Compare commits

...

16 Commits

Author SHA1 Message Date
lizzie 3826ed7191 extra fixups for cross compile 2026-07-03 17:36:58 +00:00
lizzie 9b2dd39171 force rebase 2026-07-03 17:36:58 +00:00
lizzie 2dbce2afc4 fix pkgconfig stuff 2026-07-03 17:36:58 +00:00
lizzie b8dba9507f fix debian?? 2026-07-03 17:36:58 +00:00
lizzie 7d05ab127d DISABLE FUCKING ICONV 2026-07-03 17:36:58 +00:00
lizzie 4b8f15e7a2 extra x11 notes 2026-07-03 17:36:58 +00:00
lizzie 3ba034e982 fix x11 not enabling vdpau 2026-07-03 17:36:58 +00:00
lizzie bab6eca78e fix defined threads 2026-07-03 17:36:46 +00:00
lizzie 161e4d9992 fx 2026-07-03 17:36:46 +00:00
lizzie 43969dfe1d fine apple 2026-07-03 17:36:46 +00:00
lizzie 6903f7a7fd [externals] update ffmpeg to May 16th release, fix externals build
Signed-off-by: lizzie <lizzie@eden-emu.dev>
2026-07-03 17:36:46 +00:00
lizzie 4c65780f11 Revert "[common/dynamic_library] fix AUR build error (#4156)" (#4157)
This reverts commit a9c4c8aefd.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4157
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-07-02 22:07:37 +02:00
lizzie a9c4c8aefd [common/dynamic_library] fix AUR build error (#4156)
not gonna question why there was an ifdef there

Signed-off-by: lizzie <lizzie@eden-emu.dev>

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4156
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-07-02 19:49:19 +02:00
lizzie a769505a45 [vk, ogl] Remove dedicated precomputed swizzle table; use shorter inline table in block_linear swizzle shaders (#4146)
old table = 64 * 8 * sizeof(u32) = 512 * 4 = 4096 bytes
new table = 8 * sizeof(u32) = 8 * 4 = 32 bytes

the expression
```glsl
    return ((pos & 0x0180) >> 1)
         | ((pos & 0x0040) >> 2)
         | ((pos & 0x0020) << 3)
         | ((pos & 0x0010) << 1)
         | ((pos & 0x000f) << 0);
```
is equivalent for generating the table but idk if we'd want that

Signed-off-by: lizzie <lizzie@eden-emu.dev>

Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4146
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-07-02 19:24:07 +02:00
xbzk 54c3d10b86 [vk, renderdoc] (VUID-02997) avoid vk_image_view as VK_NULL_HANDLE when feature nullDescriptor is unavailable (#4056)
This minor change in src\video_core\renderer_vulkan\pipeline_helper.h allows renderdoc capture on mhr sunbreak.
Maybe it sanitizes some crashes on old vulkan GPUs.

Device log (nvidia kepler gpu vulkan 1.1.117):
[  17.605262] Render.Vulkan <Info> video_core\vulkan_common\vulkan_device.cpp:1041:GetSuitability: Device doesn't support feature nullDescriptor

VUID-VkWriteDescriptorSet-descriptorType-02997
For sampled/combined/storage image descriptors, Khronos says imageView must be valid or VK_NULL_HANDLE
(https://docs.vulkan.org/refpages/latest/refpages/source/VkPhysicalDeviceRobustness2FeaturesKHR.html)

but then it says: if nullDescriptor is not enabled, imageView must not be VK_NULL_HANDLE.
(https://docs.vulkan.org/spec/latest/chapters/descriptorsets.html)
nullDescriptor is exactly the feature that allows descriptors to be written with null handles.

The fix relies on conditionally replacing the null sampled view with Eden’s dummy null-image-view object (ImageView constructor checks HasNullDescriptor which checks features.robustness2.nullDescriptor. cheap checks, and only for null handles, so no notable cost).

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4056
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
2026-07-02 03:23:08 +02:00
xbzk bdf60d79a0 [video_core] gate Mesa EXIT terminator detection behind non-proprietary driver (#4151)
FIXES A REGRESSION FROM 4012.

TryFindSize's Mesa fallback matched the unconditional @PT EXIT (0xE30000000007000F) that nv50_ir emits at program end.
That exact word is also emitted by NVN mid-program for unconditional early-outs, so on retail titles (e.g. Super Mario 3D World) the scan truncated at the first early EXIT and dropped the rest of the shader, making textures vanish.

Mesa's terminal EXIT and NVN's mid-program EXIT are bit-identical, so no mask can separate them. Gate the heuristic on !is_proprietary_driver:

NVN binaries (bindless cbuf slot 2) keep self-branch-only sizing, while nouveau/Mesa homebrew (single terminal EXIT, no self-branch trailer) still uses the EXIT terminator. nv50_ir emits exactly one EXIT (early returns lower to BRA-to-exit, discard to DISCARD), so a single terminal match is correct for that path.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4151
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-07-01 13:06:00 +02:00
12 changed files with 187 additions and 217 deletions
+75 -33
View File
@@ -58,16 +58,24 @@ endif()
if (PLATFORM_PS4 OR PLATFORM_MANAGARM)
# Doesn't support VA-API, don't go thru the embarrassment of trying to enable it
list(APPEND FFmpeg_HWACCEL_FLAGS --disable-vaapi)
elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
list(APPEND FFmpeg_HWACCEL_FLAGS
--disable-vaapi
--disnable-cuvid
--disnable-ffnvcodec
--disnable-nvdec
--disable-vulkan
--disable-libdrm)
elseif (UNIX)
find_package(PkgConfig REQUIRED)
pkg_check_modules(LIBVA libva)
pkg_check_modules(CUDA cuda)
pkg_check_modules(FFNVCODEC ffnvcodec)
pkg_check_modules(VDPAU vdpau)
# X11 is "optional", some distros may opt for a super-fully wayland install...
find_package(X11)
if(X11_FOUND)
# Include X11 if possible, some APIs such as VDPAU heavily depend
# upon it's existence!
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${X11_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${X11_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${X11_LDFLAGS})
if (NOT APPLE)
# In Solaris needs explicit linking for ffmpeg which links to /lib/amd64/libX11.so
if(PLATFORM_SUN)
@@ -76,32 +84,37 @@ elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
"${CMAKE_SYSROOT}/usr/lib/xorg/amd64/libdrm.so")
else()
pkg_check_modules(LIBDRM libdrm REQUIRED)
list(APPEND FFmpeg_HWACCEL_LIBRARIES
${LIBDRM_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS
${LIBDRM_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBDRM_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBDRM_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBDRM_LDFLAGS})
endif()
list(APPEND FFmpeg_HWACCEL_FLAGS
--enable-libdrm)
list(APPEND FFmpeg_HWACCEL_FLAGS --enable-libdrm)
endif()
pkg_check_modules(LIBVA libva)
if(LIBVA_FOUND)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBVA_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBVA_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBVA_LDFLAGS})
# Logically, they can only exist if libva itself exists
pkg_check_modules(LIBVA-DRM libva-drm REQUIRED)
pkg_check_modules(LIBVA-X11 libva-x11 REQUIRED)
list(APPEND FFmpeg_HWACCEL_LIBRARIES
${X11_LIBRARIES}
${LIBVA-DRM_LIBRARIES}
${LIBVA-X11_LIBRARIES}
${LIBVA_LIBRARIES})
if (LIBVA-DRM_FOUND)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBVA-DRM_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBVA-DRM_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBVA-DRM_LDFLAGS})
endif()
if (LIBVA-X11_FOUND)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBVA-X11_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBVA-X11_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBVA-X11_LDFLAGS})
endif()
list(APPEND FFmpeg_HWACCEL_FLAGS
--enable-hwaccel=h264_vaapi
--enable-hwaccel=vp8_vaapi
--enable-hwaccel=vp9_vaapi)
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS
${X11_INCLUDE_DIRS}
${LIBVA-DRM_INCLUDE_DIRS}
${LIBVA-X11_INCLUDE_DIRS}
${LIBVA_INCLUDE_DIRS}
)
message(STATUS "ffmpeg: va-api libraries version ${LIBVA_VERSION} found")
else()
list(APPEND FFmpeg_HWACCEL_FLAGS --disable-vaapi)
@@ -112,6 +125,12 @@ elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
message(WARNING "ffmpeg: X11 libraries not found, disabling VA-API...")
endif()
pkg_check_modules(CUDA cuda)
pkg_check_modules(FFNVCODEC ffnvcodec)
pkg_check_modules(VDPAU vdpau)
pkg_check_modules(VULKAN vulkan)
find_package(spirv-headers)
if (FFNVCODEC_FOUND)
list(APPEND FFmpeg_HWACCEL_FLAGS
--enable-cuvid
@@ -119,8 +138,7 @@ elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
--enable-nvdec
--enable-hwaccel=h264_nvdec
--enable-hwaccel=vp8_nvdec
--enable-hwaccel=vp9_nvdec
)
--enable-hwaccel=vp9_nvdec)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${FFNVCODEC_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${FFNVCODEC_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${FFNVCODEC_LDFLAGS})
@@ -141,8 +159,7 @@ elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
list(APPEND FFmpeg_HWACCEL_FLAGS
--enable-vdpau
--enable-hwaccel=h264_vdpau
--enable-hwaccel=vp9_vdpau
)
--enable-hwaccel=vp9_vdpau)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${VDPAU_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${VDPAU_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${VDPAU_LDFLAGS})
@@ -151,6 +168,24 @@ elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
list(APPEND FFmpeg_HWACCEL_FLAGS --disable-vdpau)
message(WARNING "ffmpeg: libvdpau-dev not found, disabling Video Decode and Presentation API for Unix (VDPAU)...")
endif()
if (VULKAN_FOUND AND NOT APPLE)
list(APPEND FFmpeg_HWACCEL_FLAGS
--enable-vulkan
--enable-hwaccel=h264_vulkan
--enable-hwaccel=vp9_vulkan)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${VULKAN_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${VULKAN_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${VULKAN_LDFLAGS})
message(STATUS "ffmpeg: vulkan libraries version ${VULKAN_VERSION} found")
endif()
if (SPIRV_HEADERS_FOUND)
list(APPEND FFmpeg_HWACCEL_LIBRARIES ${SPIRV_HEADERS_LIBRARIES})
list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${SPIRV_HEADERS_INCLUDE_DIRS})
list(APPEND FFmpeg_HWACCEL_LDFLAGS ${SPIRV_HEADERS_LDFLAGS})
message(STATUS "ffmpeg: spirv-headers libraries version ${SPIRV_HEADERS_VERSION} found")
endif()
endif()
if (PLATFORM_PS4)
@@ -169,7 +204,7 @@ if (PLATFORM_PS4)
)
elseif (PLATFORM_MANAGARM)
# Required for proper stuff
list(APPEND FFmpeg_CROSS_COMPILE_FLAGS
list(APPEND FFmpeg_CROSS_COMPILE_FLAGS
--disable-pthreads
--extra-libs="${FFmpeg_CROSS_COMPILE_LIBS}"
)
@@ -247,8 +282,14 @@ else()
# `configure` parameters builds only exactly what yuzu needs from FFmpeg
# `--disable-vdpau` is needed to avoid linking issues
set(FFmpeg_CC ${CMAKE_C_COMPILER_LAUNCHER} ${CMAKE_C_COMPILER})
set(FFmpeg_CXX ${CMAKE_CXX_COMPILER_LAUNCHER} ${CMAKE_CXX_COMPILER})
set(FFmpeg_CC ${CMAKE_C_COMPILER_LAUNCHER} ${CMAKE_C_COMPILER})
set(FFmpeg_CXX ${CMAKE_CXX_COMPILER_LAUNCHER} ${CMAKE_CXX_COMPILER})
if (FFmpeg_IS_CROSS_COMPILING)
list(APPEND FFmpeg_CROSS_COMPILE_FLAGS --ld=${CMAKE_LINKER})
endif ()
# TODO: This is a bit fragile, it depends on FFmpeg_HWACCEL_LDFLAGS being a list
# whereas CMAKE_C_FLAGS (and likewise) is meant to NOT be a list, but rather
# an elongated piece of string
add_custom_command(
OUTPUT
${FFmpeg_MAKEFILE}
@@ -262,6 +303,7 @@ else()
--disable-ffprobe
--disable-network
--disable-swresample
--disable-iconv
--enable-decoder=h264
--enable-decoder=vp8
--enable-decoder=vp9
@@ -269,10 +311,9 @@ else()
--enable-pic
--cc=${FFmpeg_CC}
--cxx=${FFmpeg_CXX}
--ld=${CMAKE_LINKER}
--extra-cflags=${CMAKE_C_FLAGS}
--extra-cxxflags=${CMAKE_CXX_FLAGS}
--extra-ldflags=${CMAKE_C_LINK_FLAGS}
--extra-ldflags="${FFmpeg_HWACCEL_LDFLAGS}"
${FFmpeg_HWACCEL_FLAGS}
${FFmpeg_CROSS_COMPILE_FLAGS}
WORKING_DIRECTORY
@@ -280,6 +321,7 @@ else()
)
unset(FFmpeg_CC)
unset(FFmpeg_CXX)
unset(FFmpeg_LD)
unset(FFmpeg_HWACCEL_FLAGS)
unset(FFmpeg_CROSS_COMPILE_FLAGS)
@@ -294,7 +336,7 @@ else()
set(FFmpeg_BUILD_LIBRARIES ${FFmpeg_LIBRARIES})
# BSD make or Solaris make don't support ffmpeg make-j8
if (PLATFORM_LINUX OR ANDROID OR APPLE OR WIN32 OR PLATFORM_FREEBSD)
if (DEFINED SYSTEM_THREADS AND (PLATFORM_LINUX OR ANDROID OR APPLE OR WIN32 OR PLATFORM_FREEBSD))
set(FFmpeg_MAKE_ARGS -j${SYSTEM_THREADS})
else()
set(FFmpeg_MAKE_ARGS "")
@@ -11,9 +11,8 @@
#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants {
#define END_PUSH_CONSTANTS };
#define UNIFORM(n)
#define BINDING_SWIZZLE_BUFFER 0
#define BINDING_INPUT_BUFFER 1
#define BINDING_OUTPUT_IMAGE 2
#define BINDING_INPUT_BUFFER 0
#define BINDING_OUTPUT_IMAGE 1
#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv
@@ -26,8 +25,7 @@
#define BEGIN_PUSH_CONSTANTS
#define END_PUSH_CONSTANTS
#define UNIFORM(n) layout (location = n) uniform
#define BINDING_SWIZZLE_BUFFER 0
#define BINDING_INPUT_BUFFER 1
#define BINDING_INPUT_BUFFER 0
#define BINDING_OUTPUT_IMAGE 0
#endif
@@ -43,10 +41,6 @@ UNIFORM(6) uint block_height;
UNIFORM(7) uint block_height_mask;
END_PUSH_CONSTANTS
layout(binding = BINDING_SWIZZLE_BUFFER, std430) readonly buffer SwizzleTable {
uint swizzle_table[];
};
#if HAS_EXTENDED_TYPES
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU8 { uint8_t u8data[]; };
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU16 { uint16_t u16data[]; };
@@ -71,9 +65,19 @@ const uint GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHI
const uvec2 SWIZZLE_MASK = uvec2(GOB_SIZE_X - 1, GOB_SIZE_Y - 1);
uint SwizzleTable(uint pos) {
const uint t[8] = uint[](
0x12100200, 0x13110301, 0x16140604, 0x17150705,
0x1a180a08, 0x1b190b09, 0x1e1c0e0c, 0x1f1d0f0d
);
const uint i = pos >> 4;
const uint h = (t[i / 4] >> ((i % 4) * 8)) & 0xff;
return (h << 4) | (pos & 0xf);
}
uint SwizzleOffset(uvec2 pos) {
pos = pos & SWIZZLE_MASK;
return swizzle_table[pos.y * 64 + pos.x];
return SwizzleTable(pos.y * 64 + pos.x);
}
uvec4 ReadTexel(uint offset) {
@@ -11,9 +11,8 @@
#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants {
#define END_PUSH_CONSTANTS };
#define UNIFORM(n)
#define BINDING_SWIZZLE_BUFFER 0
#define BINDING_INPUT_BUFFER 1
#define BINDING_OUTPUT_IMAGE 2
#define BINDING_INPUT_BUFFER 0
#define BINDING_OUTPUT_IMAGE 1
#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv
@@ -26,8 +25,7 @@
#define BEGIN_PUSH_CONSTANTS
#define END_PUSH_CONSTANTS
#define UNIFORM(n) layout (location = n) uniform
#define BINDING_SWIZZLE_BUFFER 0
#define BINDING_INPUT_BUFFER 1
#define BINDING_INPUT_BUFFER 0
#define BINDING_OUTPUT_IMAGE 0
#endif
@@ -45,10 +43,6 @@ UNIFORM(8) uint block_depth;
UNIFORM(9) uint block_depth_mask;
END_PUSH_CONSTANTS
layout(binding = BINDING_SWIZZLE_BUFFER, std430) readonly buffer SwizzleTable {
uint swizzle_table[];
};
#if HAS_EXTENDED_TYPES
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU8 { uint8_t u8data[]; };
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU16 { uint16_t u16data[]; };
@@ -73,9 +67,19 @@ const uint GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHI
const uvec2 SWIZZLE_MASK = uvec2(GOB_SIZE_X - 1, GOB_SIZE_Y - 1);
uint SwizzleTable(uint pos) {
const uint t[8] = uint[](
0x12100200, 0x13110301, 0x16140604, 0x17150705,
0x1a180a08, 0x1b190b09, 0x1e1c0e0c, 0x1f1d0f0d
);
const uint i = pos >> 4;
const uint h = (t[i / 4] >> ((i % 4) * 8)) & 0xff;
return (h << 4) | (pos & 0xf);
}
uint SwizzleOffset(uvec2 pos) {
pos = pos & SWIZZLE_MASK;
return swizzle_table[pos.y * 64 + pos.x];
return SwizzleTable(pos.y * 64 + pos.x);
}
uvec4 ReadTexel(uint offset) {
@@ -10,9 +10,8 @@
#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants {
#define END_PUSH_CONSTANTS };
#define UNIFORM(n)
#define BINDING_SWIZZLE_BUFFER 0
#define BINDING_INPUT_BUFFER 1
#define BINDING_OUTPUT_BUFFER 2
#define BINDING_INPUT_BUFFER 0
#define BINDING_OUTPUT_BUFFER 1
#else
#extension GL_NV_gpu_shader5 : enable
#ifdef GL_NV_gpu_shader5
@@ -23,7 +22,6 @@
#define BEGIN_PUSH_CONSTANTS
#define END_PUSH_CONSTANTS
#define UNIFORM(n) layout(location = n) uniform
#define BINDING_SWIZZLE_BUFFER 0
#define BINDING_INPUT_BUFFER 1
#define BINDING_OUTPUT_BUFFER 0
#endif
@@ -66,13 +64,9 @@ END_PUSH_CONSTANTS
#endif
// --- Buffers ---
layout(binding = BINDING_SWIZZLE_BUFFER, std430) readonly buffer SwizzleTable {
uint swizzle_table[];
};
#if HAS_EXTENDED_TYPES
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU8 { uint8_t u8data[]; };
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU16 { uint16_t u16data[]; };
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU8 { uint8_t u8data[]; };
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU16 { uint16_t u16data[]; };
#endif
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU32 { uint u32data[]; };
layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU64 { uvec2 u64data[]; };
@@ -96,10 +90,20 @@ const uint GOB_SIZE_Z_SHIFT = 0;
const uint GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT;
const uvec2 SWIZZLE_MASK = uvec2(GOB_SIZE_X - 1u, GOB_SIZE_Y - 1u);
uint SwizzleTable(uint pos) {
const uint t[8] = uint[](
0x12100200, 0x13110301, 0x16140604, 0x17150705,
0x1a180a08, 0x1b190b09, 0x1e1c0e0c, 0x1f1d0f0d
);
const uint i = pos >> 4;
const uint h = (t[i / 4] >> ((i % 4) * 8)) & 0xff;
return (h << 4) | (pos & 0xf);
}
// --- Helpers ---
uint SwizzleOffset(uvec2 pos) {
pos &= SWIZZLE_MASK;
return swizzle_table[pos.y * 64u + pos.x];
return SwizzleTable(pos.y * 64u + pos.x);
}
uvec4 ReadTexel(uint offset) {
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -56,10 +59,8 @@ UtilShaders::UtilShaders(ProgramManager& program_manager_)
copy_bc4_program(MakeProgram(OPENGL_COPY_BC4_COMP)),
convert_s8d24_program(MakeProgram(OPENGL_CONVERT_S8D24_COMP)),
convert_ms_to_nonms_program(MakeProgram(CONVERT_MSAA_TO_NON_MSAA_COMP)),
convert_nonms_to_ms_program(MakeProgram(CONVERT_NON_MSAA_TO_MSAA_COMP)) {
const auto swizzle_table = Tegra::Texture::MakeSwizzleTable();
swizzle_table_buffer.Create();
glNamedBufferStorage(swizzle_table_buffer.handle, sizeof(swizzle_table), &swizzle_table, 0);
convert_nonms_to_ms_program(MakeProgram(CONVERT_NON_MSAA_TO_MSAA_COMP))
{
}
UtilShaders::~UtilShaders() = default;
@@ -116,13 +117,11 @@ void UtilShaders::ASTCDecode(Image& image, const StagingBufferMap& map,
void UtilShaders::BlockLinearUpload2D(Image& image, const StagingBufferMap& map,
std::span<const SwizzleParameters> swizzles) {
static constexpr Extent3D WORKGROUP_SIZE{32, 32, 1};
static constexpr GLuint BINDING_SWIZZLE_BUFFER = 0;
static constexpr GLuint BINDING_INPUT_BUFFER = 1;
static constexpr GLuint BINDING_INPUT_BUFFER = 0;
static constexpr GLuint BINDING_OUTPUT_IMAGE = 0;
program_manager.BindComputeProgram(block_linear_unswizzle_2d_program.handle);
glFlushMappedNamedBufferRange(map.buffer, map.offset, image.guest_size_bytes);
glBindBufferBase(GL_SHADER_STORAGE_BUFFER, BINDING_SWIZZLE_BUFFER, swizzle_table_buffer.handle);
const GLenum store_format = StoreFormat(BytesPerBlock(image.info.format));
for (const SwizzleParameters& swizzle : swizzles) {
@@ -153,14 +152,11 @@ void UtilShaders::BlockLinearUpload2D(Image& image, const StagingBufferMap& map,
void UtilShaders::BlockLinearUpload3D(Image& image, const StagingBufferMap& map,
std::span<const SwizzleParameters> swizzles) {
static constexpr Extent3D WORKGROUP_SIZE{16, 8, 8};
static constexpr GLuint BINDING_SWIZZLE_BUFFER = 0;
static constexpr GLuint BINDING_INPUT_BUFFER = 1;
static constexpr GLuint BINDING_INPUT_BUFFER = 0;
static constexpr GLuint BINDING_OUTPUT_IMAGE = 0;
glFlushMappedNamedBufferRange(map.buffer, map.offset, image.guest_size_bytes);
program_manager.BindComputeProgram(block_linear_unswizzle_3d_program.handle);
glBindBufferBase(GL_SHADER_STORAGE_BUFFER, BINDING_SWIZZLE_BUFFER, swizzle_table_buffer.handle);
const GLenum store_format = StoreFormat(BytesPerBlock(image.info.format));
for (const SwizzleParameters& swizzle : swizzles) {
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -45,9 +48,6 @@ public:
private:
ProgramManager& program_manager;
OGLBuffer swizzle_table_buffer;
OGLProgram astc_decoder_program;
OGLProgram block_linear_unswizzle_2d_program;
OGLProgram block_linear_unswizzle_3d_program;
@@ -203,7 +203,11 @@ inline void PushImageDescriptors(TextureCache& texture_cache,
const VideoCommon::ImageViewId image_view_id{(views++)->id};
const VideoCommon::SamplerId sampler_id{*(samplers++)};
ImageView& image_view{texture_cache.GetImageView(image_view_id)};
const VkImageView vk_image_view{image_view.Handle(desc.type)};
VkImageView vk_image_view{image_view.Handle(desc.type)};
if (vk_image_view == VK_NULL_HANDLE) {
const VkImageView null_image_view{texture_cache.GetImageView(VideoCommon::NULL_IMAGE_VIEW_ID).Handle(desc.type)};
if (null_image_view != VK_NULL_HANDLE) vk_image_view = null_image_view;
}
const Sampler& sampler{texture_cache.GetSampler(sampler_id)};
const bool use_fallback_sampler{sampler.HasAddedAnisotropy() &&
!image_view.SupportsAnisotropy()};
@@ -653,71 +653,8 @@ void ASTCDecoderPass::Assemble(Image& image, const StagingBufferRef& map,
scheduler.Finish();
}
constexpr u32 BL3D_BINDING_SWIZZLE_TABLE = 0;
constexpr u32 BL3D_BINDING_INPUT_BUFFER = 1;
constexpr u32 BL3D_BINDING_OUTPUT_BUFFER = 2;
constexpr std::array<VkDescriptorSetLayoutBinding, 3> BL3D_DESCRIPTOR_SET_BINDINGS{{
{
.binding = BL3D_BINDING_SWIZZLE_TABLE,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, // swizzle_table[]
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
{
.binding = BL3D_BINDING_INPUT_BUFFER,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, // block-linear input
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
{
.binding = BL3D_BINDING_OUTPUT_BUFFER,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
}};
constexpr DescriptorBankInfo BL3D_BANK_INFO{
.uniform_buffers = 0,
.storage_buffers = 3,
.texture_buffers = 0,
.image_buffers = 0,
.textures = 0,
.images = 0,
.score = 3,
};
constexpr std::array<VkDescriptorUpdateTemplateEntry, 3>
BL3D_DESCRIPTOR_UPDATE_TEMPLATE_ENTRY{{
{
.dstBinding = BL3D_BINDING_SWIZZLE_TABLE,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = BL3D_BINDING_SWIZZLE_TABLE * sizeof(DescriptorUpdateEntry),
.stride = sizeof(DescriptorUpdateEntry),
},
{
.dstBinding = BL3D_BINDING_INPUT_BUFFER,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = BL3D_BINDING_INPUT_BUFFER * sizeof(DescriptorUpdateEntry),
.stride = sizeof(DescriptorUpdateEntry),
},
{
.dstBinding = BL3D_BINDING_OUTPUT_BUFFER,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = BL3D_BINDING_OUTPUT_BUFFER * sizeof(DescriptorUpdateEntry),
.stride = sizeof(DescriptorUpdateEntry),
}
}};
constexpr u32 BL3D_BINDING_INPUT_BUFFER = 0;
constexpr u32 BL3D_BINDING_OUTPUT_BUFFER = 1;
struct alignas(16) BlockLinearUnswizzle3DPushConstants {
u32 blocks_dim[3]; // Offset 0
@@ -745,11 +682,50 @@ BlockLinearUnswizzle3DPass::BlockLinearUnswizzle3DPass(
DescriptorPool& descriptor_pool_,
StagingBufferPool& staging_buffer_pool_,
ComputePassDescriptorQueue& compute_pass_descriptor_queue_)
: ComputePass(
device_, scheduler_, descriptor_pool_,
BL3D_DESCRIPTOR_SET_BINDINGS,
BL3D_DESCRIPTOR_UPDATE_TEMPLATE_ENTRY,
BL3D_BANK_INFO,
: ComputePass(device_, scheduler_, descriptor_pool_,
std::array<VkDescriptorSetLayoutBinding, 2>{{
{
.binding = BL3D_BINDING_INPUT_BUFFER,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, // block-linear input
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
{
.binding = BL3D_BINDING_OUTPUT_BUFFER,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
}},
std::array<VkDescriptorUpdateTemplateEntry, 2>{{
{
.dstBinding = BL3D_BINDING_INPUT_BUFFER,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = BL3D_BINDING_INPUT_BUFFER * sizeof(DescriptorUpdateEntry),
.stride = sizeof(DescriptorUpdateEntry),
},
{
.dstBinding = BL3D_BINDING_OUTPUT_BUFFER,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = BL3D_BINDING_OUTPUT_BUFFER * sizeof(DescriptorUpdateEntry),
.stride = sizeof(DescriptorUpdateEntry),
}
}},
DescriptorBankInfo{
.uniform_buffers = 0,
.storage_buffers = 2,
.texture_buffers = 0,
.image_buffers = 0,
.textures = 0,
.images = 0,
.score = 2,
},
COMPUTE_PUSH_CONSTANT_RANGE<sizeof(BlockLinearUnswizzle3DPushConstants)>,
BLOCK_LINEAR_UNSWIZZLE_3D_BCN_COMP_SPV),
scheduler{scheduler_},
@@ -822,8 +798,6 @@ void BlockLinearUnswizzle3DPass::UnswizzleChunk(
pc.blocks_dim[2] = z_count; // Only process the count
compute_pass_descriptor_queue.Acquire(scheduler, 3);
compute_pass_descriptor_queue.AddBuffer(*image.runtime->swizzle_table_buffer, 0,
image.runtime->swizzle_table_size);
compute_pass_descriptor_queue.AddBuffer(swizzled.buffer,
sw.buffer_offset + swizzled.offset,
image.guest_size_bytes - sw.buffer_offset);
@@ -909,40 +909,6 @@ TextureCacheRuntime::TextureCacheRuntime(const Device& device_, Scheduler& sched
bl3d_unswizzle_pass.emplace(device, scheduler, descriptor_pool,
staging_buffer_pool, compute_pass_descriptor_queue);
}
// --- Create swizzle table buffer ---
{
auto table = Tegra::Texture::MakeSwizzleTable();
swizzle_table_size = static_cast<VkDeviceSize>(table.size() * sizeof(table[0]));
auto staging = staging_buffer_pool.Request(swizzle_table_size, MemoryUsage::Upload);
std::memcpy(staging.mapped_span.data(), table.data(), static_cast<size_t>(swizzle_table_size));
VkBufferCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.size = swizzle_table_size,
.usage = VK_BUFFER_USAGE_STORAGE_BUFFER_BIT |
VK_BUFFER_USAGE_TRANSFER_DST_BIT |
VK_BUFFER_USAGE_TRANSFER_SRC_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
};
swizzle_table_buffer = memory_allocator.CreateBuffer(ci, MemoryUsage::DeviceLocal);
scheduler.RequestOutsideRenderPassOperationContext();
scheduler.Record([staging_buf = staging.buffer,
dst_buf = *swizzle_table_buffer,
size = swizzle_table_size,
src_off = staging.offset](vk::CommandBuffer cmdbuf) {
const VkBufferCopy region{
.srcOffset = src_off,
.dstOffset = 0,
.size = size,
};
cmdbuf.CopyBuffer(staging_buf, dst_buf, region);
});
}
}
void TextureCacheRuntime::Finish() {
@@ -2484,18 +2450,13 @@ void TextureCacheRuntime::AccelerateImageUpload(
if (!Settings::values.gpu_unswizzle_enabled.GetValue() || !bl3d_unswizzle_pass) {
if (IsPixelFormatBCn(image.info.format) && image.info.type == ImageType::e3D) {
ASSERT_MSG(false, "GPU unswizzle is disabled for BCn 3D texture");
ASSERT(false && "GPU unswizzle is disabled for BCn 3D texture");
}
ASSERT(false);
return;
}
if (bl3d_unswizzle_pass &&
IsPixelFormatBCn(image.info.format) &&
image.info.type == ImageType::e3D &&
image.info.resources.levels == 1 &&
image.info.resources.layers == 1) {
if (bl3d_unswizzle_pass && IsPixelFormatBCn(image.info.format) && image.info.type == ImageType::e3D && image.info.resources.levels == 1 && image.info.resources.layers == 1) {
return bl3d_unswizzle_pass->Unswizzle(image, map, swizzles, z_start, z_count);
}
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2019 yuzu Emulator Project
@@ -130,9 +130,6 @@ public:
std::optional<ASTCDecoderPass> astc_decoder_pass;
std::optional<BlockLinearUnswizzle3DPass> bl3d_unswizzle_pass;
vk::Buffer swizzle_table_buffer;
VkDeviceSize swizzle_table_size = 0;
std::optional<MSAACopyPass> msaa_copy_pass;
const Settings::ResolutionScalingInfo& resolution;
std::array<std::vector<VkFormat>, VideoCore::Surface::MaxPixelFormat> view_formats;
+2 -3
View File
@@ -255,8 +255,7 @@ std::optional<u64> GenericEnvironment::TryFindSize() {
static constexpr u64 SELF_BRANCH_A = 0xE2400FFFFF87000FULL;
static constexpr u64 SELF_BRANCH_B = 0xE2400FFFFF07000FULL;
static constexpr u64 MESA_EXIT_MASK = 0xFFF00000000F001FULL;
static constexpr u64 MESA_EXIT_VALUE = (0xE30ULL << 52) | (0x7ULL << 16) | 0xFULL;
static constexpr u64 EXIT_VALUE = 0xE30000000007000FULL;
code.resize(MAXIMUM_SIZE / INST_SIZE);
@@ -271,7 +270,7 @@ std::optional<u64> GenericEnvironment::TryFindSize() {
if (inst == SELF_BRANCH_A || inst == SELF_BRANCH_B) {
return offset + index;
}
if ((inst & MESA_EXIT_MASK) == MESA_EXIT_VALUE) {
if (!is_proprietary_driver && inst == EXIT_VALUE) {
return offset + index + INST_SIZE;
}
}
+3 -18
View File
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2018 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -23,24 +26,6 @@ constexpr u32 GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_
constexpr u32 SWIZZLE_X_BITS = 0b100101111;
constexpr u32 SWIZZLE_Y_BITS = 0b011010000;
using SwizzleTable = std::array<std::array<u32, GOB_SIZE_X>, GOB_SIZE_Y>;
/**
* This table represents the internal swizzle of a gob, in format 16 bytes x 2 sector packing.
* Calculates the offset of an (x, y) position within a swizzled texture.
* Taken from the Tegra X1 Technical Reference Manual. pages 1187-1188
*/
constexpr SwizzleTable MakeSwizzleTable() {
SwizzleTable table{};
for (u32 y = 0; y < table.size(); ++y) {
for (u32 x = 0; x < table[0].size(); ++x) {
table[y][x] = ((x % 64) / 32) * 256 + ((y % 8) / 2) * 64 + ((x % 32) / 16) * 32 +
(y % 2) * 16 + (x % 16);
}
}
return table;
}
/// Unswizzles a block linear texture into linear memory.
void UnswizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel,
u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth,