)]}'
{
  "log": [
    {
      "commit": "db02a07b9acf4e4e896d55629d070f3469d7c1e6",
      "tree": "01ffbbd846f93dc17e68536fcfc07e66c8aeabe2",
      "parents": [
        "a785e154a2b6142b93950cd618dce5b0a918cb54"
      ],
      "author": {
        "name": "Matt Stark",
        "email": "msta@google.com",
        "time": "Sun Nov 15 01:47:34 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sun Nov 15 01:53:39 2026"
      },
      "message": "build: Expose generated public headers via public_deps\n\nTargets that declare generated headers in `public` must have a\npublic_dep on said header.\n\nThis will start being enforced with gn-review.googlesource.com/c/gn/+/28781,\nso we first need to fix up chromium so `gn gen` doesn\u0027t fail.\n\nBug: 556959580\nChange-Id: Ib1474e9f5357477e7e6d2d3270541b0f6a6a6964\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8804196\nOwners-Override: Matt Stark \u003cmsta@google.com\u003e\nCommit-Queue: Matt Stark \u003cmsta@google.com\u003e\nReviewed-by: Takuto Ikuta \u003ctikuta@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1745033}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9cc30ff0b660f198576fd3dea267c72c6e972507\n"
    },
    {
      "commit": "a785e154a2b6142b93950cd618dce5b0a918cb54",
      "tree": "33c2b52e296b92b44ebf2bc5d58dce68fc781664",
      "parents": [
        "54578f20feadffb50b824e1130c184244f52dcc0"
      ],
      "author": {
        "name": "Lokbondo Kung",
        "email": "lokokung@google.com",
        "time": "Sat Nov 14 09:22:15 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 14 09:29:34 2026"
      },
      "message": "[webgpu] Defer DawnWireServices::Disconnect() on GPU context loss\n\nWhen GPU context loss occurs on the client thread,\nWebGPUImplementation::OnGpuControlLostContext() previously called\nDawnWireServices::Disconnect() synchronously before invoking\nlost_context_callback_. Because dawn::wire::client::Client::Disconnect()\nsynchronously fires spontaneous Dawn callbacks (such as device.lost and\nRequestAdapter()), those callbacks could re-enter Blink/JavaScript and\ntear down the ExecutionContext (and thus DawnControlClientHolder and\nWebGPUImplementation) while\nWebGPUImplementation::OnGpuControlLostContext() was still on the call\nstack, causing a use-after-free when OnGpuControlLostContext() resumed.\n\nFix this by:\n1. Posting DawnWireServices::Disconnect() to main_task_runner_ from\n   WebGPUImplementation::OnGpuControlLostContext() so Dawn wire\n   callbacks run in a fresh call stack.\n2. Reordering DawnWireServices::Disconnect() to disconnect serializer_\n   and memory_transfer_service_ before wire_client_.Disconnect(), and\n   guarding DawnWireServices::Disconnect() with disconnected_ so\n   synchronous teardown of WebGPUImplementation inside a\n   wire_client_.Disconnect() callback cannot access freed\n   WebGPUImplementation or WebGPUCmdHelper members.\n3. Keeping synchronous api_channel_-\u003eDisconnect() in\n   DawnControlClientHolder::Destroy() (when ExecutionContext teardown is\n   already underway and Blink callbacks early-out) while removing the\n   synchronous disconnect from\n   DawnControlClientHolder::MarkContextLost().\n\nBug: 566646405\nChange-Id: I924e0ef8261346fe9385bcc4fe513752efda564d\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8753029\nCommit-Queue: Loko Kung \u003clokokung@google.com\u003e\nReviewed-by: Kai Ninomiya \u003ckainino@chromium.org\u003e\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1744902}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8f02a7f058ce2c132a0bd200e5c4e92205825273\n"
    },
    {
      "commit": "54578f20feadffb50b824e1130c184244f52dcc0",
      "tree": "fd9d2d72292e7d03f3a314068b8d873ffb793fb7",
      "parents": [
        "8bc61a59794b0740bc41e7ccd5ab5b4242435905"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Fri Nov 13 17:12:43 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 17:17:47 2026"
      },
      "message": "gpu: Take spans in ContextState::GetStateAs*()\n\nUpdate the generated GetStateAsGLint() and GetStateAsGLfloat() to take a\nbase::span and fill it with copy_from(), dropping num_written and the\nempty-buffer size-query path (already covered by GetHelper()\u0027s fallback\nto GLES2Util::GLGetNumValuesReturned()).\n\nThis removes 134 UNSAFE_TODO macros from context_state_impl_autogen.h,\nleaving 2 at the GLES2DecoderImpl caller boundary.\n\nChange-Id: Ia2dcb10ffe92ea2aaf70abb03f9d6ab121844294\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8784180\nCommit-Queue: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1744173}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5e6c4295a474b7c69442eca23c19f3dfa91423f5\n"
    },
    {
      "commit": "8bc61a59794b0740bc41e7ccd5ab5b4242435905",
      "tree": "2f20be5e1e2863a2951d4ebeac16cc9463a1f4e0",
      "parents": [
        "68d8da2b1f562ff470e58fac83da7405972fa446"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Fri Nov 13 16:22:30 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 16:29:34 2026"
      },
      "message": "Fix and re-enable GPU.WebGPU.Support\n\nReportWebGPUSupportMetrics has been disabled by default on Windows,\nAndroid, and Linux since late 2023:\n- On Windows (crbug.com/40073612), enumerating all D3D12 adapters woke\n  sleeping secondary GPUs on hybrid laptops and hung the GPU process.\n  Use GetActiveAdapterLuid and RequestAdapterOptionsLUID so Dawn only\n  probes the GPU Chrome is already using, and record kNone when no\n  active hardware adapter LUID is available.\n- On Android and Linux (crbug.com/40075983, crbug.com/40076059,\n  crbug.com/346990068), enumeration crashed in Dawn\u0027s Vulkan and\n  OpenGLES backends before Vulkan 1.1 checks and explicit EGLDisplay\n  and GetGLProcAddress plumbing landed. In addition, CollectDawnInfo\n  returned early for collect_metrics before the absl::Cleanup guard\n  added in crbug.com/355661612 to restore the GPU thread\u0027s EGLContext.\n  Move the absl::Cleanup guard ahead of the collect_metrics early\n  return.\n\nAlso pass WebGPU Dawn toggles when enumerating metric adapters, skip\nprobing OpenGLES compatibility mode when a Core adapter is already\nsupported, reuse the enumerated Core adapters in\nReportWebGPUAdapterMetrics, and remove the obsolete\nWebGPUSupportMetrics entry from fieldtrial_testing_config.json.\n\nBug: 40073612, 40075983, 40076059, 346990068, 355661612, 578398837\nChange-Id: I2dd2240c81ef73085cd3f3bdb20138d6d9cd21e5\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8612731\nReviewed-by: Roger McFarlane \u003crogerm@chromium.org\u003e\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1744152}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c5b15a4afe7e57bcd9ec4b4960272f7da509ddec\n"
    },
    {
      "commit": "68d8da2b1f562ff470e58fac83da7405972fa446",
      "tree": "d2cc186efd17386b5748342095982b94b412c7f4",
      "parents": [
        "44ee829679c78d05b2d120d9007e2942556be558"
      ],
      "author": {
        "name": "Ilya Nikolaevskiy",
        "email": "ilnik@chromium.org",
        "time": "Fri Nov 13 16:04:04 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 16:28:44 2026"
      },
      "message": "[gpu] Add GraphiteSharedContext::RunLockedAndSubmitSync()\n\nExisting AndSubmit variants don\u0027t suit usecases of multiplane readback.\n\nReplace the AndSubmit() variants with RunLockedAndSubmitSync(), which runs\na callback with the context locked and then submits with SyncToCpu::kYes\nunder the same lock.\n\nAlso, WrappedGraphiteTextureBacking::ReadbackToMemory() now reads all\nplanes with one submit.\n\nBug: 583357780\nChange-Id: Ia1611c10d25593757ff210d894287134881df9fd\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8801756\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCommit-Queue: Ilya Nikolaevskiy \u003cilnik@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1744147}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5dfa71dd5d14e9ab16da6cb435e603e8713baa9b\n"
    },
    {
      "commit": "44ee829679c78d05b2d120d9007e2942556be558",
      "tree": "109525acd6515d185874e86acfc17dcfc1f4c72f",
      "parents": [
        "391e21763d94b8e44ae6aff9d0c5957666213eee"
      ],
      "author": {
        "name": "Satoru Takabayashi",
        "email": "satorux@google.com",
        "time": "Fri Nov 13 13:46:38 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 13:53:00 2026"
      },
      "message": "[GPU] Remove unused per-StoragePartition DawnGraphite disk cache\n\nRenderProcessHostImpl::Init() creates a per-StoragePartition GPU disk\ncache for each entry in kGpuDiskCacheTypes, including kDawnGraphite\n(added in crrev.com/c/5296841). However, Skia Graphite only uses the\nglobal kGraphiteDawnGpuDiskCacheHandle in GpuServiceImpl, and renderer\nGpuChannels only look up kGlShaders and kDawnWebGPU, leaving the\nper-StoragePartition DawnGraphiteCache instances unused.\n\nRemove kDawnGraphite from kGpuDiskCacheTypes to avoid creating an\nunused blockfile cache (~100 KiB browser PSS and 5 open FDs per active\nnon-OTR StoragePartition).\n\nBug: 325309553, b:575247202\nTest: gpu_unittests --gtest_filter\u003dGpuDiskCacheTest.*\nTest: content_unittests --gtest_filter\u003d*ShaderCacheTest.*\nTest: tast run \u003cdut\u003e multivm.Login.novm\nChange-Id: Idcbeac147fa79790c46abfa1270ed69fc011b312\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8605386\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nReviewed-by: Hidehiko Abe \u003chidehiko@chromium.org\u003e\nCommit-Queue: Satoru Takabayashi \u003csatorux@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1744084}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: de6d4c4b92478aa2aad72c348ee29541732b22e5\n"
    },
    {
      "commit": "391e21763d94b8e44ae6aff9d0c5957666213eee",
      "tree": "f5abe817fd7b2fdb7d1e42b28419d274c09a5a8e",
      "parents": [
        "b1237ed0980ea5c15a3688c9a33fb8038ab86b5c"
      ],
      "author": {
        "name": "Thiabaud Engelbrecht",
        "email": "thiabaud@google.com",
        "time": "Fri Nov 13 12:20:22 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 12:30:40 2026"
      },
      "message": "[gpu/service] Deserialize OOP-R fonts if BeginRasterCHROMIUM() fails\n\nDeserialize the font buffer in RasterDecoderImpl::DoRasterCHROMIUM()\nbefore checking `sk_surface_` or `paint_buffer`.\n\nClientFontManager::Serialize() calls SkStrikeServer::writeStrikeData(),\nwhich irreversibly marks typefaces, initial SkFontMetrics, and glyphs\nas sent to the GPU. When DoBeginRasterCHROMIUM() fails to create\n`sk_surface_` for a tile, returning early in DoRasterCHROMIUM() drops\nthat tile\u0027s session-wide font delta. Subsequent tiles using the same\ntypeface or strike then fail SkStrikeClient::readStrikeData() or miss\ncached glyphs, triggering ServiceFontManager DumpWithoutCrashing()\nreports in notifyReadFailure() and notifyCacheMiss().\n\nAdd an absl::Cleanup guard in DoRasterCHROMIUM() to unlock\n`new_locked_handles` and call SkGraphics::PurgePinnedFontCache() when\nreturning before transferring handles to `locked_handles_`, since\nDoEndRasterCHROMIUM() also returns early when `sk_surface_` is null.\n\nChange-Id: I5172b82981f294da0d61534dfc64acd727b7d6b8\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8791603\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Thiabaud Engelbrecht \u003cthiabaud@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1744024}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bb51f2ae7a58572e1bd756da6a536cbfd029e46f\n"
    },
    {
      "commit": "b1237ed0980ea5c15a3688c9a33fb8038ab86b5c",
      "tree": "2fe0e6ee34c8241160be54dbacc7355f50a919f0",
      "parents": [
        "b04b2bc036699afc9df0c6b867947dd22238e156"
      ],
      "author": {
        "name": "Le Hoang Quyen",
        "email": "lehoangquyen@chromium.org",
        "time": "Fri Nov 13 08:05:22 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 08:17:34 2026"
      },
      "message": "graphite: force submit if there are too many pending commands\n\nAdd max_pending_commands parameter to GraphiteSharedContext and\nSkiaGraphiteFeatureParams (defaulting to 500000). When the number\nof pending commands reaches or exceeds max_pending_commands in\ninsertRecording(), automatically submit and flush the backend.\n\nBug: 434070352\nChange-Id: Ib9fdeca16a13fc83081f76a6a87956111f01937f\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8773832\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCommit-Queue: Quyen Le \u003clehoangquyen@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1743884}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: bbc5b4e7649596631ab2fee817e0acf693b2fb5f\n"
    },
    {
      "commit": "b04b2bc036699afc9df0c6b867947dd22238e156",
      "tree": "9c73300b2162e4fe3492b204d50550c3343e1f57",
      "parents": [
        "b048703df26a50e81d824618208fe2d033b72fb9"
      ],
      "author": {
        "name": "Jilin Yang",
        "email": "jilinyang@google.com",
        "time": "Fri Nov 13 05:45:44 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 05:52:28 2026"
      },
      "message": "[unsafe-buffer] Remove unsafe buffer usage from RasterImplementationTest\n\nThe test wrapped memcpy/memset/memcmp and ring buffer pointer arithmetic\nin UNSAFE_TODO. Use the span accessors the involved objects already\nexpose (Buffer::as_byte_span(), ExpectedMemoryInfo::span,\nCommandBufferHelper::GetPutOffsetForTest()): the mock actions take spans\nand command comparisons use bounds-checked slices. The YUV plane spans\ncome from gfx::SkPixmapToSpan() instead of raw pointer + size. GetPut()\nnow returns a span, and CommandsMatch() gains an overload taking it so\nthe hand-written comparisons share the generated tests\u0027 helper.\n\nFixed: 490859258\nChange-Id: I3db4450d1052aec5047402af9b09153cc45ef218\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8778100\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Jilin Yang \u003cjilinyang@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1743802}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2fb19577dd7f01ce332b7c810192a5ba67e19c21\n"
    },
    {
      "commit": "b048703df26a50e81d824618208fe2d033b72fb9",
      "tree": "79009d732d5a619682e3be75bb7453b1ce7b201d",
      "parents": [
        "214e611b3e0dde9463436da3f92047308119754d"
      ],
      "author": {
        "name": "Lokbondo Kung",
        "email": "lokokung@google.com",
        "time": "Fri Nov 13 05:25:57 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 13 05:32:35 2026"
      },
      "message": "[webgpu] Pass DawnInstance into WebGPUDecoder::Create()\n\nMove creation of the DawnInstance used by WebGPU contexts out of\nWebGPUDecoder::Create() and into its callers so that the instance can\neventually be shared across WebGPU contexts. For now,\nWebGPUCommandBufferStub, InProcessCommandBuffer, and WebGPUDecoderTest\nall create it directly via DawnInstance::CreateForWebGPU().\n\nDawnInstance is now exported from the gpu_gles2 component since it is\ncreated from gpu_ipc_service.\n\nThere is no behavior change yet: each WebGPU context still gets its own\nDawnInstance.\n\nBug: 412761228\nChange-Id: I9727044cd491fda62e30022114b700f7907cd563\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8678299\nReviewed-by: Kai Ninomiya \u003ckainino@chromium.org\u003e\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Loko Kung \u003clokokung@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1743798}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a4ce34a539eea32474478b5bb0ccf0e5ddf92d69\n"
    },
    {
      "commit": "214e611b3e0dde9463436da3f92047308119754d",
      "tree": "57d9f8deabbbc8937917814bd8a8a7e51dd45985",
      "parents": [
        "8d42ecfe5894d80b6225de3ec805d651abadb4d2"
      ],
      "author": {
        "name": "Lokbondo Kung",
        "email": "lokokung@google.com",
        "time": "Thu Nov 12 23:15:58 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 23:22:15 2026"
      },
      "message": "[webgpu] Only build WebGPU decoder and stub sources when using Dawn\n\nWebGPU contexts can only be serviced when Dawn is enabled, but\nwebgpu_decoder.{h,cc} and webgpu_command_buffer_stub.{h,cc} were built\nunconditionally and relied on NOTREACHED() in WebGPUDecoder::Create()\nfor builds without Dawn.\n\nMove those sources under `use_dawn` in the BUILD files so they can use\nDawn types freely without BUILDFLAG(USE_DAWN) guards, and instead guard\nthe few places that create WebGPU contexts:\n- GpuChannel::CreateCommandBuffer() and\n  InProcessCommandBuffer::InitializeOnGpuThread() now fail WebGPU\n  context creation with kFatalFailure when Dawn is not enabled instead\n  of crashing in WebGPUDecoder::Create().\n- InProcessCommandBuffer::GetWebGPUDecoderForTest() is only declared\n  when Dawn is enabled.\n- Remove the now unreachable Fuchsia branch in\n  WebGPUCommandBufferStub::Initialize() since Fuchsia does not use\n  Dawn.\n\nBug: 412761228\nChange-Id: I8fd575eeae8a218cca4b6bd73a1bf68b7bc898b4\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8695257\nReviewed-by: Kai Ninomiya \u003ckainino@chromium.org\u003e\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Loko Kung \u003clokokung@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1743513}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a2147410f710a95c44efb85b537fdccd69757fcc\n"
    },
    {
      "commit": "8d42ecfe5894d80b6225de3ec805d651abadb4d2",
      "tree": "a7747fd9db649e282ed17d6b9ee95c3fbca1cd56",
      "parents": [
        "3880da3a48ca22e543eb62c14e058aca8de534ed"
      ],
      "author": {
        "name": "Lokbondo Kung",
        "email": "lokokung@google.com",
        "time": "Thu Nov 12 21:28:20 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 21:36:52 2026"
      },
      "message": "[webgpu] Make DawnWireServices::lock_ unconditional\n\nSince crrev.com/c/7953663 changed WebGPUImplementation\u0027s\nmain_task_runner_ to always initialize to\nSequencedTaskRunner::GetCurrentDefault(), the support_locking parameter\non WebGPUImplementation\u0027s constructor has been unused, and\nWebGPUImplementation::Initialize() (which passed\n`main_task_runner_ !\u003d nullptr`) has already been unconditionally\nenabling DawnWireServices::lock_.\n\nRemove the unused support_locking parameter from WebGPUImplementation\nand DawnWireServices and store DawnWireServices::lock_ as a non-optional\nbase::Lock directly, as multithreaded wire handling is intended to be\nenabled permanently.\n\nBug: 566646405\nChange-Id: Id91e44c351c6038071a64e26ac452235f4dd4a39\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8798665\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Loko Kung \u003clokokung@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1743386}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a4d25b86544ffd9c661423e29c45309905308034\n"
    },
    {
      "commit": "3880da3a48ca22e543eb62c14e058aca8de534ed",
      "tree": "2eece640675538fc351af1f111003633162b2153",
      "parents": [
        "6b0f514e6259a08231a98ed8d1bb3596e811cf41"
      ],
      "author": {
        "name": "Jordan Bayles",
        "email": "jophba@chromium.org",
        "time": "Thu Nov 12 18:08:23 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 18:14:11 2026"
      },
      "message": "[gpu] Cap initial command buffer wait priority boost at kNormal\n\nCommandBufferStub::WaitForTokenInRange and WaitForGetOffsetInRange\npreviously boosted any waiting command buffer sequence immediately to\nSchedulingPriority::kHigh. Because equal-priority sequences in\ngpu::Scheduler are ordered by task order_num, a renderer or worker WebGL\ncontext calling gl.readPixels() or gl.getError() after queueing heavy\nshaders would boost its entire queued shader backlog to kHigh and run\nahead of newly scheduled Viz Display::DrawAndSwap and CopyOutputRequest\ntasks.\n\nCap the initial wait priority boost at SchedulingPriority::kNormal (so\nkLow sequences boost to kNormal while kNormal sequences remain at\nkNormal) and start a timer (kCommandBufferWaitHighPriorityDelay,\ndefault 100ms) that escalates the sequence to kHigh only if the wait\nremains unresolved. This prevents routine per-frame waits from\npreempting Viz mid-frame while ensuring a command buffer blocked behind\ncontinuous kHigh GPU work cannot starve or trigger the client thread\u0027s\nHangWatcher.\n\nBug: 40287938, b/329260636\nChange-Id: Iec02c90634b96959b04d8f0fbdeac6569aa1a30e\nSkip-Clang-Tidy-Checks: modernize-use-default-member-init\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8670114\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCommit-Queue: Jordan Bayles \u003cjophba@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1743132}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 494e61ebd2fd9df8b61d8ab82f950d6395e3dc34\n"
    },
    {
      "commit": "6b0f514e6259a08231a98ed8d1bb3596e811cf41",
      "tree": "7f142258a35b8adc6b84b296e2a93591a3633969",
      "parents": [
        "6afef0160325c7abdc39e65bfd9f44e141330534"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Thu Nov 12 15:33:23 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 15:38:17 2026"
      },
      "message": "[gpu] Remove disable_accelerated_vp8_decode driver workaround\n\nThis only has entries for Windows but is checked only for Linux/CrOS\nin media VideoDecoderPipeline. This CL removes this workaround entry.\n\nBug: 570597474\nChange-Id: I00c6103f61628eb46b8aa3f67bb11dcb9618607d\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8793894\nReviewed-by: Dale Curtis \u003cdalecurtis@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1742940}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 4eefd81b0a4a012ab8a5832ab1a156b10e8ea48e\n"
    },
    {
      "commit": "6afef0160325c7abdc39e65bfd9f44e141330534",
      "tree": "73330b6a3740d605434d5866c28bb9e483002f10",
      "parents": [
        "a80f7565b7759e2f5efb840b132c1077e1b345ea"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Thu Nov 12 11:37:20 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 11:44:03 2026"
      },
      "message": "gpu: Remove unused array-state code in command buffer generator\n\nList-default state support was added in 337d59638d63 (2014) solely for\nCHROMIUM_path_rendering\u0027s 16-float matrices, which were removed in\nd6222bed782d (2020).\n\nDelete the six dead branches in build_cmd_buffer_lib.py (including four\nuncompiled UNSAFE_TODO macros) and assert that defaults are strings.\nNo change to generated files.\n\nChange-Id: I423d6f8bc270bd7b2d47f39937a98ed872cb9346\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8802313\nAuto-Submit: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nCommit-Queue: Corentin Wallez \u003ccwallez@chromium.org\u003e\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1742797}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b4fc675babb06ba9fe707466748061c01746e468\n"
    },
    {
      "commit": "a80f7565b7759e2f5efb840b132c1077e1b345ea",
      "tree": "dd928a46b2ed37aec95bd38da0726e8003dfbbda",
      "parents": [
        "64ff78eb34dc455f80c636d9bdcc71748874cde7"
      ],
      "author": {
        "name": "Michael Tang",
        "email": "tangm@microsoft.com",
        "time": "Thu Nov 12 02:04:55 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 02:11:02 2026"
      },
      "message": "Use DXGI_SCALING_NONE for render pass backing swap chains\n\nWe previously used DXGI_SCALING_STRETCH which doesn\u0027t appear to have an\neffect with CreateSwapChainForComposition, as DComp visuals do not have\nan explicit target output size.\n\nWe\u0027ll switch to DXGI_SCALING_NONE in preparation of new Windows APIs\nthat allow us to specify an explicit target size for DComp visuals. This\nAPI is guaranteed to be available when EXPERIMENTAL_IDCompositionDevice7\nis available.\n\nChange-Id: If9a370000b4e87e52e990127249f62096a6a6964\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8272679\nReviewed-by: Rafael Cintron \u003crafael.cintron@microsoft.com\u003e\nCommit-Queue: Michael Tang \u003ctangm@microsoft.com\u003e\nCr-Commit-Position: refs/heads/main@{#1742446}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 05e80d336f7e0445f0ef322ae49769bb6a71f32b\n"
    },
    {
      "commit": "64ff78eb34dc455f80c636d9bdcc71748874cde7",
      "tree": "d75f7c2f76addc8fe99e69710865ec2937bff03d",
      "parents": [
        "45995ec3711d0c5365f5c1a5be97cde44bd33d33"
      ],
      "author": {
        "name": "Bo Liu",
        "email": "boliu@chromium.org",
        "time": "Thu Nov 12 00:48:53 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 12 01:02:21 2026"
      },
      "message": "android: Support SandboxedGpuProcessService main change\n\nAdd base::Feature and chrome://flag to use SandboxedGpuProcessService.\n\nAdd a new ChildConnectionAllocator for SandboxedGpuProcessService and\nuse it for GPU process when feature is enabled.\n\nDisable ADPF since it will not work inside sandboxed process.\n\nBug: 583357788\nChange-Id: Ie9f3cc2be88ea90e86359e13f958bb673ea2c110\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8758896\nReviewed-by: Kenneth Russell \u003ckbr@chromium.org\u003e\nReviewed-by: Sam Maier \u003csmaier@google.com\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Bo Liu \u003cboliu@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1742386}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1512b693b50ed558c0a1913504a1ecd970cad983\n"
    },
    {
      "commit": "45995ec3711d0c5365f5c1a5be97cde44bd33d33",
      "tree": "eef4dc69ce496154c59384e484d1c2e466134a95",
      "parents": [
        "03d6b96fdd9035a4f9ae04dc1829bf07201c0c9d"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Wed Nov 11 23:34:59 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 23:41:04 2026"
      },
      "message": "[DCSI] Deduplicate Wrapped*CompoundImageRepresentation and Produce*\n\nCompoundImageBacking previously repeated the same representation wrapper\nboilerplate (owned_backing_, wrapped_, compound_backing(),\nSetClearedRect(), OnContextLost(), and NotifyBeginAccess/NotifyEndAccess\nforwarding) across 11 Wrapped*CompoundImageRepresentation subclasses, as\nwell as identical GetOrAllocateBacking() + Produce*() + make_unique\nplumbing across all 11 Produce*() overrides.\n\nIntroduce a common WrappedCompoundImageRepresentation\u003cBaseRep\u003e template\nbase class (with owned_backing_ declared before wrapped_ to preserve\nteardown order) and a private ProduceWrapped\u003cWrappedRep\u003e() helper on\nCompoundImageBacking so each representation wrapper and Produce*()\noverride only specifies its representation-specific access forwarding.\n\nBug: 448962784\nChange-Id: I7240eb6bfbdb9cd15d0a44f7203d784d21bfa72a\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8782775\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1742326}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 02ac197d87ea4111685035d19f8093855f9c0d97\n"
    },
    {
      "commit": "03d6b96fdd9035a4f9ae04dc1829bf07201c0c9d",
      "tree": "c12c0ab4fd682e5b3c66a7eb36a3501e99a92e26",
      "parents": [
        "cf6c23b5b634e55a5c83088669821779a8483f15"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Wed Nov 11 22:18:56 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 22:40:18 2026"
      },
      "message": "[DCSI] Unify CompoundImageBacking ctors and drop WrapExternalBacking\n\nCompoundImageBacking previously maintained two separate constructors and\nan internal WrapExternalBacking() helper left over from when\nSharedImageFactory wrapped standalone GPU backings externally. Also,\nCreateSharedMemory() and the two CreateSharedMemoryForTesting()\noverloads duplicated shared memory backing creation and setup.\n\nSimplify CompoundImageBacking construction:\n- Remove WrapExternalBacking() and unify both constructors into a single\n  CompoundImageBacking constructor taking an initial ElementHolder, with\n  CreateSingleElement() for single-element GPU/mappable backings and\n  AddLazyGpuElement() for optional lazy GPU elements.\n- Deduplicate CreateSharedMemory() and CreateSharedMemoryForTesting()\n  overloads through a shared CreateSharedMemoryInternal() helper.\n- Move internal static helpers (IsValidSharedMemoryFormat,\n  GetGpuSharedImageUsage, RecordBackingTypeUMA) out of the public/class\n  interface into the anonymous namespace in compound_image_backing.cc,\n  and remove the unimplemented SetLatestContent() declaration.\n\nCreateSingleElement() also carries over the wrapped backing\u0027s\narray_layers, which the wrapping constructor dropped, and backings the\ncompound allocates itself now get the same layer count. Current users\nsee no difference: texture-array images are only created from an\nAHardwareBuffer (WebXR on Android) and are cleared from the start, so\nEnsureClear() in the GLES2 passthrough decoder, the only reader of the\ncompound\u0027s layer count, never clears them layer by layer.\n\nBug: 448962784\nChange-Id: Idedc3b4a1430dd500657d9e3b2fb730ed7286ca4\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8783313\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1742235}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: c8320ce426ea42823507b720894779fb2ff8898b\n"
    },
    {
      "commit": "cf6c23b5b634e55a5c83088669821779a8483f15",
      "tree": "9aab305bdf4c4bb687f00fc97fdd40ccd710d11c",
      "parents": [
        "e352ae023c686459119551dbcfcd98bd328839d7"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Wed Nov 11 19:36:31 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 19:40:31 2026"
      },
      "message": "[gpu] Remove max_3d_array_texture_size_1024 workaround\n\nThis GPU driver workaround is only for validating decoder on Linux.\nPassthrough decoder is used on linux by default so this codepath is\ndead and can be removed.\n\nBug: 570597474\nChange-Id: I72aeccea51bc73628514f105e117a61a3c6cb1ef\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8786681\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1742040}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 128c0202cff48420e0b36544c2c3d92f577c04ed\n"
    },
    {
      "commit": "e352ae023c686459119551dbcfcd98bd328839d7",
      "tree": "f63ea3f126a6d0f601356f07c5a9202eb6db67d0",
      "parents": [
        "4ee3b87334a9ead7094ebae8188d2641c35fc37e"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Wed Nov 11 18:43:37 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 18:53:57 2026"
      },
      "message": "[gpu] Remove old Linux Mesa driver workaround\n\nThis workaround is for very old Mesa drivers \u003c10.6 that Chrome does\nnot support running on. Remove this workaround and related code.\n\nBug: 570597474\nChange-Id: Iacf8c78d4be5a15c5c4d55bd5593bf00b4444beb\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8778097\nReviewed-by: Kramer Ge \u003cfangzhoug@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741971}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e7c1502f8621180d0bb3ecec7048ce35181cc70d\n"
    },
    {
      "commit": "4ee3b87334a9ead7094ebae8188d2641c35fc37e",
      "tree": "8ea6f2217fb6baddf6d5feac0d0c498efeda2c59",
      "parents": [
        "946996b65ac67635d280ab6f7793c67273d06a93"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Wed Nov 11 18:06:36 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 18:14:06 2026"
      },
      "message": "gpu: Replace memcmp() of test commands with CommandsMatch()\n\nThe client implementation unittests compared the expected commands\nagainst the command buffer using memcmp() on a raw pointer, wrapped in\nUNSAFE_TODO. This pattern is emitted ~220 times by the command buffer\ncode generator, making gles2_implementation_unittest_autogen.h the file\nwith the most UNSAFE_TODO in Chromium.\n\nStore the commands as a span over the ring buffer and add a\nCommandsMatch() helper to the test fixtures, doing a bounds-checked\nbyte comparison. Update the generator and the hand-written tests to use\nit, then regenerate.\n\nNo behavior change, test-only.\n\nChange-Id: Ib993cc29d44b275d504265dad065dfed87198cf7\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8783474\nCommit-Queue: Brandon Jones \u003cbajones@chromium.org\u003e\nReviewed-by: Brandon Jones \u003cbajones@chromium.org\u003e\nAuto-Submit: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741912}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7e14e12fb11480408bc29bebc43945ea0f6338bb\n"
    },
    {
      "commit": "946996b65ac67635d280ab6f7793c67273d06a93",
      "tree": "b9334eafb366f3ab9738bda5e9d4c69afe932151",
      "parents": [
        "94f3e42cdd3e640d260f40793120ca7ec914bcc5"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Wed Nov 11 16:31:41 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 16:37:28 2026"
      },
      "message": "[gpu] Reject R8/R16F formats based on usage from clients\n\nR16F and R8 formats can only be used by CameraBufferFactory for video\ncapture on ChromeOS (MJPEG format) and for WebNN ML context. This\nchange checks for specific usages and rejects CreateSharedImage calls\nthat don\u0027t provide appropriate usage, this helps validate\nsingle-channel format usage by clients.\n\nBug: 416285370\nChange-Id: I4638e44557ba965a5309d30ae4bb96d7eb03f262\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8776953\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741827}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a340348d971a6070f80db29337b2f4c9c981241e\n"
    },
    {
      "commit": "94f3e42cdd3e640d260f40793120ca7ec914bcc5",
      "tree": "24c07de89d46c46f7cce190b8ddfdfb43dee890b",
      "parents": [
        "ba80b49ba7b307a7650c649a89a17ee96d0f47f8"
      ],
      "author": {
        "name": "Qiu Jianlin",
        "email": "jianlin.qiu@intel.com",
        "time": "Wed Nov 11 08:54:28 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 09:00:35 2026"
      },
      "message": "gpu: Support Y210 shared images and sample them via Skia YUVA\n\nAdds Y210 (packed 10-bit 4:2:2) as a single-plane shared image: the\npixel-format plumbing, D3D11 backing support (sample-only, with a\ndriver probe for the aliased-view SRV), and Skia YUVA sampling via\nSK_ENABLE_YUVA_PACKED_422 with the kYUYV plane config over the\nhalf-width R16G16B16A16 view, exposed at half width in the GL\nrepresentation.\n\nNothing produces Y210 frames or shared images yet, so the NOTREACHED()\nguards added for Y210 on the CPU conversion, software-upload and\ncapture paths (and the kYUYV cases in the Skia CPU YUVA helpers) are\nunreachable today. Follow-ups make the format producible and fill the\nremaining gaps: zero-copy decode for the D3D11 video decoder\n(https://crrev.com/c/8578317), CPU blit support for packed 4:2:2\n(https://crrev.com/c/8677020), CPU paint fallbacks (libyuv Y210-to-\nARGB), Graphite support, and GPU conversion in PaintCanvasVideoRenderer.\n\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nBug: 537818862\nChange-Id: I4070b2fce5e3b7e81bd9b8abf64df224d7fc5e5d\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8578277\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Qiu, Jianlin \u003cjianlin.qiu@intel.com\u003e\nCr-Commit-Position: refs/heads/main@{#1741566}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 20e693f921ae352a3cd184372d0cf910d72f2518\n"
    },
    {
      "commit": "ba80b49ba7b307a7650c649a89a17ee96d0f47f8",
      "tree": "f15efa9b593d101f730fa771c8d68d9355cfecdc",
      "parents": [
        "663551e0b77cca6989f1e90680d5622b04f28a8c"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Wed Nov 11 05:05:23 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 05:10:58 2026"
      },
      "message": "gpu: Fix ExternalVk release layout for images without color attachment\n\nExternalVkImageBacking::Create() only adds\nVK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT for write usages, so read-only\nimages lack it, e.g. GLES2_READ images when\nGL_ANGLE_memory_object_flags is supported (since 42425e09bfbb) and\nWEBGPU_READ-only images. However, BeginAccess() and UploadToVkImage()\nrelease VkImages to the external queue family for GL and Dawn access in\nVK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL, and GetVkImageLayoutsForGL()\nreports GL_LAYOUT_COLOR_ATTACHMENT_EXT to glWaitSemaphoreEXT() and\nglSignalSemaphoreEXT(). Skia and ANGLE then record barriers to or from\nthat layout on images without color attachment usage, violating\nVUID-VkImageMemoryBarrier-oldLayout-01208.\n\nAdd GetLayoutForExternalAccess(), which returns\nVK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL for such images\n(VK_IMAGE_USAGE_SAMPLED_BIT is always set), and use it in BeginAccess()\nand UploadToVkImage(). The USE_WEBGPU_ON_VULKAN_VIA_GL_INTEROP path of\nBeginAccess() transitioned images to\nVK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL but recorded\nVK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL; it now transitions to and\nrecords the layout from GetLayoutForExternalAccess(). Also only chain\nVkProtectedSubmitInfo in VulkanCommandBuffer::Submit() when\nallow_protected_memory is true (omitting it is equivalent to VK_FALSE\nper the Vulkan spec, and SwiftShader aborts when it is chained).\n\nDawn now also receives VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL for\nimages without color attachment usage, so\nExternalVkImageDawnImageRepresentation no longer adds RenderAttachment\ninternal usage for them.\n\nAdd unit tests covering all three release paths (Ganesh Vulkan\nBeginAccess(), WebGPU-via-GL-interop BeginAccess(), and\nUploadToVkImage() for both GL and Dawn). Filter the new tests on the\nLinux TSan SwiftShader bot (crbug.com/462477487, crbug.com/462477489).\n\nBug: 583941980\nChange-Id: Iec6d52c909e4bf6692b8e3f3e52d4e6273b6d67a\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8764978\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741441}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e17bb5187f64db43b378b7f9a6f8036e6309b5ff\n"
    },
    {
      "commit": "663551e0b77cca6989f1e90680d5622b04f28a8c",
      "tree": "6cf27ba6a7975e96e3efe989e4e05f3584e1a185",
      "parents": [
        "24fc9d0acac548fe8283614db435e69d668bc5cd"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Wed Nov 11 03:33:51 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 03:40:57 2026"
      },
      "message": "gpu: Enable texture storage for float and half-float RG formats\n\nWhen initializing feature validators with GL_EXT_texture_rg, FeatureInfo\npreviously omitted GL_R32F_EXT, GL_RG32F_EXT, GL_R16F_EXT, and\nGL_RG16F_EXT from texture_internal_format_storage even when the\ncorresponding float and half-float pixel types were supported. In\naddition, GL_RED_EXT was erroneously added to texture_internal_format\nduring ext_texture_norm16 setup.\n\nBecause SharedContextState initializes FeatureInfo with an ES2 context\ntype without calling EnableES3Validators(), GLCommonImageBackingFactory\ncomputed supports_storage as false for SinglePlaneFormat::kR_F16 (as\nwell as kRGBA_F16 and kR_F16 on ES3 contexts without\nGL_OES_texture_half_float) and fell back to glTexImage2D instead of\nglTexStorage2DEXT even when ext_texture_storage was enabled. This also\nprevented TexImageTexStorageEquivalence tests from asserting that\nTexImage2D and TexStorage2D support are equivalent across all\nuncompressed single-plane formats.\n\nThis CL:\n- Adds GL_R32F_EXT/GL_RG32F_EXT and GL_R16F_EXT/GL_RG16F_EXT to\n  texture_internal_format_storage in FeatureInfo::InitializeFeatures()\n  when ext_texture_rg and the corresponding pixel types are supported,\n  and removes the extraneous GL_RED_EXT from ext_texture_norm16.\n- Overrides GL_RED_EXT to GL_R16F in\n  GLFormatCaps::ToGLFormatDescOverrideHalfFloatType() when GL_HALF_FLOAT\n  is used on ES3 without GL_OES_texture_half_float.\n- Updates GLCommonImageBackingFactory to check ES3 half-float support\n  for GL_R16F and GL_RGBA16F when determining storage support.\n- Tightens TexImageTexStorageEquivalence tests to assert\n  supports_tex_image \u003d\u003d supports_tex_storage.\n\nBug: None\nChange-Id: Ia3df7832ac48598626be173320ca318632a80dd9\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8770898\nReviewed-by: vikas soni \u003cvikassoni@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741390}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8fced1fa1109a0346c3b8ac390b3cf7ae2de82fc\n"
    },
    {
      "commit": "24fc9d0acac548fe8283614db435e69d668bc5cd",
      "tree": "1298f0222f7070f923ff7d4cff8dcbc2c20fc8b9",
      "parents": [
        "ea90b337f76a086d88b09c70635d9bcdb1588dee"
      ],
      "author": {
        "name": "Jordan Bayles",
        "email": "jophba@chromium.org",
        "time": "Wed Nov 11 03:12:49 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 11 03:19:33 2026"
      },
      "message": "[gpu, content] Fix SchedulingPriority::kLow mojom and finch worker priority\n\nEnumTraits\u003cgpu::mojom::SchedulingPriority\u003e::ToMojom previously mapped\ngpu::SchedulingPriority::kLow to gpu::mojom::SchedulingPriority::kNormal.\nAs a result, any command buffer created by a client with kLow priority\n(such as kGpuStreamPriorityWorker) was deserialized in the GPU service\nas kNormal for the past several years.\n\nFix EnumTraits\u003cgpu::mojom::SchedulingPriority\u003e::ToMojom to map kLow to\ngpu::mojom::SchedulingPriority::kLow. Because the GPU scheduler has been\ntuned with kGpuStreamPriorityWorker effectively running at kNormal, keep\nkGpuStreamPriorityWorker at kNormal by default and add a default-disabled\nfeature (kLowPriorityGpuWorkerStream, with fieldtrial testing config) to\nroll out kLow priority for worker GPU streams via Finch.\n\nBug: 40287938, b/329260636\nChange-Id: Ic39e00159ab5fc02eaba50e1be8d21c7b9b4763b\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8674479\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Dave Tapuska \u003cdtapuska@chromium.org\u003e\nCommit-Queue: Jordan Bayles \u003cjophba@chromium.org\u003e\nReviewed-by: Fred Shih \u003cffred@chromium.org\u003e\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741371}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7bcbddea45da1189d81ff2b1873cec5851ecde97\n"
    },
    {
      "commit": "ea90b337f76a086d88b09c70635d9bcdb1588dee",
      "tree": "3352f5344df07b4087964f897338d33cae593ca7",
      "parents": [
        "9e7b0d13bf4f1a6a6199b08aff5d0513f510302b"
      ],
      "author": {
        "name": "Liza Burakova",
        "email": "liza@chromium.org",
        "time": "Tue Nov 10 23:34:21 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 10 23:40:26 2026"
      },
      "message": "Fix unshared Skia Graphite recorder cache limits (96/64 MB)\n\nA comment in GetDefaultGraphiteContextOptions() states that it should\nuse the 96MB default for the resource cache, however currently\nthat only happens on machines with \u003c 4GB RAM.\n\nThis change updates the limits so that Graphite Context now has a\nmax of 96MB and the compositor recorder has a limit of 64MB.\n\nOn local testing on MacOS, this improves Chrome\u0027s memory usage by\n~2%.\n\nBug: 576572140\nChange-Id: I97ad7a36db56c8e33b8420714e556dd75b43f2d3\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8643411\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCommit-Queue: Liza Burakova \u003cliza@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1741111}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 060ae673200560754f2c533a23298c7b7d7c00fe\n"
    },
    {
      "commit": "9e7b0d13bf4f1a6a6199b08aff5d0513f510302b",
      "tree": "0bf94c380ab4536f52861f5c6b9483ca1febb257",
      "parents": [
        "ba916b88530bc39a7d4a8d4641948f257806618e"
      ],
      "author": {
        "name": "Christopher Cameron",
        "email": "ccameron@chromium.org",
        "time": "Tue Nov 10 16:29:44 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 10 16:42:32 2026"
      },
      "message": "RasterInterface: Use SkPixmap for ARGB and image readback\n\nChange RasterInterface::ReadbackARGBPixelsAsync and\nRasterInterface::ReadbackImagePixels to take a destination SkPixmap,\nrather than a separate SkImageInfo, row bytes, and pointer (or span).\n\nUpdate the implementations (which unpack the SkPixmap and are otherwise\nunchanged) and callers. Use skia::as_writable_byte_span to get\ndestination spans in RasterImplementation, and skia::MakeSkPixmap in\ncallers that have a span. Where the destination is known to be large\nenough, crash if it is not. This changes no behavior.\n\nUse media::VideoFrame::GetWritableVisibleRGBPlanePixmap in\nBackgroundReadback and StaticBitmapImageToVideoFrameCopier (removing\nGetImageInfoForFrame).\n\nChange-Id: I733b177853c0729c62c62526d5b2eb51f4a2ed46\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8761932\nReviewed-by: Khushal Sagar \u003ckhushalsagar@chromium.org\u003e\nReviewed-by: Colin Blundell \u003cblundell@chromium.org\u003e\nCommit-Queue: ccameron chromium \u003cccameron@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1740577}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 56406bd932a7975fab5f97e625552e652d8a88cf\n"
    },
    {
      "commit": "ba916b88530bc39a7d4a8d4641948f257806618e",
      "tree": "91a90a3769e7348734a8741d9ecd7b25be9e2a12",
      "parents": [
        "2ae7763ce7e915ee9419fdc93f9a6abceef6b11b"
      ],
      "author": {
        "name": "Wez",
        "email": "wez@chromium.org",
        "time": "Tue Nov 10 15:46:58 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 10 15:52:36 2026"
      },
      "message": "[fuchsia] Expand //build/fuchsia/OWNERS coverage across feature subtrees\n\nEnsure Fuchsia-specific files and subdirectories across feature and\nplatform subtrees (//base, //chromecast, //components, //device, //gpu,\n//ipc, //media, //net, //services, //skia, //sql,\n//third_party/blink/renderer, and //ui) are covered by\n//build/fuchsia/OWNERS, and replace hardcoded individual owners in\n//media/fuchsia/OWNERS with a reference to //build/fuchsia/OWNERS.\n\nChange-Id: I2037d8a7fa02c2d44646888b029a7224ffefbf53\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8776191\nCommit-Queue: Wez \u003cwez@chromium.org\u003e\nOwners-Override: Wez \u003cwez@chromium.org\u003e\nReviewed-by: Greg Thompson \u003cgrt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1740532}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d33fdaaff804d4967a3df07d78f51b1514b25536\n"
    },
    {
      "commit": "2ae7763ce7e915ee9419fdc93f9a6abceef6b11b",
      "tree": "632c6b0142849ef9dd07c3ad409c03fc2387000c",
      "parents": [
        "1414f7796e220eef5cb70de6c5ffffd88f025dc4"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Tue Nov 10 15:06:24 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 10 15:13:51 2026"
      },
      "message": "[gpu] Remove Vivante driver workarounds\n\nGL_VIV_shader_binary extension [1] and Vivante GC1000 renderer string\nare only supported on GPUs that do not support GLES 3.0 and above.\nThese entries can be removed.\n\nAlso remove entry 313 as it is a duplicate of 370 which is less\nrestrictive.\n\n[1] https://opengles.gpuinfo.org/listreports.php?extension\u003dGL_VIV_shader_binary\n\nBug: 570597474\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: I129a276c9d42841ed1ffbc5583f2b3f59f099a05\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8772459\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1740486}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 19c312eb8ad172ec2bed14e83997e20ff806c769\n"
    },
    {
      "commit": "1414f7796e220eef5cb70de6c5ffffd88f025dc4",
      "tree": "eda7460f9cf858b43d1c553ae7e77607f9a294fc",
      "parents": [
        "ddd21b0bf6641e69d5372c68e4e6b2ab49f29128"
      ],
      "author": {
        "name": "Christopher Cameron",
        "email": "ccameron@chromium.org",
        "time": "Tue Nov 10 04:50:39 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 10 04:58:25 2026"
      },
      "message": "RasterInterface: Use SkYUVAPixmaps for YUV readback\n\nChange RasterInterface::ReadbackYUVPixelsAsync to take a destination\nSkYUVAPixmaps and SkColorSpace, rather than an output rect and per-plane\nstrides and spans. Fail if the destination is not 8-bit Y_U_V 4:2:0.\nThe SkYUVColorSpace and SkColorSpace are not yet sent to the GPU process\n(this is a TODO), so this does not change behavior.\n\nIn RasterImplementation, compute the shared memory layout for the planes\nin a loop, in anticipation of supporting more than I420. On all\nearly-out paths, post `release_mailbox` and `readback_done` (with\nfailure), rather than dropping them.\n\nCallers use media::VideoFrame::GetWritableVisibleSkYUVAPixmaps, and\npass the destination frame\u0027s color space as the SkColorSpace.\n\nBug: 40261346\nChange-Id: I6af77436516c0c52ddc6fc9c5172aafb027f4da5\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8769598\nReviewed-by: Florin Malita \u003cfmalita@chromium.org\u003e\nCommit-Queue: ccameron chromium \u003cccameron@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1740077}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5cf39c0da2a32c1128bb565e0f9a1a705512ba9b\n"
    },
    {
      "commit": "ddd21b0bf6641e69d5372c68e4e6b2ab49f29128",
      "tree": "ceb12a8793bb157c15f058890f31a4438ca49442",
      "parents": [
        "5f9154f26b69c13973304e211b75cf14addfac2a"
      ],
      "author": {
        "name": "Eugene Zemtsov",
        "email": "eugene@chromium.org",
        "time": "Tue Nov 10 01:59:39 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 10 02:05:55 2026"
      },
      "message": "Fix staging texture caching in CopyNativeBufferToSharedMemoryAsync\n\nGpuMemoryBufferHandleSharedState::staging_texture() and\nGetStagingTextureForGMBHandleCopies() returned ComPtr\u003cID3D11Texture2D\u003e\nby value, and CopyNativeBufferToSharedMemoryAsync() passed\n\u0026GetStagingTextureForGMBHandleCopies() (the address of a temporary via\nWRL\u0027s ComPtr::operator\u0026()) to CopyDXGIBufferToShMem(). As a result, a\nnew D3D11 staging texture was allocated and destroyed on every DXGI GMB\nreadback instead of reusing staging_texture_.\n\nReturn the cached ComPtr by reference, reset staging_texture_ alongside\nd3d11_device_ on device removal, and change CopyDXGIBufferToShMem() and\nCopyD3D11TexToMem() to take ComPtr\u003cID3D11Texture2D\u003e\u0026 so temporaries\ncannot bind at compile time.\n\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nBug: 581949593\nChange-Id: Iae203f93d21c0c934d978fb458fdde34fc9a53b0\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8760833\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nReviewed-by: Nasko Oskov \u003cnasko@chromium.org\u003e\nCommit-Queue: Eugene Zemtsov \u003ceugene@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1739956}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b332416260bf081e3d19e35006d8565d13a758e7\n"
    },
    {
      "commit": "5f9154f26b69c13973304e211b75cf14addfac2a",
      "tree": "345fc84eee64b4b952dace0ccd8136df11aaff60",
      "parents": [
        "695e688edf2775f9405ece84e130d4dbb212d109"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Mon Nov 09 20:09:49 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Nov 09 20:24:11 2026"
      },
      "message": "[DCSI] Decouple SharedMemoryImageBacking from CompoundImageBacking\n\nRemove SharedMemoryImageBacking-specific includes, downcasts, helper\nmethods, and state from CompoundImageBacking so CPU shared memory and\nGPU backings are synchronized uniformly via SharedImageCopyManager:\n\n- Route CopyToGpuMemoryBuffer() and CopyToGpuMemoryBufferAsync() from\n  GetElementWithLatestContent(/*exclude_element\u003d*/\u0026shm_element) into\n  shm_element (elements_[0], checked for kSharedMemory) using\n  copy_manager_-\u003eCopyImage() and CopyImageAsync(), releasing AutoLock\n  before invoking CopyImageAsync() or callbacks.\n- Delete GetSharedMemoryPixmaps(), GetShmElement(), GetGpuBacking(),\n  has_shm_backing_, and #include \"shared_memory_image_backing.h\" from\n  compound_image_backing.{h,cc}.\n- Drop the redundant CompoundImageBacking::ClearedRect() override in\n  favor of ClearTrackingSharedImageBacking::ClearedRect() (using\n  ClearedRectInternal() in NotifyEndAccess()).\n\nTAG\u003dagy\nCONV\u003dbd0df832-42e7-41af-992e-0392c5b240dc\n\nBug: 448962784,434215964\nChange-Id: I2b8ef8512c094932b114d628a9f8341c0d82e102\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8737909\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1739487}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f59e8768ba7423f11eea5e9f298c2d9732586f9a\n"
    },
    {
      "commit": "695e688edf2775f9405ece84e130d4dbb212d109",
      "tree": "75792a38b1d2671dc2ec82d94e6320f89e476658",
      "parents": [
        "313fcb0ddfabdaa74d1ed348288d085658a6852f"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Mon Nov 09 18:34:47 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Nov 09 18:40:28 2026"
      },
      "message": "[gpu] Check GrContext flush success instead of semaphores\n\nGrDirectContext flush returns a FlushResult that separates success\nfrom GrSemaphoresSubmitted now, which was not happening earlier.\nThis change updates callsites to check for fSuccess properly instead\nof implicitly checking the semaphore result.\n\nBug: 568937157\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: I4179ac2e83f190e865caa781a66e91889d57e740\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8738496\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1739361}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2303b14cb3bf6f3ab3720fe347460b3ed69c5b6a\n"
    },
    {
      "commit": "313fcb0ddfabdaa74d1ed348288d085658a6852f",
      "tree": "0138e2f37dde98124b90bc67d326d4ebd2202a3d",
      "parents": [
        "309301574c63c72a5da23abd424a3d95a6f88761"
      ],
      "author": {
        "name": "Jessica Huang",
        "email": "jessing@google.com",
        "time": "Mon Nov 09 18:18:22 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Nov 09 18:26:08 2026"
      },
      "message": "Remove dead CollectData() from VertexArrayObjectManager\n\nCollectData() has had no callers since client side vertex array support\nwas removed in crrev.com/c/6860267. Delete it along with\ncollection_buffer_ and collection_buffer_size_, which only it used.\nThis removes a std::unique_ptr\u003cint8_t[]\u003e with a separately tracked size\nand four UNSAFE_TODO()s.\n\nNo behavior change.\n\nBug: 40285824\nTest: tools/autotest.py --quiet --run-all -C out/Release gpu/command_buffer/client/vertex_array_object_manager_unittest.cc\nTest: out/Release/gpu_unittests --gtest_filter\u003d\u0027GLES2Implementation*:VertexArrayObjectManager*\u0027\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: I20ec0e6757c952c7ad831748dd535630604d2bfa\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8763065\nCommit-Queue: Jessica Huang \u003cjessing@google.com\u003e\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1739335}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7663cce533a5048f36a8389babc6083aec31d0d1\n"
    },
    {
      "commit": "309301574c63c72a5da23abd424a3d95a6f88761",
      "tree": "28da02db562f37138b7e5a638edf0c50139bbde6",
      "parents": [
        "3f687b3739fb5aaab3478e4c9cdde8ba967bc1c5"
      ],
      "author": {
        "name": "Jonathan Ross",
        "email": "jonross@chromium.org",
        "time": "Mon Nov 09 16:59:41 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Nov 09 17:08:16 2026"
      },
      "message": "gpu: Fix cleared state tracking in CompoundImageBacking\n\nA CompoundImageBacking created from shared memory starts out cleared,\nbut its lazily allocated GPU backing does not. Copying the compound\ncleared state onto that new backing let the GLES2 passthrough decoder\nbind it as cleared before any upload.\n\n- Only copy the cleared state to a new GPU backing if the latest\n  content is already on the GPU.\n- Add IsBackingCleared() and use it in EnsureClear(), which now fails\n  closed: DoCreateAndTexStorage2DSharedImageINTERNAL stops instead of\n  binding an uninitialized texture.\n- Validate readback parameters before ProduceSkia(), so invalid\n  readbacks don\u0027t allocate a GPU backing.\n- D3D: mark backings cleared after UploadFromMemory(), and wait until a\n  multiplanar texture is initialized before beginning Dawn access.\n\nBug: 573265737, 577359426\nChange-Id: Iaf2e5071377ee9e42b77094543bab75a12c736d1\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8641617\nReviewed-by: vikas soni \u003cvikassoni@chromium.org\u003e\nCommit-Queue: Jonathan Ross \u003cjonross@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1739232}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a5f0464fb44250bafdece90b3e84dba06acd470e\n"
    },
    {
      "commit": "3f687b3739fb5aaab3478e4c9cdde8ba967bc1c5",
      "tree": "7b96a226f621b2b22e72f5d86e2906ba33075cb7",
      "parents": [
        "55e72ef18de6919fc7db28b766e1e30f653516f1"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Mon Nov 09 15:47:12 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Nov 09 15:52:52 2026"
      },
      "message": "[gpu] Clean up driver bug workarounds\n\nThis change removes workaround entries and related code for entries\nthat were earlier removed [1][2]. These entries were removed because\nthey are older GPUs that do not support GLES 3.0 and above. This CL\nremoves disable_gl_rgb_format, etc1_power_of_two_only,\nuse_non_zero_size_for_client_side_stream_buffers and\nuse_client_side_arrays_for_stream_buffers workarounds and related\ncode from decoders.\n\n[1] https://crrev.com/c/8688080\n[2] https://crrev.com/c/8687563\n\nBug: 570597474\nChange-Id: Iabb911f4ac6aa33ea0b6a1b43186faef78314d05\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8749141\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Dave Tapuska \u003cdtapuska@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1739158}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: fa30d68a835c0b86da46f760c3fa3a2175198a25\n"
    },
    {
      "commit": "55e72ef18de6919fc7db28b766e1e30f653516f1",
      "tree": "68f73714dce4be08f8fcc5372e7f82feaa523721",
      "parents": [
        "618dcbb5e97c47fa0f1982c4d8677dac63f775d8"
      ],
      "author": {
        "name": "Mikihito Matsuura",
        "email": "mikt@google.com",
        "time": "Sat Nov 07 07:41:11 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 07:48:38 2026"
      },
      "message": "[gpu] Exclude ephemeral OOP-R PaintOp rasterization from SLQ\n\nExclude per-PaintOp playback and destruction in\nRasterDecoderImpl::DoRasterCHROMIUM (after PaintOp::Deserialize) and\nSkia surface flush/teardown in RasterDecoderImpl::DoEndRasterCHROMIUM\nfrom PartitionAlloc Scheduler-Loop Quarantine safety checks in the GPU\nprocess, gated behind the ExcludeOopRasterizationFromSafetyChecks\nfeature flag (disabled by default).\n\nBug: 493702418\nChange-Id: Iabddfa1353c120587dcb1280ffa3dbc804caabf2\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8713820\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCommit-Queue: Mikihito Matsuura \u003cmikt@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1738397}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ca4b9564a3a10be1669f946a1f36d73f5dec64d9\n"
    },
    {
      "commit": "618dcbb5e97c47fa0f1982c4d8677dac63f775d8",
      "tree": "d4cb04175c8299cecebc42debeb7c43fcf86e6c0",
      "parents": [
        "be20c6294d6267532be000bd21780f8efcc4e0a2"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Sat Nov 07 05:03:28 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 05:10:05 2026"
      },
      "message": "[DCSI] Add CopyImageAsync to SharedImageCopyManager\n\nAdd SharedImageCopyStrategy::CopyAsync() and\nSharedImageCopyManager::CopyImageAsync() so callers such as\nCompoundImageBacking::CopyToGpuMemoryBufferAsync() can perform\nasynchronous GPU-to-SharedMemory readbacks via SharedImageCopyManager\nwithout downcasting to SharedMemoryImageBacking:\n\n- SharedImageCopyStrategy::CopyAsync() defaults to invoking Copy()\n  synchronously and running the completion callback.\n- CPUReadbackUploadCopyStrategy::CopyAsync() delegates non-SharedMemory\n  to kSharedMemory copies directly to\n  SharedImageBacking::ReadbackToMemoryAsync() into\n  SharedMemoryImageBacking::pixmaps().\n- SharedImageCopyManager::CopyImageAsync() dispatches to the first\n  strategy supporting the backing pair, or invokes the callback with\n  false if none matches.\n\nBug: 448962784,434215964\nChange-Id: I1a7ef8512c094932b114d628a9f8341c0d82e101\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8745981\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1738348}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e72d336b09f61d197b626c2187d92ff27939b2cc\n"
    },
    {
      "commit": "be20c6294d6267532be000bd21780f8efcc4e0a2",
      "tree": "eddcbc723a4f9c9dfb8ac0f0b73b2cbb58ced35a",
      "parents": [
        "f13cb343e68ae795a168f3393271f4b1bba78ce3"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Sat Nov 07 02:27:44 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 02:33:49 2026"
      },
      "message": "gpu: Reject uncleared SharedImage import when EnsureClear fails\n\nIn GLES2DecoderPassthroughImpl::DoCreateAndTexStorage2DSharedImageINTERNAL,\nPassthroughResources::SharedImageData::EnsureClear attempts to lazily\nclear an uncleared SharedImage by attaching its texture to a temporary\nFBO and clearing it.\n\nPreviously, if EnsureClear failed (for example, when importing an\nuncleared GL_RGBA16F SharedImage into a WebGL2 context where\nGL_EXT_color_buffer_float has not been enabled, or when BeginScopedAccess\nfails or raises a GL error), DoCreateAndTexStorage2DSharedImageINTERNAL\nstill bound the uncleared service texture ID to the client texture ID.\nAdditionally, EnsureClear called glClear before checking\nglCheckFramebufferStatusEXT, emitting a spurious\nGL_INVALID_FRAMEBUFFER_OPERATION error, and called FlushErrors before\nrestoring temporary GL state and ending access.\n\nThis CL:\n1. Checks glCheckFramebufferStatusEXT(GL_FRAMEBUFFER) \u003d\u003d\n   GL_FRAMEBUFFER_COMPLETE before calling glClear in EnsureClear, and\n   scopes the temporary FBO/state resets, ScopedAccess, and\n   ScopedPixelLocalStorageInterrupt so they complete before FlushErrors.\n2. In DoCreateAndTexStorage2DSharedImageINTERNAL, checks that\n   shared_image_data.representation()-\u003eIsCleared() is true after\n   SharedImageData construction; if not, creates a placeholder texture\n   with DoGenTextures, inserts GL_INVALID_OPERATION, and does not map\n   the uncleared SharedImage texture.\n\nBug: 577339422\nChange-Id: Iee522c5adc1e55aed61b47fd7ca5f5963b7c0ade\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8758491\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Kenneth Russell \u003ckbr@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1738301}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 58a22a83bc625d168a4daea64f20a489aced294b\n"
    },
    {
      "commit": "f13cb343e68ae795a168f3393271f4b1bba78ce3",
      "tree": "d75bd1dc1f1b63c062c50375c1c7cb71d753de9c",
      "parents": [
        "9aa237bb7da94b77b000ec56b46bebde5cab3f05"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Sat Nov 07 01:58:37 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 02:10:47 2026"
      },
      "message": "Validate allocated size in ScopedSharedMemoryPtr and transfer cache\n\nWhen allocating memory via ScopedTransferBufferPtr,\nTransferBuffer::AllocUpTo() can return fewer bytes than requested if\nShrinkOrExpandRingBufferIfNecessary() shrinks the ring buffer or if\navailable space is limited. ScopedSharedMemoryPtr and\nClientTransferCache::MapTransferBufferEntry() previously checked\nvalid(), which only checks for a non-null buffer pointer, without\nverifying that size() meets the requested byte length.\n\nUpdate ScopedSharedMemoryPtr to reject zero-size or overflowing\nrequests, verify that the allocated transfer buffer meets the requested\nsize, call Discard() and fall back to MappedMemoryManager when the\nallocation is smaller than requested, and verify mapped memory\nallocation sizes.\n\nUpdate ClientTransferCache::MapTransferBufferEntry() and MapEntry() to\nreject zero-size requests and short allocations, and update\nRasterImplementation::MapTransferCacheEntry() to fall back to MapEntry()\nwhen MapTransferBufferEntry() returns an empty span.\n\nValidate byte size and row stride arithmetic against 32-bit overflow in\nWritePixels(), WritePixelsYUV(), and MappedMemoryManager::Alloc(), and\nwiden bytes_since_last_shrink_ and the shrink threshold calculation in\nTransferBuffer to 64 bits.\n\nBug: 575525279\nChange-Id: Ibaea98ac0827a7abf71be63e52242c47b5c82ad3\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8759371\nReviewed-by: vikas soni \u003cvikassoni@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1738284}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b385c7d8dd84f67d76ce1bf8904d234b94e138b6\n"
    },
    {
      "commit": "9aa237bb7da94b77b000ec56b46bebde5cab3f05",
      "tree": "17401c378a79574533d03cd9a57d224480426d0e",
      "parents": [
        "4c16b501d773bffc0b8a8a38923a36a38d1a1115"
      ],
      "author": {
        "name": "Ken Russell",
        "email": "kbr@chromium.org",
        "time": "Sat Nov 07 01:50:04 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 01:56:47 2026"
      },
      "message": "gpu: change name of limit-varying shader translator workaround.\n\nIt now applies to both vertex and fragment shaders. Also remove the\nspecific number from the name of the workaround, as later bugs\nindicate it will be necessary to change the limit again.\n\nBug: 569367111\nChange-Id: Ib047bc03730bb96c4832ed9f3d7b739bad6310c1\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8762174\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Kenneth Russell \u003ckbr@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1738276}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 4c013d7b9d29e9f37ff199079a8e81aafb265683\n"
    },
    {
      "commit": "4c16b501d773bffc0b8a8a38923a36a38d1a1115",
      "tree": "6ca730019a3c19be05d4a8bc9bb079dedfed2871",
      "parents": [
        "2940d832543e837c70dea5f2abf4884916ecb245"
      ],
      "author": {
        "name": "Nico Weber",
        "email": "thakis@chromium.org",
        "time": "Sat Nov 07 00:24:37 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 00:31:31 2026"
      },
      "message": "gpu: Remove a few unused variables\n\nNew libc++ has a mode that warns on these, and this is part of\nenabling that mode.\n\nNo behavior change.\n\nBug: 581374525\nChange-Id: I3f32809adfd380016aafd2394b08369b531ea2dc\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8746865\nAuto-Submit: Nico Weber \u003cthakis@chromium.org\u003e\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Nico Weber \u003cthakis@chromium.org\u003e\nOwners-Override: Nico Weber \u003cthakis@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1738223}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e9c6174fe690a22ba21c2f9c75ca180bbb0f564b\n"
    },
    {
      "commit": "2940d832543e837c70dea5f2abf4884916ecb245",
      "tree": "c5ce1d22cd7317aca6355d388ca4bffc61a1a15e",
      "parents": [
        "f74707ecd84c46f65fc9c627818e0c6fd2420662"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Fri Nov 06 23:48:00 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Nov 07 00:02:13 2026"
      },
      "message": "Add GLES2Util::GetTextureTargetForSamplerType()\n\nAdd a shared helper mapping a sampler uniform type to the texture target\nit samples, returning GL_NONE for non-sampler types. It covers the ES3.0\nsampler types plus the ES3.1/3.2 multisample, multisample array, cube\nmap array and buffer sampler types, and the external and rectangle\nsampler types.\n\nUse it in GLES2DecoderImpl::GetBindTargetForSamplerType(), which keeps\nits NOTREACHED() for types the validating decoder doesn\u0027t support. A\nfollow-up will use it in the passthrough decoder.\n\nBug: 582302430\nChange-Id: Ia447ac17706810b365e2ca3536eba4f58835adb5\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8750118\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Kenneth Russell \u003ckbr@chromium.org\u003e\nAuto-Submit: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1738185}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f0b4fecef37b772706ef17c0bf9bd1899d99fdb1\n"
    },
    {
      "commit": "f74707ecd84c46f65fc9c627818e0c6fd2420662",
      "tree": "a54fa4d5c94f60924a8091931ef9f397be2dfb24",
      "parents": [
        "e74cfa34a4066df3a000ce352d6baf7e1ec057c7"
      ],
      "author": {
        "name": "Patrick Monette",
        "email": "pmonette@chromium.org",
        "time": "Fri Nov 06 18:39:50 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 18:45:11 2026"
      },
      "message": "Remove unneeded MemoryPressureListenerRegistry from tests\n\nThese tests don\u0027t depend on memory pressure notifications.\n\nBug: 406578344\nChange-Id: Ia332f0de40092960b29a33c17f2674e7f1b594ef\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8757492\nOwners-Override: Francois Pierre Doray \u003cfdoray@chromium.org\u003e\nCommit-Queue: Patrick Monette \u003cpmonette@chromium.org\u003e\nReviewed-by: Francois Pierre Doray \u003cfdoray@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737822}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 27cac8fef16505965e5beba83948edd74cb50d5d\n"
    },
    {
      "commit": "e74cfa34a4066df3a000ce352d6baf7e1ec057c7",
      "tree": "597ec44651c11723a14650cc9c81ad82b931a327",
      "parents": [
        "0c8de148b09778155322e7444c737dd00ce1518b"
      ],
      "author": {
        "name": "Arthur Milchior",
        "email": "arthurmilchior@google.com",
        "time": "Fri Nov 06 17:21:48 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 17:26:36 2026"
      },
      "message": "[GPU] Re-enable `GLCubeMapTextureTest` pixel readback tests\n\n`GLCubeMapTextureTest.ReadPixels` and\n`GLCubeMapTextureTest.ReadPixelsFromIncompleteCubeTexture` were disabled\ndue to flakiness observed specifically on ASan/LSan builders\n(https://crbug.com/40246425).\n\nThe tests verify reading pixels from framebuffer attachments backed by\ncomplete and incomplete cube map textures.\n\nGuarding the tests with `#if defined(ADDRESS_SANITIZER) \u0026\u0026\ndefined(LEAK_SANITIZER)` mirrors the existing pattern used for\n`MAYBE_Compositor` in `gl_bind_uniform_location_unittest.cc`, allowing\nthe tests to run on macOS, Linux, and other standard configurations\nwhile keeping ASan/LSan disabled until the sanitizer issues are\nresolved.\n\nLocal verification on macOS and Linux confirms that both tests execute\ndeterministically without failures.\n\nThis CL introduces `MAYBE_ReadPixels` and\n`MAYBE_ReadPixelsFromIncompleteCubeTexture`, and adds documentation\ncomments.\n\nVerified on macOS.\n\nPart of http://go/arthurmilchior-fix-all-test-skill-chrome for AI week.\n\nBug: 40246425\nChange-Id: Iec9bf487e752ed268e30d445430ca4354f9f5487\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8688758\nAuto-Submit: Arthur Milchior \u003carthurmilchior@chromium.org\u003e\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCommit-Queue: Arthur Milchior \u003carthurmilchior@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737722}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f801d0c8a2e1ff6981a41bb3f97cae22c7346bee\n"
    },
    {
      "commit": "0c8de148b09778155322e7444c737dd00ce1518b",
      "tree": "510601cf20df6dcde27fb5fcd725c9e9ab6c75a0",
      "parents": [
        "6c195f0090c3bb45cd6082587e2c506341c3f405"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Fri Nov 06 07:14:00 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 07:18:52 2026"
      },
      "message": "Only allow shared images at color attachments in passthrough decoder\n\nShared image textures created with\nCreateAndTexStorage2DSharedImageCHROMIUM could be attached to any\nframebuffer attachment point, though clients only attach them to color\nattachments. Generate GL_INVALID_OPERATION from glFramebufferTexture2D,\nglFramebufferTextureLayer, glFramebufferTextureMultiviewOVR and\nglFramebufferTexture2DMultisampleEXT when attaching a shared image\ntexture to GL_DEPTH_ATTACHMENT, GL_STENCIL_ATTACHMENT or\nGL_DEPTH_STENCIL_ATTACHMENT. Attaching shared images to any color\nattachment, and detaching (texture 0), remain allowed.\n\nThis narrows what clients can do with shared images. The restriction can\nbe disabled with the kRestrictSharedImageAttachmentsToColorAttachments\nkill switch. Because of that, the shared image access checks still\nhandle shared images at depth and stencil attachments; they can be\nsimplified to only consider color attachments once the kill switch is\nremoved.\n\nBug: 582302547\nChange-Id: Iedc4ad034893cb24851ab5f59db3e271b0751b2f\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8738182\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nAuto-Submit: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737307}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 48ba8c215d61be243cb4dfc91dc37e2416a0d460\n"
    },
    {
      "commit": "6c195f0090c3bb45cd6082587e2c506341c3f405",
      "tree": "090759bb5f8db7fc136e8c455b623ae8108d0da2",
      "parents": [
        "7526d15c1caa5761b184b23e8f5289c3d594a91a"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Fri Nov 06 06:52:14 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 06:57:04 2026"
      },
      "message": "gpu: Update stale SwiftShader comments in GpuInit2\n\nGpuInit2 now falls back to WARP instead of SwiftShader on Windows.\nUpdate comments that still referred to the SwiftShader fallback.\n\nBug: 570527192\nChange-Id: Ia1b21fd9f01c7cf1bcb1a31d97b846ff8be7c1f6\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8727296\nReviewed-by: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nAuto-Submit: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737296}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ab4aedea98586f68da830892055ce8e865da299c\n"
    },
    {
      "commit": "7526d15c1caa5761b184b23e8f5289c3d594a91a",
      "tree": "1f24bc0fe6fef4b4aa2ba7842015de834ee09d30",
      "parents": [
        "22d38a04195af705e0ee9b8651a4f393c52635e6"
      ],
      "author": {
        "name": "Patrick Monette",
        "email": "pmonette@chromium.org",
        "time": "Fri Nov 06 06:29:33 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 06:36:45 2026"
      },
      "message": "Add TestMemoryConsumerRegistry::NotifyUpdateMemoryLimitAndRelease()\n\nMany tests simulate memory pressure by updating the memory limit and\nthen requesting a memory release, which mirrors what the memory\ncoordinator does in production. For async consumers, this required\npassing base::DoNothing() to NotifyUpdateMemoryLimitAsync() and relying\non main thread task ordering for the release to be observed after the\nlimit update.\n\nThis CL adds NotifyUpdateMemoryLimitAndRelease() and its async variant\nto TestMemoryConsumerRegistry, and migrates all existing callers that\nissue both notifications back-to-back. Tests that check state between\nthe two notifications are left unchanged.\n\nThis is a mechanical, test-only change.\n\nBug: 489349561\nChange-Id: Id2dc349c5f6f92e88e36bf26f398a4266ff63acf\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8689854\nOwners-Override: Francois Pierre Doray \u003cfdoray@chromium.org\u003e\nCommit-Queue: Patrick Monette \u003cpmonette@chromium.org\u003e\nReviewed-by: Francois Pierre Doray \u003cfdoray@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737285}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 471f41bdff38f38a60d94c9d454062f5e91cee07\n"
    },
    {
      "commit": "22d38a04195af705e0ee9b8651a4f393c52635e6",
      "tree": "ef92c8767b1158bd4e6148ba28ccf018165cb2d2",
      "parents": [
        "96a27c232aeeb5ca84ffdfe6e67635eca46e9c0b"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Fri Nov 06 06:03:02 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 06:30:20 2026"
      },
      "message": "[DCSI] Fold SharedMemoryCopyStrategy into CPU copy strategy\n\nSharedMemoryCopyStrategy only handled 1-hop CPU upload/readback between\na SharedMemoryImageBacking and a non-shared-memory backing using\nSharedMemoryImageBacking::pixmaps(). Fold that fast path directly into\nCPUReadbackUploadCopyStrategy and remove SharedMemoryCopyStrategy:\n\n- When either src_backing or dst_backing is kSharedMemory,\n  CPUReadbackUploadCopyStrategy::Copy() uploads from or reads back into\n  SharedMemoryImageBacking::pixmaps() directly in 1 hop without\n  allocating intermediate SkBitmap staging buffers.\n- Reject kSharedMemory-to-kSharedMemory in CanCopy() and Copy() since\n  CompoundImageBacking holds at most one shared memory backing and\n  SharedMemoryImageBacking does not implement ReadbackToMemory() or\n  UploadFromMemory().\n- Retain the 2-hop temporary SkBitmap readback + upload fallback when\n  neither backing is kSharedMemory.\n\nBug: 434215964\nChange-Id: Ib9790d5f11cf0c7f8eb7289196fe645cacbabcef\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8716108\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737263}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 70251e3aa0cf0094f1d7206c050eec997872c1f5\n"
    },
    {
      "commit": "96a27c232aeeb5ca84ffdfe6e67635eca46e9c0b",
      "tree": "380c259258440b328302ccb81e6680a6cb1ef7b6",
      "parents": [
        "0802f59451860023a3c29237958b7832c23d58fb"
      ],
      "author": {
        "name": "Ted Meyer",
        "email": "tmathmeyer@chromium.org",
        "time": "Fri Nov 06 05:47:32 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 05:55:10 2026"
      },
      "message": "[AV2] Add VideoCodec::kAV2 enum entries\n\nBug: 576731818\nChange-Id: I10932090a1580308cc851e35cf8f61720120f13c\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8637950\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Ted (Chromium) Meyer \u003ctmathmeyer@chromium.org\u003e\nReviewed-by: Dale Curtis \u003cdalecurtis@chromium.org\u003e\nReviewed-by: Joel Hockey \u003cjoelhockey@chromium.org\u003e\nReviewed-by: Evan Liu \u003cevliu@google.com\u003e\nReviewed-by: Elly \u003cellyjones@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737249}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 22bc7ac7ebd7949b408e331cffd8aefb8092a8dd\n"
    },
    {
      "commit": "0802f59451860023a3c29237958b7832c23d58fb",
      "tree": "a81266500acaacc288f7166847aeccd3a5f21b66",
      "parents": [
        "ee34f6d575c20123077b62e9df88f965d1f4d05d"
      ],
      "author": {
        "name": "kylechar",
        "email": "kylechar@chromium.org",
        "time": "Fri Nov 06 00:20:04 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 00:25:28 2026"
      },
      "message": "Treat recorderer snap() failure as a failure\n\nGraphiteFlush() was treating recorder-\u003esnap() returning null as\nsuccess, even thoguh it only returns null for an error. Treat it as a\nfailure to flush instead.\n\nFixed: 574795295\nChange-Id: I1fc9363c26c92884273240a9f3d952bb21a5c547\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8717309\nReviewed-by: Greg Daniel \u003cegdaniel@google.com\u003e\nReviewed-by: Michael Ludwig \u003cmichaelludwig@google.com\u003e\nCommit-Queue: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737029}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 75c83fdef57820ed3939a64e07a1085746627d95\n"
    },
    {
      "commit": "ee34f6d575c20123077b62e9df88f965d1f4d05d",
      "tree": "cb6593fbb3980d2979e9564447bd300370c4c3c7",
      "parents": [
        "4be20a6f8009575ef485919c3d89f3c5fa3a5a81"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Fri Nov 06 00:12:36 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Nov 06 00:18:18 2026"
      },
      "message": "[DCSI] Handle external-sampler sources in CPU copy strategy\n\nWhen a multiplanar SharedImage uses PrefersExternalSampler(), GPU\nreadback samples the image into a single RGBA plane, and\nCompoundImageBacking dynamically allocates a single-plane RGBA secondary\nbacking. Previously, CPUReadbackUploadCopyStrategy::CanCopy rejected\nthis pair because src_backing-\u003eformat() !\u003d dst_backing-\u003eformat(), and\nCopy() allocated per-plane bitmaps from src_backing rather than matching\nthe single-plane RGBA destination.\n\nExtract GetEffectiveCopyFormat() into shared_image_format_service_utils\nand update CPUReadbackUploadCopyStrategy to allow external-sampler\nsource -\u003e single-plane RGBA destination copies while rejecting uploads\ninto external-sampler destinations.\n\nBug: 434215964\nChange-Id: I078c0bd23341e76d19b4aee64447be0fd3122c8a\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8710133\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1737023}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: fb437a514a6c79b27c4df8a498f02dd4649a479e\n"
    },
    {
      "commit": "4be20a6f8009575ef485919c3d89f3c5fa3a5a81",
      "tree": "2f77278548c19da69656ef2bb5284b6bfa3d126a",
      "parents": [
        "a0284e45bb95c4db51565c90e8eb52e330b8e71b"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Thu Nov 05 19:56:38 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 05 20:03:09 2026"
      },
      "message": "Fix BlitRGBAToYUVA() edge coverage on odd-sized subsampled targets\n\nSubsampled planes of odd-sized images are allocated with ceiling-rounded\ndimensions (via SkYUVAInfo::PlaneDimensions()), while dividing dst_rect\nby nominal subsampling factors in float produces fractional bounds (e.g.\n2047.5 on a 2048-texel plane). Non-AA draws with fractional right/bottom\nedges do not rasterize the last column/row under standard half-open\nrasterization rules, leaving edge texels unwritten even though callers\nmark the destination SharedImage as fully cleared.\n\nRound plane_dst_rect out to whole texels with roundOut() and clamp to\nthe plane surface bounds with intersect(). Also guard src_image before\nquerying isOpaque() to avoid a null pointer dereference when clearing to\nblack with nullptr source on multiplanar formats with alpha.\n\nAdd unit tests in RGBAToYUVATest.OddDimensionsSubsampled and\nRasterDecoderOOPTest.CopySharedImageOddSizeMultiplanar.\n\nBug: 574790796\nChange-Id: Ib64cec52c4f32de2b803b3cf252b365af886ccdc\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8725593\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Michael Ludwig \u003cmichaelludwig@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1736739}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d37c03f266b977e5ec808c055d4317860f0feb5a\n"
    },
    {
      "commit": "a0284e45bb95c4db51565c90e8eb52e330b8e71b",
      "tree": "a428d3c75f944830b00b597af75a76e6dc3be3fc",
      "parents": [
        "4e217c93e4abd983ca4185a4349aad82f595fc5d"
      ],
      "author": {
        "name": "Vasiliy Telezhnikov",
        "email": "vasilyt@chromium.org",
        "time": "Thu Nov 05 17:02:57 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 05 17:09:27 2026"
      },
      "message": "Make SkImageInfo optional in ClientSharedImage::GetSkPixmapForPlane\n\nIt was intended as a migration path in case metadata mismatch.\n\nWe also update first call-site in ExportedCanvasResource with always\nmatch metadata [1].\n\n[1] https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/platform/graphics/canvas_resource.cc;drc\u003d4da5380189b9bc77d21c5a59e0658405c980f149;l\u003d185\n\nChange-Id: I60d2018d2f2dd5f1126a913114bf792a4180dc53\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8706492\nReviewed-by: Colin Blundell \u003cblundell@chromium.org\u003e\nCommit-Queue: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1736517}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e0292bf1023f4c16d28113293cbe953002bda8fb\n"
    },
    {
      "commit": "4e217c93e4abd983ca4185a4349aad82f595fc5d",
      "tree": "84813436d6c26eefb6e80045f58c564c84c86611",
      "parents": [
        "99783a52f181c439a538e37a9ea43ca5ac7cea74"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Thu Nov 05 16:40:27 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 05 16:48:17 2026"
      },
      "message": "Require shared image access for GL reads in passthrough decoder\n\nShared image textures may only be accessed through GL between\nBeginSharedImageAccessDirectCHROMIUM and\nEndSharedImageAccessDirectCHROMIUM, but the passthrough decoder did not\nenforce this. Reading a shared image outside of an access is not\nsynchronized with other users of the backing, and on some backings the\nread can be silently dropped by the driver.\n\nGenerate GL_INVALID_OPERATION from glReadPixels, glCopyTexImage2D,\nglCopyTexSubImage2D/3D and glBlitFramebufferCHROMIUM when an attachment\nof the read framebuffer that would be read is a shared image texture\nwithout an active access, and from glCopyTextureCHROMIUM and\nglCopySubTextureCHROMIUM when the source texture is such a texture.\n\nTo look up shared images by texture service id without scanning all of\nthem, PassthroughResources now keeps an index from service id to the\nclient ids a shared image is imported as.\n\nThe enforcement can be disabled with the\nkEnforceSharedImageAccessForGLReads kill switch.\n\nPotential followup CLs:\n\n1) Draw calls that sample a shared image without an access (and\n   rendering into one) aren\u0027t covered.\n\n2) Only allow shared images to be attached to GL_COLOR_ATTACHMENT0,\n   which would also remove the need to query GL_READ_BUFFER.\n\nBug: 577345414\nChange-Id: I0deaaa928c72e916ba7b540fe1d512d1294f525e\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8721597\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1736486}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a7bd7703182886094195cd85ded9c629d01ea803\n"
    },
    {
      "commit": "99783a52f181c439a538e37a9ea43ca5ac7cea74",
      "tree": "ad7ae7e7def6d02c8d6385952244eab0f4f11242",
      "parents": [
        "e713e6d326f8646084c8cafc5af9026c58873355"
      ],
      "author": {
        "name": "Arthur Milchior",
        "email": "arthurmilchior@google.com",
        "time": "Thu Nov 05 13:54:14 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 05 14:00:52 2026"
      },
      "message": "[GPU] Re-enable `GLEXTWindowRectanglesTest.Defaults`\n\n`GLEXTWindowRectanglesTest.Defaults` was disabled due to flakiness\nobserved specifically on ASan/LSan builders\n(https://crbug.com/40246425).\n\nThe test verifies default state and limits for the window rectangles\nextension.\n\nGuarding the test with `#if defined(ADDRESS_SANITIZER) \u0026\u0026\ndefined(LEAK_SANITIZER)` mirrors the existing pattern used for\n`MAYBE_Compositor` in `gl_bind_uniform_location_unittest.cc`, allowing\nthe test to run on macOS, Linux, and other standard configurations while\nkeeping ASan/LSan disabled until the sanitizer issues are resolved.\n\nLocal verification on macOS and Linux confirms that the test executes\ndeterministically without failures.\n\nThis CL introduces `MAYBE_Defaults` and adds a documentation comment.\n\nVerified on macOS.\n\nPart of http://go/arthurmilchior-fix-all-test-skill-chrome for AI week.\n\nBug: 40246425\nChange-Id: If4af01e716a69e6e382f65c6abae05e53444bbd5\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8688378\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCommit-Queue: Arthur Milchior \u003carthurmilchior@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1736323}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9136d9cd363929220ccba4e6e1547d9c1a08f7f1\n"
    },
    {
      "commit": "e713e6d326f8646084c8cafc5af9026c58873355",
      "tree": "6f43234bdad26d19156575a1d1a221b01b1acdf6",
      "parents": [
        "4f7fe92fe8770d66b2c539f31010a6acb47296dc"
      ],
      "author": {
        "name": "Michael Tang",
        "email": "tangm@microsoft.com",
        "time": "Wed Nov 04 23:55:36 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Nov 05 00:00:08 2026"
      },
      "message": "Add missing field to GPUInfoKnownFields\n\nThis one was missing and escaped the static assert, possibly due to\nstruct packing and field alignment oddities.\n\nChange-Id: I958c915171a5c65f82f34138a1fc34d96a6a6964\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8617858\nReviewed-by: Rafael Cintron \u003crafael.cintron@microsoft.com\u003e\nCommit-Queue: Michael Tang \u003ctangm@microsoft.com\u003e\nReviewed-by: Michael Tang \u003ctangm@microsoft.com\u003e\nCr-Commit-Position: refs/heads/main@{#1735829}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: b9dad246cc3281d701141d8ce98567d5763a6ab8\n"
    },
    {
      "commit": "4f7fe92fe8770d66b2c539f31010a6acb47296dc",
      "tree": "0111da5947dd88c721bf2cbbfe83395442018793",
      "parents": [
        "76c9f6635418d8e88a2e5e1dd3e0c70b1f1aee43"
      ],
      "author": {
        "name": "Jonathan Ross",
        "email": "jonross@chromium.org",
        "time": "Wed Nov 04 19:40:25 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 19:49:33 2026"
      },
      "message": "[gpu] Use empty Size for offscreen GL surfaces\n\nIn https://crrev.com/c/8677084, the 0x0 size DCHECKs in\nGLOzoneEGLGbm::CreateOffscreenGLSurface were converted to CHECK_EQ with\nbase::NotFatalUntil::M163, which is fatal on non-official release\nbuilders such as ChromeOS FYI Release (amd64-generic).\n\nUpdate GPU unit tests that create a dummy offscreen GLSurface only to\nmake a GLContext current to pass gfx::Size() instead of gfx::Size(4, 4),\nmatching production callers and SharedImageTestBase.\n\nBug: 581806371\nTest: gl_tests, gpu_unittests\nChange-Id: I99cddda1238c823771b710b8983c5a148c155dc3\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8727507\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Jonathan Ross \u003cjonross@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1735504}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 90fafcffedf6f0cd4fd36030aa9e941af84f2ba3\n"
    },
    {
      "commit": "76c9f6635418d8e88a2e5e1dd3e0c70b1f1aee43",
      "tree": "d1a20fdda7fa8ddeb787a7d5208a064c3b8efc2b",
      "parents": [
        "3843030013c05ed19e9cb81cbbe082cb2d4f060e"
      ],
      "author": {
        "name": "Nathan Hebert",
        "email": "nhebert@chromium.org",
        "time": "Wed Nov 04 15:48:19 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 15:52:49 2026"
      },
      "message": "media/gpu/chromeos: fix V4L2 global decoder limitations\n\nIn the V4L2 stateful backend, there is a limit placed on the total\nnumber of simultaneous decoders: 15 per decoder utility process. The\ncode was added before decoders were moved out of the renderer process\nand to utility processes.\n\nSince decoding no longer happens on one process, Chrome lost the ability\nto globally limit the number of decoders. The device limitations should\nbe system-wide.\n\nOn ChromeOS, add a global 15 decoder limit to the Qualcomm ChromeOS\ndevices, which use stateful video decoders. This could help QC ChromeOS\nvideo stability.\n\nAlso, add a limit of 15 stateful V4L2 decoders per utility process, to\nVideoDecoderPipeline. This should not have any impact, but lets a future\nCL clean up some code without regressing Linux.\n\nBug: b:573316672\nTest: checked chrome://gpu on trogdor for workaround\n\nChange-Id: I4ad3d82c9315d83b8ac6bb4bd5fe73b3b69de26f\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8600325\nCommit-Queue: Nathan Hebert \u003cnhebert@chromium.org\u003e\nReviewed-by: Andres Calderon Jaramillo \u003candrescj@chromium.org\u003e\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nAuto-Submit: Nathan Hebert \u003cnhebert@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1735198}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ce7c83397f1c2eb1c30805099c0da6127059e0e2\n"
    },
    {
      "commit": "3843030013c05ed19e9cb81cbbe082cb2d4f060e",
      "tree": "c411dae99848a3eea7c2d41781367c4c9c114aaf",
      "parents": [
        "a52864a3751f749088b58763588431dd7e763c38"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Wed Nov 04 15:28:37 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 15:46:47 2026"
      },
      "message": "gpu: Clean up GpuInit2 helpers\n\nFollow-up cleanups in gpu_init_2.cc, with no intended behavior change:\n- Add CollectContextInfoAndRecomputeFeatures() for the repeated\n  \"collect context info, set crash keys, recompute features\" sequence.\n- Add GpuWatchdogInitScope::ScopedPause, so the watchdog is resumed on\n  every return path, and make Pause()/Resume() private.\n- Extract DisableVulkanFeatures() and ApplyVulkanDeviceToGpuInfo(), the\n  latter shared by InitializeVulkan() and the Android Dawn adapter\n  validation. Reuse the local vulkan_info and return early when the\n  Vulkan instance fails to initialize.\n- Extract ToSkiaBackendType() from SetSkiaBackendType().\n- Share ResolveInProcessGrContextType() between the Android and\n  non-Android InitializeInProcess() paths.\n- Take const InitState\u0026 where the state is only read, and references in\n  DisableInProcessGpuVulkan().\n- Fix the stale \"InitializeDawn() failed!\" message.\n- Remove unused includes, add missing ones (base/feature_list.h,\n  \u003cunistd.h\u003e, \u003cmemory\u003e, ui/gl/gpu_preference.h), and simplify the\n  CanAccessDeviceFile() build guard.\n\nBug: 570527192\nChange-Id: I16f60b5fe713a5b293d59679124329279b209595\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8719128\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nAuto-Submit: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1735176}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7870217e2a53f84c60c5f55c76102525f039c5de\n"
    },
    {
      "commit": "a52864a3751f749088b58763588431dd7e763c38",
      "tree": "acf3f42a78396a0872071738edc6eacd3805ee51",
      "parents": [
        "cf947c179becaaf174ddcd05e0ed93e78e036972"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Wed Nov 04 07:32:50 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 07:40:02 2026"
      },
      "message": "gpu: Skip the GL context for Mac GPU info collection\n\nOn macOS, when kGpuInitOptimization is enabled and ANGLE runs on Metal,\ncollect context graphics info in the new CollectMetalGraphicsInfo()\ninstead of creating a GL context in CollectGraphicsInfoGL(), similar to\nCollectD3D11GraphicsInfo() on Windows:\n- display_type, gl_vendor, gl_renderer, gl_version, gl_ws_* and\n  gl_reset_notification_strategy come from ANGLE\u0027s EGL display via\n  CollectGraphicsInfoFromANGLEDisplay().\n- pixel/vertex_shader_version are \"3.00\", matching what\n  CollectGraphicsInfoGL() parses out of ANGLE\u0027s\n  GL_SHADING_LANGUAGE_VERSION for its ES 3.0 context.\n- The active GPU is the one whose IORegistry entry ID matches ANGLE\u0027s\n  MTLDevice.registryID (GPUDevice::system_device_id on macOS), otherwise\n  the primary GPU, as on Windows. MarkPrimaryGpuActive() moves to\n  gpu_info_collector.{h,cc} to be shared.\n- The active GPU\u0027s vendor_string/device_string are set from the GL\n  strings. They identify GPUs without a PCI device ID (e.g. Apple\n  Silicon).\n- driver_vendor and driver_version are parsed from the GL strings with\n  GLVersionInfo, as CollectGraphicsInfoGL() does.\n\ngl_extensions and max_msaa_samples remain empty in this path. Neither\nthe GPU blocklist nor the driver bug list has macOS (or OS-independent)\nentries that depend on them.\n\nAlso add a GetGLWindowSystemBindingInfoEGL() overload that takes an\nEGLDisplay, so CollectGraphicsInfoFromANGLEDisplay() can reuse it\nwithout a current context.\n\nIf ANGLE isn\u0027t on Metal, CollectGraphicsInfoGL() is still used.\n\nBug: 570527192\nChange-Id: I161a3d09295abd52f7ddcbdfc70ae6e94e402f5f\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8693268\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734837}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 485b9c8e74e7164dbbc75f1e66b9bccc9f97ba70\n"
    },
    {
      "commit": "cf947c179becaaf174ddcd05e0ed93e78e036972",
      "tree": "ca719249e435960b17cda9eadfd9d6c9104172b2",
      "parents": [
        "3ea090db797c719687ae6e321360c2601cdad80d"
      ],
      "author": {
        "name": "Nathan Hebert",
        "email": "nhebert@chromium.org",
        "time": "Wed Nov 04 05:37:55 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 05:43:35 2026"
      },
      "message": "media/v4l2: migrate MT8173 decoder limits to GPU workaround\n\nMigrates MT8173 (ChromeOS Hana/Elm) decoder limits determination from a\nV4L2 query to GPU workaround. There is already a workaround to limit the\nnumber of decoders to 10 which works around an issue with AMD devices.\n\nThis should be a pure refactor, but it may change behavior on ChromeOS\ndue to the use of OOPVideoDecoder since it can now correctly limit the\nnumber of decoders globally.\n\nBug: b:573316672\nTest: checked chrome://gpu on elm for workaround\n\nChange-Id: I954c274cb703e180219339e591a13c1ce9f9bc1e\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8600324\nCommit-Queue: Nathan Hebert \u003cnhebert@chromium.org\u003e\nAuto-Submit: Nathan Hebert \u003cnhebert@chromium.org\u003e\nReviewed-by: Andres Calderon Jaramillo \u003candrescj@chromium.org\u003e\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734773}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: ab4f06f05b43275841ae8243f539c3dd62f0c760\n"
    },
    {
      "commit": "3ea090db797c719687ae6e321360c2601cdad80d",
      "tree": "01bb64092203864eb5f955fe6cdde2977fbea114",
      "parents": [
        "6f8f1ea6fc7da78cdb0794c70c1f3df4a07200b1"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Wed Nov 04 05:02:13 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 05:17:45 2026"
      },
      "message": "gpu: Share GL finalization helpers with GpuInit2::InitializeInProcess()\n\nInitializeInProcess() repeated, inline, most of what the helpers used by\nInitializeAndStartSandbox() do: collecting context info and recomputing\nfeatures (including the Windows WARP fallback), extension setup, the\ndefault offscreen surface, extra info collection, the Linux core profile\nrecollection, the WARP info adjustment, disabling WebNN when GL is\ndisabled, and the native pixmap query.\n\nLet FinalizeFeaturesWithContextInfo() and FinalizeGLInitialization()\nhandle the in-process case instead: failures are only logged and crash\nkeys are not set, as before. InitializeInProcess() now calls the helpers.\n\nAlso compute InitState::gl_disabled once in\nInitializeGLBindingsAndDisplay(), and set\nGPUInfo::gl_implementation_parts at the end of\nFinalizeGLInitialization() for both paths.\n\nMinor in-process behavior differences:\n- CollectGpuExtraInfo() now runs before extension setup, and\n  InitializePlatformOverlaySettings() runs after the WARP info\n  adjustment, matching InitializeAndStartSandbox().\n- gl_implementation_parts is set after a WARP reinitialization, so it\n  reflects WARP instead of the hardware GL implementation.\n\nBug: 570527192\nChange-Id: I546e2793423ece371613be12ab37a3052bec66fb\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8718989\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734748}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7551f1b5bd5bd8c9196938060aab0ac8da08643d\n"
    },
    {
      "commit": "6f8f1ea6fc7da78cdb0794c70c1f3df4a07200b1",
      "tree": "4b14e3ef1e9ad1d01463d7217cf8e132bc85b39b",
      "parents": [
        "9877b79e4231ad132b687be0fa9d0dd3ead09447"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Wed Nov 04 01:18:24 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Wed Nov 04 01:24:58 2026"
      },
      "message": "gpu: Remove GPUInfo::max_msaa_samples\n\nGPUInfo::max_msaa_samples is no longer consumed by any feature/workaround\ndecision logic, UMA metrics (GPU.MaxMSAASampleCount was removed in\ncrrev.com/c/5120557), or rendering pipeline, and is only populated when\ncreating a temporary GL context in CollectGraphicsInfoGL() (which is\nskipped under kGpuInitOptimization and does not reflect non-GL rendering\nbackends such as Vulkan or Skia Graphite).\n\nRemove max_msaa_samples from GPUInfo, its Mojo struct/traits,\nCollectGraphicsInfoGL(), and chrome://gpu.\n\nBug: 570527192\nChange-Id: I3ba0149b88d9d0f191df359ebe54c2e507dcb41e\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8711480\nAuto-Submit: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Elias Klim \u003celklm@chromium.org\u003e\nReviewed-by: Tom Sepez \u003ctsepez@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCommit-Queue: Tom Sepez \u003ctsepez@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734641}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: cd0e892c11c91fa07e2c0c21d302321cab35b3cd\n"
    },
    {
      "commit": "9877b79e4231ad132b687be0fa9d0dd3ead09447",
      "tree": "a34048bdbbde914a0986e2e4a5404e288925bc01",
      "parents": [
        "bfc84b129baf294e5a6910a7cb2817286a25696b"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Tue Nov 03 21:15:42 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 21:23:14 2026"
      },
      "message": "[gpu] Remove PowerVR SGX 5* and GLES\u003c3.0 driver workaround entries\n\nPowerVR SGX 5* [1] renderer strings do not support GLES 3.0\nonwards which is now required. This change removes these entries\nalong with few others which were for GLES \u003c 3.0 from driver bug list,\nbut keeps corresponding features. I will remove feature code in\nfollow-up.\n\n[1] https://opengles.gpuinfo.org/\n\nBug: 570597474\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: Ifc6ef6f386ccfbb111cb88e37ba660410c9206e7\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8687563\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734398}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 7df3e4136ec23b144e9aaa84a07a120ba18d6725\n"
    },
    {
      "commit": "bfc84b129baf294e5a6910a7cb2817286a25696b",
      "tree": "382b7c890af7e1b6144208f1cd16f92520eadcf3",
      "parents": [
        "1ac653db2af8bc582b8e9c4a6f03f692763e3d33"
      ],
      "author": {
        "name": "Free Debreuil",
        "email": "freedebreuil@google.com",
        "time": "Tue Nov 03 20:50:55 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 20:59:15 2026"
      },
      "message": "[cc] Bundle raster-inducing scroll offsets into RasterInducingValues\n\nWrap raster-inducing scroll offsets in a RasterInducingValues struct so\nother compositor-sampled values applied at raster time can share the\nsame plumbing.\n\nTile raster, picture quads, playback and serialization now pass\nRasterInducingValues instead of a ScrollOffsetMap. Raster tasks and\npicture quads own a copy, and PlaybackParams and SerializeOptions hold a\npointer to it. The tiling client\u0027s GetRasterInducingScrollOffsets()\nbecomes GetRasterInducingValues().\n\nNo behavior change.\n\nBug: 545152722\nChange-Id: If929b2b893513e021425a0be07203413397ee1f8\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8588023\nAuto-Submit: Free Debreuil \u003cfreedebreuil@google.com\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734363}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: d7932f8c4969bcf7c508df7191430c5d391a03c6\n"
    },
    {
      "commit": "1ac653db2af8bc582b8e9c4a6f03f692763e3d33",
      "tree": "0a3a6069e90acf71a8a78195c37d14b52e3fe8a9",
      "parents": [
        "90e0c10e4b9d6ef5e2655c2915d07fb073ff476c"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Tue Nov 03 20:47:07 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 20:54:13 2026"
      },
      "message": "[gpu] Check SubmitIfNecessary before clear in RasterDecoder/CopySIHelper\n\nCheck return value of SubmitIfNecessary before calling SetCleared in\nplaces where clear happens in RasterDecoder and CopySharedImageHelper.\nThis translates to updating success bool in all these places as it\nalready stores result of FlushWriteAccess.\n\nBug: 568937157\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: Ie1039ab31d68d28c47bbb58b557ecb51edf3730e\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8704260\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734357}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 25a7e6c752822d0459e8565ab4984dd97d9bf314\n"
    },
    {
      "commit": "90e0c10e4b9d6ef5e2655c2915d07fb073ff476c",
      "tree": "c64834b8c0109772ab192df4f6e993e9e52c0481",
      "parents": [
        "2d786bc3a6d76b3e41d9a58b69c25eb114ab26ab"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Tue Nov 03 18:44:50 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 18:50:23 2026"
      },
      "message": "[gpu] Remove Mali-4* and MuMu workaround entries\n\nMuMu GL and Mali-4* [1] renderer strings do not support GLES 3.0\nonwards which is now required. This change removes these entries\nfrom driver bug list, but keeps corresponding features. Most of them\nare needed for other workarounds except disable_gl_rgb_format which\nI can remove in follow-up.\n\n[1] https://opengles.gpuinfo.org/\n\nBug: 570597474\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: I1da87fdd0e0fa09594a9741da5a5649234914638\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8688080\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734225}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2eeca240aa71837623ceb94d553899ff72829e24\n"
    },
    {
      "commit": "2d786bc3a6d76b3e41d9a58b69c25eb114ab26ab",
      "tree": "3fc673fe59b9bae232f538f1fe8799cb048b4f83",
      "parents": [
        "dc17138514ed74cdc187776165d7f5ff1d70b564"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Tue Nov 03 18:35:47 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 18:42:08 2026"
      },
      "message": "[gpu] Reject R16, RG16 and RG8 formats at SharedImageStub\n\nThese formats are not used for creating SharedImages and should be\nrejected at the SharedImageStub.\n\nBug: 416285370\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nChange-Id: I62b89519a0612b8af40e9c1c230544ae4711bf85\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8687763\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734212}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 72c9200f25a390340846d667cabf27312550afa4\n"
    },
    {
      "commit": "dc17138514ed74cdc187776165d7f5ff1d70b564",
      "tree": "d44851d6b28f6cb3f93dcbc1d315c7842ff39030",
      "parents": [
        "5b772a7ebe948382e6f5340cce1e38666e9b7a95"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Tue Nov 03 17:03:49 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 17:12:00 2026"
      },
      "message": "Fix GPU.BlocklistFeatureTestResults.Webgpu user-disabled check\n\nUpdateFeatureStats() treated WebGPU as user-disabled when\n--enable-unsafe-webgpu was not passed. Normal users do not pass it, so\n99.9% of samples were recorded in the User-disabled bucket (2).\n\nWebGPU is turned off via the WebGPUService feature rather than a\nswitch. In that case, gpu::GetWebGPUFeatureStatus() already returns\nkGpuFeatureStatusDisabled. So never override it.\n\nThis was found while estimating WebGL vs. WebGPU reach, see\ngo/webgl-webgl-reach.\n\nAlso:\n- Extend the histogram expiry, which had already passed.\n- Rename bucket 2 from User-disabled to Disabled, add the missing\n  Software (3) and Undefined (4) labels, and keep the enum in sync with\n  gpu::GpuFeatureStatus via LINT.IfChange.\n\nBug: 578398837\nChange-Id: I7bd902b830a72060a06181954a3a44452eb07a76\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8618574\nReviewed-by: Colin Blundell \u003cblundell@chromium.org\u003e\nReviewed-by: Mark Pearson \u003cmpearson@chromium.org\u003e\nCommit-Queue: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nAuto-Submit: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1734087}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: a2b42767d487bcd474763f5e14a43887a9b9d7f8\n"
    },
    {
      "commit": "5b772a7ebe948382e6f5340cce1e38666e9b7a95",
      "tree": "55dba7fdb833a27c86b051446e049d5bd91b724a",
      "parents": [
        "b16d74a2479471e30c7f43ef0d5e1883e129e577"
      ],
      "author": {
        "name": "kylechar",
        "email": "kylechar@chromium.org",
        "time": "Tue Nov 03 14:57:40 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 15:04:45 2026"
      },
      "message": "Disable WrappedSkiaBackingExclusiveAccess\n\nThere is also overlapping access on Android that needs to be fixed. Turn\nthe kill switch off on all platforms as a precaution.\n\nThe dawn-win10-x86-deps-rel bot is failing every run and equivalent CI\nbot is also failing so exclude it.\n\nCq-Exclude-Trybots: luci.chromium.try:dawn-win10-x86-deps-rel\nBug: b/579626809\nBug: 574220812\nChange-Id: I6f3c38cc862380a5c44860cdfeca92dce291fbda\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8693805\nCommit-Queue: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nReviewed-by: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1733950}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 45fa5582526bfabfa970c782eb2ef974338e3867\n"
    },
    {
      "commit": "b16d74a2479471e30c7f43ef0d5e1883e129e577",
      "tree": "7c30d18161fce610cc133120f19a91f0aacb60d7",
      "parents": [
        "37e1a418df983f45e55cc571651a31248911fe3b"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Tue Nov 03 14:42:03 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 14:49:07 2026"
      },
      "message": "spanification: Pass a span to GLES2Util::GetQualifiedEnumString\n\nTake the enum string table as a base::span instead of a pointer and\ncount, so the lookup is a range-based loop with no pointer arithmetic.\nUpdate the code generator and regenerate the autogen callers.\n\nThis CL was written by an AI agent (autospan-agent) and reviewed\nby me before sending.\n\nBug: 40285824\nChange-Id: I5fd06e00d88113cc82a422bf9e55fa68d637c505\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8692453\nReviewed-by: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nAuto-Submit: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nReviewed-by: Kai Ninomiya \u003ckainino@chromium.org\u003e\nCommit-Queue: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1733933}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8b581560c568a1ac832a2239cf4710a84d892980\n"
    },
    {
      "commit": "37e1a418df983f45e55cc571651a31248911fe3b",
      "tree": "c07518229e54aa8ddbde68564e8a7c4a9ee84adf",
      "parents": [
        "f36068feedec5bc5729020b75bbb1a9ef9107c53"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Tue Nov 03 10:27:46 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 10:34:20 2026"
      },
      "message": "[DCSI] Skip copying from uninitialized backing in NotifyBeginAccess\n\nWhen CompoundImageBacking wraps an initial GPU backing created without\ninitial pixel data (so its ClearedRect() is empty and no shared memory\nbacking is present), its constructor still initializes\nelement.content_id_ \u003d latest_content_id_.\n\nWhen kUseDynamicBackingAllocations is enabled and the first access\nstream is incompatible with that initial backing (such as WebGL\naccessing kGL on Fuchsia QEMU where ExternalVkImageBacking requires a\nseparate GL texture), GetOrAllocateBacking() dynamically allocates a\nsecond backing (GLTextureImageBacking) and NotifyBeginAccess() attempts\nto copy from the uninitialized ExternalVkImageBacking into the new\nGLTextureImageBacking, which fails.\n\nSkip copy_manager_-\u003eCopyImage() in NotifyBeginAccess() when the source\nbacking\u0027s ClearedRect() is empty, as there is no initialized pixel data\nto synchronize, and remove the redundant updated_backing flag.\n\nTAG\u003dagy\nCONV\u003da1774bb9-9b74-4b02-a1a1-52f383e8ccb9\n\nFixed: 578202968\nChange-Id: I03ee99bdca81833713df41599a94ea5d04947569\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8665146\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1733778}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9a2395158b59d672ee4bd2db8a45fb0bbd7bd06b\n"
    },
    {
      "commit": "f36068feedec5bc5729020b75bbb1a9ef9107c53",
      "tree": "210b484ec1b7996328ee196acf7cff0a3a22c45c",
      "parents": [
        "56be178760b5f8873a7340f01b56972a2f0cfb23"
      ],
      "author": {
        "name": "vikas soni",
        "email": "vikassoni@chromium.org",
        "time": "Tue Nov 03 10:27:16 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Tue Nov 03 10:33:34 2026"
      },
      "message": "[DCSI] Fix cleared rect propagation across CompoundImageBacking\n\nKeep underlying backings\u0027 cleared rects in sync with their actual pixel\ncontent instead of broadcasting non-empty cleared rects to stale\nbackings:\n1. In NotifyEndAccess(), propagate the cleared rect to each destination\n   backing after a proactive copy from a transient backing succeeds.\n2. In NotifyBeginAccess(), copy the source backing\u0027s cleared rect\n   directly to the destination backing instead of looping over all\n   elements_ (which missed transient backings and marked unrelated stale\n   backings as cleared).\n3. In SetClearedRect(), only propagate non-empty rects to elements with\n   the latest content, while still resetting all elements when the rect\n   is empty (e.g. WEBGPU_MAILBOX_DISCARD before BeginAccess).\n\nTAG\u003dagy\nCONV\u003da1774bb9-9b74-4b02-a1a1-52f383e8ccb9\n\nBug: 475210055,578202968\nChange-Id: I15b2bd203f31870a1960a5dba72495251bd3532f\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8665763\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: vikas soni \u003cvikassoni@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1733777}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 84db535cac121703b6069c79db160f951a370089\n"
    },
    {
      "commit": "56be178760b5f8873a7340f01b56972a2f0cfb23",
      "tree": "6b5803644b3f824c6023964c4de401e9d4f815ef",
      "parents": [
        "d8f2368f14db2857ec45b69af183b4fa2ed71c45"
      ],
      "author": {
        "name": "Lokbondo Kung",
        "email": "lokokung@google.com",
        "time": "Mon Nov 02 14:05:55 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Mon Nov 02 14:10:57 2026"
      },
      "message": "[webgpu] Make DawnInstance ref-counted and own its DawnPlatform\n\nMake DawnInstance inherit from base::RefCountedThreadSafe\u003cDawnInstance\u003e\nso that a single instance can eventually be shared across WebGPU\ncontexts.\n\n- DawnInstance now owns the DawnPlatform used by the underlying Dawn\n  instance and disconnects it in ~DawnInstance(). Since the Dawn\n  instance may outlive the wrapper, this replaces the manual\n  DisconnectDawnPlatform() calls in WebGPUDecoderImpl::Destroy() and\n  ~DawnSharedContext(), which would otherwise break once the instance\n  is shared.\n- Replace DawnInstance::Create() with CreateForWebGPU() and\n  CreateForGraphite(), which create the appropriately configured\n  DawnPlatform internally. DawnInstance also records the SafetyLevel it\n  was created with, and WebGPUDecoderImpl now queries it from the\n  instance instead of recomputing it from GpuPreferences so that the\n  two always match.\n- Create the WebGPU DawnInstance in WebGPUDecoder::Create() and pass it\n  into CreateWebGPUDecoderImpl() and WebGPUDecoderImpl() as a\n  scoped_refptr\u003cDawnInstance\u003e. WebGPUDecoderImpl no longer holds a\n  separate DawnPlatform.\n- DawnSharedContext no longer holds a DawnPlatform and instead accesses\n  it through its DawnInstance. The gl::ProgressReporter used to create\n  the platform is now passed to DawnSharedContext::Initialize() instead\n  of the constructor.\n\nBug: 412761228\nChange-Id: I3e86722c959331263d32dabe978a279a385b42c2\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8615801\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCommit-Queue: Loko Kung \u003clokokung@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1732810}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: e563556ec298cc96243af97a60db4345570e0583\n"
    },
    {
      "commit": "d8f2368f14db2857ec45b69af183b4fa2ed71c45",
      "tree": "f0845c5f4b0a9aa92d8af59ee2c6a6532f6b1bd3",
      "parents": [
        "c2cef25fc93c2e86b32baf405a79d6ad92916904"
      ],
      "author": {
        "name": "arthur milchior",
        "email": "arthurmilchior@google.com",
        "time": "Sun Nov 01 10:41:27 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sun Nov 01 10:45:53 2026"
      },
      "message": "[GPU] Re-enable `PointCoordTest.RenderTo` on Linux and ChromeOS\n\n`PointCoordTest.RenderTo` was disabled in release builds on Linux (and\nlater ChromeOS via https://crrev.com/c/2351281) due to flakiness on an\nold Linux ATI bot (https://crbug.com/40299409).\n\nGit history was investigated and the exact repairing commit could not be\npinpointed, though `gl_tests` now runs on ANGLE/SwANGLE and modern Mesa\ndrivers. Re-enable `PointCoordTest.RenderTo`. Verified on Linux.\n\nPart of http://go/arthurmilchior-fix-all-test-skill-chrome for AI week.\n\nBug: 40299409\nChange-Id: I436f039e21ce7e86fb9874908bcfe75428bc0f10\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8690291\nReviewed-by: Corentin Wallez \u003ccwallez@chromium.org\u003e\nAuto-Submit: Arthur Milchior \u003carthurmilchior@chromium.org\u003e\nCommit-Queue: Corentin Wallez \u003ccwallez@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1732402}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9e5d54c27f060455c3a3d0ca49a51eca368f0b25\n"
    },
    {
      "commit": "c2cef25fc93c2e86b32baf405a79d6ad92916904",
      "tree": "09ebf436f8b56efbcdeee7f921a26293ce7d0da9",
      "parents": [
        "3045604d6faa654e0e33ffe3a124a38d42bbbcc3"
      ],
      "author": {
        "name": "Thomas Smith",
        "email": "thomsmit@google.com",
        "time": "Sat Oct 31 15:37:26 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Oct 31 15:42:26 2026"
      },
      "message": "gpu: Flush active/pending queries on virtual context switch and FBO change\n\nWhen flush_queries_before_deleting_or_unbinding_fbo is enabled (ARM Mali\nValhall/Avalon), a user FBO must not be unbound or deleted while GPU\njobs or queries referencing it remain unflushed.\n\nPreviously:\n1. GLES2DecoderImpl::FlushQueriesBeforeDeletingOrUnbindingFboWorkaround\n   only checked HasPendingQueries(), missing queries that had been begun\n   (in active_queries_) but not yet ended when the FBO was unbound or\n   deleted.\n2. With virtualized GL contexts, switching out a decoder\u0027s virtual\n   context allows the state restorer, Skia (whose workaround only tracks\n   Skia-issued timer queries), or another decoder on the same real GL\n   context to rebind or delete FBOs while the outgoing decoder still has\n   unflushed queries.\n\nFix both by:\n- Adding QueryManager::HaveActiveQueries() and checking both active and\n  pending queries in FlushQueriesBeforeDeletingOrUnbindingFboWorkaround.\n- Overriding PauseQueries() in GLES2QueryManager to flush the query job\n  chain when switching out a virtual context that had active or pending\n  queries.\n\nAssisted-by: Gemini 4 Argon (Medium)\nBug: 534468209\nChange-Id: If306f8afcc5287b9c615ee1a3606f233b367604e\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8667961\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nReviewed-by: Brandon Jones \u003cbajones@chromium.org\u003e\nCommit-Queue: Thomas Smith \u003cthomsmit@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1732281}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 27ae271326a88775cfacb72e3cd7cc5de92b845e\n"
    },
    {
      "commit": "3045604d6faa654e0e33ffe3a124a38d42bbbcc3",
      "tree": "7c38c45ba8383a00d4fbb0b01f3ac9602a423071",
      "parents": [
        "82c11eb2f2c4c088dd9b473be914578fe0033015"
      ],
      "author": {
        "name": "Le Hoang Quyen",
        "email": "lehoangquyen@chromium.org",
        "time": "Sat Oct 31 08:29:14 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Oct 31 08:34:46 2026"
      },
      "message": "gpu: Disallow SHARED_IMAGE_USAGE_MIPMAP from client requests\n\nMove SHARED_IMAGE_USAGE_MIPMAP to service-side-only usages (after\nLAST_CLIENT_USAGE), as it is only used on the GPU side by Viz\u0027s\nSkiaRenderer for render pass backings. Also validate\nIsValidClientUsage() in SharedImageStub so invalid or service-only\nclient usages are rejected over IPC.\n\nBug: 574216398\nChange-Id: Ib4d89fb13a33a01ff04f0a46f60427f003be01c0\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8689167\nCommit-Queue: Quyen Le \u003clehoangquyen@google.com\u003e\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1732216}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5c04ddb341e1931b19651bd0fc3385a17d4bfa09\n"
    },
    {
      "commit": "82c11eb2f2c4c088dd9b473be914578fe0033015",
      "tree": "1f19435852fa1f26b14f6d4eab22d9370cd5ecd2",
      "parents": [
        "cbce038c99542a1131f9ebf70a0dc8f2d2c88cfe"
      ],
      "author": {
        "name": "Achuith Bhandarkar",
        "email": "achuith@chromium.org",
        "time": "Sat Oct 31 04:39:41 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Oct 31 04:43:34 2026"
      },
      "message": "gpu: enable exit-time destructors check in cc, vulkan, and gl tests\n\nEliminate static objects with non-trivial destructors across cc\nperftests, vulkan test support, and gl unit tests by moving\nTaskEnvironment ownership onto CCTestSuite, managing native test\nwindow state through trivially destructible pointers cleaned up on\nwindow teardown, and storing fake GL extension strings on the\nGLApiTest fixture. Enable the exit-time destructors warning on those\ntargets while keeping the suppression on cc_unittests.\n\nTAG\u003dagy\nCONV\u003d24fb0bac-49f1-4a9f-9459-e507cfd2dbfd\n\nBug: 40031409, 522379416\nChange-Id: I425c397d2c2c8f59c83e6c8f6ffdfee9842ccf68\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8655639\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Achuith Bhandarkar \u003cachuith@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1732157}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 13d6347754fee49e2f9069f3704ed27553abb6d3\n"
    },
    {
      "commit": "cbce038c99542a1131f9ebf70a0dc8f2d2c88cfe",
      "tree": "3216275606727ce54dfcebe17eb2c6a9ef96506a",
      "parents": [
        "6a83bd382037f64a03103d9f5b4b1e8e8ce9e291"
      ],
      "author": {
        "name": "Zhenyao Mo",
        "email": "zmo@chromium.org",
        "time": "Sat Oct 31 01:44:35 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Sat Oct 31 01:50:47 2026"
      },
      "message": "gpu: Optimize CollectContextGraphicsInfo for kGpuInitOptimization\n\nOn Windows, when ANGLE runs on D3D11 (not WARP), kGpuInitOptimization\nis enabled and ANGLE\u0027s ID3D11Device is available, collect context\ngraphics info directly from the D3D11 device in the new\nCollectD3D11GraphicsInfo() and skip CollectGraphicsInfoGL():\n- d3d11_feature_level comes from ID3D11Device::GetFeatureLevel().\n- vertex/pixel_shader_version are derived from the feature level,\n  mirroring ANGLE\u0027s Renderer11::getMajor/MinorShaderModel(), instead of\n  parsing GL_RENDERER. GPU.D3DShaderModel is recorded from it as well.\n- The active GPU is identified via IdentifyActiveGPUWithLuid(), falling\n  back to the primary GPU.\n- display_type, gl_vendor (ANGLE\u0027s GL_VENDOR equals its EGL_VENDOR),\n  gl_ws_vendor/version/extensions and gl_reset_notification_strategy\n  come from the EGL display via the new\n  CollectGraphicsInfoFromANGLEDisplay(), without a current context.\n  It\u0027s platform independent so other ANGLE platforms can reuse it.\n- gl_renderer and gl_version also come from the EGL display, through\n  ANGLE\u0027s new EGL_ANGLE_query_renderer_strings extension. gl_version is\n  built from the ANGLE version and the default GLContextAttribs ES\n  version, which is what the context in CollectGraphicsInfoGL() uses.\n  Chromium\u0027s EGL bindings support for the extension is added in the\n  parent CL.\n\nCollectDriverInfoD3D() now also sets driver_vendor for every adapter,\nderived from the PCI vendor id like ANGLE\u0027s GL_RENDERER does (or\nANGLE\u0027s hex format for unknown vendors). It already set driver_version\nfrom the adapter\u0027s UMD version, so both are known without GL strings.\n\ngl_extensions and max_msaa_samples remain empty in this path.\n\nAll other configurations (WARP, SwiftShader, other ANGLE backends, or\nthe feature disabled) keep going through CollectGraphicsInfoGL(). In\nthat path, GPU.D3DShaderModel is now recorded from the D3D11 feature\nlevel.\n\nBug: 570527192\nChange-Id: Id03a50c29d8a87af3fd8cff966e10fdb89cce460\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8670349\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Zhenyao Mo \u003czmo@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1732105}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 621ca775a43aad45d9a25b84108fe5124b73a6a7\n"
    },
    {
      "commit": "6a83bd382037f64a03103d9f5b4b1e8e8ce9e291",
      "tree": "b355717dc1a6a969a6e2089437553bf8b56b923f",
      "parents": [
        "cfddf5cde54d0f8361a0f229e8e09680938d3861"
      ],
      "author": {
        "name": "Patrick Monette",
        "email": "pmonette@chromium.org",
        "time": "Fri Oct 30 23:41:49 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 23:47:42 2026"
      },
      "message": "[MemoryCoordinator] Skip MemoryConsumer registration where suppressed\n\nThe SuppressMemoryListeners experiment showed that reacting to memory\npressure is not beneficial for some listeners on some platforms. Since\nthen, these listeners were migrated to base::MemoryConsumer, which\nsilently made the suppression mask a no-op for them.\n\nThis CL restores the experiment\u0027s behavior by not registering the\nfollowing classes as MemoryConsumers on the platforms where they were\nsuppressed:\n- DiscardableSharedMemoryManager: Mac and Windows.\n- GpuChannelManager: Mac and Windows.\n- SSLClientSessionCache: Mac and Windows.\n- PrerenderHostRegistry: Android.\n\nFrameEvictionManager is intentionally not included, as it is now\ncovered by the kScalableFrameEviction experiment.\n\nNote that not registering also opts these classes out of all memory\ncoordinator policies (e.g. ZeroLimitPolicy), not just memory pressure.\n\nTests that exercise the MemoryConsumer path are compiled out on\nplatforms where the class does not register.\n\nBug: 441951621\nChange-Id: I7f79b5087f00329e7619b9381afa1acc53b6fc85\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8675464\nReviewed-by: Francois Pierre Doray \u003cfdoray@chromium.org\u003e\nCommit-Queue: Patrick Monette \u003cpmonette@chromium.org\u003e\nOwners-Override: Francois Pierre Doray \u003cfdoray@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1732051}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 2f2b3fa082872ad2e67eb55c68a3f3c61dcd496f\n"
    },
    {
      "commit": "cfddf5cde54d0f8361a0f229e8e09680938d3861",
      "tree": "5ce656724dd8489907b796485ad968b0d15db9b6",
      "parents": [
        "43a498d64d5e3bcc131b776cdb8edbe720dd7c71"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Fri Oct 30 17:08:27 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 17:12:57 2026"
      },
      "message": "Reland \"[gpu] Reject multiplanar format with Premul alpha at mojom\"\n\nThis is a reland of commit 8fdd6a879fdec1f385bc3b3369947c4cbc54b14f\n\nDiff from original (PS2-PS1):\nFix to CompositingStructTraitsTest.CompositorFrame to not pass\npremul for multiplanar TransferableResource.\n\nOriginal change\u0027s description:\n\u003e [gpu] Reject multiplanar format with Premul alpha at mojom\n\u003e\n\u003e This change adds rejection at SharedImageMetadata mojo level for\n\u003e multiplanar formats without external sampler that have premul\n\u003e alpha. This is follow-up from CHECK added in [1] that is now causing\n\u003e some fuzzers to be unhappy.\n\u003e\n\u003e [1] https://crrev.com/c/8356445\n\u003e\n\u003e Bug: 565560506\n\u003e Change-Id: I68ee155cc3c11931ff173cd7bd35eba2d6ed57be\n\u003e Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8465102\n\u003e Auto-Submit: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\n\u003e Reviewed-by: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\n\u003e Commit-Queue: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\n\u003e Commit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\n\u003e Reviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\n\u003e Cr-Commit-Position: refs/heads/main@{#1730957}\n\nBug: 565560506\nChange-Id: Ia0aa1555c310854ef56e814a4cbd465728e9a9fd\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8687238\nReviewed-by: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1731614}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 4c7c3799a7179cd379ead550ed44dce787921983\n"
    },
    {
      "commit": "43a498d64d5e3bcc131b776cdb8edbe720dd7c71",
      "tree": "caa7e553173d786f515a35d21f06d0d4b25a1ab0",
      "parents": [
        "3a2252cb7b7ca4b7ff9dfc2f73286f43ce8c4904"
      ],
      "author": {
        "name": "Kramer Ge",
        "email": "fangzhoug@chromium.org",
        "time": "Fri Oct 30 16:22:33 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 16:26:40 2026"
      },
      "message": "[media] Disable D3D11 shared handle decoder on AMD\n\nEnabling D3D11VideoDecoderUseSharedHandle causes single-texture D3D11\nvideo decoders to allocate NV12 output textures with\nD3D11_RESOURCE_MISC_SHARED_NTHANDLE | D3D11_RESOURCE_MISC_SHARED. On\nWindows with modern AMD GPUs (where VP9 and AV1 use single-texture\ndecoding), allocating D3D11_BIND_DECODER textures with\nD3D11_RESOURCE_MISC_SHARED triggers a driver bug that corrupts the top\nrows of decoded video frames with green and purple glare.\n\nAdd the disable_d3d11_video_decoder_shared_handle GPU driver bug\nworkaround for Windows AMD GPUs (0x1002), gate use_shared_handle_ on it\nin D3DVideoDecoder, and re-enable D3D11VideoDecoderUseSharedHandle by\ndefault on non-AMD GPUs.\n\nBug: 402171652, 40270305\nFixed: 577963606\nTest: pixel_skia_gold_passthrough_test on gpu-fyi-try-win11-amd-rel-64\nChange-Id: I59a3720a6643452f1d7585e78fa3dd2030891c71\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8684923\nCommit-Queue: Kramer Ge \u003cfangzhoug@chromium.org\u003e\nReviewed-by: Dale Curtis \u003cdalecurtis@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1731569}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 5229121256d0e987311e19b99ef3ef3710c30a7d\n"
    },
    {
      "commit": "3a2252cb7b7ca4b7ff9dfc2f73286f43ce8c4904",
      "tree": "dad320d0e9cd12e4aee785c91842c39399120b11",
      "parents": [
        "896b2a345ddf03526256c75d523edbadf8dd3097"
      ],
      "author": {
        "name": "Arthur Sonzogni",
        "email": "arthursonzogni@chromium.org",
        "time": "Fri Oct 30 15:07:31 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 15:14:34 2026"
      },
      "message": "Split the GpuControlClient lock in CommandBufferProxyImpl\n\nSince https://crrev.com/c/7953663, OnReturnData() holds\nshared_state_lock_ on the IO thread while calling the client. This\ndeadlocks with the main thread:\n- DawnWireServices::Commit() holds the Dawn lock, then takes\n  shared_state_lock_. OnReturnData() does the opposite.\n- WaitFor*InRange() hold shared_state_lock_ while waiting for a [Sync]\n  reply, which the IO thread can\u0027t forward.\n\nProtect gpu_control_client_ on the IO thread with a new\ngpu_control_client_lock_ instead. The IO thread no longer takes\nshared_state_lock_.\n\nFound by the WebGL dEQP tests with WebGLOnWebGPU\n(https://crrev.com/c/8641193): timeouts went from 5-50% to 0 in 934\nruns.\n\nBug: 413078308, 574515421\nChange-Id: If4cd6d03f425aa56add2f8b0130c48ebd7ff228c\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8657244\nCommit-Queue: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nReviewed-by: Loko Kung \u003clokokung@google.com\u003e\nAuto-Submit: Arthur Sonzogni \u003carthursonzogni@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1731474}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 60cf978a88d18dab8cc29ca95d7cb9f9daabd7df\n"
    },
    {
      "commit": "896b2a345ddf03526256c75d523edbadf8dd3097",
      "tree": "1c0fbf4e3c98fb4aaa7c3e362f4529b960518988",
      "parents": [
        "e3b660fea6c91f2517d5c9b9ed337e7b99984036"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Fri Oct 30 14:46:14 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 14:52:13 2026"
      },
      "message": "[gl] Remove disable_metal entry from workarounds list:\n\nThe driver bug workaround entry was removed a while ago [1], but the\nentry in list and related code is leftover. This change cleans it up.\n\n[1] https://crrev.com/c/4436150\n\nBug: 570597474\nChange-Id: Ic3ba12fae5b887b472cffe211724455c7860d2af\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8664574\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Kenneth Russell \u003ckbr@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1731449}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 79b4eb453663b1e54fe1b788c4e3f96417bc82e3\n"
    },
    {
      "commit": "e3b660fea6c91f2517d5c9b9ed337e7b99984036",
      "tree": "042e85ea18d22a9e5d876b20a2ed2c9a2fad846b",
      "parents": [
        "38c5bc6648609b3effb515fb5eb11e3af6fd6b61"
      ],
      "author": {
        "name": "Achuith Bhandarkar",
        "email": "achuith@chromium.org",
        "time": "Fri Oct 30 10:59:12 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 11:03:58 2026"
      },
      "message": "gpu: Skip sync DestroyCommandBuffer wait during browser shutdown\n\nWhen VizProcessTransportFactory is destroyed during browser teardown,\nits main and worker ContextProviderCommandBuffers destroy their\nCommandBufferProxyImpl instances right before BrowserGpuChannelHostFactory\ncloses the GPU channel. CommandBufferProxyImpl::DisconnectChannel issues\na synchronous GpuChannel::DestroyCommandBuffer Mojo call that blocks the\nbrowser UI thread for 19-33 ms on ChromeOS logout traces and accounts\nfor 0.9% of browser [Shutdown hang] crashes.\n\nAdd CommandBufferProxyImpl::SetChannelShuttingDown() and call it from\n~VizProcessTransportFactory() after shutting down the raster worker\nthread so DisconnectChannel() flushes deferred requests and invokes the\nasynchronous DestroyCommandBuffer overload instead of blocking on the\nGPU reply.\n\nTAG\u003dagy\nCONV\u003dd52e364b-43fb-4209-8dcc-548616246e32\n\nBug: 575822472, 572927279\nChange-Id: Ie20734f7fab2f318ebcd4bca1c946c305f67bcab\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8647894\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCommit-Queue: Achuith Bhandarkar \u003cachuith@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1731314}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: de2e1e816e02b099141d6a4b7b87d15e2706d4ab\n"
    },
    {
      "commit": "38c5bc6648609b3effb515fb5eb11e3af6fd6b61",
      "tree": "1fb72f48d85cb08251942ef239d132589de48524",
      "parents": [
        "0a5b788b6d727c2002ea4142cadca29a343cd5fc"
      ],
      "author": {
        "name": "Markus Handell",
        "email": "handellm@google.com",
        "time": "Fri Oct 30 07:41:01 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 07:45:37 2026"
      },
      "message": "mac: Reach IOSurfaces in VTVideoEncodeAccelerator via ProduceVideo\n\nVTVideoEncodeAccelerator wraps SharedImage-backed frames in a\nCVPixelBuffer by producing an overlay representation and reading its\nIOSurface. SharedImageManager::ProduceOverlay enforces\nSHARED_IMAGE_USAGE_SCANOUT, so SharedImages tagged only with\nMACOS_VIDEO_TOOLBOX (a standalone 2D OffscreenCanvas under\nkWebRtcMacSharedImageEncode) trip the usage enforcement. The encoder\nis not an overlay consumer: it only needs the IOSurface and read\nsynchronization, not the overlay interface or its SCANOUT requirement.\n\nFix this by giving IOSurfaceImageBacking a video representation that\nexposes the IOSurface with the same read synchronization as the overlay\npath (backing lock, read-only BeginAccess, WaitForCommandsToBeScheduled)\nand switching VTVideoEncodeAccelerator to ProduceVideo, which enforces\nno usage. VideoImageRepresentation::ScopedReadAccess gains an IS_APPLE\nGetIOSurface() accessor, forwarded through CompoundImageBacking for\nmappable (capture, video frame pool) images. ProduceOverlay and its\nenforcement are unchanged.\n\nBug: 576960275\nChange-Id: I429e59a011daeafb08610905ce745974244fab19\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8683080\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nReviewed-by: Eugene Zemtsov \u003ceugene@chromium.org\u003e\nCommit-Queue: Markus Handell \u003chandellm@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1731189}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9cdd7dc895376a524cc8f9a20a6b9eebb5c365d8\n"
    },
    {
      "commit": "0a5b788b6d727c2002ea4142cadca29a343cd5fc",
      "tree": "4647ddd6ea506d7f3b670060cd8e7a031dfbac87",
      "parents": [
        "ae3581bec157f8e8e2b70b3d7873ded9faab3c7c"
      ],
      "author": {
        "name": "luci-bisection@appspot.gserviceaccount.com",
        "email": "luci-bisection@appspot.gserviceaccount.com",
        "time": "Fri Oct 30 04:10:37 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 04:24:41 2026"
      },
      "message": "Revert \"[gpu] Reject multiplanar format with Premul alpha at mojom\"\n\nThis reverts commit 8fdd6a879fdec1f385bc3b3369947c4cbc54b14f.\n\nReason for revert:\nLUCI Bisection has identified this change as the cause of a test failure. See the analysis: https://ci.chromium.org/ui/p/chromium/bisection/test-analysis/b/6053149132455936\n\nSample build with failed test: https://ci.chromium.org/b/8666561493776930577\nAffected test(s):\n[://services\\:services_unittests!gtest::CompositingStructTraitsTest#CompositorFrame](https://ci.chromium.org/ui/test/chromium/:%2F%2Fservices%5C:services_unittests%21gtest::CompositingStructTraitsTest%23CompositorFrame?q\u003dVHash%3Aab00eba16afa230f)\n\nIf this is a false positive, please report it at http://b.corp.google.com/createIssue?component\u003d1199205\u0026description\u003dAnalysis%3A+https%3A%2F%2Fci.chromium.org%2Fui%2Fp%2Fchromium%2Fbisection%2Ftest-analysis%2Fb%2F6053149132455936\u0026format\u003dPLAIN\u0026priority\u003dP3\u0026title\u003dWrongly+blamed+https%3A%2F%2Fchromium-review.googlesource.com%2Fc%2Fchromium%2Fsrc%2F%2B%2F8465102\u0026type\u003dBUG\n\nOriginal change\u0027s description:\n\u003e [gpu] Reject multiplanar format with Premul alpha at mojom\n\u003e\n\u003e This change adds rejection at SharedImageMetadata mojo level for\n\u003e multiplanar formats without external sampler that have premul\n\u003e alpha. This is follow-up from CHECK added in [1] that is now causing\n\u003e some fuzzers to be unhappy.\n\u003e\n\u003e [1] https://crrev.com/c/8356445\n\u003e\n\u003e Bug: 565560506\n\u003e Change-Id: I68ee155cc3c11931ff173cd7bd35eba2d6ed57be\n\u003e Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8465102\n\u003e Auto-Submit: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\n\u003e Reviewed-by: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\n\u003e Commit-Queue: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\n\u003e Commit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\n\u003e Reviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\n\u003e Cr-Commit-Position: refs/heads/main@{#1730957}\n\u003e\n\nBug: 565560506\nNo-Presubmit: true\nNo-Tree-Checks: true\nNo-Try: true\nChange-Id: Ia81d26fce31511fdd8547bf8c55decd159480c22\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8684693\nReviewed-by: Darren Shen \u003cshend@chromium.org\u003e\nBot-Commit: rubber-stamper@appspot.gserviceaccount.com \u003crubber-stamper@appspot.gserviceaccount.com\u003e\nOwners-Override: Darren Shen \u003cshend@chromium.org\u003e\nCommit-Queue: Darren Shen \u003cshend@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1731097}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6c14e12d254bb6f90df512be410a058a6b9a7e39\n"
    },
    {
      "commit": "ae3581bec157f8e8e2b70b3d7873ded9faab3c7c",
      "tree": "be02299b6028c6bec558656f201bd3cc4052c345",
      "parents": [
        "00b34aa4b81d66caed179e813bfdbe2a30593f9b"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Fri Oct 30 00:15:32 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Fri Oct 30 00:21:37 2026"
      },
      "message": "[gpu] Reject multiplanar format with Premul alpha at mojom\n\nThis change adds rejection at SharedImageMetadata mojo level for\nmultiplanar formats without external sampler that have premul\nalpha. This is follow-up from CHECK added in [1] that is now causing\nsome fuzzers to be unhappy.\n\n[1] https://crrev.com/c/8356445\n\nBug: 565560506\nChange-Id: I68ee155cc3c11931ff173cd7bd35eba2d6ed57be\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8465102\nAuto-Submit: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\nCommit-Queue: Giovanni Ortuno Urquidi \u003cortuno@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730957}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 8fdd6a879fdec1f385bc3b3369947c4cbc54b14f\n"
    },
    {
      "commit": "00b34aa4b81d66caed179e813bfdbe2a30593f9b",
      "tree": "4647ddd6ea506d7f3b670060cd8e7a031dfbac87",
      "parents": [
        "b256c38ae294ba3a54d1272814ed4cc1aa57184a"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Thu Oct 29 23:16:55 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 23:26:55 2026"
      },
      "message": "[gpu] Allow fuzzer to lose context on out of memory for decoders\n\nAllow losing context on OOM, so that textures with id 0 are not\ngenerated.\n\nBug: 579142416\nChange-Id: I01890dbf8087614b8dc041c47a974abc1ecdc5b6\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8682447\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730901}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 37ca7ee4df9e051cd4357268464bbd05c0c166f4\n"
    },
    {
      "commit": "b256c38ae294ba3a54d1272814ed4cc1aa57184a",
      "tree": "b207d0e82b842c9e9f557827890b7e8385e6363f",
      "parents": [
        "193e438a78bfdeb193c6c5d1c8cf8b12401ff8f7"
      ],
      "author": {
        "name": "kylechar",
        "email": "kylechar@chromium.org",
        "time": "Thu Oct 29 21:54:17 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 22:03:05 2026"
      },
      "message": "Fix WrappedSkImageBackingFactoryTest.AccessExclusivity\n\nThe test was failing as access exclusivity was disabled on Chrome OS due\nto overlapping access found in production. Move the feature to a header\nand skip the test when feature is disabled.\n\nBug: 574220812\nChange-Id: Ief7f988a0850eccfe85323d2ce62ade48a4541f5\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8686291\nCommit-Queue: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nReviewed-by: Kramer Ge \u003cfangzhoug@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730790}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 17956d13ebba29ca98ebbad23967e00f2402b7d7\n"
    },
    {
      "commit": "193e438a78bfdeb193c6c5d1c8cf8b12401ff8f7",
      "tree": "134b5dd8e01620b06e5dbb5e3e5a18c085db5cd7",
      "parents": [
        "5ec9d5041ec03ec3bb771b3d84eb19550506a9e6"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Thu Oct 29 20:40:07 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 20:46:26 2026"
      },
      "message": "[gpu] Use GpuFeatureInfo::IsFeatureEnabled in gpu init 2\n\nIsFeatureEnabled on GpuFeatureInfo can check for status as\nkGpuFeatureStatusEnabled. Use the helper in gpu_init_2 so that it\nis simpler.\n\nBug: 570527192\nChange-Id: Ia438cdf30e07aecb3f71fa8072e051cad989dea9\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8668549\nReviewed-by: Zhenyao Mo \u003czmo@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730707}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 996204957e319e7a5b612af0512fb72ced6540c4\n"
    },
    {
      "commit": "5ec9d5041ec03ec3bb771b3d84eb19550506a9e6",
      "tree": "967690c8e60a8e143ab74fc0b47d03a19708c9cc",
      "parents": [
        "0127c047cc5105544ab3352baede3d3e6549fa52"
      ],
      "author": {
        "name": "Le Hoang Quyen",
        "email": "lehoangquyen@chromium.org",
        "time": "Thu Oct 29 18:15:26 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 18:26:13 2026"
      },
      "message": "Apply check_graphics_reset_status_after_readback to all Mali GPUs\n\nBroaden gl_renderer matching for\ncheck_graphics_reset_status_after_readback from Mali-G.* to Mali.* so\nthe workaround applies to all Mali GPUs.\n\nFixed: 558111285\nChange-Id: Id8814e4557fa6450444de4a43a7cceb2da61d184\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8684861\nReviewed-by: Colin Blundell \u003cblundell@chromium.org\u003e\nCommit-Queue: Quyen Le \u003clehoangquyen@google.com\u003e\nCr-Commit-Position: refs/heads/main@{#1730544}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: f6a48fca88424de3d2ffef86ae2e859ec6916895\n"
    },
    {
      "commit": "0127c047cc5105544ab3352baede3d3e6549fa52",
      "tree": "6578be4d4e203334da1d6c52d405764317d3211c",
      "parents": [
        "13b1172cb74b7a0ce45008be14d2b8648741243e"
      ],
      "author": {
        "name": "Vasiliy Telezhnikov",
        "email": "vasilyt@chromium.org",
        "time": "Thu Oct 29 17:50:01 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 17:56:21 2026"
      },
      "message": "Disallow SharedImage texture redefinition\n\nBug: 574216445\nChange-Id: Ia043db083457ee2a4f826a540a979dbef619b630\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8671208\nReviewed-by: Kyle Charbonneau \u003ckylechar@chromium.org\u003e\nCommit-Queue: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730515}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 57275c36ae9dec207d1892a775a83cfc084cb989\n"
    },
    {
      "commit": "13b1172cb74b7a0ce45008be14d2b8648741243e",
      "tree": "e08275c1600c730211af9d018b9d8ebf54d2cc65",
      "parents": [
        "d07c09f1150e3f9a460758bc894a144fd6c02c6a"
      ],
      "author": {
        "name": "Ilya Nikolaevskiy",
        "email": "ilnik@chromium.org",
        "time": "Thu Oct 29 14:27:54 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 14:33:03 2026"
      },
      "message": "[gpu] Run bottom left origin ReadPixelsYUV test on Graphite\n\nBug: 570097637\nChange-Id: I93c2cdb0eae34ba9ed7245016fd38d24a341d0c0\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8616339\nReviewed-by: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCommit-Queue: Ilya Nikolaevskiy \u003cilnik@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730296}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 28624b0f36f5d2e5974693ffcdf9239c0caf5e93\n"
    },
    {
      "commit": "d07c09f1150e3f9a460758bc894a144fd6c02c6a",
      "tree": "dadab5391d986b8b9bf9dd2d41d4e7b60ee02a53",
      "parents": [
        "59ffe19a8e8b0c31cb6d62baa63142b517c68fb4"
      ],
      "author": {
        "name": "Vasiliy Telezhnikov",
        "email": "vasilyt@chromium.org",
        "time": "Thu Oct 29 13:54:59 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 14:00:30 2026"
      },
      "message": "Remove unnecessary CanvasResource::PrepareTransferableResource from tests\n\n* EnsureResourceRecycled doesn\u0027t use the result and\n  PrepareTransferableResource doesn\u0027t modify any state.\n\n* CanvasResourceTest essentially test that SahredImagePool provides\n  correct shared image and that TransferableResource matches metadata.\n\n  SharedImagePool already has tests, add alpha type check there.\n  TransferableResource testing is pointless, as it just calls\n  shared image getters at this point.\n\nChange-Id: I5d4c7c66ec01a572c261b278a8a23883b77879ed\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8673444\nReviewed-by: Colin Blundell \u003cblundell@chromium.org\u003e\nCommit-Queue: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730262}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 6409eb9e31c90c54650df781438b937fbf1bf37b\n"
    },
    {
      "commit": "59ffe19a8e8b0c31cb6d62baa63142b517c68fb4",
      "tree": "e1fb7ac65d87235e440e0542294fb0551e272c0d",
      "parents": [
        "0b54161ee0b9af42e5a9c4827cd9dd94c7373403"
      ],
      "author": {
        "name": "Saifuddin Hitawala",
        "email": "hitawala@chromium.org",
        "time": "Thu Oct 29 13:53:37 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 13:59:22 2026"
      },
      "message": "[media] Use SharedImage alpha for TransferableResource in VRU\n\nThe Check in SIInfo has had no issues in 156 and can be fatal now.\nUpdate VideoResourceUpdater to no longer override alpha type, as\nwe can use that from SharedImage instead behind a killswitch.\n\nBug: 390568190\nChange-Id: I09e3faae5013716dadc73d2e43d622559918116a\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8664116\nReviewed-by: Vasiliy Telezhnikov \u003cvasilyt@chromium.org\u003e\nCommit-Queue: Saifuddin Hitawala \u003chitawala@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730260}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 9f2c6a67d3886e821d2ef5c817cdd637a5503e84\n"
    },
    {
      "commit": "0b54161ee0b9af42e5a9c4827cd9dd94c7373403",
      "tree": "a8c376c9867ee0453ffe351ea6f9dc2603b22da5",
      "parents": [
        "d8b4d517ae45983516f5c9e8327217b430a7ecad"
      ],
      "author": {
        "name": "Qiu Jianlin",
        "email": "jianlin.qiu@intel.com",
        "time": "Thu Oct 29 13:22:11 2026"
      },
      "committer": {
        "name": "Copybara-Service",
        "email": "copybara-worker@google.com",
        "time": "Thu Oct 29 13:31:07 2026"
      },
      "message": "viz/media: Add Y410 single-plane shared image and packed pixel format\n\nAdd viz::SinglePlaneFormat::kY410, media::PIXEL_FORMAT_Y410 and the\nsurrounding plumbing for packed 10-bit 4:4:4 YUV shared images. The\ncorresponding Y410 support has been added to Skia/ANGLE for Ganesh and\nGraphite.\n\nY410 is sampled through the aliased R10G10B10A2 DXGI view and rendered\nvia the Skia YUVA path (kUYVA) in follow-up CLs. No producer or\nconsumer is added here, so this is plumbing with no behavior change.\n\nY210 (packed 10-bit 4:2:2) is added in a follow-up CL together with\nSK_ENABLE_YUVA_PACKED_422, since it can only be sampled and sized\ncorrectly through the half-width kYUYV plane config.\n\nBug: 537818862\nChange-Id: Iacbc89be1c5e0b11ec7b4ca69ef7843649627457\nReviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8578257\nCommit-Queue: Qiu, Jianlin \u003cjianlin.qiu@intel.com\u003e\nReviewed-by: Dale Curtis \u003cdalecurtis@chromium.org\u003e\nReviewed-by: Jonathan Ross \u003cjonross@chromium.org\u003e\nReviewed-by: Ilya Nikolaevskiy \u003cilnik@chromium.org\u003e\nReviewed-by: Johannes Kron \u003ckron@chromium.org\u003e\nReviewed-by: Joe Mason \u003cjoenotcharles@google.com\u003e\nReviewed-by: Maggie Chen \u003cmagchen@chromium.org\u003e\nCr-Commit-Position: refs/heads/main@{#1730234}\nNOKEYCHECK\u003dTrue\nGitOrigin-RevId: 1056cfa6700e8344b1e55875c565b8ff326b739c\n"
    }
  ],
  "next": "d8b4d517ae45983516f5c9e8327217b430a7ecad"
}
