)]}'
{
  "commit": "915b5e33bc568f06182fece3dc51962ffb1cc1c0",
  "tree": "2ae610d4490b1318d1d6b07ec518dae1ab1d3ed2",
  "parents": [
    "ebf49e8218939929018dad8a54a6c109a08e6e5b"
  ],
  "author": {
    "name": "Ryan Harrison",
    "email": "rharrison@chromium.org",
    "time": "Tue Nov 03 05:51:31 2026 -0800"
  },
  "committer": {
    "name": "dawn-scoped@luci-project-accounts.iam.gserviceaccount.com",
    "email": "dawn-scoped@luci-project-accounts.iam.gserviceaccount.com",
    "time": "Tue Nov 03 05:51:31 2026 -0800"
  },
  "message": "Roll third_party/mesa/src/ f950de848..b5b73e596 (681 commits)\n\nhttps://chromium.googlesource.com/external/gitlab.freedesktop.org/mesa/mesa/+log/f950de848de8..b5b73e5964be\n\n$ git log f950de848..b5b73e596 --date\u003dshort --no-merges --format\u003d\u0027%ad %ae %s\u0027\n2026-11-01 popcornmix v3d: don\u0027t drop the last row in the sand30 blit\n2026-10-29 jianxun.zhang iris: Support sample_d_performance dri option (GFX35+)\n2026-10-29 jianxun.zhang intel: Add a dri option for sample_d_performance (GFX35+)\n2026-10-28 jianxun.zhang anv: Add an option for sample_d_performance (GFX35+)\n2026-10-27 jianxun.zhang intel/genxml: Add SAMPLER_STATE::sample_d_performance (GFX35+)\n2026-10-26 triang3l r600: Implement alpha to coverage dither control on pre-EG\n2025-12-10 nanley.g.chery isl: Choose linear for single row 2D images on Xe3p+\n2022-04-25 nanley.g.chery iris: Drop the extra 4K alignment on surfaces\n2026-11-02 michael.blumenkrantz zink: fuse mul and add into ffma_weak on RADV\n2026-11-02 kocdemir gfxstream: only reset test connection when the last VkInstance is destroyed\n2026-10-23 zzyiwei vulkan/wsi/x11: drop x11_requires_mailbox_image_count helper\n2026-10-23 zzyiwei vulkan/wsi/x11: drop x11_needs_wait_for_fences helper\n2026-10-23 zzyiwei vulkan/wsi/x11: use resolved present mode to calculate image count\n2026-11-02 samuel.pitoiset vulkan: update spec to 1.4.367\n2026-09-02 caio.oliveira brw: Pass brw_compile_params down to variant compilation\n2026-10-08 caio.oliveira jay: Fail if scratch size after RA is too large\n2026-08-31 caio.oliveira jay: Propagate shader compilation errors\n2026-10-15 emreleno jay: restrict accumulator modifiers before Gfx12.5\n2026-10-14 emreleno jay: restrict accumulator combinations before Gfx12.5\n2026-10-14 emreleno jay: restrict accumulator source positions before Gfx12.5\n2026-10-14 emreleno jay: limit accumulator allocation before Gfx12.5\n2026-10-28 kenneth jay: Lower ALU to scalar before filling out prog_data\n2026-10-16 calder.young jay: Use Xe2 immediate offset bits for spills and fills\n2026-09-29 calder.young jay: Encode the mlen and ex_mlen portions of descriptors post-RA\n2026-10-02 calder.young jay: Add a variant of jay_shrink_sources for ISA sources only\n2026-09-29 calder.young jay: move lowering call out of nested if statement\n2026-10-16 calder.young gen: Add helper for encoding bindless ex_desc immediates\n2026-10-23 alyssa jay: fix uniform stores in divergent frag shaders\n2026-10-01 calder.young util: Add more ways to delete from dynarray\n2026-10-20 cgmeiner asahi: Use util_dynarray_grow_zero(..)\n2026-10-20 cgmeiner agx: Use util_dynarray_grow_zero(..)\n2026-10-19 cgmeiner jay: Use util_dynarray_grow_zero(..)\n2026-10-19 cgmeiner panfrost: Use util_dynarray_grow_zero(..)\n2026-10-19 cgmeiner pan/compiler: Use util_dynarray_grow_zero(..)\n2026-10-19 cgmeiner util/dynarray: Add util_dynarray_grow_zero(..)\n2026-08-28 david.rosca va: Unmap coded buffer on encode failure\n2026-10-29 david.rosca radeonsi/mm: Disable unsupported EFC conversions on VCN5 with old FW\n2026-10-29 david.rosca radv/video: Use EFC range bug cap\n2026-10-29 david.rosca ac/video: Add EFC range bug cap\n2026-10-30 alyssa panfrost: don\u0027t use helpers for memory barriers\n2026-11-01 genie1997.dev tu/kgsl: Force 8-channel macrotile mode on A635\n2026-10-29 dtgs1208 panvk: Add labels to internal driver BOs\n2026-11-02 samuel.pitoiset zink/ci: mark one test as a regression on POLARIS10/NAVI10\n2026-10-15 arjob.mukherjee pvr: Add support for linear transfer for ETC2 formats\n2026-09-25 arjob.mukherjee pvr: Add support for VK_KHR_load_store_op_none\n2026-10-26 kocdemir gfxstream: simplify guest Vulkan codegen and marshaling\n2026-10-28 christoph.pillmayer kraid: Add Swizzle::bits_read\n2026-10-20 christoph.pillmayer kraid/ra: Fix round robin search start going OOB\n2026-10-14 christoph.pillmayer kraid: Tightly pack FAU values\n2026-09-15 christoph.pillmayer kraid: FAU allocator\n(...)\n2026-10-22 michael.blumenkrantz util: restore the launch affinity in threads that shed a pinned one\n2026-10-22 kocdemir gfxstream: rerun codegen for new extensions\n2026-10-19 natsu gfxstream: introduce guest capabilities struct\n2026-10-19 caio.oliveira spirv: Fix is_null_constant annotation for aggregates with pointers inside\n2026-08-21 aleksi.sapon llvmpipe: avoid running helper invocations when not needed\n2026-10-20 aleksi.sapon llvmpipe: improve exec-mask-is-not-zero logic\n2026-10-19 dylan.c.baker meson: imgui-sdl3-vulkan needs the vulkan loader headers\n2026-10-19 gustavo zink: keep the renderpass atom dirty for zink_batch_rp()\n2026-06-16 karmjit.mahil tu: Set prefer_gmem for Sweet Surrender VR\n2025-10-21 mark tu: Prefer GMEM on Beat Saber\n2026-10-19 karmjit.mahil tu: Add back float16 3D path for multisampled copies\n2026-10-22 erik.faye-lund panvk: write_image_view_desc deals with images, not samplers\n2026-10-06 mark util/u_trace: Add tests for u_trace_clone_append\n2026-08-28 mark util/u_trace: Copy only the live timestamp range when cloning\n2026-10-19 anholt meson: Enable split-dwarf on subprojects, as well.\n2026-10-21 dpiliaiev freedreno: Fixup incorrectly copied A6XX_TPL1_DBG_ECO_CNTL for A650\n2026-10-21 dpiliaiev tu/kgsl: Fix BDA capture/replay failures due to deferred VA frees\n2026-09-09 dpiliaiev tu: Apply tu_enable_ssbo_emulation workaround to all UE5 games\n2026-09-22 dpiliaiev driconf: Support \"application_versions\" attribute on \u003cengine\u003e\n2026-09-22 dpiliaiev driconf: Allow ranges to match a single value\n2026-10-01 cgmeiner panvk: Advertise VK_KHR_extended_flags\n2026-10-19 cgmeiner panvk: Implement VK_KHR_extended_flags\n2026-10-19 cgmeiner panvk: Convert image flags to flags2\n2026-09-30 cgmeiner vulkan: Add vk_image_format_info_2_stencil_usage(..)\n2026-10-16 martin.roukala zink/ci: add piglit cl tests when running on RADV\n2026-10-21 martin.roukala zink/ci: update nvk expectations\n2026-10-21 martin.roukala zink/ci: limit cl testing to one device for NVK\n2026-10-21 martin.roukala zink/ci: fix the indentation of the [piglit.env] in suites\n2026-10-10 maraeo radv: fix gang submit with protected command buffers\n2026-09-20 maraeo radv: use the VRAM suballocator for gfx12 streamout state\n2026-09-20 maraeo radv: use the VRAM suballocator for the unaligned indirect dispatch workaround\n2026-09-20 maraeo radv: use the VRAM suballocator for indirect copies\n2026-09-20 maraeo radv: use the VRAM suballocator for CP DMA realignment\n2026-09-20 maraeo radv: use the VRAM suballocator for EOP buffers\n2026-10-10 maraeo radv: add a linear VRAM suballocator without CPU access\n2026-10-10 maraeo radv: always set GTT_USWC for VRAM\n2026-09-20 maraeo radv: use the correct SMEM cache line size\n2026-10-10 maraeo radv: remove radv_cmd_buffer::upload in favor of the array\n2026-09-20 maraeo radv: return a GPU address from upload functions instead of the offset\n2026-09-20 maraeo radv: never pass alignment 0 to radv_cmd_buffer_upload_alloc\n2026-09-20 maraeo radv: prepare for an arbitrary number of linear suballocators\n2026-09-20 maraeo radv: add alignment param to all upload functions\n2026-09-20 maraeo radv: pass radv_linear_suballoc_buffer* to suballoc functions\n2026-09-20 maraeo radv: rename radv_cmd_buffer_upload -\u003e radv_linear_suballoc_buffer\n2026-10-10 maraeo radv: rename radv_cmd_buffer_upload::map and upload_bo\n2026-10-19 samuel.pitoiset radv: set radv_max_clip_discard_watermark\u003d6.0 for Zink\n2026-10-19 samuel.pitoiset radv: add radv_max_clip_discard_watermark drirc\n2026-10-19 samuel.pitoiset radv: dirty the guardband state when the clip discard distance changes\n2026-10-16 osmnyagc util/u_trace: only create the queue when traces need processing\n2026-10-19 dpiliaiev anv: Call anv_queue_trace only where there is something to process\n\nCreated with:\n  roll-dep third_party/mesa/src\n\nChange-Id: I144155cc809e20dec59fe61760716edf052ad800\nReviewed-on: https://dawn-review.googlesource.com/c/dawn/+/355576\nAuto-Submit: Ryan Harrison \u003crharrison@chromium.org\u003e\nReviewed-by: James Price \u003cjrprice@google.com\u003e\nCommit-Queue: Ryan Harrison \u003crharrison@chromium.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "bdead01f36c6534e01534d3bf75a596bd0055459",
      "old_mode": 33188,
      "old_path": "DEPS",
      "new_id": "c2849a85f4bb54e5de1e91113675d3509c62f6e8",
      "new_mode": 33188,
      "new_path": "DEPS"
    },
    {
      "type": "modify",
      "old_id": "f950de848de8bbb4a626ee62ff444e0d87cc911c",
      "old_mode": 57344,
      "old_path": "third_party/mesa/src",
      "new_id": "b5b73e5964bee5d9cdc2b7e813e791d948094e75",
      "new_mode": 57344,
      "new_path": "third_party/mesa/src"
    }
  ]
}