Merge branch 'master' into Texture2DArray

This commit is contained in:
greggameplayer 2018-11-02 14:26:32 +01:00 committed by GitHub
commit cb8e4a4633
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
27 changed files with 1480 additions and 1069 deletions

View file

@ -30,8 +30,8 @@
namespace OpenGL {
using Maxwell = Tegra::Engines::Maxwell3D::Regs;
using PixelFormat = SurfaceParams::PixelFormat;
using SurfaceType = SurfaceParams::SurfaceType;
using PixelFormat = VideoCore::Surface::PixelFormat;
using SurfaceType = VideoCore::Surface::SurfaceType;
MICROPROFILE_DEFINE(OpenGL_VAO, "OpenGL", "Vertex Array Setup", MP_RGB(128, 128, 192));
MICROPROFILE_DEFINE(OpenGL_Shader, "OpenGL", "Shader Setup", MP_RGB(128, 128, 192));
@ -104,7 +104,7 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
}
ASSERT_MSG(has_ARB_separate_shader_objects, "has_ARB_separate_shader_objects is unsupported");
OpenGLState::ApplyDefaultState();
// Clipping plane 0 is always enabled for PICA fixed clip plane z <= 0
state.clip_distance[0] = true;
@ -115,8 +115,6 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
state.draw.shader_program = 0;
state.Apply();
glEnable(GL_BLEND);
glGetIntegerv(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT, &uniform_buffer_alignment);
LOG_CRITICAL(Render_OpenGL, "Sync fixed function OpenGL state here!");
@ -703,7 +701,8 @@ bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config,
// 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)};
const auto& pixel_format{
VideoCore::Surface::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");

View file

@ -17,15 +17,20 @@
#include "video_core/engines/maxwell_3d.h"
#include "video_core/renderer_opengl/gl_rasterizer_cache.h"
#include "video_core/renderer_opengl/utils.h"
#include "video_core/surface.h"
#include "video_core/textures/astc.h"
#include "video_core/textures/decoders.h"
#include "video_core/utils.h"
namespace OpenGL {
using SurfaceType = SurfaceParams::SurfaceType;
using PixelFormat = SurfaceParams::PixelFormat;
using ComponentType = SurfaceParams::ComponentType;
using VideoCore::Surface::ComponentTypeFromDepthFormat;
using VideoCore::Surface::ComponentTypeFromRenderTarget;
using VideoCore::Surface::ComponentTypeFromTexture;
using VideoCore::Surface::PixelFormatFromDepthFormat;
using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
using VideoCore::Surface::PixelFormatFromTextureFormat;
using VideoCore::Surface::SurfaceTargetFromTextureType;
struct FormatTuple {
GLint internal_format;
@ -35,46 +40,6 @@ struct FormatTuple {
bool compressed;
};
static bool IsPixelFormatASTC(PixelFormat format) {
switch (format) {
case PixelFormat::ASTC_2D_4X4:
case PixelFormat::ASTC_2D_5X4:
case PixelFormat::ASTC_2D_8X8:
case PixelFormat::ASTC_2D_8X5:
case PixelFormat::ASTC_2D_4X4_SRGB:
case PixelFormat::ASTC_2D_5X4_SRGB:
case PixelFormat::ASTC_2D_8X8_SRGB:
case PixelFormat::ASTC_2D_8X5_SRGB:
return true;
default:
return false;
}
}
static std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) {
switch (format) {
case PixelFormat::ASTC_2D_4X4:
return {4, 4};
case PixelFormat::ASTC_2D_5X4:
return {5, 4};
case PixelFormat::ASTC_2D_8X8:
return {8, 8};
case PixelFormat::ASTC_2D_8X5:
return {8, 5};
case PixelFormat::ASTC_2D_4X4_SRGB:
return {4, 4};
case PixelFormat::ASTC_2D_5X4_SRGB:
return {5, 4};
case PixelFormat::ASTC_2D_8X8_SRGB:
return {8, 8};
case PixelFormat::ASTC_2D_8X5_SRGB:
return {8, 5};
default:
LOG_CRITICAL(HW_GPU, "Unhandled format: {}", static_cast<u32>(format));
UNREACHABLE();
}
}
void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
const auto cpu_addr{memory_manager.GpuToCpuAddress(gpu_addr_)};
@ -267,7 +232,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
return params;
}
static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_format_tuples = {{
static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{
{GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV, ComponentType::UNorm, false}, // ABGR8U
{GL_RGBA8, GL_RGBA, GL_BYTE, ComponentType::SNorm, false}, // ABGR8S
{GL_RGBA8UI, GL_RGBA_INTEGER, GL_UNSIGNED_BYTE, ComponentType::UInt, false}, // ABGR8UI
@ -355,19 +320,19 @@ static constexpr std::array<FormatTuple, SurfaceParams::MaxPixelFormat> tex_form
ComponentType::Float, false}, // Z32FS8
}};
static GLenum SurfaceTargetToGL(SurfaceParams::SurfaceTarget target) {
static GLenum SurfaceTargetToGL(SurfaceTarget target) {
switch (target) {
case SurfaceParams::SurfaceTarget::Texture1D:
case SurfaceTarget::Texture1D:
return GL_TEXTURE_1D;
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
return GL_TEXTURE_2D;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceTarget::Texture3D:
return GL_TEXTURE_3D;
case SurfaceParams::SurfaceTarget::Texture1DArray:
case SurfaceTarget::Texture1DArray:
return GL_TEXTURE_1D_ARRAY;
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture2DArray:
return GL_TEXTURE_2D_ARRAY;
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceTarget::TextureCubemap:
return GL_TEXTURE_CUBE_MAP;
}
LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
@ -392,31 +357,10 @@ MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
return {0, actual_height, MipWidth(mip_level), 0};
}
/// Returns true if the specified PixelFormat is a BCn format, e.g. DXT or DXN
static bool IsFormatBCn(PixelFormat format) {
switch (format) {
case PixelFormat::DXT1:
case PixelFormat::DXT23:
case PixelFormat::DXT45:
case PixelFormat::DXN1:
case PixelFormat::DXN2SNORM:
case PixelFormat::DXN2UNORM:
case PixelFormat::BC7U:
case PixelFormat::BC6H_UF16:
case PixelFormat::BC6H_SF16:
case PixelFormat::DXT1_SRGB:
case PixelFormat::DXT23_SRGB:
case PixelFormat::DXT45_SRGB:
case PixelFormat::BC7U_SRGB:
return true;
}
return false;
}
template <bool morton_to_gl, PixelFormat format>
void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
std::size_t gl_buffer_size, VAddr addr) {
constexpr u32 bytes_per_pixel = SurfaceParams::GetBytesPerPixel(format);
constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
// With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
// pixel values.
@ -435,7 +379,7 @@ void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 d
}
using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
SurfaceParams::MaxPixelFormat>;
VideoCore::Surface::MaxPixelFormat>;
static constexpr GLConversionArray morton_to_gl_fns = {
// clang-format off
@ -575,7 +519,7 @@ static constexpr GLConversionArray gl_to_morton_fns = {
void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
std::vector<u8>& gl_buffer, u32 mip_level) {
u32 depth = params.MipDepth(mip_level);
if (params.target == SurfaceParams::SurfaceTarget::Texture2D) {
if (params.target == SurfaceTarget::Texture2D) {
// TODO(Blinkhawk): Eliminate this condition once all texture types are implemented.
depth = 1U;
}
@ -622,13 +566,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
if (src_params.type == SurfaceType::ColorTexture) {
switch (src_params.target) {
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
GL_TEXTURE_2D, src_surface->Texture().handle, 0);
glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
0, 0);
break;
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceTarget::TextureCubemap:
glFramebufferTexture2D(
GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@ -637,12 +581,12 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
break;
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture2DArray:
glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
src_surface->Texture().handle, 0, 0);
glFramebufferTextureLayer(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
break;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceTarget::Texture3D:
glFramebufferTexture3D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + src_attachment,
SurfaceTargetToGL(src_params.target),
src_surface->Texture().handle, 0, 0);
@ -658,13 +602,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
}
switch (dst_params.target) {
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
GL_TEXTURE_2D, dst_surface->Texture().handle, 0);
glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D,
0, 0);
break;
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceTarget::TextureCubemap:
glFramebufferTexture2D(
GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face),
@ -673,13 +617,13 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT,
static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + cubemap_face), 0, 0);
break;
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture2DArray:
glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
dst_surface->Texture().handle, 0, 0);
glFramebufferTextureLayer(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, 0, 0, 0);
break;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceTarget::Texture3D:
glFramebufferTexture3D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + dst_attachment,
SurfaceTargetToGL(dst_params.target),
dst_surface->Texture().handle, 0, 0);
@ -800,21 +744,21 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
UNREACHABLE();
} else {
switch (dst_params.target) {
case SurfaceParams::SurfaceTarget::Texture1D:
case SurfaceTarget::Texture1D:
glTextureSubImage1D(dst_surface->Texture().handle, 0, 0, width, dest_format.format,
dest_format.type, nullptr);
break;
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
glTextureSubImage2D(dst_surface->Texture().handle, 0, 0, 0, width, height,
dest_format.format, dest_format.type, nullptr);
break;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture3D:
case SurfaceTarget::Texture2DArray:
glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0, 0, width, height,
static_cast<GLsizei>(dst_params.depth), dest_format.format,
dest_format.type, nullptr);
break;
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceTarget::TextureCubemap:
glTextureSubImage3D(dst_surface->Texture().handle, 0, 0, 0,
static_cast<GLint>(cubemap_face), width, height, 1,
dest_format.format, dest_format.type, nullptr);
@ -851,17 +795,17 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
if (!format_tuple.compressed) {
// Only pre-create the texture for non-compressed textures.
switch (params.target) {
case SurfaceParams::SurfaceTarget::Texture1D:
case SurfaceTarget::Texture1D:
glTexStorage1D(SurfaceTargetToGL(params.target), params.max_mip_level,
format_tuple.internal_format, rect.GetWidth());
break;
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceTarget::Texture2D:
case SurfaceTarget::TextureCubemap:
glTexStorage2D(SurfaceTargetToGL(params.target), params.max_mip_level,
format_tuple.internal_format, rect.GetWidth(), rect.GetHeight());
break;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture3D:
case SurfaceTarget::Texture2DArray:
glTexStorage3D(SurfaceTargetToGL(params.target), params.max_mip_level,
format_tuple.internal_format, rect.GetWidth(), rect.GetHeight(),
params.depth);
@ -916,7 +860,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo
S8Z24 s8z24_pixel{};
Z24S8 z24s8_pixel{};
constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::S8Z24)};
constexpr auto bpp{GetBytesPerPixel(PixelFormat::S8Z24)};
for (std::size_t y = 0; y < height; ++y) {
for (std::size_t x = 0; x < width; ++x) {
const std::size_t offset{bpp * (y * width + x)};
@ -936,7 +880,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height, bo
}
static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
constexpr auto bpp{SurfaceParams::GetBytesPerPixel(PixelFormat::G8R8U)};
constexpr auto bpp{GetBytesPerPixel(PixelFormat::G8R8U)};
for (std::size_t y = 0; y < height; ++y) {
for (std::size_t x = 0; x < width; ++x) {
const std::size_t offset{bpp * (y * width + x)};
@ -1042,7 +986,7 @@ void CachedSurface::FlushGLBuffer() {
const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
// Ensure no bad interactions with GL_UNPACK_ALIGNMENT
ASSERT(params.width * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 == 0);
ASSERT(params.width * GetBytesPerPixel(params.pixel_format) % 4 == 0);
glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.width));
ASSERT(!tuple.compressed);
glBindBuffer(GL_PIXEL_PACK_BUFFER, 0);
@ -1074,7 +1018,7 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
std::size_t buffer_offset =
static_cast<std::size_t>(static_cast<std::size_t>(y0) * params.MipWidth(mip_map) +
static_cast<std::size_t>(x0)) *
SurfaceParams::GetBytesPerPixel(params.pixel_format);
GetBytesPerPixel(params.pixel_format);
const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
const GLuint target_tex = texture.handle;
@ -1090,35 +1034,34 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
cur_state.Apply();
// Ensure no bad interactions with GL_UNPACK_ALIGNMENT
ASSERT(params.MipWidth(mip_map) * SurfaceParams::GetBytesPerPixel(params.pixel_format) % 4 ==
0);
ASSERT(params.MipWidth(mip_map) * GetBytesPerPixel(params.pixel_format) % 4 == 0);
glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map)));
GLsizei image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false));
glActiveTexture(GL_TEXTURE0);
if (tuple.compressed) {
switch (params.target) {
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
glCompressedTexImage2D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
static_cast<GLsizei>(params.MipWidth(mip_map)),
static_cast<GLsizei>(params.MipHeight(mip_map)), 0, image_size,
&gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceTarget::Texture3D:
glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
static_cast<GLsizei>(params.MipWidth(mip_map)),
static_cast<GLsizei>(params.MipHeight(mip_map)),
static_cast<GLsizei>(params.MipDepth(mip_map)), 0, image_size,
&gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture2DArray:
glCompressedTexImage3D(SurfaceTargetToGL(params.target), mip_map, tuple.internal_format,
static_cast<GLsizei>(params.MipWidth(mip_map)),
static_cast<GLsizei>(params.MipHeight(mip_map)),
static_cast<GLsizei>(params.depth), 0, image_size,
&gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::TextureCubemap: {
case SurfaceTarget::TextureCubemap: {
GLsizei layer_size = static_cast<GLsizei>(params.LayerSizeGL(mip_map));
for (std::size_t face = 0; face < params.depth; ++face) {
glCompressedTexImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face),
@ -1143,30 +1086,30 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle,
} else {
switch (params.target) {
case SurfaceParams::SurfaceTarget::Texture1D:
case SurfaceTarget::Texture1D:
glTexSubImage1D(SurfaceTargetToGL(params.target), mip_map, x0,
static_cast<GLsizei>(rect.GetWidth()), tuple.format, tuple.type,
&gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
glTexSubImage2D(SurfaceTargetToGL(params.target), mip_map, x0, y0,
static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
&gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceTarget::Texture3D:
glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), params.MipDepth(mip_map),
tuple.format, tuple.type, &gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::Texture2DArray:
case SurfaceTarget::Texture2DArray:
glTexSubImage3D(SurfaceTargetToGL(params.target), mip_map, x0, y0, 0,
static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), params.depth, tuple.format,
tuple.type, &gl_buffer[mip_map][buffer_offset]);
break;
case SurfaceParams::SurfaceTarget::TextureCubemap: {
case SurfaceTarget::TextureCubemap: {
std::size_t start = buffer_offset;
for (std::size_t face = 0; face < params.depth; ++face) {
glTexSubImage2D(static_cast<GLenum>(GL_TEXTURE_CUBE_MAP_POSITIVE_X + face), mip_map,
@ -1341,8 +1284,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
// For compatible surfaces, we can just do fast glCopyImageSubData based copy
if (old_params.target == new_params.target && old_params.type == new_params.type &&
old_params.depth == new_params.depth && old_params.depth == 1 &&
SurfaceParams::GetFormatBpp(old_params.pixel_format) ==
SurfaceParams::GetFormatBpp(new_params.pixel_format)) {
GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format)) {
FastCopySurface(old_surface, new_surface);
return new_surface;
}
@ -1355,13 +1297,14 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
const bool is_blit{old_params.pixel_format == new_params.pixel_format};
switch (new_params.target) {
case SurfaceParams::SurfaceTarget::Texture2D:
case SurfaceTarget::Texture2D:
if (is_blit) {
BlitSurface(old_surface, new_surface, read_framebuffer.handle, draw_framebuffer.handle);
} else {
CopySurface(old_surface, new_surface, copy_pbo.handle);
}
break;
case SurfaceParams::SurfaceTarget::TextureCubemap:
case SurfaceParams::SurfaceTarget::Texture3D:
case SurfaceParams::SurfaceTarget::Texture2DArray:
@ -1374,7 +1317,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
}
return new_surface;
} // namespace OpenGL
}
Surface RasterizerCacheOpenGL::TryFindFramebufferSurface(VAddr addr) const {
return TryGet(addr);

View file

@ -7,6 +7,7 @@
#include <array>
#include <map>
#include <memory>
#include <string>
#include <vector>
#include "common/alignment.h"
@ -18,6 +19,7 @@
#include "video_core/rasterizer_cache.h"
#include "video_core/renderer_opengl/gl_resource_manager.h"
#include "video_core/renderer_opengl/gl_shader_gen.h"
#include "video_core/surface.h"
#include "video_core/textures/decoders.h"
#include "video_core/textures/texture.h"
@ -27,135 +29,12 @@ class CachedSurface;
using Surface = std::shared_ptr<CachedSurface>;
using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
using SurfaceTarget = VideoCore::Surface::SurfaceTarget;
using SurfaceType = VideoCore::Surface::SurfaceType;
using PixelFormat = VideoCore::Surface::PixelFormat;
using ComponentType = VideoCore::Surface::ComponentType;
struct SurfaceParams {
enum class PixelFormat {
ABGR8U = 0,
ABGR8S = 1,
ABGR8UI = 2,
B5G6R5U = 3,
A2B10G10R10U = 4,
A1B5G5R5U = 5,
R8U = 6,
R8UI = 7,
RGBA16F = 8,
RGBA16U = 9,
RGBA16UI = 10,
R11FG11FB10F = 11,
RGBA32UI = 12,
DXT1 = 13,
DXT23 = 14,
DXT45 = 15,
DXN1 = 16, // This is also known as BC4
DXN2UNORM = 17,
DXN2SNORM = 18,
BC7U = 19,
BC6H_UF16 = 20,
BC6H_SF16 = 21,
ASTC_2D_4X4 = 22,
G8R8U = 23,
G8R8S = 24,
BGRA8 = 25,
RGBA32F = 26,
RG32F = 27,
R32F = 28,
R16F = 29,
R16U = 30,
R16S = 31,
R16UI = 32,
R16I = 33,
RG16 = 34,
RG16F = 35,
RG16UI = 36,
RG16I = 37,
RG16S = 38,
RGB32F = 39,
RGBA8_SRGB = 40,
RG8U = 41,
RG8S = 42,
RG32UI = 43,
R32UI = 44,
ASTC_2D_8X8 = 45,
ASTC_2D_8X5 = 46,
ASTC_2D_5X4 = 47,
BGRA8_SRGB = 48,
DXT1_SRGB = 49,
DXT23_SRGB = 50,
DXT45_SRGB = 51,
BC7U_SRGB = 52,
ASTC_2D_4X4_SRGB = 53,
ASTC_2D_8X8_SRGB = 54,
ASTC_2D_8X5_SRGB = 55,
ASTC_2D_5X4_SRGB = 56,
MaxColorFormat,
// Depth formats
Z32F = 57,
Z16 = 58,
MaxDepthFormat,
// DepthStencil formats
Z24S8 = 59,
S8Z24 = 60,
Z32FS8 = 61,
MaxDepthStencilFormat,
Max = MaxDepthStencilFormat,
Invalid = 255,
};
static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max);
enum class ComponentType {
Invalid = 0,
SNorm = 1,
UNorm = 2,
SInt = 3,
UInt = 4,
Float = 5,
};
enum class SurfaceType {
ColorTexture = 0,
Depth = 1,
DepthStencil = 2,
Fill = 3,
Invalid = 4,
};
enum class SurfaceTarget {
Texture1D,
Texture2D,
Texture3D,
Texture1DArray,
Texture2DArray,
TextureCubemap,
};
static SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type) {
switch (texture_type) {
case Tegra::Texture::TextureType::Texture1D:
return SurfaceTarget::Texture1D;
case Tegra::Texture::TextureType::Texture2D:
case Tegra::Texture::TextureType::Texture2DNoMipmap:
return SurfaceTarget::Texture2D;
case Tegra::Texture::TextureType::Texture3D:
return SurfaceTarget::Texture3D;
case Tegra::Texture::TextureType::TextureCubemap:
return SurfaceTarget::TextureCubemap;
case Tegra::Texture::TextureType::Texture1DArray:
return SurfaceTarget::Texture1DArray;
case Tegra::Texture::TextureType::Texture2DArray:
return SurfaceTarget::Texture2DArray;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type));
UNREACHABLE();
return SurfaceTarget::Texture2D;
}
}
static std::string SurfaceTargetName(SurfaceTarget target) {
switch (target) {
case SurfaceTarget::Texture1D:
@ -177,660 +56,8 @@ struct SurfaceParams {
}
}
static bool SurfaceTargetIsLayered(SurfaceTarget target) {
switch (target) {
case SurfaceTarget::Texture1D:
case SurfaceTarget::Texture2D:
case SurfaceTarget::Texture3D:
return false;
case SurfaceTarget::Texture1DArray:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubemap:
return true;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target));
UNREACHABLE();
return false;
}
}
/**
* 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.
*/
static constexpr u32 GetCompressionFactor(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> compression_factor_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
4, // ASTC_2D_8X8
4, // ASTC_2D_8X5
4, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
4, // ASTC_2D_8X8_SRGB
4, // ASTC_2D_8X5_SRGB
4, // ASTC_2D_5X4_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < compression_factor_table.size());
return compression_factor_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetDefaultBlockHeight(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
1, // ABGR8U
1, // ABGR8S
1, // ABGR8UI
1, // B5G6R5U
1, // A2B10G10R10U
1, // A1B5G5R5U
1, // R8U
1, // R8UI
1, // RGBA16F
1, // RGBA16U
1, // RGBA16UI
1, // R11FG11FB10F
1, // RGBA32UI
4, // DXT1
4, // DXT23
4, // DXT45
4, // DXN1
4, // DXN2UNORM
4, // DXN2SNORM
4, // BC7U
4, // BC6H_UF16
4, // BC6H_SF16
4, // ASTC_2D_4X4
1, // G8R8U
1, // G8R8S
1, // BGRA8
1, // RGBA32F
1, // RG32F
1, // R32F
1, // R16F
1, // R16U
1, // R16S
1, // R16UI
1, // R16I
1, // RG16
1, // RG16F
1, // RG16UI
1, // RG16I
1, // RG16S
1, // RGB32F
1, // RGBA8_SRGB
1, // RG8U
1, // RG8S
1, // RG32UI
1, // R32UI
8, // ASTC_2D_8X8
5, // ASTC_2D_8X5
4, // ASTC_2D_5X4
1, // BGRA8_SRGB
4, // DXT1_SRGB
4, // DXT23_SRGB
4, // DXT45_SRGB
4, // BC7U_SRGB
4, // ASTC_2D_4X4_SRGB
8, // ASTC_2D_8X8_SRGB
5, // ASTC_2D_8X5_SRGB
4, // ASTC_2D_5X4_SRGB
1, // Z32F
1, // Z16
1, // Z24S8
1, // S8Z24
1, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < block_height_table.size());
return block_height_table[static_cast<std::size_t>(format)];
}
static constexpr u32 GetFormatBpp(PixelFormat format) {
if (format == PixelFormat::Invalid)
return 0;
constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
32, // ABGR8U
32, // ABGR8S
32, // ABGR8UI
16, // B5G6R5U
32, // A2B10G10R10U
16, // A1B5G5R5U
8, // R8U
8, // R8UI
64, // RGBA16F
64, // RGBA16U
64, // RGBA16UI
32, // R11FG11FB10F
128, // RGBA32UI
64, // DXT1
128, // DXT23
128, // DXT45
64, // DXN1
128, // DXN2UNORM
128, // DXN2SNORM
128, // BC7U
128, // BC6H_UF16
128, // BC6H_SF16
32, // ASTC_2D_4X4
16, // G8R8U
16, // G8R8S
32, // BGRA8
128, // RGBA32F
64, // RG32F
32, // R32F
16, // R16F
16, // R16U
16, // R16S
16, // R16UI
16, // R16I
32, // RG16
32, // RG16F
32, // RG16UI
32, // RG16I
32, // RG16S
96, // RGB32F
32, // RGBA8_SRGB
16, // RG8U
16, // RG8S
64, // RG32UI
32, // R32UI
16, // ASTC_2D_8X8
16, // ASTC_2D_8X5
32, // ASTC_2D_5X4
32, // BGRA8_SRGB
64, // DXT1_SRGB
128, // DXT23_SRGB
128, // DXT45_SRGB
128, // BC7U
32, // ASTC_2D_4X4_SRGB
16, // ASTC_2D_8X8_SRGB
16, // ASTC_2D_8X5_SRGB
32, // ASTC_2D_5X4_SRGB
32, // Z32F
16, // Z16
32, // Z24S8
32, // S8Z24
64, // Z32FS8
}};
ASSERT(static_cast<std::size_t>(format) < bpp_table.size());
return bpp_table[static_cast<std::size_t>(format)];
}
u32 GetFormatBpp() const {
return GetFormatBpp(pixel_format);
}
static PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
switch (format) {
case Tegra::DepthFormat::S8_Z24_UNORM:
return PixelFormat::S8Z24;
case Tegra::DepthFormat::Z24_S8_UNORM:
return PixelFormat::Z24S8;
case Tegra::DepthFormat::Z32_FLOAT:
return PixelFormat::Z32F;
case Tegra::DepthFormat::Z16_UNORM:
return PixelFormat::Z16;
case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
return PixelFormat::Z32FS8;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
switch (format) {
// TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
// gamma.
case Tegra::RenderTargetFormat::RGBA8_SRGB:
return PixelFormat::RGBA8_SRGB;
case Tegra::RenderTargetFormat::RGBA8_UNORM:
return PixelFormat::ABGR8U;
case Tegra::RenderTargetFormat::RGBA8_SNORM:
return PixelFormat::ABGR8S;
case Tegra::RenderTargetFormat::RGBA8_UINT:
return PixelFormat::ABGR8UI;
case Tegra::RenderTargetFormat::BGRA8_SRGB:
return PixelFormat::BGRA8_SRGB;
case Tegra::RenderTargetFormat::BGRA8_UNORM:
return PixelFormat::BGRA8;
case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
return PixelFormat::A2B10G10R10U;
case Tegra::RenderTargetFormat::RGBA16_FLOAT:
return PixelFormat::RGBA16F;
case Tegra::RenderTargetFormat::RGBA16_UNORM:
return PixelFormat::RGBA16U;
case Tegra::RenderTargetFormat::RGBA16_UINT:
return PixelFormat::RGBA16UI;
case Tegra::RenderTargetFormat::RGBA32_FLOAT:
return PixelFormat::RGBA32F;
case Tegra::RenderTargetFormat::RG32_FLOAT:
return PixelFormat::RG32F;
case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
return PixelFormat::R11FG11FB10F;
case Tegra::RenderTargetFormat::B5G6R5_UNORM:
return PixelFormat::B5G6R5U;
case Tegra::RenderTargetFormat::BGR5A1_UNORM:
return PixelFormat::A1B5G5R5U;
case Tegra::RenderTargetFormat::RGBA32_UINT:
return PixelFormat::RGBA32UI;
case Tegra::RenderTargetFormat::R8_UNORM:
return PixelFormat::R8U;
case Tegra::RenderTargetFormat::R8_UINT:
return PixelFormat::R8UI;
case Tegra::RenderTargetFormat::RG16_FLOAT:
return PixelFormat::RG16F;
case Tegra::RenderTargetFormat::RG16_UINT:
return PixelFormat::RG16UI;
case Tegra::RenderTargetFormat::RG16_SINT:
return PixelFormat::RG16I;
case Tegra::RenderTargetFormat::RG16_UNORM:
return PixelFormat::RG16;
case Tegra::RenderTargetFormat::RG16_SNORM:
return PixelFormat::RG16S;
case Tegra::RenderTargetFormat::RG8_UNORM:
return PixelFormat::RG8U;
case Tegra::RenderTargetFormat::RG8_SNORM:
return PixelFormat::RG8S;
case Tegra::RenderTargetFormat::R16_FLOAT:
return PixelFormat::R16F;
case Tegra::RenderTargetFormat::R16_UNORM:
return PixelFormat::R16U;
case Tegra::RenderTargetFormat::R16_SNORM:
return PixelFormat::R16S;
case Tegra::RenderTargetFormat::R16_UINT:
return PixelFormat::R16UI;
case Tegra::RenderTargetFormat::R16_SINT:
return PixelFormat::R16I;
case Tegra::RenderTargetFormat::R32_FLOAT:
return PixelFormat::R32F;
case Tegra::RenderTargetFormat::R32_UINT:
return PixelFormat::R32UI;
case Tegra::RenderTargetFormat::RG32_UINT:
return PixelFormat::RG32UI;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format,
Tegra::Texture::ComponentType component_type,
bool is_srgb) {
// TODO(Subv): Properly implement this
switch (format) {
case Tegra::Texture::TextureFormat::A8R8G8B8:
if (is_srgb) {
return PixelFormat::RGBA8_SRGB;
}
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::ABGR8U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::ABGR8S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::ABGR8UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::B5G6R5:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::B5G6R5U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::A2B10G10R10:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::A2B10G10R10U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::A1B5G5R5:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::A1B5G5R5U;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R8:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::R8U;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R8UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::G8R8:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::G8R8U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::G8R8S;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R16_G16_B16_A16:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::RGBA16U;
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGBA16F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::BF10GF11RF11:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R11FG11FB10F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32_B32_A32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGBA32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RGBA32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RG32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RG32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32_G32_B32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RGB32F;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R16:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R16F;
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::R16U;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::R16S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R16UI;
case Tegra::Texture::ComponentType::SINT:
return PixelFormat::R16I;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::R32:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::R32F;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::R32UI;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::ZF32:
return PixelFormat::Z32F;
case Tegra::Texture::TextureFormat::Z16:
return PixelFormat::Z16;
case Tegra::Texture::TextureFormat::Z24S8:
return PixelFormat::Z24S8;
case Tegra::Texture::TextureFormat::DXT1:
return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1;
case Tegra::Texture::TextureFormat::DXT23:
return is_srgb ? PixelFormat::DXT23_SRGB : PixelFormat::DXT23;
case Tegra::Texture::TextureFormat::DXT45:
return is_srgb ? PixelFormat::DXT45_SRGB : PixelFormat::DXT45;
case Tegra::Texture::TextureFormat::DXN1:
return PixelFormat::DXN1;
case Tegra::Texture::TextureFormat::DXN2:
switch (component_type) {
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::DXN2UNORM;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::DXN2SNORM;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
case Tegra::Texture::TextureFormat::BC7U:
return is_srgb ? PixelFormat::BC7U_SRGB : PixelFormat::BC7U;
case Tegra::Texture::TextureFormat::BC6H_UF16:
return PixelFormat::BC6H_UF16;
case Tegra::Texture::TextureFormat::BC6H_SF16:
return PixelFormat::BC6H_SF16;
case Tegra::Texture::TextureFormat::ASTC_2D_4X4:
return is_srgb ? PixelFormat::ASTC_2D_4X4_SRGB : PixelFormat::ASTC_2D_4X4;
case Tegra::Texture::TextureFormat::ASTC_2D_5X4:
return is_srgb ? PixelFormat::ASTC_2D_5X4_SRGB : PixelFormat::ASTC_2D_5X4;
case Tegra::Texture::TextureFormat::ASTC_2D_8X8:
return is_srgb ? PixelFormat::ASTC_2D_8X8_SRGB : PixelFormat::ASTC_2D_8X8;
case Tegra::Texture::TextureFormat::ASTC_2D_8X5:
return is_srgb ? PixelFormat::ASTC_2D_8X5_SRGB : PixelFormat::ASTC_2D_8X5;
case Tegra::Texture::TextureFormat::R16_G16:
switch (component_type) {
case Tegra::Texture::ComponentType::FLOAT:
return PixelFormat::RG16F;
case Tegra::Texture::ComponentType::UNORM:
return PixelFormat::RG16;
case Tegra::Texture::ComponentType::SNORM:
return PixelFormat::RG16S;
case Tegra::Texture::ComponentType::UINT:
return PixelFormat::RG16UI;
case Tegra::Texture::ComponentType::SINT:
return PixelFormat::RG16I;
}
LOG_CRITICAL(HW_GPU, "Unimplemented component_type={}",
static_cast<u32>(component_type));
UNREACHABLE();
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}, component_type={}",
static_cast<u32>(format), static_cast<u32>(component_type));
UNREACHABLE();
}
}
static ComponentType ComponentTypeFromTexture(Tegra::Texture::ComponentType type) {
// TODO(Subv): Implement more component types
switch (type) {
case Tegra::Texture::ComponentType::UNORM:
return ComponentType::UNorm;
case Tegra::Texture::ComponentType::FLOAT:
return ComponentType::Float;
case Tegra::Texture::ComponentType::SNORM:
return ComponentType::SNorm;
case Tegra::Texture::ComponentType::UINT:
return ComponentType::UInt;
case Tegra::Texture::ComponentType::SINT:
return ComponentType::SInt;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented component type={}", static_cast<u32>(type));
UNREACHABLE();
}
}
static ComponentType ComponentTypeFromRenderTarget(Tegra::RenderTargetFormat format) {
// TODO(Subv): Implement more render targets
switch (format) {
case Tegra::RenderTargetFormat::RGBA8_UNORM:
case Tegra::RenderTargetFormat::RGBA8_SRGB:
case Tegra::RenderTargetFormat::BGRA8_UNORM:
case Tegra::RenderTargetFormat::BGRA8_SRGB:
case Tegra::RenderTargetFormat::RGB10_A2_UNORM:
case Tegra::RenderTargetFormat::R8_UNORM:
case Tegra::RenderTargetFormat::RG16_UNORM:
case Tegra::RenderTargetFormat::R16_UNORM:
case Tegra::RenderTargetFormat::B5G6R5_UNORM:
case Tegra::RenderTargetFormat::BGR5A1_UNORM:
case Tegra::RenderTargetFormat::RG8_UNORM:
case Tegra::RenderTargetFormat::RGBA16_UNORM:
return ComponentType::UNorm;
case Tegra::RenderTargetFormat::RGBA8_SNORM:
case Tegra::RenderTargetFormat::RG16_SNORM:
case Tegra::RenderTargetFormat::R16_SNORM:
case Tegra::RenderTargetFormat::RG8_SNORM:
return ComponentType::SNorm;
case Tegra::RenderTargetFormat::RGBA16_FLOAT:
case Tegra::RenderTargetFormat::R11G11B10_FLOAT:
case Tegra::RenderTargetFormat::RGBA32_FLOAT:
case Tegra::RenderTargetFormat::RG32_FLOAT:
case Tegra::RenderTargetFormat::RG16_FLOAT:
case Tegra::RenderTargetFormat::R16_FLOAT:
case Tegra::RenderTargetFormat::R32_FLOAT:
return ComponentType::Float;
case Tegra::RenderTargetFormat::RGBA32_UINT:
case Tegra::RenderTargetFormat::RGBA16_UINT:
case Tegra::RenderTargetFormat::RG16_UINT:
case Tegra::RenderTargetFormat::R8_UINT:
case Tegra::RenderTargetFormat::R16_UINT:
case Tegra::RenderTargetFormat::RG32_UINT:
case Tegra::RenderTargetFormat::R32_UINT:
case Tegra::RenderTargetFormat::RGBA8_UINT:
return ComponentType::UInt;
case Tegra::RenderTargetFormat::RG16_SINT:
case Tegra::RenderTargetFormat::R16_SINT:
return ComponentType::SInt;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat format) {
switch (format) {
case Tegra::FramebufferConfig::PixelFormat::ABGR8:
return PixelFormat::ABGR8U;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static ComponentType ComponentTypeFromDepthFormat(Tegra::DepthFormat format) {
switch (format) {
case Tegra::DepthFormat::Z16_UNORM:
case Tegra::DepthFormat::S8_Z24_UNORM:
case Tegra::DepthFormat::Z24_S8_UNORM:
return ComponentType::UNorm;
case Tegra::DepthFormat::Z32_FLOAT:
case Tegra::DepthFormat::Z32_S8_X24_FLOAT:
return ComponentType::Float;
default:
LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
UNREACHABLE();
}
}
static SurfaceType GetFormatType(PixelFormat pixel_format) {
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxColorFormat)) {
return SurfaceType::ColorTexture;
}
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxDepthFormat)) {
return SurfaceType::Depth;
}
if (static_cast<std::size_t>(pixel_format) <
static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) {
return SurfaceType::DepthStencil;
}
// TODO(Subv): Implement the other formats
ASSERT(false);
return SurfaceType::Invalid;
}
/// Returns the sizer in bytes of the specified pixel format
static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) {
if (pixel_format == SurfaceParams::PixelFormat::Invalid) {
return 0;
}
return GetFormatBpp(pixel_format) / CHAR_BIT;
return VideoCore::Surface::GetFormatBpp(pixel_format);
}
/// Returns the rectangle corresponding to this surface

View file

@ -373,6 +373,7 @@ public:
if (sets_cc) {
const std::string zero_condition = "( " + ConvertIntegerSize(value, size) + " == 0 )";
SetInternalFlag(InternalFlag::ZeroFlag, zero_condition);
LOG_WARNING(HW_GPU, "Control Codes Imcomplete.");
}
}
@ -1525,6 +1526,10 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1,
instr.alu.saturate_d, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FMUL Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::FADD_C:
@ -1535,6 +1540,10 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1,
instr.alu.saturate_d, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FADD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::MUFU: {
@ -1588,6 +1597,10 @@ private:
'(' + condition + ") ? min(" + parameters + ") : max(" +
parameters + ')',
1, 1, false, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FMNMX Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::RRO_C:
@ -1618,6 +1631,10 @@ private:
regs.GetRegisterAsFloat(instr.gpr8) + " * " +
GetImmediate32(instr),
1, 1, instr.fmul32.saturate, 0, true);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "FMUL32 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::FADD32I: {
@ -1641,6 +1658,10 @@ private:
}
regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1, false, 0, true);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "FADD32 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
}
@ -1661,6 +1682,10 @@ private:
std::to_string(instr.bfe.GetLeftShiftValue() + instr.bfe.shift_position) + ')';
regs.SetRegisterToInteger(instr.gpr0, true, 0, outer_shift, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "BFE Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
@ -1698,12 +1723,20 @@ private:
// Cast to int is superfluous for arithmetic shift, it's only for a logical shift
regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(" + op_a + " >> " + op_b + ')',
1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "SHR Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::SHL_C:
case OpCode::Id::SHL_R:
case OpCode::Id::SHL_IMM:
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " << " + op_b, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "SHL Generates an unhandled Control Code");
UNREACHABLE();
}
break;
default: {
LOG_CRITICAL(HW_GPU, "Unhandled shift instruction: {}", opcode->get().GetName());
@ -1723,6 +1756,10 @@ private:
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
instr.iadd32i.saturate != 0);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "IADD32 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
case OpCode::Id::LOP32I: {
if (instr.alu.lop32i.invert_a)
@ -1734,6 +1771,10 @@ private:
WriteLogicOperation(instr.gpr0, instr.alu.lop32i.operation, op_a, op_b,
Tegra::Shader::PredicateResultMode::None,
Tegra::Shader::Pred::UnusedIndex);
if (instr.op_32.generates_cc) {
LOG_CRITICAL(HW_GPU, "LOP32I Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
@ -1770,6 +1811,10 @@ private:
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1,
instr.alu.saturate_d);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "IADD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::IADD3_C:
@ -1831,6 +1876,11 @@ private:
}
regs.SetRegisterToInteger(instr.gpr0, true, 0, result, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "IADD3 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::ISCADD_C:
@ -1846,6 +1896,10 @@ private:
regs.SetRegisterToInteger(instr.gpr0, true, 0,
"((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "ISCADD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::POPC_C:
@ -1877,6 +1931,10 @@ private:
WriteLogicOperation(instr.gpr0, instr.alu.lop.operation, op_a, op_b,
instr.alu.lop.pred_result_mode, instr.alu.lop.pred48);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "LOP Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::LOP3_C:
@ -1892,6 +1950,10 @@ private:
}
WriteLop3Instruction(instr.gpr0, op_a, op_b, op_c, lut);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "LOP3 Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::IMNMX_C:
@ -1906,6 +1968,10 @@ private:
'(' + condition + ") ? min(" + parameters + ") : max(" +
parameters + ')',
1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "IMNMX Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::LEA_R2:
@ -2107,6 +2173,10 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, "fma(" + op_a + ", " + op_b + ", " + op_c + ')',
1, 1, instr.alu.saturate_d, 0, true);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "FFMA Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
@ -2212,6 +2282,11 @@ private:
}
regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "I2F Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::F2F_R: {
@ -2250,6 +2325,11 @@ private:
}
regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1, instr.alu.saturate_d);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "F2F Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::F2I_R:
@ -2299,6 +2379,10 @@ private:
regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_output_signed, 0, op_a, 1,
1, false, 0, instr.conversion.dest_size);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "F2I Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
@ -3107,6 +3191,11 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, value, 1, 1);
}
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "PSET Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Type::PredicateSetPredicate: {
@ -3372,6 +3461,10 @@ private:
}
regs.SetRegisterToInteger(instr.gpr0, is_signed, 0, sum, 1, 1);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "XMAD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
default: {
@ -3381,6 +3474,12 @@ private:
EmitFragmentOutputsWrite();
}
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "EXIT Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
switch (instr.flow.cond) {
case Tegra::Shader::FlowCondition::Always:
shader.AddLine("return true;");
@ -3410,6 +3509,11 @@ private:
// Enclose "discard" in a conditional, so that GLSL compilation does not complain
// about unexecuted instructions that may follow this.
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "KIL Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
shader.AddLine("if (true) {");
++shader.scope;
shader.AddLine("discard;");
@ -3467,6 +3571,11 @@ private:
case OpCode::Id::BRA: {
ASSERT_MSG(instr.bra.constant_buffer == 0,
"BRA with constant buffers are not implemented");
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "BRA Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
const u32 target = offset + instr.bra.GetBranchTarget();
shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
break;
@ -3507,13 +3616,21 @@ private:
}
case OpCode::Id::SYNC: {
// The SYNC opcode jumps to the address previously set by the SSY opcode
ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "SYNC Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
EmitPopFromFlowStack();
break;
}
case OpCode::Id::BRK: {
// The BRK opcode jumps to the address previously set by the PBK opcode
ASSERT(instr.flow.cond == Tegra::Shader::FlowCondition::Always);
const Tegra::Shader::ControlCode cc = instr.flow_control_code;
if (cc != Tegra::Shader::ControlCode::T) {
LOG_CRITICAL(HW_GPU, "BRK Control Code used: {}", static_cast<u32>(cc));
UNREACHABLE();
}
EmitPopFromFlowStack();
break;
}
@ -3543,6 +3660,11 @@ private:
regs.SetRegisterToInteger(instr.gpr0, result_signed, 1, result, 1, 1,
instr.vmad.saturate == 1, 0, Register::Size::Word,
instr.vmad.cc);
if (instr.generates_cc) {
LOG_CRITICAL(HW_GPU, "VMAD Generates an unhandled Control Code");
UNREACHABLE();
}
break;
}
case OpCode::Id::VSETP: {

View file

@ -89,7 +89,18 @@ OpenGLState::OpenGLState() {
point.size = 1;
}
void OpenGLState::Apply() const {
void OpenGLState::ApplyDefaultState() {
glDisable(GL_FRAMEBUFFER_SRGB);
glDisable(GL_CULL_FACE);
glDisable(GL_DEPTH_TEST);
glDisable(GL_PRIMITIVE_RESTART);
glDisable(GL_STENCIL_TEST);
glEnable(GL_BLEND);
glDisable(GL_COLOR_LOGIC_OP);
glDisable(GL_SCISSOR_TEST);
}
void OpenGLState::ApplySRgb() const {
// sRGB
if (framebuffer_srgb.enabled != cur_state.framebuffer_srgb.enabled) {
if (framebuffer_srgb.enabled) {
@ -100,96 +111,122 @@ void OpenGLState::Apply() const {
glDisable(GL_FRAMEBUFFER_SRGB);
}
}
}
void OpenGLState::ApplyCulling() const {
// Culling
if (cull.enabled != cur_state.cull.enabled) {
const bool cull_changed = cull.enabled != cur_state.cull.enabled;
if (cull_changed) {
if (cull.enabled) {
glEnable(GL_CULL_FACE);
} else {
glDisable(GL_CULL_FACE);
}
}
if (cull.enabled) {
if (cull_changed || cull.mode != cur_state.cull.mode) {
glCullFace(cull.mode);
}
if (cull.mode != cur_state.cull.mode) {
glCullFace(cull.mode);
}
if (cull.front_face != cur_state.cull.front_face) {
glFrontFace(cull.front_face);
if (cull_changed || cull.front_face != cur_state.cull.front_face) {
glFrontFace(cull.front_face);
}
}
}
void OpenGLState::ApplyDepth() const {
// Depth test
if (depth.test_enabled != cur_state.depth.test_enabled) {
const bool depth_test_changed = depth.test_enabled != cur_state.depth.test_enabled;
if (depth_test_changed) {
if (depth.test_enabled) {
glEnable(GL_DEPTH_TEST);
} else {
glDisable(GL_DEPTH_TEST);
}
}
if (depth.test_func != cur_state.depth.test_func) {
if (depth.test_enabled &&
(depth_test_changed || depth.test_func != cur_state.depth.test_func)) {
glDepthFunc(depth.test_func);
}
// Depth mask
if (depth.write_mask != cur_state.depth.write_mask) {
glDepthMask(depth.write_mask);
}
// Depth range
if (depth.depth_range_near != cur_state.depth.depth_range_near ||
depth.depth_range_far != cur_state.depth.depth_range_far) {
glDepthRange(depth.depth_range_near, depth.depth_range_far);
}
}
// Primitive restart
if (primitive_restart.enabled != cur_state.primitive_restart.enabled) {
void OpenGLState::ApplyPrimitiveRestart() const {
const bool primitive_restart_changed =
primitive_restart.enabled != cur_state.primitive_restart.enabled;
if (primitive_restart_changed) {
if (primitive_restart.enabled) {
glEnable(GL_PRIMITIVE_RESTART);
} else {
glDisable(GL_PRIMITIVE_RESTART);
}
}
if (primitive_restart.index != cur_state.primitive_restart.index) {
if (primitive_restart_changed ||
(primitive_restart.enabled &&
primitive_restart.index != cur_state.primitive_restart.index)) {
glPrimitiveRestartIndex(primitive_restart.index);
}
}
// Color mask
if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
color_mask.green_enabled != cur_state.color_mask.green_enabled ||
color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled,
color_mask.alpha_enabled);
}
// Stencil test
if (stencil.test_enabled != cur_state.stencil.test_enabled) {
void OpenGLState::ApplyStencilTest() const {
const bool stencil_test_changed = stencil.test_enabled != cur_state.stencil.test_enabled;
if (stencil_test_changed) {
if (stencil.test_enabled) {
glEnable(GL_STENCIL_TEST);
} else {
glDisable(GL_STENCIL_TEST);
}
}
auto config_stencil = [](GLenum face, const auto& config, const auto& prev_config) {
if (config.test_func != prev_config.test_func || config.test_ref != prev_config.test_ref ||
config.test_mask != prev_config.test_mask) {
glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask);
}
if (config.action_depth_fail != prev_config.action_depth_fail ||
config.action_depth_pass != prev_config.action_depth_pass ||
config.action_stencil_fail != prev_config.action_stencil_fail) {
glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail,
config.action_depth_pass);
}
if (config.write_mask != prev_config.write_mask) {
glStencilMaskSeparate(face, config.write_mask);
}
};
config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front);
config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
if (stencil.test_enabled) {
auto config_stencil = [stencil_test_changed](GLenum face, const auto& config,
const auto& prev_config) {
if (stencil_test_changed || config.test_func != prev_config.test_func ||
config.test_ref != prev_config.test_ref ||
config.test_mask != prev_config.test_mask) {
glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask);
}
if (stencil_test_changed || config.action_depth_fail != prev_config.action_depth_fail ||
config.action_depth_pass != prev_config.action_depth_pass ||
config.action_stencil_fail != prev_config.action_stencil_fail) {
glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail,
config.action_depth_pass);
}
if (config.write_mask != prev_config.write_mask) {
glStencilMaskSeparate(face, config.write_mask);
}
};
config_stencil(GL_FRONT, stencil.front, cur_state.stencil.front);
config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
}
}
// Blending
if (blend.enabled != cur_state.blend.enabled) {
void OpenGLState::ApplyScissorTest() const {
const bool scissor_changed = scissor.enabled != cur_state.scissor.enabled;
if (scissor_changed) {
if (scissor.enabled) {
glEnable(GL_SCISSOR_TEST);
} else {
glDisable(GL_SCISSOR_TEST);
}
}
if (scissor_changed || scissor_changed || scissor.x != cur_state.scissor.x ||
scissor.y != cur_state.scissor.y || scissor.width != cur_state.scissor.width ||
scissor.height != cur_state.scissor.height) {
glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
}
}
void OpenGLState::ApplyBlending() const {
const bool blend_changed = blend.enabled != cur_state.blend.enabled;
if (blend_changed) {
if (blend.enabled) {
ASSERT(!logic_op.enabled);
glEnable(GL_BLEND);
@ -197,29 +234,32 @@ void OpenGLState::Apply() const {
glDisable(GL_BLEND);
}
}
if (blend.enabled) {
if (blend_changed || blend.color.red != cur_state.blend.color.red ||
blend.color.green != cur_state.blend.color.green ||
blend.color.blue != cur_state.blend.color.blue ||
blend.color.alpha != cur_state.blend.color.alpha) {
glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
}
if (blend.color.red != cur_state.blend.color.red ||
blend.color.green != cur_state.blend.color.green ||
blend.color.blue != cur_state.blend.color.blue ||
blend.color.alpha != cur_state.blend.color.alpha) {
glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
if (blend_changed || blend.src_rgb_func != cur_state.blend.src_rgb_func ||
blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
blend.src_a_func != cur_state.blend.src_a_func ||
blend.dst_a_func != cur_state.blend.dst_a_func) {
glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func,
blend.dst_a_func);
}
if (blend_changed || blend.rgb_equation != cur_state.blend.rgb_equation ||
blend.a_equation != cur_state.blend.a_equation) {
glBlendEquationSeparate(blend.rgb_equation, blend.a_equation);
}
}
}
if (blend.src_rgb_func != cur_state.blend.src_rgb_func ||
blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
blend.src_a_func != cur_state.blend.src_a_func ||
blend.dst_a_func != cur_state.blend.dst_a_func) {
glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func,
blend.dst_a_func);
}
if (blend.rgb_equation != cur_state.blend.rgb_equation ||
blend.a_equation != cur_state.blend.a_equation) {
glBlendEquationSeparate(blend.rgb_equation, blend.a_equation);
}
// Logic Operation
if (logic_op.enabled != cur_state.logic_op.enabled) {
void OpenGLState::ApplyLogicOp() const {
const bool logic_op_changed = logic_op.enabled != cur_state.logic_op.enabled;
if (logic_op_changed) {
if (logic_op.enabled) {
ASSERT(!blend.enabled);
glEnable(GL_COLOR_LOGIC_OP);
@ -228,11 +268,13 @@ void OpenGLState::Apply() const {
}
}
if (logic_op.operation != cur_state.logic_op.operation) {
if (logic_op.enabled &&
(logic_op_changed || logic_op.operation != cur_state.logic_op.operation)) {
glLogicOp(logic_op.operation);
}
}
// Textures
void OpenGLState::ApplyTextures() const {
for (std::size_t i = 0; i < std::size(texture_units); ++i) {
const auto& texture_unit = texture_units[i];
const auto& cur_state_texture_unit = cur_state.texture_units[i];
@ -251,28 +293,29 @@ void OpenGLState::Apply() const {
glTexParameteriv(texture_unit.target, GL_TEXTURE_SWIZZLE_RGBA, mask.data());
}
}
}
// Samplers
{
bool has_delta{};
std::size_t first{}, last{};
std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
for (std::size_t i = 0; i < std::size(samplers); ++i) {
samplers[i] = texture_units[i].sampler;
if (samplers[i] != cur_state.texture_units[i].sampler) {
if (!has_delta) {
first = i;
has_delta = true;
}
last = i;
void OpenGLState::ApplySamplers() const {
bool has_delta{};
std::size_t first{}, last{};
std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers;
for (std::size_t i = 0; i < std::size(samplers); ++i) {
samplers[i] = texture_units[i].sampler;
if (samplers[i] != cur_state.texture_units[i].sampler) {
if (!has_delta) {
first = i;
has_delta = true;
}
}
if (has_delta) {
glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
samplers.data());
last = i;
}
}
if (has_delta) {
glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
samplers.data());
}
}
void OpenGLState::Apply() const {
// Framebuffer
if (draw.read_framebuffer != cur_state.draw.read_framebuffer) {
glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer);
@ -305,27 +348,12 @@ void OpenGLState::Apply() const {
if (draw.program_pipeline != cur_state.draw.program_pipeline) {
glBindProgramPipeline(draw.program_pipeline);
}
// Scissor test
if (scissor.enabled != cur_state.scissor.enabled) {
if (scissor.enabled) {
glEnable(GL_SCISSOR_TEST);
} else {
glDisable(GL_SCISSOR_TEST);
}
}
if (scissor.x != cur_state.scissor.x || scissor.y != cur_state.scissor.y ||
scissor.width != cur_state.scissor.width || scissor.height != cur_state.scissor.height) {
glScissor(scissor.x, scissor.y, scissor.width, scissor.height);
}
// Viewport
if (viewport.x != cur_state.viewport.x || viewport.y != cur_state.viewport.y ||
viewport.width != cur_state.viewport.width ||
viewport.height != cur_state.viewport.height) {
glViewport(viewport.x, viewport.y, viewport.width, viewport.height);
}
// Clip distance
for (std::size_t i = 0; i < clip_distance.size(); ++i) {
if (clip_distance[i] != cur_state.clip_distance[i]) {
@ -336,12 +364,28 @@ void OpenGLState::Apply() const {
}
}
}
// Color mask
if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
color_mask.green_enabled != cur_state.color_mask.green_enabled ||
color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
glColorMask(color_mask.red_enabled, color_mask.green_enabled, color_mask.blue_enabled,
color_mask.alpha_enabled);
}
// Point
if (point.size != cur_state.point.size) {
glPointSize(point.size);
}
ApplyScissorTest();
ApplyStencilTest();
ApplySRgb();
ApplyCulling();
ApplyDepth();
ApplyPrimitiveRestart();
ApplyBlending();
ApplyLogicOp();
ApplyTextures();
ApplySamplers();
cur_state = *this;
}

View file

@ -173,7 +173,8 @@ public:
}
/// Apply this state as the current OpenGL state
void Apply() const;
/// Set the initial OpenGL state
static void ApplyDefaultState();
/// Resets any references to the given resource
OpenGLState& UnbindTexture(GLuint handle);
OpenGLState& ResetSampler(GLuint handle);
@ -188,6 +189,16 @@ private:
// Workaround for sRGB problems caused by
// QT not supporting srgb output
static bool s_rgb_used;
void ApplySRgb() const;
void ApplyCulling() const;
void ApplyDepth() const;
void ApplyPrimitiveRestart() const;
void ApplyStencilTest() const;
void ApplyScissorTest() const;
void ApplyBlending() const;
void ApplyLogicOp() const;
void ApplyTextures() const;
void ApplySamplers() const;
};
} // namespace OpenGL