citra/src/common/color.h

353 lines
10 KiB
C++
Raw Normal View History

2014-12-31 15:17:07 +01:00
// Copyright 2014 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <cstring>
2014-12-31 15:17:07 +01:00
#include "common/common_types.h"
2015-05-06 09:06:12 +02:00
#include "common/swap.h"
#include "common/vector_math.h"
2014-12-31 15:17:07 +01:00
namespace Common::Color {
2014-12-31 15:17:07 +01:00
/// Convert a 1-bit color component to 8 bit
[[nodiscard]] constexpr u8 Convert1To8(u8 value) {
2014-12-31 15:17:07 +01:00
return value * 255;
}
/// Convert a 4-bit color component to 8 bit
[[nodiscard]] constexpr u8 Convert4To8(u8 value) {
2014-12-31 15:17:07 +01:00
return (value << 4) | value;
}
/// Convert a 5-bit color component to 8 bit
[[nodiscard]] constexpr u8 Convert5To8(u8 value) {
2014-12-31 15:17:07 +01:00
return (value << 3) | (value >> 2);
}
/// Convert a 6-bit color component to 8 bit
[[nodiscard]] constexpr u8 Convert6To8(u8 value) {
2014-12-31 15:17:07 +01:00
return (value << 2) | (value >> 4);
}
/// Convert a 8-bit color component to 1 bit
[[nodiscard]] constexpr u8 Convert8To1(u8 value) {
return value >> 7;
}
/// Convert a 8-bit color component to 4 bit
[[nodiscard]] constexpr u8 Convert8To4(u8 value) {
return value >> 4;
}
/// Convert a 8-bit color component to 5 bit
[[nodiscard]] constexpr u8 Convert8To5(u8 value) {
return value >> 3;
}
/// Convert a 8-bit color component to 6 bit
[[nodiscard]] constexpr u8 Convert8To6(u8 value) {
return value >> 2;
}
2014-12-31 15:17:07 +01:00
Rasterizer cache refactor (#6375) * rasterizer_cache: Remove custom texture code * It's a hacky buggy mess, will be reimplemented later when the cache is in a better state * rasterizer_cache: Refactor surface upload/download * Switch to the texture_codec header which was written as part of the vulkan backend by steveice and me * Move most of the upload logic to the rasterizer cache and out of the surface object * Scaled uploads/downloads have been disabled for now since they require more runtime infrastructure * rasterizer_cache: Refactor runtime interface * Remove aspect enum which is the same as SurfaceType * Replace Subresource with specific structures for each operation (blit/copy/clear). This mimics moderns APIs vulkan much better * Pass the surface to the runtime instead of the texture * Implement CopyTextures with glCopyImageSubData which is available on 4.3 and gles. This function also has an overload for cubes which will be removed later. * rasterizer_cache: Move texture allocation to the runtime * renderer_opengl: Remove TextureDownloaderES * It's overly compilcated and unused at the moment. Will be replaced with a simple compute shader in a later commit * rasterizer_cache: Split CachedSurface * This commit splits CachedSurface into two classes, SurfaceBase which contains the backend agnostic functions and Surface which is the opengl specific part * For now the cache uses the opengl surface directly and there are a few ugly casts with watchers, those will be taken care of when the template convertion and watcher removal are added respectively * rasterizer_cache: Move reinterpreters to the runtime * rasterizer_cache: Move some pixel format function to the cpp file * rasterizer_cache: Common texture acceleration functions * They don't contain any backend specific code so they shouldn't be duplicated * rasterizer_cache: Remove BlitSurfaces * It's better to prefer copy/blit in the caller anyway * rasterizer_cache: Only allocate needed levels * rasterizer_cache: Move texture runtime out of common dir * Also shorten the util header filename * surface_params: Cleanup code * Add more comments, organize it a bit etc * rasterizer_cache: Move texture filtering to the runtime * rasterizer_cache: Move to VideoCore * renderer_opengl: Reimplement scaled uploads/downloads * Instead of looking up for temporary textures, each allocation now contains both a scaled and unscaled handle This allows the scale operations to be done inside the surface object itself and improves performance in general * In particular the scaled download code has been expanded to use ARB_get_texture_sub_image when possible which is faster and more convenient than glReadPixels. The latter is still relevant for OpenGLES though. * Finally allocations are now given a handy debug name that can be viewed from renderdoc. * rasterizer_cache: Remove global state * gl_rasterizer: Abstract common draw operations to Framebuffer * This also allows to cache framebuffer objects instead of always swapping the textures, something that particularly benefits mali gpus * rasterizer_cache: Implement multi-level surfaces * With this commit the cache can now directly upload and use mipmaps without needing to sync them with watchers. By using native mimaps directly this also adds support for mipmap for cube * Texture cubes have also been updated to drop the watcher requirement * host_shaders: Add CMake integration for string shaders * Improves build time shader generation making it much less prone to errors. Also moves the presentation shaders here to avoid embedding them to the cpp file. * Texture filter shaders now make explicit use of uniform bindings for better vulkan compatibility * renderer_opengl: Emulate lod bias in the shader * This way opengles can emulate it correctly * gl_rasterizer: Respect GL_MAX_TEXTURE_BUFFER_SIZE * Older Bifrost Mali GPUs only support up to 64kb texture buffers. Citra would try to allocate a much larger buffer the first 64kb of which would work fine but after that the driver starts misbehaving and showing various graphical glitches * rasterizer_cache: Cleanup CopySurface * renderer_opengl: Keep frames synchronized when using a GPU debugger * rasterizer_cache: Rename Surface to SurfaceRef * Makes it clear that surface is a shared_ptr and not an object * rasterizer_cache: Cleanup * Move constructor to the top of the file * Move FindMatch to the top as well and remove the Invalid flag which was redudant; all FindMatch calls used it expect from MatchFlags::Copy which ignores it anyway * gl_texture_runtime: Make driver const * gl_texture_runtime: Fix RGB8 format handling * The texture_codec header, being written with vulkan in mind converts RGB8 to RGBA8. The backend wasn't adjusted to account for this though and treated the data as RGB8. * Also remove D16 convertions, both opengl and vulkan are required to support this format so these are not needed * gl_texture_runtime: Reduce state switches during FBO blits * glBlitFramebuffer is only affected by the scissor rectangle so just disable scissor testing instead of resetting our entire state * surface_params: Prevent texcopy that spans multiple levels * It would have failed before as well, with multi-level surfaces it triggers the assert though * renderer_opengl: Centralize texture filters * A lot of code is shared between the filters thus is makes it sense to centralize them * Also fix an issue with partial texture uploads * Address review comments * rasterizer_cache: Use leading return types * rasterizer_cache: Cleanup null checks * renderer_opengl: Add additional logging * externals: Actually downgrade glad * For some reason I missed adding the files to git * surface_params: Do not check for levels in exact match * Some games will try to use the base level of a multi level surface. Checking for levels forces another surface to be created and a copy to be made which is both unncessary and breaks custom textures --------- Co-authored-by: bunnei <bunneidev@gmail.com>
2023-04-21 09:14:55 +02:00
/// Averages the RGB components of a color
[[nodiscard]] constexpr u8 AverageRgbComponents(const Common::Vec4<u8>& color) {
return (static_cast<u32>(color.r()) + color.g() + color.b()) / 3;
}
/**
* Decode a color stored in RGBA8 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeRGBA8(const u8* bytes) {
return {bytes[3], bytes[2], bytes[1], bytes[0]};
}
/**
* Decode a color stored in RGB8 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeRGB8(const u8* bytes) {
return {bytes[2], bytes[1], bytes[0], 255};
}
2015-05-14 22:39:46 +02:00
/**
* Decode a color stored in RG8 (aka HILO8) format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
2015-05-14 22:39:46 +02:00
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeRG8(const u8* bytes) {
return {bytes[1], bytes[0], 0, 255};
2015-05-14 22:39:46 +02:00
}
/**
* Decode a color stored in RGB565 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeRGB565(const u8* bytes) {
u16_le pixel;
std::memcpy(&pixel, bytes, sizeof(pixel));
return {Convert5To8((pixel >> 11) & 0x1F), Convert6To8((pixel >> 5) & 0x3F),
Convert5To8(pixel & 0x1F), 255};
}
/**
* Decode a color stored in RGB5A1 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
u16_le pixel;
std::memcpy(&pixel, bytes, sizeof(pixel));
return {Convert5To8((pixel >> 11) & 0x1F), Convert5To8((pixel >> 6) & 0x1F),
Convert5To8((pixel >> 1) & 0x1F), Convert1To8(pixel & 0x1)};
}
/**
* Decode a color stored in RGBA4 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeRGBA4(const u8* bytes) {
u16_le pixel;
std::memcpy(&pixel, bytes, sizeof(pixel));
return {Convert4To8((pixel >> 12) & 0xF), Convert4To8((pixel >> 8) & 0xF),
Convert4To8((pixel >> 4) & 0xF), Convert4To8(pixel & 0xF)};
}
Rasterizer cache refactor (#6375) * rasterizer_cache: Remove custom texture code * It's a hacky buggy mess, will be reimplemented later when the cache is in a better state * rasterizer_cache: Refactor surface upload/download * Switch to the texture_codec header which was written as part of the vulkan backend by steveice and me * Move most of the upload logic to the rasterizer cache and out of the surface object * Scaled uploads/downloads have been disabled for now since they require more runtime infrastructure * rasterizer_cache: Refactor runtime interface * Remove aspect enum which is the same as SurfaceType * Replace Subresource with specific structures for each operation (blit/copy/clear). This mimics moderns APIs vulkan much better * Pass the surface to the runtime instead of the texture * Implement CopyTextures with glCopyImageSubData which is available on 4.3 and gles. This function also has an overload for cubes which will be removed later. * rasterizer_cache: Move texture allocation to the runtime * renderer_opengl: Remove TextureDownloaderES * It's overly compilcated and unused at the moment. Will be replaced with a simple compute shader in a later commit * rasterizer_cache: Split CachedSurface * This commit splits CachedSurface into two classes, SurfaceBase which contains the backend agnostic functions and Surface which is the opengl specific part * For now the cache uses the opengl surface directly and there are a few ugly casts with watchers, those will be taken care of when the template convertion and watcher removal are added respectively * rasterizer_cache: Move reinterpreters to the runtime * rasterizer_cache: Move some pixel format function to the cpp file * rasterizer_cache: Common texture acceleration functions * They don't contain any backend specific code so they shouldn't be duplicated * rasterizer_cache: Remove BlitSurfaces * It's better to prefer copy/blit in the caller anyway * rasterizer_cache: Only allocate needed levels * rasterizer_cache: Move texture runtime out of common dir * Also shorten the util header filename * surface_params: Cleanup code * Add more comments, organize it a bit etc * rasterizer_cache: Move texture filtering to the runtime * rasterizer_cache: Move to VideoCore * renderer_opengl: Reimplement scaled uploads/downloads * Instead of looking up for temporary textures, each allocation now contains both a scaled and unscaled handle This allows the scale operations to be done inside the surface object itself and improves performance in general * In particular the scaled download code has been expanded to use ARB_get_texture_sub_image when possible which is faster and more convenient than glReadPixels. The latter is still relevant for OpenGLES though. * Finally allocations are now given a handy debug name that can be viewed from renderdoc. * rasterizer_cache: Remove global state * gl_rasterizer: Abstract common draw operations to Framebuffer * This also allows to cache framebuffer objects instead of always swapping the textures, something that particularly benefits mali gpus * rasterizer_cache: Implement multi-level surfaces * With this commit the cache can now directly upload and use mipmaps without needing to sync them with watchers. By using native mimaps directly this also adds support for mipmap for cube * Texture cubes have also been updated to drop the watcher requirement * host_shaders: Add CMake integration for string shaders * Improves build time shader generation making it much less prone to errors. Also moves the presentation shaders here to avoid embedding them to the cpp file. * Texture filter shaders now make explicit use of uniform bindings for better vulkan compatibility * renderer_opengl: Emulate lod bias in the shader * This way opengles can emulate it correctly * gl_rasterizer: Respect GL_MAX_TEXTURE_BUFFER_SIZE * Older Bifrost Mali GPUs only support up to 64kb texture buffers. Citra would try to allocate a much larger buffer the first 64kb of which would work fine but after that the driver starts misbehaving and showing various graphical glitches * rasterizer_cache: Cleanup CopySurface * renderer_opengl: Keep frames synchronized when using a GPU debugger * rasterizer_cache: Rename Surface to SurfaceRef * Makes it clear that surface is a shared_ptr and not an object * rasterizer_cache: Cleanup * Move constructor to the top of the file * Move FindMatch to the top as well and remove the Invalid flag which was redudant; all FindMatch calls used it expect from MatchFlags::Copy which ignores it anyway * gl_texture_runtime: Make driver const * gl_texture_runtime: Fix RGB8 format handling * The texture_codec header, being written with vulkan in mind converts RGB8 to RGBA8. The backend wasn't adjusted to account for this though and treated the data as RGB8. * Also remove D16 convertions, both opengl and vulkan are required to support this format so these are not needed * gl_texture_runtime: Reduce state switches during FBO blits * glBlitFramebuffer is only affected by the scissor rectangle so just disable scissor testing instead of resetting our entire state * surface_params: Prevent texcopy that spans multiple levels * It would have failed before as well, with multi-level surfaces it triggers the assert though * renderer_opengl: Centralize texture filters * A lot of code is shared between the filters thus is makes it sense to centralize them * Also fix an issue with partial texture uploads * Address review comments * rasterizer_cache: Use leading return types * rasterizer_cache: Cleanup null checks * renderer_opengl: Add additional logging * externals: Actually downgrade glad * For some reason I missed adding the files to git * surface_params: Do not check for levels in exact match * Some games will try to use the base level of a multi level surface. Checking for levels forces another surface to be created and a copy to be made which is both unncessary and breaks custom textures --------- Co-authored-by: bunnei <bunneidev@gmail.com>
2023-04-21 09:14:55 +02:00
/**
* Decode a color stored in IA8 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeIA8(const u8* bytes) {
return {bytes[1], bytes[1], bytes[1], bytes[0]};
}
/**
* Decode a color stored in I8 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeI8(const u8* bytes) {
return {bytes[0], bytes[0], bytes[0], 255};
}
/**
* Decode a color stored in A8 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeA8(const u8* bytes) {
return {0, 0, 0, bytes[0]};
}
/**
* Decode a color stored in IA4 format
* @param bytes Pointer to encoded source color
* @return Result color decoded as Common::Vec4<u8>
*/
[[nodiscard]] inline Common::Vec4<u8> DecodeIA4(const u8* bytes) {
u8 i = Common::Color::Convert4To8((bytes[0] & 0xF0) >> 4);
u8 a = Common::Color::Convert4To8(bytes[0] & 0x0F);
return {i, i, i, a};
}
/**
* Decode a depth value stored in D16 format
* @param bytes Pointer to encoded source value
* @return Depth value as an u32
*/
[[nodiscard]] inline u32 DecodeD16(const u8* bytes) {
u16_le data;
std::memcpy(&data, bytes, sizeof(data));
return data;
}
/**
* Decode a depth value stored in D24 format
* @param bytes Pointer to encoded source value
* @return Depth value as an u32
*/
[[nodiscard]] inline u32 DecodeD24(const u8* bytes) {
return (bytes[2] << 16) | (bytes[1] << 8) | bytes[0];
}
/**
* Decode a depth value and a stencil value stored in D24S8 format
* @param bytes Pointer to encoded source values
* @return Resulting values stored as a Common::Vec2
*/
[[nodiscard]] inline Common::Vec2<u32> DecodeD24S8(const u8* bytes) {
return {static_cast<u32>((bytes[2] << 16) | (bytes[1] << 8) | bytes[0]), bytes[3]};
}
/**
* Encode a color as RGBA8 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeRGBA8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[3] = color.r();
bytes[2] = color.g();
bytes[1] = color.b();
bytes[0] = color.a();
}
/**
* Encode a color as RGB8 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeRGB8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[2] = color.r();
bytes[1] = color.g();
bytes[0] = color.b();
}
2015-05-14 22:39:46 +02:00
/**
* Encode a color as RG8 (aka HILO8) format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeRG8(const Common::Vec4<u8>& color, u8* bytes) {
2015-05-14 22:39:46 +02:00
bytes[1] = color.r();
bytes[0] = color.g();
}
Rasterizer cache refactor (#6375) * rasterizer_cache: Remove custom texture code * It's a hacky buggy mess, will be reimplemented later when the cache is in a better state * rasterizer_cache: Refactor surface upload/download * Switch to the texture_codec header which was written as part of the vulkan backend by steveice and me * Move most of the upload logic to the rasterizer cache and out of the surface object * Scaled uploads/downloads have been disabled for now since they require more runtime infrastructure * rasterizer_cache: Refactor runtime interface * Remove aspect enum which is the same as SurfaceType * Replace Subresource with specific structures for each operation (blit/copy/clear). This mimics moderns APIs vulkan much better * Pass the surface to the runtime instead of the texture * Implement CopyTextures with glCopyImageSubData which is available on 4.3 and gles. This function also has an overload for cubes which will be removed later. * rasterizer_cache: Move texture allocation to the runtime * renderer_opengl: Remove TextureDownloaderES * It's overly compilcated and unused at the moment. Will be replaced with a simple compute shader in a later commit * rasterizer_cache: Split CachedSurface * This commit splits CachedSurface into two classes, SurfaceBase which contains the backend agnostic functions and Surface which is the opengl specific part * For now the cache uses the opengl surface directly and there are a few ugly casts with watchers, those will be taken care of when the template convertion and watcher removal are added respectively * rasterizer_cache: Move reinterpreters to the runtime * rasterizer_cache: Move some pixel format function to the cpp file * rasterizer_cache: Common texture acceleration functions * They don't contain any backend specific code so they shouldn't be duplicated * rasterizer_cache: Remove BlitSurfaces * It's better to prefer copy/blit in the caller anyway * rasterizer_cache: Only allocate needed levels * rasterizer_cache: Move texture runtime out of common dir * Also shorten the util header filename * surface_params: Cleanup code * Add more comments, organize it a bit etc * rasterizer_cache: Move texture filtering to the runtime * rasterizer_cache: Move to VideoCore * renderer_opengl: Reimplement scaled uploads/downloads * Instead of looking up for temporary textures, each allocation now contains both a scaled and unscaled handle This allows the scale operations to be done inside the surface object itself and improves performance in general * In particular the scaled download code has been expanded to use ARB_get_texture_sub_image when possible which is faster and more convenient than glReadPixels. The latter is still relevant for OpenGLES though. * Finally allocations are now given a handy debug name that can be viewed from renderdoc. * rasterizer_cache: Remove global state * gl_rasterizer: Abstract common draw operations to Framebuffer * This also allows to cache framebuffer objects instead of always swapping the textures, something that particularly benefits mali gpus * rasterizer_cache: Implement multi-level surfaces * With this commit the cache can now directly upload and use mipmaps without needing to sync them with watchers. By using native mimaps directly this also adds support for mipmap for cube * Texture cubes have also been updated to drop the watcher requirement * host_shaders: Add CMake integration for string shaders * Improves build time shader generation making it much less prone to errors. Also moves the presentation shaders here to avoid embedding them to the cpp file. * Texture filter shaders now make explicit use of uniform bindings for better vulkan compatibility * renderer_opengl: Emulate lod bias in the shader * This way opengles can emulate it correctly * gl_rasterizer: Respect GL_MAX_TEXTURE_BUFFER_SIZE * Older Bifrost Mali GPUs only support up to 64kb texture buffers. Citra would try to allocate a much larger buffer the first 64kb of which would work fine but after that the driver starts misbehaving and showing various graphical glitches * rasterizer_cache: Cleanup CopySurface * renderer_opengl: Keep frames synchronized when using a GPU debugger * rasterizer_cache: Rename Surface to SurfaceRef * Makes it clear that surface is a shared_ptr and not an object * rasterizer_cache: Cleanup * Move constructor to the top of the file * Move FindMatch to the top as well and remove the Invalid flag which was redudant; all FindMatch calls used it expect from MatchFlags::Copy which ignores it anyway * gl_texture_runtime: Make driver const * gl_texture_runtime: Fix RGB8 format handling * The texture_codec header, being written with vulkan in mind converts RGB8 to RGBA8. The backend wasn't adjusted to account for this though and treated the data as RGB8. * Also remove D16 convertions, both opengl and vulkan are required to support this format so these are not needed * gl_texture_runtime: Reduce state switches during FBO blits * glBlitFramebuffer is only affected by the scissor rectangle so just disable scissor testing instead of resetting our entire state * surface_params: Prevent texcopy that spans multiple levels * It would have failed before as well, with multi-level surfaces it triggers the assert though * renderer_opengl: Centralize texture filters * A lot of code is shared between the filters thus is makes it sense to centralize them * Also fix an issue with partial texture uploads * Address review comments * rasterizer_cache: Use leading return types * rasterizer_cache: Cleanup null checks * renderer_opengl: Add additional logging * externals: Actually downgrade glad * For some reason I missed adding the files to git * surface_params: Do not check for levels in exact match * Some games will try to use the base level of a multi level surface. Checking for levels forces another surface to be created and a copy to be made which is both unncessary and breaks custom textures --------- Co-authored-by: bunnei <bunneidev@gmail.com>
2023-04-21 09:14:55 +02:00
/**
* Encode a color as RGB565 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeRGB565(const Common::Vec4<u8>& color, u8* bytes) {
const u16_le data =
(Convert8To5(color.r()) << 11) | (Convert8To6(color.g()) << 5) | Convert8To5(color.b());
std::memcpy(bytes, &data, sizeof(data));
}
/**
* Encode a color as RGB5A1 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeRGB5A1(const Common::Vec4<u8>& color, u8* bytes) {
const u16_le data = (Convert8To5(color.r()) << 11) | (Convert8To5(color.g()) << 6) |
(Convert8To5(color.b()) << 1) | Convert8To1(color.a());
std::memcpy(bytes, &data, sizeof(data));
}
/**
* Encode a color as RGBA4 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeRGBA4(const Common::Vec4<u8>& color, u8* bytes) {
const u16 data = (Convert8To4(color.r()) << 12) | (Convert8To4(color.g()) << 8) |
(Convert8To4(color.b()) << 4) | Convert8To4(color.a());
std::memcpy(bytes, &data, sizeof(data));
}
Rasterizer cache refactor (#6375) * rasterizer_cache: Remove custom texture code * It's a hacky buggy mess, will be reimplemented later when the cache is in a better state * rasterizer_cache: Refactor surface upload/download * Switch to the texture_codec header which was written as part of the vulkan backend by steveice and me * Move most of the upload logic to the rasterizer cache and out of the surface object * Scaled uploads/downloads have been disabled for now since they require more runtime infrastructure * rasterizer_cache: Refactor runtime interface * Remove aspect enum which is the same as SurfaceType * Replace Subresource with specific structures for each operation (blit/copy/clear). This mimics moderns APIs vulkan much better * Pass the surface to the runtime instead of the texture * Implement CopyTextures with glCopyImageSubData which is available on 4.3 and gles. This function also has an overload for cubes which will be removed later. * rasterizer_cache: Move texture allocation to the runtime * renderer_opengl: Remove TextureDownloaderES * It's overly compilcated and unused at the moment. Will be replaced with a simple compute shader in a later commit * rasterizer_cache: Split CachedSurface * This commit splits CachedSurface into two classes, SurfaceBase which contains the backend agnostic functions and Surface which is the opengl specific part * For now the cache uses the opengl surface directly and there are a few ugly casts with watchers, those will be taken care of when the template convertion and watcher removal are added respectively * rasterizer_cache: Move reinterpreters to the runtime * rasterizer_cache: Move some pixel format function to the cpp file * rasterizer_cache: Common texture acceleration functions * They don't contain any backend specific code so they shouldn't be duplicated * rasterizer_cache: Remove BlitSurfaces * It's better to prefer copy/blit in the caller anyway * rasterizer_cache: Only allocate needed levels * rasterizer_cache: Move texture runtime out of common dir * Also shorten the util header filename * surface_params: Cleanup code * Add more comments, organize it a bit etc * rasterizer_cache: Move texture filtering to the runtime * rasterizer_cache: Move to VideoCore * renderer_opengl: Reimplement scaled uploads/downloads * Instead of looking up for temporary textures, each allocation now contains both a scaled and unscaled handle This allows the scale operations to be done inside the surface object itself and improves performance in general * In particular the scaled download code has been expanded to use ARB_get_texture_sub_image when possible which is faster and more convenient than glReadPixels. The latter is still relevant for OpenGLES though. * Finally allocations are now given a handy debug name that can be viewed from renderdoc. * rasterizer_cache: Remove global state * gl_rasterizer: Abstract common draw operations to Framebuffer * This also allows to cache framebuffer objects instead of always swapping the textures, something that particularly benefits mali gpus * rasterizer_cache: Implement multi-level surfaces * With this commit the cache can now directly upload and use mipmaps without needing to sync them with watchers. By using native mimaps directly this also adds support for mipmap for cube * Texture cubes have also been updated to drop the watcher requirement * host_shaders: Add CMake integration for string shaders * Improves build time shader generation making it much less prone to errors. Also moves the presentation shaders here to avoid embedding them to the cpp file. * Texture filter shaders now make explicit use of uniform bindings for better vulkan compatibility * renderer_opengl: Emulate lod bias in the shader * This way opengles can emulate it correctly * gl_rasterizer: Respect GL_MAX_TEXTURE_BUFFER_SIZE * Older Bifrost Mali GPUs only support up to 64kb texture buffers. Citra would try to allocate a much larger buffer the first 64kb of which would work fine but after that the driver starts misbehaving and showing various graphical glitches * rasterizer_cache: Cleanup CopySurface * renderer_opengl: Keep frames synchronized when using a GPU debugger * rasterizer_cache: Rename Surface to SurfaceRef * Makes it clear that surface is a shared_ptr and not an object * rasterizer_cache: Cleanup * Move constructor to the top of the file * Move FindMatch to the top as well and remove the Invalid flag which was redudant; all FindMatch calls used it expect from MatchFlags::Copy which ignores it anyway * gl_texture_runtime: Make driver const * gl_texture_runtime: Fix RGB8 format handling * The texture_codec header, being written with vulkan in mind converts RGB8 to RGBA8. The backend wasn't adjusted to account for this though and treated the data as RGB8. * Also remove D16 convertions, both opengl and vulkan are required to support this format so these are not needed * gl_texture_runtime: Reduce state switches during FBO blits * glBlitFramebuffer is only affected by the scissor rectangle so just disable scissor testing instead of resetting our entire state * surface_params: Prevent texcopy that spans multiple levels * It would have failed before as well, with multi-level surfaces it triggers the assert though * renderer_opengl: Centralize texture filters * A lot of code is shared between the filters thus is makes it sense to centralize them * Also fix an issue with partial texture uploads * Address review comments * rasterizer_cache: Use leading return types * rasterizer_cache: Cleanup null checks * renderer_opengl: Add additional logging * externals: Actually downgrade glad * For some reason I missed adding the files to git * surface_params: Do not check for levels in exact match * Some games will try to use the base level of a multi level surface. Checking for levels forces another surface to be created and a copy to be made which is both unncessary and breaks custom textures --------- Co-authored-by: bunnei <bunneidev@gmail.com>
2023-04-21 09:14:55 +02:00
/**
* Encode a color as IA8 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeIA8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[1] = AverageRgbComponents(color);
bytes[0] = color.a();
}
/**
* Encode a color as I8 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeI8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[0] = AverageRgbComponents(color);
}
/**
* Encode a color as A8 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeA8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[0] = color.a();
}
/**
* Encode a color as IA4 format
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
inline void EncodeIA4(const Common::Vec4<u8>& color, u8* bytes) {
bytes[0] = (Convert8To4(AverageRgbComponents(color)) << 4) | Convert8To4(color.a());
}
/**
* Encode a 16 bit depth value as D16 format
* @param value 16 bit source depth value to encode
* @param bytes Pointer where to store the encoded value
*/
inline void EncodeD16(u32 value, u8* bytes) {
const u16_le data = static_cast<u16>(value);
std::memcpy(bytes, &data, sizeof(data));
}
/**
* Encode a 24 bit depth value as D24 format
* @param value 24 bit source depth value to encode
* @param bytes Pointer where to store the encoded value
*/
inline void EncodeD24(u32 value, u8* bytes) {
bytes[0] = value & 0xFF;
bytes[1] = (value >> 8) & 0xFF;
bytes[2] = (value >> 16) & 0xFF;
}
/**
* Encode a 24 bit depth and 8 bit stencil values as D24S8 format
* @param depth 24 bit source depth value to encode
* @param stencil 8 bit source stencil value to encode
* @param bytes Pointer where to store the encoded value
*/
inline void EncodeD24S8(u32 depth, u8 stencil, u8* bytes) {
2015-05-21 16:51:31 +02:00
bytes[0] = depth & 0xFF;
bytes[1] = (depth >> 8) & 0xFF;
bytes[2] = (depth >> 16) & 0xFF;
bytes[3] = stencil;
}
/**
* Encode a 24 bit depth value as D24X8 format (32 bits per pixel with 8 bits unused)
* @param depth 24 bit source depth value to encode
* @param bytes Pointer where to store the encoded value
* @note unused bits will not be modified
*/
inline void EncodeD24X8(u32 depth, u8* bytes) {
bytes[0] = depth & 0xFF;
bytes[1] = (depth >> 8) & 0xFF;
bytes[2] = (depth >> 16) & 0xFF;
}
/**
* Encode an 8 bit stencil value as X24S8 format (32 bits per pixel with 24 bits unused)
* @param stencil 8 bit source stencil value to encode
* @param bytes Pointer where to store the encoded value
* @note unused bits will not be modified
*/
inline void EncodeX24S8(u8 stencil, u8* bytes) {
bytes[3] = stencil;
}
} // namespace Common::Color