a
-     * compressed image. This is used for maintaining proper surface sizes for compressed
-     * texture formats.
+     * compressed image. This is used for maintaining proper surface sizes for compressed texture
+     * formats.
      */
-    static constexpr u32 GetCompressionFactor(PixelFormat format) {
+    static constexpr u32 GetCompresssionFactor(PixelFormat format) {
         if (format == PixelFormat::Invalid)
             return 0;
 
@@ -81,12 +112,15 @@ struct SurfaceParams {
             4, // DXT23
             4, // DXT45
             4, // DXN1
-            4, // ASTC_2D_4X4
+            1, // ASTC_2D_4X4
         }};
 
         ASSERT(static_cast<size_t>(format) < compression_factor_table.size());
         return compression_factor_table[static_cast<size_t>(format)];
     }
+    u32 GetCompresssionFactor() const {
+        return GetCompresssionFactor(pixel_format);
+    }
 
     static constexpr u32 GetFormatBpp(PixelFormat format) {
         if (format == PixelFormat::Invalid)
@@ -131,6 +165,25 @@ struct SurfaceParams {
         }
     }
 
+    static bool IsFormatASTC(PixelFormat format) {
+        switch (format) {
+        case PixelFormat::ASTC_2D_4X4:
+            return true;
+        default:
+            return false;
+        }
+    }
+
+    static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
+        switch (format) {
+        case Tegra::FramebufferConfig::PixelFormat::ABGR8:
+            return PixelFormat::ABGR8;
+        default:
+            NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
+            UNREACHABLE();
+        }
+    }
+
     static PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format) {
         // TODO(Subv): Properly implement this
         switch (format) {
@@ -223,16 +276,36 @@ struct SurfaceParams {
         }
     }
 
-    static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
+    static ComponentType ComponentTypeFromGPUPixelFormat(
+        Tegra::FramebufferConfig::PixelFormat format) {
         switch (format) {
         case Tegra::FramebufferConfig::PixelFormat::ABGR8:
-            return PixelFormat::ABGR8;
+            return ComponentType::UNorm;
         default:
             NGLOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
             UNREACHABLE();
         }
     }
 
+    static bool CheckFormatsBlittable(PixelFormat pixel_format_a, PixelFormat pixel_format_b) {
+        SurfaceType a_type = GetFormatType(pixel_format_a);
+        SurfaceType b_type = GetFormatType(pixel_format_b);
+
+        if (a_type == SurfaceType::ColorTexture && b_type == SurfaceType::ColorTexture) {
+            return true;
+        }
+
+        if (a_type == SurfaceType::Depth && b_type == SurfaceType::Depth) {
+            return true;
+        }
+
+        if (a_type == SurfaceType::DepthStencil && b_type == SurfaceType::DepthStencil) {
+            return true;
+        }
+
+        return false;
+    }
+
     static SurfaceType GetFormatType(PixelFormat pixel_format) {
         if (static_cast<size_t>(pixel_format) < MaxPixelFormat) {
             return SurfaceType::ColorTexture;
@@ -244,101 +317,168 @@ struct SurfaceParams {
         return SurfaceType::Invalid;
     }
 
-    /// Returns the rectangle corresponding to this surface
-    MathUtil::Rectangle<u32> GetRect() const;
-
-    /// Returns the size of this surface in bytes, adjusted for compression
-    size_t SizeInBytes() const {
-        const u32 compression_factor{GetCompressionFactor(pixel_format)};
-        ASSERT(width % compression_factor == 0);
-        ASSERT(height % compression_factor == 0);
-        return (width / compression_factor) * (height / compression_factor) *
-               GetFormatBpp(pixel_format) / CHAR_BIT;
+    /// Update the params "size", "end" and "type" from the already set "addr", "width", "height"
+    /// and "pixel_format"
+    void UpdateParams() {
+        if (stride == 0) {
+            stride = width;
+        }
+        type = GetFormatType(pixel_format);
+        size = !is_tiled ? BytesInPixels(stride * (height - 1) + width)
+                         : BytesInPixels(stride * 8 * (height / 8 - 1) + width * 8);
+        end = addr + size;
+    }
+
+    SurfaceInterval GetInterval() const {
+        return SurfaceInterval::right_open(addr, end);
+    }
+
+    // Returns the outer rectangle containing "interval"
+    SurfaceParams FromInterval(SurfaceInterval interval) const;
+
+    SurfaceInterval GetSubRectInterval(MathUtil::Rectangle<u32> unscaled_rect) const;
+
+    // Returns the region of the biggest valid rectange within interval
+    SurfaceInterval GetCopyableInterval(const Surface& src_surface) const;
+
+    /**
+     * Gets the actual width (in pixels) of the surface. This is provided because `width` is used
+     * for tracking the surface region in memory, which may be compressed for certain formats. In
+     * this scenario, `width` is actually the compressed width.
+     */
+    u32 GetActualWidth() const {
+        return width * GetCompresssionFactor();
+    }
+
+    /**
+     * Gets the actual height (in pixels) of the surface. This is provided because `height` is used
+     * for tracking the surface region in memory, which may be compressed for certain formats. In
+     * this scenario, `height` is actually the compressed height.
+     */
+    u32 GetActualHeight() const {
+        return height * GetCompresssionFactor();
+    }
+
+    u32 GetScaledWidth() const {
+        return width * res_scale;
+    }
+
+    u32 GetScaledHeight() const {
+        return height * res_scale;
+    }
+
+    MathUtil::Rectangle<u32> GetRect() const {
+        return {0, height, width, 0};
+    }
+
+    MathUtil::Rectangle<u32> GetScaledRect() const {
+        return {0, GetScaledHeight(), GetScaledWidth(), 0};
+    }
+
+    u64 PixelsInBytes(u64 size) const {
+        return size * CHAR_BIT / GetFormatBpp(pixel_format);
+    }
+
+    u64 BytesInPixels(u64 pixels) const {
+        return pixels * GetFormatBpp(pixel_format) / CHAR_BIT;
     }
 
-    /// Returns the CPU virtual address for this surface
     VAddr GetCpuAddr() const;
 
-    /// Returns true if the specified region overlaps with this surface's region in Switch memory
-    bool IsOverlappingRegion(Tegra::GPUVAddr region_addr, size_t region_size) const {
-        return addr <= (region_addr + region_size) && region_addr <= (addr + size_in_bytes);
-    }
+    bool ExactMatch(const SurfaceParams& other_surface) const;
+    bool CanSubRect(const SurfaceParams& sub_surface) const;
+    bool CanExpand(const SurfaceParams& expanded_surface) const;
+    bool CanTexCopy(const SurfaceParams& texcopy_params) const;
 
-    /// Creates SurfaceParams from a texture configation
-    static SurfaceParams CreateForTexture(const Tegra::Texture::FullTextureInfo& config);
+    MathUtil::Rectangle<u32> GetSubRect(const SurfaceParams& sub_surface) const;
+    MathUtil::Rectangle<u32> GetScaledSubRect(const SurfaceParams& sub_surface) const;
 
-    /// Creates SurfaceParams from a framebuffer configation
-    static SurfaceParams CreateForFramebuffer(
-        const Tegra::Engines::Maxwell3D::Regs::RenderTargetConfig& config);
+    Tegra::GPUVAddr addr = 0;
+    Tegra::GPUVAddr end = 0;
+    boost::optional<VAddr> cpu_addr;
+    u64 size = 0;
 
-    Tegra::GPUVAddr addr;
-    bool is_tiled;
-    u32 block_height;
-    PixelFormat pixel_format;
-    ComponentType component_type;
-    SurfaceType type;
-    u32 width;
-    u32 height;
-    u32 unaligned_height;
-    size_t size_in_bytes;
+    u32 width = 0;
+    u32 height = 0;
+    u32 stride = 0;
+    u32 block_height = 0;
+    u16 res_scale = 1;
+
+    bool is_tiled = false;
+    PixelFormat pixel_format = PixelFormat::Invalid;
+    SurfaceType type = SurfaceType::Invalid;
+    ComponentType component_type = ComponentType::Invalid;
 };
 
-/// Hashable variation of SurfaceParams, used for a key in the surface cache
-struct SurfaceKey : Common::HashableStruct<SurfaceParams> {
-    static SurfaceKey Create(const SurfaceParams& params) {
-        SurfaceKey res;
-        res.state = params;
-        return res;
-    }
-};
+struct CachedSurface : SurfaceParams {
+    bool CanFill(const SurfaceParams& dest_surface, SurfaceInterval fill_interval) const;
+    bool CanCopy(const SurfaceParams& dest_surface, SurfaceInterval copy_interval) const;
 
-namespace std {
-template <>
-struct hash<SurfaceKey> {
-    size_t operator()(const SurfaceKey& k) const {
-        return k.Hash();
-    }
-};
-} // namespace std
-
-class CachedSurface final {
-public:
-    CachedSurface(const SurfaceParams& params);
-
-    const OGLTexture& Texture() const {
-        return texture;
+    bool IsRegionValid(SurfaceInterval interval) const {
+        return (invalid_regions.find(interval) == invalid_regions.end());
     }
 
-    static constexpr unsigned int GetGLBytesPerPixel(SurfaceParams::PixelFormat format) {
-        if (format == SurfaceParams::PixelFormat::Invalid)
+    bool IsSurfaceFullyInvalid() const {
+        return (invalid_regions & GetInterval()) == SurfaceRegions(GetInterval());
+    }
+
+    bool registered = false;
+    SurfaceRegions invalid_regions;
+
+    u64 fill_size = 0; /// Number of bytes to read from fill_data
+    std::array<u8, 4> fill_data;
+
+    OGLTexture texture;
+
+    static constexpr unsigned int GetGLBytesPerPixel(PixelFormat format) {
+        if (format == PixelFormat::Invalid)
             return 0;
 
         return SurfaceParams::GetFormatBpp(format) / CHAR_BIT;
     }
 
-    const SurfaceParams& GetSurfaceParams() const {
-        return params;
-    }
+    std::unique_ptr<u8[]> gl_buffer;
+    size_t gl_buffer_size = 0;
 
     // Read/Write data in Switch memory to/from gl_buffer
-    void LoadGLBuffer();
-    void FlushGLBuffer();
+    void LoadGLBuffer(Tegra::GPUVAddr load_start, Tegra::GPUVAddr load_end);
+    void FlushGLBuffer(Tegra::GPUVAddr flush_start, Tegra::GPUVAddr flush_end);
 
     // Upload/Download data in gl_buffer in/to this surface's texture
-    void UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle);
-    void DownloadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle);
-
-private:
-    OGLTexture texture;
-    std::vector<u8> gl_buffer;
-    SurfaceParams params;
+    void UploadGLTexture(const MathUtil::Rectangle<u32>& rect, GLuint read_fb_handle,
+                         GLuint draw_fb_handle);
+    void DownloadGLTexture(const MathUtil::Rectangle<u32>& rect, GLuint read_fb_handle,
+                           GLuint draw_fb_handle);
 };
 
-class RasterizerCacheOpenGL final : NonCopyable {
+class RasterizerCacheOpenGL : NonCopyable {
 public:
     RasterizerCacheOpenGL();
     ~RasterizerCacheOpenGL();
 
+    /// Blit one surface's texture to another
+    bool BlitSurfaces(const Surface& src_surface, const MathUtil::Rectangle<u32>& src_rect,
+                      const Surface& dst_surface, const MathUtil::Rectangle<u32>& dst_rect);
+
+    void ConvertD24S8toABGR(GLuint src_tex, const MathUtil::Rectangle<u32>& src_rect,
+                            GLuint dst_tex, const MathUtil::Rectangle<u32>& dst_rect);
+
+    /// Copy one surface's region to another
+    void CopySurface(const Surface& src_surface, const Surface& dst_surface,
+                     SurfaceInterval copy_interval);
+
+    /// Load a texture from Switch memory to OpenGL and cache it (if not already cached)
+    Surface GetSurface(const SurfaceParams& params, ScaleMatch match_res_scale,
+                       bool load_if_create);
+
+    /// Tries to find a framebuffer GPU address based on the provided CPU address
+    boost::optional<Tegra::GPUVAddr> TryFindFramebufferGpuAddress(VAddr cpu_addr) const;
+
+    /// Attempt to find a subrect (resolution scaled) of a surface, otherwise loads a texture from
+    /// Switch memory to OpenGL and caches it (if not already cached)
+    SurfaceRect_Tuple GetSurfaceSubRect(const SurfaceParams& params, ScaleMatch match_res_scale,
+                                        bool load_if_create);
+
     /// Get a surface based on the texture configuration
     Surface GetTextureSurface(const Tegra::Texture::FullTextureInfo& config);
 
@@ -346,21 +486,29 @@ public:
     SurfaceSurfaceRect_Tuple GetFramebufferSurfaces(bool using_color_fb, bool using_depth_fb,
                                                     const MathUtil::Rectangle<s32>& viewport);
 
-    /// Marks the specified surface as "dirty", in that it is out of sync with Switch memory
-    void MarkSurfaceAsDirty(const Surface& surface);
+    /// Get a surface that matches the fill config
+    Surface GetFillSurface(const void* config);
 
-    /// Tries to find a framebuffer GPU address based on the provided CPU address
-    Surface TryFindFramebufferSurface(VAddr cpu_addr) const;
+    /// Get a surface that matches a "texture copy" display transfer config
+    SurfaceRect_Tuple GetTexCopySurface(const SurfaceParams& params);
 
     /// Write any cached resources overlapping the region back to memory (if dirty)
-    void FlushRegion(Tegra::GPUVAddr addr, size_t size);
+    void FlushRegion(Tegra::GPUVAddr addr, u64 size, Surface flush_surface = nullptr);
 
-    /// Mark the specified region as being invalidated
-    void InvalidateRegion(Tegra::GPUVAddr addr, size_t size);
+    /// Mark region as being invalidated by region_owner (nullptr if Switch memory)
+    void InvalidateRegion(Tegra::GPUVAddr addr, u64 size, const Surface& region_owner);
+
+    /// Flush all cached resources tracked by this cache manager
+    void FlushAll();
 
 private:
-    void LoadSurface(const Surface& surface);
-    Surface GetSurface(const SurfaceParams& params);
+    void DuplicateSurface(const Surface& src_surface, const Surface& dest_surface);
+
+    /// Update surface's texture for given region when necessary
+    void ValidateSurface(const Surface& surface, Tegra::GPUVAddr addr, u64 size);
+
+    /// Create a new surface
+    Surface CreateSurface(const SurfaceParams& params);
 
     /// Register surface into the cache
     void RegisterSurface(const Surface& surface);
@@ -371,9 +519,18 @@ private:
     /// Increase/decrease the number of surface in pages touching the specified region
     void UpdatePagesCachedCount(Tegra::GPUVAddr addr, u64 size, int delta);
 
-    std::unordered_map<SurfaceKey, Surface> surface_cache;
+    SurfaceCache surface_cache;
     PageMap cached_pages;
+    SurfaceMap dirty_regions;
+    SurfaceSet remove_surfaces;
 
     OGLFramebuffer read_framebuffer;
     OGLFramebuffer draw_framebuffer;
+
+    OGLVertexArray attributeless_vao;
+    OGLBuffer d24s8_abgr_buffer;
+    GLsizeiptr d24s8_abgr_buffer_size;
+    OGLProgram d24s8_abgr_shader;
+    GLint d24s8_abgr_tbo_size_u_id;
+    GLint d24s8_abgr_viewport_u_id;
 };

src/video_core/renderer_opengl/gl_resource_manager.h

@@ -38,7 +38,7 @@ public:
         if (handle == 0)
             return;
         glDeleteTextures(1, &handle);
-        OpenGLState::GetCurState().UnbindTexture(handle).Apply();
+        OpenGLState::GetCurState().ResetTexture(handle).Apply();
         handle = 0;
     }
 

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -9,7 +9,6 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "video_core/engines/shader_bytecode.h"
-#include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_shader_decompiler.h"
 
 namespace GLShader {
@@ -398,8 +397,7 @@ public:
     /// Generates code representing a uniform (C buffer) register, interpreted as the input type.
     std::string GetUniform(u64 index, u64 offset, GLSLRegister::Type type) {
         declr_const_buffers[index].MarkAsUsed(index, offset, stage);
-        std::string value = 'c' + std::to_string(index) + '[' + std::to_string(offset / 4) + "][" +
-                            std::to_string(offset % 4) + ']';
+        std::string value = 'c' + std::to_string(index) + '[' + std::to_string(offset) + ']';
 
         if (type == GLSLRegister::Type::Float) {
             return value;
@@ -413,12 +411,8 @@ public:
     std::string GetUniformIndirect(u64 index, s64 offset, const Register& index_reg,
                                    GLSLRegister::Type type) {
         declr_const_buffers[index].MarkAsUsedIndirect(index, stage);
-
-        std::string final_offset = "((floatBitsToInt(" + GetRegister(index_reg, 0) + ") + " +
-                                   std::to_string(offset) + ") / 4)";
-
-        std::string value =
-            'c' + std::to_string(index) + '[' + final_offset + " / 4][" + final_offset + " % 4]";
+        std::string value = 'c' + std::to_string(index) + "[(floatBitsToInt(" +
+                            GetRegister(index_reg, 0) + ") + " + std::to_string(offset) + ") / 4]";
 
         if (type == GLSLRegister::Type::Float) {
             return value;
@@ -460,10 +454,9 @@ public:
 
         unsigned const_buffer_layout = 0;
         for (const auto& entry : GetConstBuffersDeclarations()) {
-            declarations.AddLine("layout(std140) uniform " + entry.GetName());
+            declarations.AddLine("layout(std430) buffer " + entry.GetName());
             declarations.AddLine('{');
-            declarations.AddLine("    vec4 c" + std::to_string(entry.GetIndex()) +
-                                 "[MAX_CONSTBUFFER_ELEMENTS];");
+            declarations.AddLine("    float c" + std::to_string(entry.GetIndex()) + "[];");
             declarations.AddLine("};");
             declarations.AddNewLine();
             ++const_buffer_layout;
@@ -1720,10 +1713,7 @@ private:
 }; // namespace Decompiler
 
 std::string GetCommonDeclarations() {
-    std::string declarations = "bool exec_shader();\n";
-    declarations += "#define MAX_CONSTBUFFER_ELEMENTS " +
-                    std::to_string(RasterizerOpenGL::MaxConstbufferSize / (sizeof(GLvec4)));
-    return declarations;
+    return "bool exec_shader();";
 }
 
 boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code, u32 main_offset,

src/video_core/renderer_opengl/gl_state.cpp

@@ -48,9 +48,24 @@ OpenGLState::OpenGLState() {
     logic_op = GL_COPY;
 
     for (auto& texture_unit : texture_units) {
-        texture_unit.Reset();
+        texture_unit.texture_2d = 0;
+        texture_unit.sampler = 0;
+        texture_unit.swizzle.r = GL_RED;
+        texture_unit.swizzle.g = GL_GREEN;
+        texture_unit.swizzle.b = GL_BLUE;
+        texture_unit.swizzle.a = GL_ALPHA;
     }
 
+    lighting_lut.texture_buffer = 0;
+
+    fog_lut.texture_buffer = 0;
+
+    proctex_lut.texture_buffer = 0;
+    proctex_diff_lut.texture_buffer = 0;
+    proctex_color_map.texture_buffer = 0;
+    proctex_alpha_map.texture_buffer = 0;
+    proctex_noise_lut.texture_buffer = 0;
+
     draw.read_framebuffer = 0;
     draw.draw_framebuffer = 0;
     draw.vertex_array = 0;
@@ -208,12 +223,54 @@ void OpenGLState::Apply() const {
             if (current.enabled != new_state.enabled || current.bindpoint != new_state.bindpoint ||
                 current.ssbo != new_state.ssbo) {
                 if (new_state.enabled) {
-                    glBindBufferBase(GL_UNIFORM_BUFFER, new_state.bindpoint, new_state.ssbo);
+                    glBindBufferBase(GL_SHADER_STORAGE_BUFFER, new_state.bindpoint, new_state.ssbo);
                 }
             }
         }
     }
 
+    // Lighting LUTs
+    if (lighting_lut.texture_buffer != cur_state.lighting_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::LightingLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, lighting_lut.texture_buffer);
+    }
+
+    // Fog LUT
+    if (fog_lut.texture_buffer != cur_state.fog_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::FogLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, fog_lut.texture_buffer);
+    }
+
+    // ProcTex Noise LUT
+    if (proctex_noise_lut.texture_buffer != cur_state.proctex_noise_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexNoiseLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_noise_lut.texture_buffer);
+    }
+
+    // ProcTex Color Map
+    if (proctex_color_map.texture_buffer != cur_state.proctex_color_map.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexColorMap.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_color_map.texture_buffer);
+    }
+
+    // ProcTex Alpha Map
+    if (proctex_alpha_map.texture_buffer != cur_state.proctex_alpha_map.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexAlphaMap.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_alpha_map.texture_buffer);
+    }
+
+    // ProcTex LUT
+    if (proctex_lut.texture_buffer != cur_state.proctex_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_lut.texture_buffer);
+    }
+
+    // ProcTex Diff LUT
+    if (proctex_diff_lut.texture_buffer != cur_state.proctex_diff_lut.texture_buffer) {
+        glActiveTexture(TextureUnits::ProcTexDiffLUT.Enum());
+        glBindTexture(GL_TEXTURE_BUFFER, proctex_diff_lut.texture_buffer);
+    }
+
     // Framebuffer
     if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
         glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@@ -281,12 +338,26 @@ void OpenGLState::Apply() const {
     cur_state = *this;
 }
 
-OpenGLState& OpenGLState::UnbindTexture(GLuint handle) {
+OpenGLState& OpenGLState::ResetTexture(GLuint handle) {
     for (auto& unit : texture_units) {
         if (unit.texture_2d == handle) {
-            unit.Unbind();
+            unit.texture_2d = 0;
         }
     }
+    if (lighting_lut.texture_buffer == handle)
+        lighting_lut.texture_buffer = 0;
+    if (fog_lut.texture_buffer == handle)
+        fog_lut.texture_buffer = 0;
+    if (proctex_noise_lut.texture_buffer == handle)
+        proctex_noise_lut.texture_buffer = 0;
+    if (proctex_color_map.texture_buffer == handle)
+        proctex_color_map.texture_buffer = 0;
+    if (proctex_alpha_map.texture_buffer == handle)
+        proctex_alpha_map.texture_buffer = 0;
+    if (proctex_lut.texture_buffer == handle)
+        proctex_lut.texture_buffer = 0;
+    if (proctex_diff_lut.texture_buffer == handle)
+        proctex_diff_lut.texture_buffer = 0;
     return *this;
 }
 

src/video_core/renderer_opengl/gl_state.h

@@ -91,21 +91,36 @@ public:
             GLint b; // GL_TEXTURE_SWIZZLE_B
             GLint a; // GL_TEXTURE_SWIZZLE_A
         } swizzle;
-
-        void Unbind() {
-            texture_2d = 0;
-            swizzle.r = GL_RED;
-            swizzle.g = GL_GREEN;
-            swizzle.b = GL_BLUE;
-            swizzle.a = GL_ALPHA;
-        }
-
-        void Reset() {
-            Unbind();
-            sampler = 0;
-        }
     } texture_units[32];
 
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } lighting_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } fog_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_noise_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_color_map;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_alpha_map;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_lut;
+
+    struct {
+        GLuint texture_buffer; // GL_TEXTURE_BINDING_BUFFER
+    } proctex_diff_lut;
+
     struct {
         GLuint read_framebuffer; // GL_READ_FRAMEBUFFER_BINDING
         GLuint draw_framebuffer; // GL_DRAW_FRAMEBUFFER_BINDING
@@ -150,7 +165,7 @@ public:
     void Apply() const;
 
     /// Resets any references to the given resource
-    OpenGLState& UnbindTexture(GLuint handle);
+    OpenGLState& ResetTexture(GLuint handle);
     OpenGLState& ResetSampler(GLuint handle);
     OpenGLState& ResetProgram(GLuint handle);
     OpenGLState& ResetPipeline(GLuint handle);

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -150,6 +150,7 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf
                                          screen_info)) {
         // Reset the screen info's display texture to its own permanent texture
         screen_info.display_texture = screen_info.texture.resource.handle;
+        screen_info.display_texcoords = MathUtil::Rectangle<float>(0.f, 0.f, 1.f, 1.f);
 
         Memory::RasterizerFlushVirtualRegion(framebuffer_addr, size_in_bytes,
                                              Memory::FlushMode::Flush);

src/video_core/renderer_opengl/renderer_opengl.h

@@ -27,7 +27,7 @@ struct TextureInfo {
 /// Structure used for storing information about the display target for the Switch screen
 struct ScreenInfo {
     GLuint display_texture;
-    const MathUtil::Rectangle<float> display_texcoords{0.0f, 0.0f, 1.0f, 1.0f};
+    MathUtil::Rectangle<float> display_texcoords;
     TextureInfo texture;
 };
 

src/yuzu/configuration/config.cpp

@@ -84,8 +84,6 @@ void Config::ReadValues() {
     qt_config->beginGroup("Renderer");
     Settings::values.resolution_factor = qt_config->value("resolution_factor", 1.0).toFloat();
     Settings::values.toggle_framelimit = qt_config->value("toggle_framelimit", true).toBool();
-    Settings::values.use_accurate_framebuffers =
-        qt_config->value("use_accurate_framebuffers", false).toBool();
 
     Settings::values.bg_red = qt_config->value("bg_red", 0.0).toFloat();
     Settings::values.bg_green = qt_config->value("bg_green", 0.0).toFloat();
@@ -186,7 +184,6 @@ void Config::SaveValues() {
     qt_config->beginGroup("Renderer");
     qt_config->setValue("resolution_factor", (double)Settings::values.resolution_factor);
     qt_config->setValue("toggle_framelimit", Settings::values.toggle_framelimit);
-    qt_config->setValue("use_accurate_framebuffers", Settings::values.use_accurate_framebuffers);
 
     // Cast to double because Qt's written float values are not human-readable
     qt_config->setValue("bg_red", (double)Settings::values.bg_red);

src/yuzu/configuration/configure_graphics.cpp

@@ -59,13 +59,11 @@ void ConfigureGraphics::setConfiguration() {
     ui->resolution_factor_combobox->setCurrentIndex(
         static_cast<int>(FromResolutionFactor(Settings::values.resolution_factor)));
     ui->toggle_framelimit->setChecked(Settings::values.toggle_framelimit);
-    ui->use_accurate_framebuffers->setChecked(Settings::values.use_accurate_framebuffers);
 }
 
 void ConfigureGraphics::applyConfiguration() {
     Settings::values.resolution_factor =
         ToResolutionFactor(static_cast<Resolution>(ui->resolution_factor_combobox->currentIndex()));
     Settings::values.toggle_framelimit = ui->toggle_framelimit->isChecked();
-    Settings::values.use_accurate_framebuffers = ui->use_accurate_framebuffers->isChecked();
     Settings::Apply();
 }

src/yuzu/configuration/configure_graphics.ui

@@ -29,13 +29,6 @@
           </property>
          </widget>
         </item>
-        <item>
-         <widget class="QCheckBox" name="use_accurate_framebuffers">
-          <property name="text">
-           <string>Use accurate framebuffers (slow)</string>
-          </property>
-         </widget>
-        </item>
         <item>
          <layout class="QHBoxLayout" name="horizontalLayout">
           <item>

src/yuzu/main.cpp

@@ -334,6 +334,8 @@ bool GMainWindow::SupportsRequiredGLExtensions() {
         unsupported_ext.append("ARB_program_interface_query");
     if (!GLAD_GL_ARB_separate_shader_objects)
         unsupported_ext.append("ARB_separate_shader_objects");
+    if (!GLAD_GL_ARB_shader_storage_buffer_object)
+        unsupported_ext.append("ARB_shader_storage_buffer_object");
     if (!GLAD_GL_ARB_vertex_attrib_binding)
         unsupported_ext.append("ARB_vertex_attrib_binding");
 

src/yuzu_cmd/config.cpp

@@ -98,8 +98,6 @@ void Config::ReadValues() {
         (float)sdl2_config->GetReal("Renderer", "resolution_factor", 1.0);
     Settings::values.toggle_framelimit =
         sdl2_config->GetBoolean("Renderer", "toggle_framelimit", true);
-    Settings::values.use_accurate_framebuffers =
-        sdl2_config->GetBoolean("Renderer", "use_accurate_framebuffers", false);
 
     Settings::values.bg_red = (float)sdl2_config->GetReal("Renderer", "bg_red", 0.0);
     Settings::values.bg_green = (float)sdl2_config->GetReal("Renderer", "bg_green", 0.0);

src/yuzu_cmd/default_ini.h

@@ -102,10 +102,6 @@ resolution_factor =
 # 0 (default): Off, 1: On
 use_vsync =
 
-# Whether to use accurate framebuffers
-# 0 (default): Off (fast), 1 : On (slow)
-use_accurate_framebuffers =
-
 # The clear color for the renderer. What shows up on the sides of the bottom screen.
 # Must be in range of 0.0-1.0. Defaults to 1.0 for all.
 bg_red =