Oopenharmony_ciCVE-2026-7923 patch修复
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[graphite][vulkan] Support input attachment reads with MSAA Input attachment reads (subpassLoad() in SkSL) in SPIR-V take a sample (in OpImageRead) when MSAA. This means the SPIR-V for the same pipeline is slightly different from single-sampled when used with a multisampled render pass. To avoid recompiling the shader when MSAA, this change applies a small SPIR-V transformation after the fact to make the SPIR-V compatible with MSAA. The transformer itself is designed after ANGLE's equivalent, allowing easy additions in the future, if more transformations are found to be useful. The change additionally includes a concept of "reserved ids" in the SPIR-V generator such that the transformer can easily find the instructions it needs / rely on certain basic types to be always available. An example shader from the PlusMergesAA dm test shows the following diff of SPIR-V before and after the transformation: ; SPIR-V ; Version: 1.0 ; Generator: Google Skia SkSL; 0 -; Bound: 37 +; Bound: 39 ; Schema: 0 + OpCapability SampleRateShading OpCapability InputAttachment OpCapability Shader %5 = OpExtInstImport "GLSL.std.450" OpMemoryModel Logical GLSL450 - OpEntryPoint Fragment %6 "main" %15 + OpEntryPoint Fragment %6 "main" %15 %37 OpExecutionMode %6 OriginUpperLeft ; Debug Information OpName %10 "FSUniforms" ; id %10 OpMemberName %10 0 "color_0" OpName %13 "IntrinsicUniforms" ; id %13 OpMemberName %13 0 "viewport" OpMemberName %13 1 "dstReadBounds" OpName %15 "sk_FragColor" ; id %15 OpName %4 "DstTextureInput" ; id %4 OpName %6 "main" ; id %6 OpName %21 "outColor_0" ; id %21 OpName %27 "dstColor" ; id %27 OpName %32 "outColor_1" ; id %32 ; Annotations OpMemberDecorate %10 0 Offset 0 + OpDecorate %37 RelaxedPrecision + OpDecorate %37 Flat + OpDecorate %37 BuiltIn SampleId OpDecorate %10 Block OpDecorate %7 Binding 1 OpDecorate %7 DescriptorSet 1 OpMemberDecorate %13 0 Offset 0 OpMemberDecorate %13 1 Offset 16 OpDecorate %13 Block OpDecorate %15 RelaxedPrecision OpDecorate %15 Location 0 OpDecorate %15 Index 0 OpDecorate %4 Binding 0 OpDecorate %4 DescriptorSet 0 OpDecorate %4 InputAttachmentIndex 0 OpDecorate %21 RelaxedPrecision OpDecorate %27 RelaxedPrecision OpDecorate %28 RelaxedPrecision OpDecorate %29 RelaxedPrecision OpDecorate %32 RelaxedPrecision OpDecorate %33 RelaxedPrecision OpDecorate %34 RelaxedPrecision ; Types, variables and constants %1 = OpTypeInt 32 1 %2 = OpTypePointer Input %1 + %37 = OpVariable %2 Input ; RelaxedPrecision, Flat, BuiltIn SampleId %8 = OpTypeFloat 32 %9 = OpTypeVector %8 4 %10 = OpTypeStruct %9 ; Block %11 = OpTypePointer Uniform %10 %7 = OpVariable %11 Uniform ; Binding 1, DescriptorSet 1 %13 = OpTypeStruct %9 %9 ; Block %14 = OpTypePointer PushConstant %13 %12 = OpVariable %14 PushConstant %16 = OpTypePointer Output %9 %15 = OpVariable %16 Output ; RelaxedPrecision, Location 0, Index 0 - %3 = OpTypeImage %8 SubpassData 0 0 0 2 Unknown + %3 = OpTypeImage %8 SubpassData 0 0 1 2 Unknown %17 = OpTypePointer UniformConstant %3 %4 = OpVariable %17 UniformConstant ; Binding 0, DescriptorSet 0, InputAttachmentIndex 0 %18 = OpTypeVoid %19 = OpTypeFunction %18 %22 = OpTypePointer Function %9 %23 = OpConstant %1 0 %25 = OpTypePointer Uniform %9 %30 = OpTypeVector %1 2 %31 = OpConstantComposite %30 %23 %23 %35 = OpConstant %8 1 %36 = OpConstantComposite %9 %35 %35 %35 %35 ; Function 6 %6 = OpFunction %18 None %19 %20 = OpLabel %21 = OpVariable %22 Function ; RelaxedPrecision %27 = OpVariable %22 Function ; RelaxedPrecision %32 = OpVariable %22 Function ; RelaxedPrecision %24 = OpAccessChain %25 %7 %23 %26 = OpLoad %9 %24 OpStore %21 %26 %29 = OpLoad %3 %4 ; RelaxedPrecision - %28 = OpImageRead %9 %29 %31 ; RelaxedPrecision + %38 = OpLoad %1 %37 + %28 = OpImageRead %9 %29 %31 Sample %38 ; RelaxedPrecision OpStore %27 %28 %34 = OpFAdd %9 %26 %28 ; RelaxedPrecision %33 = OpExtInst %9 %5 FMin %34 %36 ; RelaxedPrecision OpStore %32 %33 OpStore %15 %33 OpReturn OpFunctionEnd Using SPIR-V Cross, the diff in GLSL looks straightforward: 17c17 < layout(input_attachment_index = 0, set = 0, binding = 0) uniform subpassInput DstTextureInput; --- > layout(input_attachment_index = 0, set = 0, binding = 0) uniform subpassInputMS DstTextureInput; 26c26 < mediump vec4 _28 = subpassLoad(DstTextureInput); --- > mediump vec4 _28 = subpassLoad(DstTextureInput, gl_SampleID); Bug: b/390458117 Change-Id: Ia851495b73614298ab04a2b676f8b196bfeb776c Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1008420 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Nicolette Prevost <nicolettep@google.com> | 11 个月前 | |
Remove SK_USE_LEGACY_MIPMAP_LOD_BIAS guard Chromium was updated in https://crrev.com/c/6332723 Change-Id: Ieaae4bd9e0606e2e71665ba73d5c055891113763 Bug: b/40044183 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/962116 Auto-Submit: Kaylee Lubick <kjlubick@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> Commit-Queue: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
2026.03.30 arkweb_144代码蓝黄同步 Signed-off-by: jiujiaoxiaogula <sujiahao10@huawei.com> Change-Id: Id6b8f8e08796118f0ec7d7f82b8f2430ef3448e1 | 3 个月前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
Reland "Move most KnownRuntimeEffects into the sksl_rt_shader module." This reverts commit a4d59988bde01b17fed7d8f1be603c42c68763ef. Reason for revert: adding Chrome flag Original change's description: > Revert "Move most KnownRuntimeEffects into the sksl_rt_shader module." > > This reverts commit 4db44b0c04d56caeb9f46075b889037fe7e66265. > > Reason for revert: possibly breaking Chrome roll > > Original change's description: > > Move most KnownRuntimeEffects into the sksl_rt_shader module. > > > > The remaining KnownRuntimeEffects are assembled programmatically, > > or are so simple that they wouldn't benefit from existing in a > > module. > > > > Bug: b/328051548 > > Change-Id: I449e56143b6d622d413fc624b24ad486bba71352 > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/883098 > > Commit-Queue: John Stiles <johnstiles@google.com> > > Commit-Queue: Robert Phillips <robertphillips@google.com> > > Auto-Submit: John Stiles <johnstiles@google.com> > > Reviewed-by: Robert Phillips <robertphillips@google.com> > > Bug: b/328051548 > Change-Id: I1c57ab02cfa76ad58557121edc1fabcf00c71995 > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/883878 > Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> > Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> > Auto-Submit: John Stiles <johnstiles@google.com> Bug: b/328051548 Change-Id: Ide1eeb67271601aae7b75c968ca517fcc0cb0cdf Reviewed-on: https://skia-review.googlesource.com/c/skia/+/883937 Auto-Submit: John Stiles <johnstiles@google.com> Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> | 1 年前 | |
Add refactors/deprecations to support Ganesh modules Having a function with many possible implementations chosen at compile time (e.g. GrGLMakeNativeInterface) is not compatible with modular Bazel build rules, so I've deprecated it. Clients will need to directly instantiate what they need. Additionally, this removes some PNG encoding from GrAtlasManager (and other test-only functions) and putting those in a tools helper class, breaking a debug-only dependency. There are a few other small cleanups too. A follow-on CL will add the Bazel rules for the ganesh gl module. No-Try: true Change-Id: Ibd1309d04aa08c77d68625af6bc1d0d237d4aa22 Bug: b/293490566 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/788798 Reviewed-by: Greg Daniel <egdaniel@google.com> Commit-Queue: Kevin Lubick <kjlubick@google.com> | 2 年前 | |
CVE-2026-7923 patch修复 TicketNo: DTS2026052234994 Description: CVE-2026-7923 patch修复 Team:OTHERS Feature or Bugfix:Bugfix Binary Source: sync from gitcode PrivateCode(Yes/No):No ------ DO NOT MODIFY, AUTO-GENERATED! ------ gitcode-Issue: 131 Time: "2026-05-13T03:10:55.284530Z" PR-Num: 216 gitcode-PR: https://gitcode.com/openharmony-tpc/chromium_third_party_skia/merge_requests/216 UserId:842199 Co-Authored-By: Agent cherry picked from commit 243a3f1f8500d9afdbcaf4df85ecdecae32e257c cherry picked from commit 1370d28c2f017351def224d94b82aaf4582bb9d4 Change-Id: I1778641851542ee2d33b6624479938f1d8520aaa Reviewed-by: w30064525 Approved-by: f00619500 Merged-on: https://open.codehub.huawei.com/OpenSourceCenter_CR/openharmony-tpc/chromium_third_party_skia/-/change_requests/188 Merged-by: public pjenkins Signed-off-by: lxinjie <liuxinjie6@huawei.com> | 1 个月前 | |
[SkSL:RP] Prevent overflow when computing slot allocation size Bug: 355465305 Change-Id: Ife25289f7b3489701c67b7dc5d30e473019a1193 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/888376 Reviewed-by: Julia Lavrova <jlavrova@google.com> Commit-Queue: Brian Osman <brianosman@google.com> | 1 年前 | |
Prepare more call-sites for std::span Change-Id: Ieccf7ff710b32994fa16cc2fde3f428619925e54 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1099836 Reviewed-by: Florin Malita <fmalita@google.com> Commit-Queue: Mike Reed <mike@reedtribe.org> Reviewed-by: Eric Boren <borenet@google.com> | 7 个月前 | |
Add design doc links for go/sksl-rp. The SkRP doc is now world-visible and this will make it easier for newcomers to find. Change-Id: I728125c4c9eabfd4743b49e60529fc8149e939fa Reviewed-on: https://skia-review.googlesource.com/c/skia/+/884019 Commit-Queue: John Stiles <johnstiles@google.com> Auto-Submit: John Stiles <johnstiles@google.com> Reviewed-by: Kaylee Lubick <kjlubick@google.com> Commit-Queue: Kaylee Lubick <kjlubick@google.com> | 1 年前 | |
[graphite][vulkan] Support input attachment reads with MSAA Input attachment reads (subpassLoad() in SkSL) in SPIR-V take a sample (in OpImageRead) when MSAA. This means the SPIR-V for the same pipeline is slightly different from single-sampled when used with a multisampled render pass. To avoid recompiling the shader when MSAA, this change applies a small SPIR-V transformation after the fact to make the SPIR-V compatible with MSAA. The transformer itself is designed after ANGLE's equivalent, allowing easy additions in the future, if more transformations are found to be useful. The change additionally includes a concept of "reserved ids" in the SPIR-V generator such that the transformer can easily find the instructions it needs / rely on certain basic types to be always available. An example shader from the PlusMergesAA dm test shows the following diff of SPIR-V before and after the transformation: ; SPIR-V ; Version: 1.0 ; Generator: Google Skia SkSL; 0 -; Bound: 37 +; Bound: 39 ; Schema: 0 + OpCapability SampleRateShading OpCapability InputAttachment OpCapability Shader %5 = OpExtInstImport "GLSL.std.450" OpMemoryModel Logical GLSL450 - OpEntryPoint Fragment %6 "main" %15 + OpEntryPoint Fragment %6 "main" %15 %37 OpExecutionMode %6 OriginUpperLeft ; Debug Information OpName %10 "FSUniforms" ; id %10 OpMemberName %10 0 "color_0" OpName %13 "IntrinsicUniforms" ; id %13 OpMemberName %13 0 "viewport" OpMemberName %13 1 "dstReadBounds" OpName %15 "sk_FragColor" ; id %15 OpName %4 "DstTextureInput" ; id %4 OpName %6 "main" ; id %6 OpName %21 "outColor_0" ; id %21 OpName %27 "dstColor" ; id %27 OpName %32 "outColor_1" ; id %32 ; Annotations OpMemberDecorate %10 0 Offset 0 + OpDecorate %37 RelaxedPrecision + OpDecorate %37 Flat + OpDecorate %37 BuiltIn SampleId OpDecorate %10 Block OpDecorate %7 Binding 1 OpDecorate %7 DescriptorSet 1 OpMemberDecorate %13 0 Offset 0 OpMemberDecorate %13 1 Offset 16 OpDecorate %13 Block OpDecorate %15 RelaxedPrecision OpDecorate %15 Location 0 OpDecorate %15 Index 0 OpDecorate %4 Binding 0 OpDecorate %4 DescriptorSet 0 OpDecorate %4 InputAttachmentIndex 0 OpDecorate %21 RelaxedPrecision OpDecorate %27 RelaxedPrecision OpDecorate %28 RelaxedPrecision OpDecorate %29 RelaxedPrecision OpDecorate %32 RelaxedPrecision OpDecorate %33 RelaxedPrecision OpDecorate %34 RelaxedPrecision ; Types, variables and constants %1 = OpTypeInt 32 1 %2 = OpTypePointer Input %1 + %37 = OpVariable %2 Input ; RelaxedPrecision, Flat, BuiltIn SampleId %8 = OpTypeFloat 32 %9 = OpTypeVector %8 4 %10 = OpTypeStruct %9 ; Block %11 = OpTypePointer Uniform %10 %7 = OpVariable %11 Uniform ; Binding 1, DescriptorSet 1 %13 = OpTypeStruct %9 %9 ; Block %14 = OpTypePointer PushConstant %13 %12 = OpVariable %14 PushConstant %16 = OpTypePointer Output %9 %15 = OpVariable %16 Output ; RelaxedPrecision, Location 0, Index 0 - %3 = OpTypeImage %8 SubpassData 0 0 0 2 Unknown + %3 = OpTypeImage %8 SubpassData 0 0 1 2 Unknown %17 = OpTypePointer UniformConstant %3 %4 = OpVariable %17 UniformConstant ; Binding 0, DescriptorSet 0, InputAttachmentIndex 0 %18 = OpTypeVoid %19 = OpTypeFunction %18 %22 = OpTypePointer Function %9 %23 = OpConstant %1 0 %25 = OpTypePointer Uniform %9 %30 = OpTypeVector %1 2 %31 = OpConstantComposite %30 %23 %23 %35 = OpConstant %8 1 %36 = OpConstantComposite %9 %35 %35 %35 %35 ; Function 6 %6 = OpFunction %18 None %19 %20 = OpLabel %21 = OpVariable %22 Function ; RelaxedPrecision %27 = OpVariable %22 Function ; RelaxedPrecision %32 = OpVariable %22 Function ; RelaxedPrecision %24 = OpAccessChain %25 %7 %23 %26 = OpLoad %9 %24 OpStore %21 %26 %29 = OpLoad %3 %4 ; RelaxedPrecision - %28 = OpImageRead %9 %29 %31 ; RelaxedPrecision + %38 = OpLoad %1 %37 + %28 = OpImageRead %9 %29 %31 Sample %38 ; RelaxedPrecision OpStore %27 %28 %34 = OpFAdd %9 %26 %28 ; RelaxedPrecision %33 = OpExtInst %9 %5 FMin %34 %36 ; RelaxedPrecision OpStore %32 %33 OpStore %15 %33 OpReturn OpFunctionEnd Using SPIR-V Cross, the diff in GLSL looks straightforward: 17c17 < layout(input_attachment_index = 0, set = 0, binding = 0) uniform subpassInput DstTextureInput; --- > layout(input_attachment_index = 0, set = 0, binding = 0) uniform subpassInputMS DstTextureInput; 26c26 < mediump vec4 _28 = subpassLoad(DstTextureInput); --- > mediump vec4 _28 = subpassLoad(DstTextureInput, gl_SampleID); Bug: b/390458117 Change-Id: Ia851495b73614298ab04a2b676f8b196bfeb776c Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1008420 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Nicolette Prevost <nicolettep@google.com> | 11 个月前 | |
[vulkan] Generate SPIR-V directly in vector<uint32_t> SPIR-V is naturally defined in 32-bit words and contains lots of zeros, using a string as storage is awkward and less efficient (especially through OutputStream). In this change, SPIR-V is directly constructed in std::vector<uint32_t>. This change improves SPIR-V compilation time by about 6% on Pixel 8, comparing compile times of the first 4000 shaders built by dm tests. Bug: b/390458117 Change-Id: Ifc844bc4819641bf144cfe97e720b0d9985ea310 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006640 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
Do not emit chromium_experimental_immediate enable This is now a language feature and not an enable flag since https://dawn-review.googlesource.com/c/dawn/+/273614 The feature is automatically enabled when the UnsafeWebGPU toggle is enabled, which Skia enables. Bug: 366291600 Bug: 460481195 Change-Id: Ief8c97bf184098055d116616016f6fe20d8d636b Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1112616 Commit-Queue: Antonio Maiorano <amaiorano@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 7 个月前 | |
[vulkan] Use std::vector<uint32_t> for SPIR-V Instead of std::string. SPIR-V is defined in 32-bit words and includes lots of zeros. Use of std::string (due to other backends) is awkward. This change makes the "native shader" be a type that can either contain text or binary. As a bonus, it makes it less error-prone to mix up shader input and output std::strings. As a small drive-by, the skia_print_native_shaders feature now works with SPIR-V as well. Bug: b/390458117 Change-Id: I65326b043e9634651071b66403906a2295fbaacc Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006716 Commit-Queue: Shahbaz Youssefi <syoussefi@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
Manual roll Dawn from 3a5d9945a673 to 3e03a4fc77a2 (188 revisions) Manual roll requested by michaelludwig@google.com https://dawn.googlesource.com/dawn.git/+log/3a5d9945a673..3e03a4fc77a2 2025-07-01 lokokung@google.com [dawn][native] Remove anonymous struct in Buffer. 2025-07-01 jiawei.shao@intel.com Vulkan: Support subgroup matrix with i8 and u8 as element type 2025-07-01 jiawei.shao@intel.com [spirv] Handle offset in subgroupMatrixLoad for i8 u8 matrices 2025-07-01 kainino@chromium.org [dawn] Add missing include <concepts> to ComboLimits.h 2025-06-30 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from d90e098a4598 to 9a4e8ded41e3 (1 revision) 2025-06-29 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll Depot Tools from d255a8d41e7a to 5cc29c7a2178 (30 revisions) 2025-06-29 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from b39b1685ca27 to d90e098a4598 (2 revisions) 2025-06-28 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from d3f4b5f78306 to b39b1685ca27 (12 revisions) 2025-06-28 kainino@chromium.org [dawn] Fix more diffs with upstream webgpu.h, roll webgpu-headers 2025-06-28 kainino@chromium.org [dawn] Return WGPUStatus from wgpuSurfacePresent 2025-06-28 lokokung@google.com [dawn][native] Use a mutex for native event completion. 2025-06-27 jiawei.shao@intel.com [spirv] Handle stride in subgroupMatrixLoad for i8 u8 subgroup matrices 2025-06-27 jiawei.shao@intel.com Add end2end tests on the subgroup matrices in column major 2025-06-27 jimblackler@google.com Kotlin: Include the license in the generated files. 2025-06-27 jiawei.shao@intel.com Vulkan: Filter SubgroupMatrixConfig with F16 when shader-f16 isn't enabled 2025-06-27 jrprice@google.com Suppress some new CTS failures 2025-06-27 dneto@google.com ir-to-program: support bitcast in root block 2025-06-27 kainino@chromium.org [dawn][emscripten] Refactor instance capabilities into features+limits 2025-06-27 dneto@google.com spirv-reader: support spec op SNegate 2025-06-27 dneto@google.com spirv-reader: support fconvert 2025-06-27 dneto@google.com spirv-reader: spec constant op Sconvert, UConvert not supported 2025-06-27 dneto@google.com ir-to-program: support unary op in root block 2025-06-27 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 56e9bfff2ec8 to ee071b31d1c7 (3 revisions) 2025-06-27 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 8a9f8820723d to 94abfe972ad8 (1 revision) 2025-06-27 amaiorano@google.com d3d: Clean up TextureCopySplitter API 2025-06-27 kainino@chromium.org [dawn] Make ComboLimits chain only standard limits extensions by default 2025-06-27 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from e9ab451bc13c to d3f4b5f78306 (12 revisions) 2025-06-27 jiawei.shao@intel.com [dawn] Use C++20 requires expression and concepts library 2025-06-27 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from b3ec3a5f5448 to 56e9bfff2ec8 (8 revisions) 2025-06-27 dsinclair@chromium.org [spirv-reader][ir] Fix user call texture handling 2025-06-26 dsinclair@chromium.org Reset count when clearing the hash map/set. 2025-06-26 jrprice@google.com [ir] Fix offset of struct members 2025-06-26 petermcneeley@google.com [dawn] MacOS use relaxed math where possible 2025-06-26 jrprice@google.com [spirv-reader] Handle ArrayStride on runtime arrays 2025-06-26 jrprice@google.com [spirv-reader] Implement DecomposeStridedArray transform 2025-06-26 jrprice@google.com [ir] Validate construct for array types 2025-06-26 rharrison@chromium.org [tint][ir] Cleanup usage of ValidateBindingOptions 2025-06-26 jrprice@google.com Revert "[Win] Call CompareObjectHandles directly" 2025-06-26 dsinclair@chromium.org Combine generated enums to single file in wgsl/ 2025-06-26 dsinclair@chromium.org Combine generated enums to single file in core/ 2025-06-26 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 93c7c2c8e623 to 8a9f8820723d (1 revision) 2025-06-26 kainino@chromium.org [emscripten] Clarify non-package Emdawnwebgpu readme 2025-06-26 dsinclair@chromium.org [spirv-reader][ir] Ignore RelaxedPrecision on vars. 2025-06-26 rafael.cintron@microsoft.com [Win] Call CompareObjectHandles directly 2025-06-26 dsinclair@chromium.org [spirv-reader][ir] Ignore more struct member decorations. 2025-06-26 dsinclair@chromium.org [spirv-reader][ir] Add validation for struct member matrix annotations. 2025-06-26 amaiorano@google.com d3d: Add missing unit test coverage of Compute2DTextureCopySplits to CopySplitTests 2025-06-26 dsinclair@chromium.org [spirv-reader][ir] Add support for MatrixStride annotation 2025-06-26 dsinclair@chromium.org [spirv-reader][ir] Add support for RowMajor annotation 2025-06-26 dsinclair@chromium.org Remove unused methods. 2025-06-26 dsinclair@chromium.org [spirv] Create depth and array matcher helpers 2025-06-26 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from beae1b4a81c9 to b3ec3a5f5448 (3 revisions) 2025-06-26 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from b390fb19adc5 to 93c7c2c8e623 (2 revisions) 2025-06-26 jrprice@google.com [dawn] Print the subgroup matrix configurations in E2E tests 2025-06-26 dsinclair@chromium.org [spirv-reader][ir] Add support for ColMajor annotation 2025-06-25 cwallez@chromium.org [tint][glsl] Move TextureBuiltinFromUniform after binding remapping. 2025-06-25 lokokung@google.com [dawn][common] Add FetchMax helpers for atomic max. 2025-06-25 shrekshao@google.com [Compat GL] Add render pipeline tests for TextureShaderBuiltinTests 2025-06-25 petermcneeley@google.com [dawn] Change enum test from assert to check 2025-06-25 rharrison@chromium.org [tint][fuzz] Use printer capabilities in writer_fuzz passes 2025-06-25 dsinclair@chromium.org [hlsl][ir] Rename expected files to drop .ir. 2025-06-25 cwallez@chromium.org [tint][glsl] Support binding_array in texture_polyfill 2025-06-25 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from 02a9ce0423c4 to e9ab451bc13c (25 revisions) 2025-06-25 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from ed01d9931de3 to 436722b39118 (1 revision) 2025-06-25 lokokung@google.com Reland "[dawn][metal] Makes AllowSpontaneous work for Metal queue events." 2025-06-25 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 14960950b20c to beae1b4a81c9 (1 revision) 2025-06-25 dsinclair@chromium.org [ir] Remove the use ir flags. 2025-06-25 bajones@chromium.org Reduce allocations in BindGroupTracker 2025-06-25 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 8f5595872e15 to b390fb19adc5 (3 revisions) 2025-06-25 jimblackler@google.com Remove legacy callback handling. 2025-06-25 cwallez@chromium.org [dawn][opengl] Simplify GenerateTextureBuiltinFromUniformData 2025-06-25 tikuta@google.com Add a missing include for int32_t 2025-06-25 dsinclair@chromium.org [spirv-reader][ir] Add support for OpImage 2025-06-25 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from dd725c203c3a to 8f5595872e15 (1 revision) 2025-06-25 jrprice@google.com [spirv] Show stride when printing ExplicitLayoutArray 2025-06-25 dsinclair@chromium.org Remove the TINT_IR_ICE macro. 2025-06-24 jrprice@google.com [spirv-reader] Add support for Coherent decoration 2025-06-24 dsinclair@chromium.org [spirv-reader][ir] Support OpImageDrefGather. 2025-06-24 dsinclair@chromium.org [hlsl] Remove AST expectations. 2025-06-24 senorblanco@chromium.org Add Compat test: different aspects, same texture. 2025-06-24 dsinclair@chromium.org [spirv-reader][ir] Enabling more texture tests. 2025-06-24 dsinclair@chromium.org [spirv-reader] Move reader specific AST transforms into reader folder. 2025-06-24 dsinclair@chromium.org Remove unused AST transforms. 2025-06-24 dsinclair@chromium.org [spirv-reader][ir] Add support for ProjDref instructions. 2025-06-24 dsinclair@chromium.org [spirv-reader][ir] Support OpImageSampleDrefExplicitLod 2025-06-24 dsinclair@chromium.org [spirv-reader][ir] Support OpImageSampleDrefImplicitLod 2025-06-24 dsinclair@chromium.org [hlsl] Remove the AST backend. 2025-06-24 dsinclair@chromium.org [hlsl] Use IR by default for CLI and E2E tests. 2025-06-24 dsinclair@chromium.org [hlsl][ir] Remove the use_tint_ir flag from the HLSL backend 2025-06-24 gman@chromium.org Add link to webgpu node package 2025-06-24 chrome-branch-day@chops-service-accounts.iam.gserviceaccount.com Activate dawn M139 2025-06-24 amaiorano@google.com d3d12: Clean up CopySplitTests 2025-06-24 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 79ac1a8cd767 to 14960950b20c (3 revisions) 2025-06-24 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from d1d0a31a7a6a to dd725c203c3a (1 revision) 2025-06-24 beaufort.francois@gmail.com Revert "Disable texture_1d support for textureSampleLevel by default" 2025-06-24 jimblackler@google.com Kotlin: New integration test for Adapter functionality. 2025-06-24 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 2102183af046 to 79ac1a8cd767 (7 revisions) 2025-06-24 dsinclair@chromium.org [spirv-reader][ir] Add SPV_KHR_vulkan_memory_model as an accepted extension. 2025-06-23 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from 0ce9bc68d3fc to ed01d9931de3 (1 revision) 2025-06-23 amaiorano@google.com d3d12: Make TextureCopySplitter public API use strong types 2025-06-23 lokokung@google.com [dawn][common] Adds nullptr constructor for ResultOrError<Ref<T>, E>. 2025-06-23 dsinclair@chromium.org [spirv-reader][ir] Pre-register all functions. 2025-06-23 alanbaker@google.com Enable use of SPIR-V 1.4 in Dawn 2025-06-23 lokokung@google.com [dawn][native] Updates buffer mapping w.r.t device lock. 2025-06-23 petermcneeley@google.com Remove unrollConstEvalLoops flag from CTS runner 2025-06-23 dawn-automated-expectations@chops-service-accounts.iam.gserviceaccount.com Roll third_party/webgpu-cts/ 97733807d..2a8d4a83f (1 commit) 2025-06-23 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from 317a1e4780ef to 02a9ce0423c4 (2 revisions) 2025-06-23 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 5220e634850e to 2102183af046 (1 revision) 2025-06-23 cwallez@chromium.org [dawn][d3d] Support BGLArraySize in D3D11 and D3D12 2025-06-23 cwallez@chromium.org [tint][glsl] Directly create all replacements in texture_polyfill 2025-06-23 shanxing.mei@intel.com Enable multisample and resolve for Snorm formats in TextureFormatsTier1 2025-06-21 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from b4baabb7da9e to d1d0a31a7a6a (1 revision) 2025-06-21 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from 316ed08fbc9a to 317a1e4780ef (11 revisions) 2025-06-21 zhaoming.jiang@microsoft.com Dawn: Dump shader module source before checking cache if required 2025-06-21 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from b78ac5059324 to b4baabb7da9e (1 revision) 2025-06-21 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 9e629bbb0f73 to 5220e634850e (6 revisions) 2025-06-20 sunnyps@chromium.org Mark WebGPU-WebGPU canvas copy tests as flaky on Nvidia Win 2025-06-20 lokokung@google.com [webgpu][headers] Fixes missing FreeMember code-gen. 2025-06-20 kainino@chromium.org [emscripten] Fix emdawnwebgpu_tests_jspi 2025-06-20 amaiorano@google.com d3d12: Move functions in TextureCopySplitter to anonymous namespace 2025-06-20 cwallez@chromium.org [dawn][native] Add end2end tests passing binding_array as argument. 2025-06-20 amaiorano@google.com d3d12: Rework TextureCopySplitter to compute in blocks 2025-06-20 jrprice@google.com [msl] Add new capability to ModuleConstant transform 2025-06-20 cwallez@chromium.org [dawn][opengl] Simplify GenerateCombinedSamplerInfo 2025-06-20 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 649514930e65 to 9e629bbb0f73 (4 revisions) 2025-06-20 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 5aec1ec4e4d0 to b78ac5059324 (1 revision) 2025-06-20 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from f51bb730689f to 316ed08fbc9a (9 revisions) 2025-06-20 cwallez@chromium.org [tint][spirv-writer] Add a KeepBindingArrayAsPointer transform. 2025-06-19 dsinclair@chromium.org [spirv-reader][ast] Move SPIR-V specific transform to lower folder. 2025-06-19 dsinclair@chromium.org [spirv-reader][ir] Break non-scalar/vector Output apart. 2025-06-19 dsinclair@chromium.org [spirv-reader][ir] Error on non-array stride. 2025-06-19 dsinclair@chromium.org [spirv-reader][ir] Handle BufferBlock annotation. 2025-06-19 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 5928133450bb to 649514930e65 (6 revisions) 2025-06-19 dsinclair@chromium.org [spirv-reader][ir] Support Block struct annotation. 2025-06-19 dsinclair@chromium.org [spirv-reader][ir] Support ArrayStride annotation 2025-06-19 amaiorano@google.com d3d12: Add BlockCount and TexelCount typed integers 2025-06-19 petermcneeley@google.com [tint] N versions of fmax and fmin 2025-06-19 rafael.cintron@microsoft.com [Win] Fix GetModulePath to return correct Dawn DLL path 2025-06-19 jrprice@google.com [msl] Avoid generating packed_bool3 types 2025-06-19 dsinclair@chromium.org [spirv-reader][ir] Guard against unhandled decorations for types. 2025-06-19 dsinclair@chromium.org [ir] Add validation for core types only 2025-06-19 petermcneeley@google.com [dawn] Remove expectation now that all ftoi changes have landed 2025-06-19 jrprice@google.com [msl] Add convert overloads for vec3<bool> 2025-06-19 petermcneeley@google.com [dawn] Add bug for concurrent dispatching msl 2025-06-19 shanxing.mei@intel.com Enable and test Snorm blending in TextureFormatsTier1 2025-06-19 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 9d70781f00fe to 5928133450bb (2 revisions) 2025-06-19 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from ecfc59d5cfd1 to f51bb730689f (12 revisions) 2025-06-19 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 978a6d3f13ee to 5aec1ec4e4d0 (2 revisions) 2025-06-19 kainino@chromium.org Reland "[dawn] Move Compat limits into an extension struct" 2025-06-18 shrekshao@google.com [Compat GL] Fix emulated uniform setup on pipeline change 2025-06-18 lokokung@google.com Roll third_party/webgpu-headers/src/ 60cd90203..4f617851d (2 commits) 2025-06-18 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 9e0008570681 to 9d70781f00fe (2 revisions) 2025-06-18 beaufort.francois@gmail.com Disable texture_1d support for textureSampleLevel by default 2025-06-18 cwallez@chromium.org [tint][glsl] Remove unused options to glsl::texture_polyfill. 2025-06-18 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from 47399fc146fa to 0ce9bc68d3fc (1 revision) 2025-06-18 dawn-automated-expectations@chops-service-accounts.iam.gserviceaccount.com Roll third_party/webgpu-cts/ 605c2fec8..97733807d (2 commits) 2025-06-18 amaiorano@google.com TypedInteger: all explicit casts to any integral type and add missing arithmetic operations 2025-06-18 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 8a77b0c714d0 to 978a6d3f13ee (1 revision) 2025-06-18 dsinclair@chromium.org [spirv-reader] Remove infinite loop test. 2025-06-18 rafael.cintron@microsoft.com Always Load KernelBase.dll from System32 directory 2025-06-18 shaoboyan@microsoft.com Use BitSetRangeIterator for immediate data uploading in vulkan 2025-06-18 jimblackler@google.com Kotlin: Have image test output generated image for reference. 2025-06-18 beaufort.francois@gmail.com Add texture_1d support to textureSampleLevel 2025-06-18 lokokung@google.com [webgpu][headers] Other minor fixes to reduce header diff. 2025-06-18 lokokung@google.com [webgpu][headers] Updates generated FreeMembers functions to match upstream. 2025-06-18 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from b72a76c990d9 to ecfc59d5cfd1 (6 revisions) 2025-06-18 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 45ac6404179e to 9e0008570681 (4 revisions) 2025-06-18 petermcneeley@google.com [tint] Basic scalarization support (max min clamp) 2025-06-18 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from f94396ddffa8 to 8a77b0c714d0 (1 revision) 2025-06-18 lokokung@google.com [webgpu][headers] Adds forward declaration for callback info structs. 2025-06-18 lokokung@google.com [webgpu][headers] Adds true and false constants for Dawn's headers. 2025-06-17 dsinclair@chromium.org [spirv-reader][ir] Split input structs apart. 2025-06-17 dsinclair@chromium.org [spirv-reader][ir] Remove struct attributes after processing inputs 2025-06-17 petermcneeley@google.com [tint] New ftoi conversion maximums in const eval 2025-06-17 dsinclair@chromium.org [spirv-reader][ir] Break non-scalar/vector Input apart. 2025-06-17 rharrison@chromium.org [tint][ir] Validate that discard is not directly in a non-fragment 2025-06-17 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from dfefcbc68cc3 to 45ac6404179e (3 revisions) 2025-06-17 zhaoming.jiang@microsoft.com Dawn: Implement shader module front-end blob cache 2025-06-17 jiawei.shao@intel.com [spirv] Support i8 and u8 in subgroupMatrixMultiply{Accumulate} 2025-06-17 dneto@google.com [vulkan] Set VK_PIPELINE_SHADER_STAGE_CREATE_ALLOW_VARYING_SUBGROUP_SIZE_BIT 2025-06-17 jimblackler@google.com Kotlin: Correctly construct default=zero member structures 2025-06-17 zhaoming.jiang@microsoft.com Dawn: Add abs class LogEmitter and make DeviceBase impl it 2025-06-17 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from 7b7b93aa0b85 to 47399fc146fa (1 revision) 2025-06-17 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll vulkan-deps from 5949c46ca858 to b72a76c990d9 (9 revisions) 2025-06-17 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll DirectX Shader Compiler from 57177f77a4dc to f94396ddffa8 (1 revision) 2025-06-17 jrprice@google.com [dawn] Improve E2E test for subgroup matrices 2025-06-17 dawn-autoroll@skia-public.iam.gserviceaccount.com Roll ANGLE from 79ec8b3400ce to dfefcbc68cc3 (3 revisions) 2025-06-16 lokokung@google.com Revert "Cache VkFramebuffers" If this roll has caused a breakage, revert this CL and stop the roller using the controls here: https://autoroll.skia.org/r/dawn-skia-autoroll Please CC cwallez@google.com,dneto@google.com,michaelludwig@google.com on the revert to ensure that a human is aware of the problem. To file a bug in Dawn: https://bugs.chromium.org/p/dawn/issues/entry To file a bug in Skia: https://bugs.chromium.org/p/skia/issues/entry To report a problem with the AutoRoller itself, please file a bug: https://issues.skia.org/issues/new?component=1389291&template=1850622 Documentation for the AutoRoller is here: https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md Cq-Include-Trybots: skia/skia.primary:Build-Ubuntu24.04-Clang-x86_64-Debug-Graphite_Dawn_Vulkan;skia/skia.primary:Test-Mac14-Clang-MacMini9.1-GPU-AppleM1-arm64-Debug-All-ASAN_Graphite_Dawn_Metal;skia/skia.primary:Test-Win10-MSVC-Golo-GPU-QuadroP400-x86_64-Release-All-Graphite_Dawn_D3D12;skia/skia.primary:Test-Win10-MSVC-Golo-GPU-QuadroP400-x86_64-Release-All-Graphite_Dawn_D3D11 Bug: None Tbr: dneto@google.com,michaelludwig@google.com Test: Test: AdapterTest Test: Test: ImageTest Test: Test: Manual Test: Test: StructuresTest Test: Test: dawn_end2end_tests Test: Test: tint_unittests Change-Id: I691598b6a5b1a66dc175aefee81f701b412eaa1b Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1014504 Reviewed-by: Max Hudnell <maxhudnell@google.com> Commit-Queue: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Michael Ludwig <michaelludwig@google.com> | 1 年前 | |
Build skslc and sksl-minify from the modular Bazel build. This removes the //src/sksl:enable_skslc bool flag, much of the code guarded by SKSL_STANDALONE and all the code defined by the SK_ENABLE_SPIRV_VALIDATION, SK_ENABLE_WGSL_VALIDATION, and SK_ENABLE_SPIRV_CROSS defines. To achieve this, a few helper enums, callbacks, and new files have been introduced. To avoid a dependency cycle, e.g. sksl-minify -> core -> minified SkSL -> sksl-minify, we do add a Bazel flag to //src/core, which allows the source files to be loaded from disk instead of compiled in. For GN, we split up the .gni file lists so we can selectively include new groups instead of needing to subtract out the default module. This also moves //tools/SkGetExecutablePath.h to src/utils just because it's a bit awkward to have something in //src depend on //tools, even if it's only for our test/helper binaries. The get_executable_path has no deps, so it's ~free to move. Suggested review order: - tools/skslc/Main.cpp to see how extra params are passed in to get the "generator only" behavior. - C++ changes in src/sksl/codegen/ to remove nearly all the #ifdefs that guarded "generator only" behavior. - gn/sksl.gni to see that I've split up skia_sksl_sources and skia_sksl_gpu_sources. We'll want to update clients to use the more granular rules (Chromium and Flutter) so we can remove the coarser file groups. - BUILD.gn to see these groups be used. - BUILD.bazel changes to encapsulate the new rules. - all other files in any order. Change-Id: I695d25398d5fdf90c2656fe5a59739c014ef46f8 Bug: b/345537973 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/862416 Commit-Queue: Kaylee Lubick <kjlubick@google.com> Reviewed-by: Brian Osman <brianosman@google.com> | 2 年前 |