)]}' { "log": [ { "commit": "5bb4c7e83acea0fc73e3333625e4f20542e584d0", "tree": "f17ff30057695894269d5fb533b86b95cdc1a322", "parents": [ "a201f446aa067319cb2b2f8b221bb40a2779a7fa", "a6107dc3bd05f3eebd49d497d0535291c4bb6f2c" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 13 09:44:45 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 13 09:44:45 2026" }, "message": "Merge remote-tracking branch \u0027vk-gl-cts/dev/VK_EXT_vertex_address_binding_unrestricted\u0027 into main\n\nChange-Id: I8ebb5cd84a0de1fdee04675183aa94252f5cb03c\n" }, { "commit": "a201f446aa067319cb2b2f8b221bb40a2779a7fa", "tree": "e72abec21e031f720a81a7ec2cef12785a271693", "parents": [ "107d495abb70dcc8a0639ae96d5f91a2a357f470", "9851acaba2c342cb496eb18dbeccaacdeefbd14a" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 13 08:10:52 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 13 08:10:52 2026" }, "message": "Merge vk-gl-cts/opengl-cts-4.6.8-es-3.2.14 into vk-gl-cts/main\n\nChange-Id: Iafbd2e70f22c38ec9513c953e7b99838f838f2cd\n" }, { "commit": "9851acaba2c342cb496eb18dbeccaacdeefbd14a", "tree": "c7512bcbacbddc4efc982061e580fa4c5163a65f", "parents": [ "a4d91d846d441cecb09c9b0ab04e5d8dd711ab60" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 23 15:29:24 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 13 08:10:40 2026" }, "message": "Fail texture view reference counting on a glTexSubImage2D error (clang-analyzer-core.NonNullParamChecker)\n\nWhen glTexSubImage2D failed, the test freed its texel buffer and checked a second glGetError, which returns GL_NO_ERROR, so it went on and wrote the next mip level into the freed buffer, now a null pointer. It now fails with TCU_FAIL, like the glGetTexImage check in the same file.\n\nAffects: KHR-GL4*.texture_view.reference_counting\n\nComponents: OpenGL\n\nVK-GL-CTS issue: 7141\n\nChange-Id: Ifbd85d561e4aff7801663279756cf35d44724949\n" }, { "commit": "a4d91d846d441cecb09c9b0ab04e5d8dd711ab60", "tree": "cf3118b8eba6370a7a9414a53952fea20e589535", "parents": [ "fc08192de01d0d3b629d2e1a5a162b8d9a5abf07" ], "author": { "name": "nikolaysmirnov", "email": "nikolaysmirnov@google.com", "time": "Fri Nov 08 12:08:20 2024" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 13 08:10:37 2026" }, "message": "Fix result verification in framebuffer_fetch test\n\nThe test was checking correctness of buffers, but was ignoring the\noutput and was overwritting it in subsequent checks. This could cause\ntest to pass even though some checks failed.\n\nVK-GL-CTS issue: 5451\nComponents: OpenGL ES\n\nAffects:\ndEQP-GLES31.functional.shaders.framebuffer_fetch.basic.\nmultiple_render_targets\n\nChange-Id: I3800f5eb2464b84c33062baa08079a7218cf283f\n" }, { "commit": "107d495abb70dcc8a0639ae96d5f91a2a357f470", "tree": "55f1bb10dc69df393b0fb1b57c427eac3000db1b", "parents": [ "303742475b9b1340ba1ab8919b94c4954f2354d3" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Tue Oct 27 11:08:09 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 19:16:38 2026" }, "message": "Test advanced blend with msaa\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7130\n\nNew and affected tests:\ndEQP-VK.pipeline.*.blend_operation_advanced.*\ndEQP-VKSC.pipeline.monolithic.blend_operation_advanced.*\n\nChange-Id: Ie66e858761371f5d092bca4c7135254d2cc71160\n" }, { "commit": "303742475b9b1340ba1ab8919b94c4954f2354d3", "tree": "f686b595b985ccd7a953373436fad8926c5e66c0", "parents": [ "c9458ce10f021da5cefdb1e1c19704633cb0249b", "0ca6ff0d4927c5ff7a23a1b9b25ebb9e06d1566f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:57:11 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:57:11 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.7 into vk-gl-cts/main\n\nChange-Id: I4efbcdda9eab7a445fdb2054f1b3e33b284cf990\n" }, { "commit": "0ca6ff0d4927c5ff7a23a1b9b25ebb9e06d1566f", "tree": "c0c8cfef4751df0b33ce2af31b3dfe34f7e3330a", "parents": [ "04f7830d1fa9bfa029e478df3a7102ac9348f46f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Sun Nov 08 21:54:33 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:57:05 2026" }, "message": "Write no binary registry for spirv-binaries fractions\n\nIn fraction mode every vk-build-programs process wrote a binary registry that nothing reads. An empty --dst-path now makes vk-build-programs skip the registry, and build_spirv_binaries.py passes it for fractions. Measured locally, 32 fractions on 16 CPUs: 4% less time and CPU, 10% less peak memory per process.\n\nAffects: none\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7224\n\nChange-Id: I24f7818c221b8fd5804d194e08c8004007edf103\n" }, { "commit": "04f7830d1fa9bfa029e478df3a7102ac9348f46f", "tree": "05916169c2bf64d3467cc85de4d5d3947aaae88c", "parents": [ "f621697fb7b60b48b72c29ff336be69b922e497f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 19:51:20 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:57:02 2026" }, "message": "Fix host memory use and size limits in the 64-bit indexing tests\n\nThe SSBO layout tests write zeros to their host reference buffers when they create them, two buffers of about 4 GB for each 64-bit indexing case, although only the last elements of each array are used. Take the buffers from calloc, which returns memory that is already zero, so only the pages that are used get memory.\n\nThe SSBO, unsized array length, compute and cooperative matrix 64-bit indexing tests create buffers larger than 4 GB, but checked at most the heap size. A buffer larger than maxBufferSize is invalid (VUID-VkBufferCreateInfo-size-06409), and an allocation larger than maxMemoryAllocationSize may fail. Report NotSupported when either limit is smaller than the buffer the test creates.\n\nAffects:\ndEQP-VK.ssbo.layout.*\ndEQP-VK.ssbo.phys.layout.*\ndEQP-VK.ssbo.readonly.layout.*\ndEQP-VK.ssbo.unsized_array_length.*64b*\ndEQP-VK.compute.*.64b_indexing.copy_ssbo_64b*\ndEQP-VK.compute.*.cooperative_matrix.64b_indexing.*\ndEQP-VKSC.ssbo.layout.*\ndEQP-VKSC.ssbo.phys.layout.*\ndEQP-VKSC.ssbo.readonly.layout.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7010\n\nChange-Id: I04cdb2e2ae442ba71c669255edefa9c4858df467\n" }, { "commit": "f621697fb7b60b48b72c29ff336be69b922e497f", "tree": "07e3e512324aab9c14f0996fdb5755f89889c8e8", "parents": [ "91d9281d2484e713ae22f1525278b3f1ecaca13b" ], "author": { "name": "Maciej Jesionowski", "email": "Maciej.Jesionowski@amd.com", "time": "Mon Oct 26 17:14:38 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:59 2026" }, "message": "Avoid invalid OpConvertFToU usage in CMM1 transpose tests\n\nConverting from a negative float to an unsigned integer\nis undefined behavior.\n\nAffects:\ndEQP-VK.compute.*.cooperative_matrix.*transpose_maint1_*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7143\n\nChange-Id: I0e51e68b486d162692004efb6ebdf9b9164eb8d0\n" }, { "commit": "91d9281d2484e713ae22f1525278b3f1ecaca13b", "tree": "31783db61d67403429048b372af1e3a1fda7707b", "parents": [ "7e35ca17e6db4d894ac9d1376c7d92a074377ffe" ], "author": { "name": "Embla Flatlandsmo", "email": "embla.flatlandsmo@arm.com", "time": "Thu Oct 29 14:27:12 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:56 2026" }, "message": "Fix validation layer errors for cooperative matrix tests\n\nComponents: Vulkan\n\nAffects: dEQP-VK.compute.*.cooperative_matrix.*\n\nVK-GL-CTS issue: 7163\n\nChange-Id: I6a5e53c8ad656ba087925332008cb46a0ba76096\n" }, { "commit": "7e35ca17e6db4d894ac9d1376c7d92a074377ffe", "tree": "4bf74a3f0aedddbd7eadf3b56cdca58fc7c58b3f", "parents": [ "ee8f0d6ad1cb67159800b1667df699fafd79bef0" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 05:08:02 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:53 2026" }, "message": "Enable the extensions of the present modes in swapchain tests\n\nThe swapchain present_mode tests created a swapchain with every present mode the surface reports, but used the shared present modes without enabling VK_KHR_shared_presentable_image and with a minImageCount other than 1, and the protected memory variant used VK_PRESENT_MODE_FIFO_LATEST_READY_KHR without enabling VK_KHR_present_mode_fifo_latest_ready. Enable VK_KHR_shared_presentable_image when supported, use a minImageCount of 1 for the shared present modes, and enable VK_KHR_present_mode_fifo_latest_ready or VK_EXT_present_mode_fifo_latest_ready with its feature on the protected memory device when supported.\n\nAffects:\ndEQP-VK.wsi.*.swapchain.*\ndEQP-VK.protected_memory.interaction.wsi.*.swapchain.create.present_mode\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7082\n\nChange-Id: I60e358b04273ec197807375433e299895b56154e\n" }, { "commit": "ee8f0d6ad1cb67159800b1667df699fafd79bef0", "tree": "ab3674ff11e4d92fb330bd5847ba91f709d010dc", "parents": [ "839ab9e84a68e27d7be9d1a7d78124529e2237b2" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 21:15:02 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:50 2026" }, "message": "Do not chain the input attachment index info twice in frag_shader_only_color\n\nBuilding the read pipeline links its VkRenderingInputAttachmentIndexInfo to its VkRenderingAttachmentLocationInfo, so the inheritance chain of the read secondary command buffer, which adds the same input attachment index info after a copy of the location info, holds VkRenderingAttachmentLocationInfo twice (VUID-VkCommandBufferInheritanceInfo-sType-unique). Chain a copy without pNext.\n\nAffects: dEQP-VK.renderpasses.dynamic_rendering.partial_secondary_cmd_buff.local_read.frag_shader_only_color\nComponents: Vulkan\nVK-GL-CTS issue: 7128\nChange-Id: Id3384fb5107f3a3f36bae78c567c2d229f8e27e5\n" }, { "commit": "839ab9e84a68e27d7be9d1a7d78124529e2237b2", "tree": "5167b1963e2e4f10b452d695272023aed73c31cc", "parents": [ "ceb7737575fadb71079c9462a7cbe1bd780ed1bb" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Sun Oct 18 16:03:48 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:47 2026" }, "message": "Use divisor 1 for per-vertex bindings in vkCmdSetVertexInputEXT\n\nThe legacy vertex attribute, DGC dynamic vertex input and mesh shader vertex state tests set divisor 0 on VK_VERTEX_INPUT_RATE_VERTEX bindings, which VUID-VkVertexInputBindingDescription2EXT-divisor-06227 forbids.\n\nAffects:\ndEQP-VK.pipeline.*.vertex_input.legacy_vertex_attributes.*\ndEQP-VK.dgc.ext.graphics.misc.dynamic_vertex_input_*\ndEQP-VK.mesh_shader.ext.misc.vertex_state_*dynamic_vtx_input*\nComponents: Vulkan\nVK-GL-CTS issue: 7114\nChange-Id: I09fbcafc63d0d316c5e60286e17669c2d76f9004\n" }, { "commit": "ceb7737575fadb71079c9462a7cbe1bd780ed1bb", "tree": "3a351fff1d2ee0ec6c5ea0f6d9ba2692e19831b3", "parents": [ "439bf62b60b21b02608f3a308fa443fc8ce4f4d3" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 09 21:29:54 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:44 2026" }, "message": "Stop the device fault async query thread on every exit path\n\nWhen the forced fault does not return VK_ERROR_DEVICE_LOST, or throws, the test leaves while its query thread keeps polling, and the destructor of the std::future waits for the thread forever. Clear the flag the thread reads in a scope guard, so the thread stops on every exit path, and make the flag atomic, as it is shared between threads.\n\nAffects:\ndEQP-VK.postmortem.device_fault.*.async_query\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7011\n\nChange-Id: I331412669af74aa1ae5512b7c60f96d0af4d18c0\n" }, { "commit": "439bf62b60b21b02608f3a308fa443fc8ce4f4d3", "tree": "81a0927615167f5336dd09e58c1be3c6420c3619", "parents": [ "bdca76faace48566b401c33a3b1ac936b4cfff66" ], "author": { "name": "Mengyang Liu", "email": "MengYang.Liu@amd.com", "time": "Thu Oct 29 06:13:42 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:56:41 2026" }, "message": "Fix descriptor stride for texel buffer in descriptor heap tests\n\nWe should use imageDescriptorSize instead of bufferDescriptorSize for texel buffer.\n\nAffects:\ndEQP-VK.sparse_resources.buffer.misc.null_address_*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7160\n\nChange-Id: I5ddb8ffbe0d45973392c20b90125196aa0003d8a\n" }, { "commit": "bdca76faace48566b401c33a3b1ac936b4cfff66", "tree": "d174fa1372671d20ac9f603b53d3aa02b9ad3bb4", "parents": [ "9f5a7bd7eb11fa51604ee7bcc3050f406b52e496", "ea50b86c693c41d5be842316dd8a82687edb56c7" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:46:29 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:46:29 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.6 into vk-gl-cts/vulkan-cts-1.4.7\n\nChange-Id: I5777e8480fe9827c2952f29a4c39e2e8fc8bbcae\n" }, { "commit": "ea50b86c693c41d5be842316dd8a82687edb56c7", "tree": "00dfdaede9192fca99bfb5c5425213d736397a24", "parents": [ "bf7595b943f2dd5d6e917a2631c372d033efdbbb" ], "author": { "name": "MateuszGawrysiak", "email": "mateusz.gawrysiak@intel.com", "time": "Thu Nov 05 13:21:55 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:46:22 2026" }, "message": "Fixing dEQP-VK.binding_model.descriptor_buffer push_descriptor tests\n\nFixing currentBuffer.setOffset for push descriptor buffer creation\n\nVK-GL-CTS issue: 7205\n\nAffects: dEQP-VK.binding_model.descriptor_buffer.*.push_descriptor.*\n\nComponents: Vulkan\nChange-Id: Ifae5283662f7df9bff3d2b976a4ea49b9cac9205\n" }, { "commit": "bf7595b943f2dd5d6e917a2631c372d033efdbbb", "tree": "bb4c1d2dd1504d7ea81a49bd7906a8593a5158bf", "parents": [ "429fb9329ae6aff865cf8bb7c0097170b75cc54a" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Fri Oct 23 13:55:28 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:46:20 2026" }, "message": "Fix multiple issues in multisample conservative rasterization tests\n\nThis commit fixes two issues:\n\n* Sync validation issues related to layout transitions in attachments,\n fixed by serializing subpasses through dependencies and tuning stage\n and access masks. These are write-after-write and write-racing-write.\n\n* Precision issue with a D16_UNORM depth buffer that is cleared to\n 0.5, which cannot be represented correctly and leads to test failures\n under some circumstances.\n\nAffects:\ndEQP-VK.pipeline.*.multisample.*conservative*\n\nAssisted-by: Claude Code\nChange-Id: I72eef1181732d32b118af1e526807b26f946fa5f\n" }, { "commit": "429fb9329ae6aff865cf8bb7c0097170b75cc54a", "tree": "771331938a1cd5964e9e7efeedb0c1225e9878e8", "parents": [ "92c2133aca5f6b3e046d1b113bc202a40d6a8d78" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 12:18:56 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:46:17 2026" }, "message": "Fix leaks on failure in data graph cache and shader binary tests\n\nThe data graph cache and incompatible shader binary tests leaked a pipeline or shader object when they failed, so VUID-vkDestroyDevice-device-05137 was reported at process end against an unrelated case.\n\nAffects: dEQP-VK.data_graph.cache.create_pipeline.*\nAffects: dEQP-VK.shader_object.binary.incompatible.*\nComponents: Vulkan\nVK-GL-CTS issue: 7092\nChange-Id: I2b4c8b5ab89ac56962385124a8c6adba124d53de\n" }, { "commit": "92c2133aca5f6b3e046d1b113bc202a40d6a8d78", "tree": "263eaf259cf02b179a95f639510980ea79c18ba7", "parents": [ "43a7bdbd3d976672690e15c1347fc6be8ecdb1d3" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Oct 15 16:57:25 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:46:13 2026" }, "message": "Fix 64-bit multisample image fill and drop r64ui texel buffers\n\nThe fill shader for 64-bit multisample sampled images wrote them as storage images although they had no storage usage, through the 32-bit view made for the sampler. Add storage usage, give the fill shader a view with the 64-bit format and check the sample count. GLSL has no 64-bit texture buffer type, so the r64ui uniform texel buffer cases declared a 32-bit one like the r64i cases removed earlier; remove them too.\n\nAffects:\ndEQP-VK*.robustness.robustness2.*.samples_4.*\ndEQP-VK*.robustness.image_robustness.*.samples_4.*\n\nRemoved tests:\ndEQP-VK*.robustness.robustness2.*.r64ui.*.uniform_texel_buffer.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7073\n\nChange-Id: I445d3828f212d85bd8ae1fb01ce155e56e2cfb72\n" }, { "commit": "43a7bdbd3d976672690e15c1347fc6be8ecdb1d3", "tree": "4ad5e72996c47f119ea3f96131680644025e59b6", "parents": [ "4d4bb716d016b554455274b6a810caeedd8368c8", "d86995799be75cec1e0c7fcfea6749aed70e5274" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:32:25 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:32:25 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.5 into vk-gl-cts/vulkan-cts-1.4.6\n\nChange-Id: Ic6cc76f501276b1818413d713d57f76aed5d1aa1\n" }, { "commit": "d86995799be75cec1e0c7fcfea6749aed70e5274", "tree": "1c764f9cab75af01b32c430862b9b016aebe9ca6", "parents": [ "7848d1551ddc2457169bdb35684a67488916aa04" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 10:10:34 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:32:17 2026" }, "message": "Enable the extensions VK_NV_low_latency2 depends on\n\nThe internally synchronized queue out of band cases enabled VK_NV_low_latency2 without VK_KHR_present_id or VK_KHR_present_id2, one of which it requires (VUID-vkCreateDevice-ppEnabledExtensionNames-01387), and those require VK_KHR_swapchain.\n\nAffects: dEQP-VK.synchronization2.internally_synchronized_queues.*out_of_band*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7090\n\nChange-Id: I67d44434f30c216e63af8c5ebe6e60dadb6c4fad\n" }, { "commit": "7848d1551ddc2457169bdb35684a67488916aa04", "tree": "fc6e7913c15996d940ed6814711e2e4438c5bf51", "parents": [ "43e73a3f8382c4aac8456f8fea1a80f84d8883a8" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 21:29:28 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:32:14 2026" }, "message": "Set color state for all attachments in custom resolve shader object cases\n\nThe unused_attachments cases trim the trailing undefined formats of the resolve pipeline, so with shader objects the resolve rendering can have more color attachments than the color write enable, blend enable and write mask calls cover (VUID-vkCmdDraw-attachmentCount-07750, VUID-vkCmdDraw-firstAttachment-07476 and VUID-vkCmdDraw-firstAttachment-07478). Set them for every rendering attachment, calling vkCmdSetColorWriteEnableEXT only if the colorWriteEnable feature is enabled.\n\nAffects: dEQP-VK.renderpasses.dynamic_rendering.*.custom_resolve.shader_objects.*\nComponents: Vulkan\nVK-GL-CTS issue: 6186\nChange-Id: I5844a19f2cc61a3284071af64786f80aa09f9972\n" }, { "commit": "43e73a3f8382c4aac8456f8fea1a80f84d8883a8", "tree": "b3c1036dcf88c1f32518caec87f8f26670d65f27", "parents": [ "9dd1f3f9c7094aca4fae16cb928c00b09599270c", "c7eeffe4c3c583aaf1845c3a1f35204070136398" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:20:35 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:20:35 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.4 into vk-gl-cts/vulkan-cts-1.4.5\n\nChange-Id: Ia7edbef193b45355c6b44f596bd899a5d4a780e5\n" }, { "commit": "c7eeffe4c3c583aaf1845c3a1f35204070136398", "tree": "efb18b6212464cf53ecffa467e8b690e229e68a9", "parents": [ "6bae9ca326b824c1f3390bfb82d473abf93d0f22" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 18:06:02 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:11:44 2026" }, "message": "Write all components of E5B9G9R9 attachments in dual blend tests\n\nThe dual source multi_attachments cases build the color write mask from the channels of the format, which gives R, G and B for E5B9G9R9; that format needs either all four components or none in the write mask (VUID-VkGraphicsPipelineCreateInfo-None-09043, VUID-vkCmdDraw-None-09116 with shader objects).\n\nAffects: dEQP-VK.pipeline.*.blend.dual_source.multi_attachments.e5b9g9r9_ufloat_pack32\nComponents: Vulkan\nVK-GL-CTS issue: 7107\nChange-Id: I6d3288a600fe0a2017287596fcced41c753658f6\n" }, { "commit": "6bae9ca326b824c1f3390bfb82d473abf93d0f22", "tree": "74e9bbae7206e7c41f2a2e017bf5bea10ece97d1", "parents": [ "10c7e2d8eb4005c459ad1595be5b6195a6e636ad", "4c48fad37aaae27bbe25d40ce987575991bc1223" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:01:26 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:01:26 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.3 into vk-gl-cts/vulkan-cts-1.4.4\n\nChange-Id: I14a2a0789f94f240508c76ba1b5dcc6262c1ae32\n" }, { "commit": "4c48fad37aaae27bbe25d40ce987575991bc1223", "tree": "bc6e0c242d259e5dca26bad30ef6063e1940549f", "parents": [ "4be0860ba5adbdb6aa06a3921a33d3b89a527f09" ], "author": { "name": "Valentine Burley", "email": "valentine.burley@collabora.com", "time": "Sat Oct 24 09:58:15 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:01:18 2026" }, "message": "Ignore padding bits when verifying packed format copies\n\nSeveral copy and sparse tests already masked unused LSB padding bits, but\nimage.queue_transfer, image.subresource_layout, and\nsparse_resources.mipmap_sparse_residency performed strict bitwise\ncomparisons.\n\nUpdate these tests to treat the low 6 or 4 padding bits as don\u0027t-care\nlike the other tests.\n\nAffects:\ndEQP-VK.image.queue_transfer.*\ndEQP-VK.image.subresource_layout.*\ndEQP-VK.sparse_resources.mipmap_sparse_residency.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 4621\n\nChange-Id: Id83bdfabccc85048757e2386f50db2a70c1d342a\n" }, { "commit": "4be0860ba5adbdb6aa06a3921a33d3b89a527f09", "tree": "322a8913962470401c7fa2d8c4724ae11872d8ac", "parents": [ "ac4b67e7091002ca7fcd159785d82b5e8e9aaa6f" ], "author": { "name": "Valentine Burley", "email": "valentine.burley@collabora.com", "time": "Sat Oct 24 12:45:18 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 18:01:15 2026" }, "message": "Fix missing depth barrier in early_fragment sample count tests\n\nThe test uses the same depth image across two render passes but only\nsynchronizes the color/resolve attachments between them. Add a\ndepth write-after-write barrier to cover the depth attachment.\n\nAffects:\ndEQP-VK.fragment_operations.early_fragment.sample_count_*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 6894\n\nChange-Id: I4212a278a798c89762494bfc2b62b315ea94c09e\n" }, { "commit": "c9458ce10f021da5cefdb1e1c19704633cb0249b", "tree": "43ab584925fdb4f539229dae4d92fe2c5d1f8576", "parents": [ "7cf701852c7c2a59fb98663371adbfe2bf2796a0" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Fri Oct 30 14:52:44 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:46 2026" }, "message": "Silence -Wundef warnings in shaderc\n\nshaderc is built as part of Vulkan-Video-Samples and inherits the CTS\nwarning flags. The fetched shaderc version produces -Wundef warnings in\nlibshaderc_util. Disable that warning for the shaderc_util target only.\n\nComponents: Vulkan\nVK-GL-CTS issue: 7178\n\nCo-Authored-By: Claude Code\nChange-Id: Ib6bf5a5db153842ea74c35633222a313317d224b\n" }, { "commit": "7cf701852c7c2a59fb98663371adbfe2bf2796a0", "tree": "98422475edb220102314321459bb474474f067e1", "parents": [ "1d845abcb9866428de5a87996d9df7fd6cefec55" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Thu Nov 12 08:49:11 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:43 2026" }, "message": "Fix file permissions\n\nRemove the executable bit from files that are not meant to be run\n(CMakeLists.txt files, a C++ source file and GLSL shader data files),\nand add it to Python scripts that have a shebang line but were not\nexecutable.\n\ngen_framework.py was executable but lacked a shebang line, so add one.\n\nComponents: Framework, Vulkan, OpenGL\nVK-GL-CTS issue: 7241\n\nCo-Authored-By: Claude Code\nChange-Id: Ie0821e163d1d5ed334aa494f35e03af82d5e21ad\n" }, { "commit": "1d845abcb9866428de5a87996d9df7fd6cefec55", "tree": "39476ed58b4a006616c33aa84e3abd9775411d3f", "parents": [ "a7ecc5e0c91c48b91a782eac9b49d85bea8009da" ], "author": { "name": "Derek Lesho", "email": "dlesho@codeweavers.com", "time": "Tue Oct 13 13:35:11 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:41 2026" }, "message": "Add valid GL_MAP_CLIENT_POINTER_BIT_MESA bit to storageFlags\n\nPublic PR: https://github.com/KhronosGroup/VK-GL-CTS/pull/641\n\nAffects:\nKHR-GL45.direct_state_access.buffers_errors\nKHR-GL46.direct_state_access.buffers_errors\n\nComponents: OpenGL\n\nChange-Id: I9e673798c6823d6cdb37dd7c594efced7658243f\n" }, { "commit": "a7ecc5e0c91c48b91a782eac9b49d85bea8009da", "tree": "f20d4bbd3d61b0455f36d32fd1df9f135e3a940a", "parents": [ "0fdfe78b93cce030c6743cd4cb1171fe202c909e" ], "author": { "name": "Tomek Bury", "email": "tomek.bury@broadcom.com", "time": "Thu Oct 22 16:02:36 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:38 2026" }, "message": "Fix intermittent libwayland assert failures in xdg window setup\n\nwl_display_dispatch() asserts if it returns 0, which happens when\nseveral threads wait on the default queue concurrently and another\none dispatched the events first. Serialise the waiters so only one\nthread at a time dispatches the default queue.\n\nComponents: EGL\nVK-GL-CTS issue: 7158\nAffects:\ndEQP-EGL.functional.multithread.*\n\nChange-Id: I5a50745c0f17999d0af33273989147c2f293b460\n" }, { "commit": "0fdfe78b93cce030c6743cd4cb1171fe202c909e", "tree": "afe9bd89020c4a75cfc9081ebccd6bd3a4d86f88", "parents": [ "affa096306083ce3d4423e93c89dd92a66b86e5d" ], "author": { "name": "Yuxin Hu", "email": "yuxinhu@google.com", "time": "Tue Apr 22 21:28:50 2025" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:35 2026" }, "message": "Add activities that export other GLES level test run params\n\nThese activities will be launched by Android test infrastructure\nto collect test run parameters for different GLES levels.\n\nAffected tests: None\n\nComponents: Framework\n\nVK-GL-CTS issue: 5024\n\nChange-Id: Ia9d34273b7505590e0619fea44dec6431600d6c3\n" }, { "commit": "affa096306083ce3d4423e93c89dd92a66b86e5d", "tree": "818a76470132bef91b4c420b80d385c1235bfb63", "parents": [ "ac13af0ac44f06a5fa83e89f89346e4bbb0fb10f" ], "author": { "name": "sourav", "email": "souravp@nvidia.com", "time": "Wed Oct 28 23:46:24 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:32 2026" }, "message": "Fix compute chain texture readback tolerance\n\nUse an 8-bit tolerance for GL_UNSIGNED_BYTE texture readback, matching\nthe GLES test. Using the default framebuffer precision makes the tolerance\ntoo small in 10-bit RGB contexts without alpha, causing valid results\nto fail.\n\nAffects:\nKHR-GL4*.compute_shader.pipeline-compute-chain\n\nComponents: OpenGL\n\nVK-GL-CTS issue: 7159\n\nChange-Id: I2bedb88c75265e04bee488e2862a70cd3483dc7b\n" }, { "commit": "ac13af0ac44f06a5fa83e89f89346e4bbb0fb10f", "tree": "0400d528127c016bc9100717987001bf06af16aa", "parents": [ "f4e5d144e0614028928c739c14e6a6362f1af461" ], "author": { "name": "sourav", "email": "souravp@nvidia.com", "time": "Mon Nov 02 22:33:38 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:30 2026" }, "message": "Fix zero-alpha inputs in advanced blend tests\n\nZero destination RGB when alpha becomes zero after framebuffer\nquantization. Formats such as RGB10_A2 can otherwise leave nonzero RGB\nwith zero alpha, making the corresponding RGB blend results undefined.\n\nApply the correction in getTestColors() so the framebuffer clear and\nreference calculation use the same valid premultiplied color.\n\nAffects:\nKHR-GL4*.blend_equation_advanced.blend_*.*\nKHR-GLES31.core.blend_equation_advanced.blend_*.*\n\nComponents: OpenGL\n\nVK-GL-CTS issue: 7185\n\nChange-Id: I219457ae39659570276b768a818667195b286597\n" }, { "commit": "f4e5d144e0614028928c739c14e6a6362f1af461", "tree": "f2e25e39be341ebcf24fadd722868038d3b02aba", "parents": [ "4f460f8d523a681464ea6dd7c24680d38b2497b5" ], "author": { "name": "Madeeha Javed", "email": "javed@igalia.com", "time": "Wed Jul 01 11:46:57 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:27 2026" }, "message": "Add tests for stats queries with more topologies\n\nDefault input assembly statistics tests use a single primitive topology\nwhich is VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP. New variants test more\ntopologies.\n\nNew tests:\ndEQP-VK.query_pool.statistics_query.*.input_assembly_vertices.*restart*\ndEQP-VKSC.query_pool.statistics_query.*.input_assembly_vertices.*restart*\n\nAffects:\ndEQP-VK.query_pool.statistics_query.*.input_assembly_vertices.*\ndEQP-VKSC.query_pool.statistics_query.*.input_assembly_vertices.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 6679\n\nChange-Id: I593218beb92fdbca0988a27969f798b78716b317\n" }, { "commit": "4f460f8d523a681464ea6dd7c24680d38b2497b5", "tree": "3f9b967833740bc5484a89f2debbc81703ce9743", "parents": [ "124a7d9e4417da60f7134c097c151377580ed53e" ], "author": { "name": "Piotr Byszewski", "email": "piotr.byszewski@cognizant.com", "time": "Fri Oct 09 13:21:23 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:21 2026" }, "message": "Test depthClampZeroOne with depth clip and clamp disabled\n\nWhen:\n* depthClampEnable is false\n* depthClipEnable is false\n* EXT_depth_range_unrestricted is disabled\nbut\n* depthClampZeroOne is true\nthen depth values should be clamped to [0,1].\n\nVK-GL-CTS issue: 6941\n\nComponents: Vulkan\n\nAffects:\ndEQP-VK.pipeline.*.depth.depth_clamp_zero_one.*\n\nChange-Id: I39574cc4e45f264e369e0b3a496e208e68f7a295\n" }, { "commit": "124a7d9e4417da60f7134c097c151377580ed53e", "tree": "befec17dadb2c87648839caf6c829d975ca301bb", "parents": [ "1c68e8455407a999334d5adeb0dddbeedb5f207b" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Fri Oct 23 15:17:46 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:18 2026" }, "message": "Improve run time of reconvergence tests\n\nSome of the reconvergence tests are among the slowest tests in Vulkan\nCTS. One such example is:\n\ndEQP-VK.reconvergence.subgroup_uniform_control_flow_ballot.compute.nesting4.0.40\n\nWhich took 15 seconds to run on my computer. Most of time time appeared\nto be spent in ComputeRandomProgram::simulateBallot/simulateStore. They\nare relatively simple methods, using only a few dozen lines of code\neach. subgroupSizeToMask also appeared to be a bottleneck.\n\nThis commit changes a few things:\n\n1) subgroupSizeToMask built the mask by setting up to 128 bits one at a\ntime. It now sets all bits and shifts once.\n\n2) simulateBallot and simulateStore for compute now loop per subgroup\ninstead of per invocation and skip subgroups with no active\ninvocations. Values that don\u0027t change inside the loop are computed once:\nthe expected ballot, once per subgroup, the 0x12345678 \"don\u0027t validate\"\nvalue, once per call, and the store value.\n\n3) The prerequisites parameter of all 10 simulateBallot/simulateStore\ndeclarations is now a const reference, and the compute versions use\nstatic_cast on .get(). This removes a lot of atomic reference-count\nchanges.\n\n4) calculateAndLogResult now converts the expected masks to UVec4 once\ninstead of on every comparison.\n\nWith these changes, the run time of that test goes down to around ~7\nseconds on my computer, a 2x speedup.\n\nA selection of 640 compute reconvergence tests, covering all five test\ntypes and nesting levels 1 to 5, give the same results as before and run\nin ~13 seconds instead of ~40.\n\nAffects:\ndEQP-VK.reconvergence.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 6401\n\nCo-authored-by: Claude Code\nChange-Id: I4f4c4c3bcb1a5c0f7801039a7c9eb726291e1ac2\n" }, { "commit": "1c68e8455407a999334d5adeb0dddbeedb5f207b", "tree": "2312caf9dd480d10b3eeb6bb308198f935532965", "parents": [ "5d01f425b18027478dffa1f34743a551aa8bf5aa" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Mon Oct 26 15:34:00 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:14 2026" }, "message": "Add GPL tests with mixed depth export pipelines\n\nEnabling and disabling depth exports caused issues in some\nimplementations.\n\nNew tests:\ndEQP-VK.pipeline.pipeline_library.graphics_library.misc.depth_export_switch.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 6939\n\nChange-Id: Id2021beda9a3e71617048654d8b6bafaae003118\nCo-Authored-By: Claude Code\n" }, { "commit": "5d01f425b18027478dffa1f34743a551aa8bf5aa", "tree": "4119768fa9a8cdb7ae2bf25f593e880a3f59019e", "parents": [ "b3317385730b071f744aee0f9c1863d227015649" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Tue Oct 27 16:33:44 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:03 2026" }, "message": "Test descriptor buffer offsets with empty set layout\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7144\n\nNew tests:\ndEQP-VK.binding_model.descriptor_buffer.traditional_buffer.empty_set_layout.*\n\nChange-Id: I4ae7c26ff82f76e5dda804d3e45567fbf50e4591\n" }, { "commit": "b3317385730b071f744aee0f9c1863d227015649", "tree": "478cf535664a43348ec4319dcc2e135c17b7db4d", "parents": [ "e969fc7bac02bba03ba371c0123e75948b3d3139" ], "author": { "name": "michal_jakubek", "email": "michal.jakubek@cognizant.com", "time": "Tue Oct 27 08:01:08 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:57:00 2026" }, "message": "Test non-uniform indexing of acceleration structures without descriptor heap\n\nVK-GL-CTS Issue: 6930\n\nComponents: Vulkan\n\nNew tests:\ndEQP-VK.ray_tracing_pipeline.non_uniform_as_indexing.*\n\nChange-Id: I7be90de81c7147e70114b567bfe3a61bfc2a2ebe\n" }, { "commit": "e969fc7bac02bba03ba371c0123e75948b3d3139", "tree": "c68e0edca2e5d985e58216bf3f41379ee9d149e9", "parents": [ "e95ef07b9cd94bf44ec838483613a57be54b44e7" ], "author": { "name": "Sanjeevi Suresh", "email": "sanjeevisuresh@google.com", "time": "Fri Oct 16 09:08:13 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:57 2026" }, "message": "Add compute test for dynamic buffer descriptor sets with unbound intermediate set\n\nAdds a new test case `unbound_dynamic_buffer_descriptor_set` under the\n`compute.*.basic` group to verify that leaving an intermediate descriptor set\nunbound does not shift or corrupt dynamic storage buffer bindings in higher-numbered\ndescriptor sets.\n\nTo validate the behavior, the test creates a pipeline layout with three descriptor\nset layouts (sets 0, 1, and 2), each declaring a single `VK_DESCRIPTOR_TYPE_STORAGE_BUFFER_DYNAMIC`\nbinding at binding 0. The compute shader statically references only set 0 (`data0`,\ninitialized to 10u) and set 2 (`data2`, initialized to 20u) and computes\n`data0 \u003d data0 + data2`. During command buffer recording, the test binds descriptor\nset 0 and descriptor set 2 with dynamic offset 0 while leaving descriptor set 1\nunbound. If a driver compacts or indexes dynamic buffer offsets by the number of\nbound sets rather than the pipeline layout\u0027s set indices, set 2\u0027s dynamic buffer\nbinding is shifted or reads from set 0\u0027s buffer, producing an incorrect sum instead\nof 30u.\n\nNew tests:\ndEQP-VK.compute.pipeline.basic.unbound_dynamic_buffer_descriptor_set\ndEQP-VK.compute.shader_object_binary.basic.unbound_dynamic_buffer_descriptor_set\ndEQP-VK.compute.shader_object_spirv.basic.unbound_dynamic_buffer_descriptor_set\ndEQP-VKSC.compute.pipeline.basic.unbound_dynamic_buffer_descriptor_set\n\nComponents: Vulkan\nVK-GL-CTS Issue: 7088\n\nChange-Id: Idba8fde04e737be67286ab833de509ab07fa54e4\n" }, { "commit": "e95ef07b9cd94bf44ec838483613a57be54b44e7", "tree": "f54ff2ded55803289f7e40775feb20a9d1e510e5", "parents": [ "bcedcd00d405cb79e2fe7ce0b4cd43f73f07e565" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Wed Oct 21 20:05:24 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:52 2026" }, "message": "Test descriptor heap OpImageFetch with both image resource masks\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7097\n\nNew tests:\ndEQP-VK.binding_model.descriptor_heap.resource_masking.image_fetch\n\nChange-Id: I067c27cd5317d89a4cd8fa784b7279d0a94e6968\n" }, { "commit": "bcedcd00d405cb79e2fe7ce0b4cd43f73f07e565", "tree": "311ca70fd23f765e5207cf1f960f4762f1a18d47", "parents": [ "3223cbd6cebdd2936e172f6a6f1ef90cc3861920" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Wed Oct 28 16:14:01 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:49 2026" }, "message": "Test FSR with layered depth/stencil attachments\n\nAdd depth variants of the layered and multiview fragment shading rate\ntest groups, using a depth/stencil attachment with 2 layers. These are\nuseful to verify that the rates from the shading rate attachment are\napplied to all layers.\n\nLike other depth/stencil groups, the new variants are only generated\nwith a shading rate attachment, a single sample and render passes.\n\nNew tests:\ndEQP-VK.fragment_shading_rate.renderpass2.monolithic.*_depth.*\ndEQP-VKSC.fragment_shading_rate.renderpass2.monolithic.*_depth.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7154\n\nCo-Authored-By: Claude Code\nChange-Id: I803bb810fa29dc2d07957acbb0561a3d97feccf6\n" }, { "commit": "3223cbd6cebdd2936e172f6a6f1ef90cc3861920", "tree": "88650c1303e39382c226ef07bfd7ff07c37602b5", "parents": [ "3bd906293d95933465d5787610e0be4a5886036f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 18:10:12 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:46 2026" }, "message": "Remove the DGC push index heap case with compute stage flags\n\nPush data tokens must update VK_SHADER_STAGE_ALL (VUID-VkIndirectCommandsLayoutTokenEXT-type-11333), so the stage_compute case, which uses VK_SHADER_STAGE_COMPUTE_BIT, is invalid on every implementation; stage_all covers the valid usage.\n\nRemoved tests:\ndEQP-VK.dgc.ext.compute.push_index_heap.stage_compute\n\nComponents: Vulkan\nVK-GL-CTS issue: 7127\nChange-Id: I5477a343520bf1616556a9d8e6994fe7f7778c78\n" }, { "commit": "3bd906293d95933465d5787610e0be4a5886036f", "tree": "d79a2e248eb7ac22737498bb74fd53ec3d1b927e", "parents": [ "a03d851fa40c5942aba2869c1772134b9fe7edf1" ], "author": { "name": "Stéphane Cerveau", "email": "scerveau@igalia.com", "time": "Mon Oct 12 13:32:31 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:43 2026" }, "message": "Video: use VVS enabling logger system\n\nIn order to avoid extra logs due to fence timeout in VVS,\nuse the new release of VVS conditioning it by a logger\nsystem.\n\nKeep error logs\n\nAffects: dEQP-VK.video.encode.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 6259\n\nChange-Id: I191703f66bf7d5cb068c8f72297fe036ec967409\n" }, { "commit": "a03d851fa40c5942aba2869c1772134b9fe7edf1", "tree": "048f0f31841b8d2a49884d608e262e0705dc9b71", "parents": [ "12139a4110c41fb141af804efcb68b93c296074e" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 06:27:23 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:40 2026" }, "message": "Synchronize the swapchain image in frame boundary wsi tests\n\nThe wsi test waited for the acquire semaphore at the color attachment output stage although it clears the image with a transfer, started the clear barrier at the top of pipe stage, so the layout transition did not wait for the semaphore, presented the image in VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL (VUID-VkPresentInfoKHR-pImageIndices-01430), and destroyed the swapchain right after the present (VUID-vkDestroySwapchainKHR-swapchain-01282). Wait at the transfer stage, start the clear barrier there, transition the image to VK_IMAGE_LAYOUT_PRESENT_SRC_KHR before the present, and wait for the queue before the teardown.\n\nAffects:\ndEQP-VK.api.frame_boundary.wsi.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7086\n\nChange-Id: I71fa49b0ffdb3d6b7fb2f8dc993e0eeb922fd3d5\n" }, { "commit": "12139a4110c41fb141af804efcb68b93c296074e", "tree": "7ab6d89c1d317dc008f2d48191edd788537b26a7", "parents": [ "8a5aee1110502678097ba9b51291ed7cf2e09b51" ], "author": { "name": "Graeme Leese", "email": "graeme.leese@broadcom.com", "time": "Thu Oct 29 15:16:38 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:38 2026" }, "message": "Allow higher accuracy in \u0027degrees\u0027 tests\n\nThe denorm_preserve test for the \u0027degrees\u0027 function was only allowing\nthe single result obtained using an fp16 approximation to 180/pi. Also\nallow the correctly rounded result of converting the input to degrees.\n\nComponent: Vulkan\nAffects: dEQP-VK.spirv_assembly.instruction.compute.float_controls.*.degrees_denorm_preserve\n\nChange-Id: I800d267795c443a59c7e8f8daa7582e4aeab1818\n" }, { "commit": "8a5aee1110502678097ba9b51291ed7cf2e09b51", "tree": "efb2fd4d8d1cd0b29c68af027d44830d17c52616", "parents": [ "acdeb2e88a815c16be544c8c76538f5fe23e4541" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 09 15:48:49 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:35 2026" }, "message": "Enable VK_HUAWEI_invocation_mask on the default device\n\nThe invocation mask usage cases were NotSupported on every device because the default device did not enable this extension.\n\nAffects:\ndEQP-VK.image.extended_usage_bit_compatibility.*invocation_mask_bit_huawei\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7017\n\nChange-Id: I73baa54a0ef0a5dfdb43f13aa1c1185142e13c19\n\n" }, { "commit": "acdeb2e88a815c16be544c8c76538f5fe23e4541", "tree": "a19a7d933db4847f14a890bdc57d6a581d9f9252", "parents": [ "a17b2ea469b1192fefc82ac856e8296cb409f68f" ], "author": { "name": "Shiva Chen", "email": "shiva.chen@imgtec.com", "time": "Tue Oct 20 23:07:54 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Nov 12 17:56:33 2026" }, "message": "Fix CooperativeMatrixTestInstance reference comparison\n\nFor signed types, sign-extend the truncated value\nso it matches getDataInt()\u0027s sign-extension of the actual output\nbuffer contents (e.g. (uint32_t)(int8_t)value), rather than comparing\na zero-extended unsigned bit pattern against a sign-extended one.\n\nComponents: Vulkan\n\nAffects:\ndEQP-VK.compute.pipeline.cooperative_matrix.khr_a.subgroupscope.matrixmuladd_cross.sint8_sint8.*\n\nChange-Id: I5ab223603dc5d81d4b742a50af587bfff6ffa37c\n" }, { "commit": "a6107dc3bd05f3eebd49d497d0535291c4bb6f2c", "tree": "3b4b2035c0591b77f6706993f4425017484314cb", "parents": [ "34b298fecf807eaeae3dbc08a9096ccd80d8750f" ], "author": { "name": "Madeeha Javed", "email": "javed@igalia.com", "time": "Thu Oct 22 09:44:08 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Tue Nov 10 11:39:42 2026" }, "message": "Tests for VK_EXT_vertex_address_binding_unrestricted\n\nNew tests:\ndEQP-VK.draw.renderpass.unrestricted_vertex_binding.*\n\nComponents: Framework, Vulkan\nVK-GL-CTS issue: 6418\n\nChange-Id: I3619b528b81a22e0cae9f351de1b30bc3d059424\n" }, { "commit": "34b298fecf807eaeae3dbc08a9096ccd80d8750f", "tree": "dab7c535374e32b432c7ee95b1bb5b2008976c82", "parents": [ "c28b3629309398dc2641c8d4d40079a952160cba", "a17b2ea469b1192fefc82ac856e8296cb409f68f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Tue Nov 10 11:27:09 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Tue Nov 10 11:27:09 2026" }, "message": "Merge main into dev/VK_EXT_vertex_address_binding_unrestricted\n\nChange-Id: I82514dbe0c06e255e1f1be50ba30722baf05832e\n" }, { "commit": "a17b2ea469b1192fefc82ac856e8296cb409f68f", "tree": "dab7c535374e32b432c7ee95b1bb5b2008976c82", "parents": [ "c9835145e8ee464ad35cea4ddc117cb1df574127" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Mon Nov 09 10:57:20 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Mon Nov 09 11:27:41 2026" }, "message": "Update external sources (Vulkan 1.4.368)\n\nChange-Id: I4863d7ee15782a52188bbf909aa86f5b9511bb2d\n" }, { "commit": "9f5a7bd7eb11fa51604ee7bcc3050f406b52e496", "tree": "6b59d85a3ff629085de98d4897fda56e6a8c1938", "parents": [ "1b527b7e90e87b0cd2d19f20d229232d2ffe65ce" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 21:10:28 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 21:10:28 2026" }, "message": "Give each parallel vk-build-programs process its own shader cache file\n\nWhen build_spirv_binaries runs fractions in parallel, all vk-build-programs processes use shadercache.bin in the same directory and each empties it when it starts, so a process can read an entry another one overwrote and abort with std::bad_alloc. Give each fraction its own cache file in its output directory.\n\nAffects: none\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7222\n\nChange-Id: I4435c436398b22e1d9f9e6cb83cf1d5f10eeac42\n" }, { "commit": "c9835145e8ee464ad35cea4ddc117cb1df574127", "tree": "2274e2ac98d274326c136f5f7a194575c423d4a9", "parents": [ "524672ade270aa0ef94eef19fda7552e7495b2d7" ], "author": { "name": "Piotr Byszewski", "email": "piotr.byszewski@cognizant.com", "time": "Tue Oct 13 08:21:49 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 17:22:10 2026" }, "message": "Test device extension enumeration, part 1\n\nThis change modifies the device_extension_dependencies test case so\nthat device extensions no longer require dependencies that are\ninstance extensions to be enabled. This portion of the coverage is\ngenerated by gen_framework.py, which had to be modified accordingly.\n\nNew test has been added that verifies consistency between a\nzero-extension probe instance and a full instance with all\nextensions enabled.\n\nComponents: Framework, Vulkan\n\nVK-GL-CTS issue: 6955\n\nAffects:\ndEQP-VK.info.device_extension_dependencies\ndEQP-VK.api.info.device_extension_enumeration.*\n\nChange-Id: Ib3da0b4adf7ed858f80a65a45b1e7dc389725024\n" }, { "commit": "524672ade270aa0ef94eef19fda7552e7495b2d7", "tree": "fc09849f68a2b60c01f08deaa015f786d6623011", "parents": [ "5b16d2fb3649626f78d5aa47eeda086138f476ed" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Fri Oct 23 07:45:18 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 17:22:06 2026" }, "message": "Convert mesh shader tests to use delayed creation\n\nNote not every test subgroup has been converted. Those using seeds with\nincreasing values will need to be converted first to a seed calculated\nfrom the parameters before they can be trivially converted to delayed\ncreation.\n\nNo test result should be affected by these changes, but the glob below\ncovers the tests with creation changes.\n\nThis commit also adds a convenience addTestGroup2 function, similar to\ncreateTestGroup2.\n\nAffects:\ndEQP-VK.mesh_shader.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7135\n\nAssisted-by: Claude Code\nChange-Id: Ic2e4982d9f3ece9b3e91d0820a580eea757b0e56\n" }, { "commit": "5b16d2fb3649626f78d5aa47eeda086138f476ed", "tree": "941519402e767cf4b8f8aca1af0f21c8ad278321", "parents": [ "30eb12c81b99dd8c6880ae361df80c8add8ec229" ], "author": { "name": "Ria Singhla", "email": "rsinghla@google.com", "time": "Thu Oct 15 09:17:43 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 17:22:03 2026" }, "message": "Add mixed_block_padding test case in compute module\n\nAdds a new test case `mixed_block_padding` under the compute pipeline\nbasic group to verify compiler correctness and stability when copying\ndata between buffer blocks with differing memory layouts and matrix\norientations.\n\nThe test defines a struct containing a mat3x2 matrix and copies it from\na uniform buffer (std140, row_major) to a storage buffer (std430,\ncolumn_major). This verifies that the shader compiler correctly handles\nthe layout conversion and padding differences between std140 (where each\nrow vector is padded to 16 bytes) and std430 (where column vectors are\ntightly packed to 8 bytes), and properly transposes the matrix values\nwithout compiler crashes or data corruption.\n\nNew Test: dEQP-VK.compute.pipeline.basic.mixed_block_padding\nComponents: Vulkan\nVK-GL-CTS issue: https://gitlab.khronos.org/Tracker/vk-gl-cts/-/work_items/7050\n\nChange-Id: Ia0a7ecb41b59cb12161a011e705efaa99889e2d9\n" }, { "commit": "30eb12c81b99dd8c6880ae361df80c8add8ec229", "tree": "928dda7bd89e0e124d4a46417a857a8f64491d03", "parents": [ "27010f31fd14f50ed7c14ef8c7d314e2186c75bf" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Sat Oct 17 11:58:06 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 17:21:59 2026" }, "message": "Copy query result with a stride larger than the buffer\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7069\n\nNew tests:\ndEQP-VK.pipeline.monolithic.timestamp.misc_tests.reset_query_before_copy_large_stride_*\n\nChange-Id: I16f5de083430358ba2c3633c9cd6d081e6191016\n" }, { "commit": "27010f31fd14f50ed7c14ef8c7d314e2186c75bf", "tree": "fc1fc435fbb780fcdd4a3cc1c0e4f245e6510c30", "parents": [ "8ce9fe01f9705ce6e90476678f718020100d1ba5" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Tue Oct 06 13:42:36 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 17:21:56 2026" }, "message": "Add DGC tests to rebind pipeline/shaders after using an IES\n\nSome implementations were failing to rebind the pipeline when the one\nbound after the indirect dispatch matches the one bound before the\ndispatch.\n\nNew tests:\ndEQP-VK.dgc.ext.compute.misc.rebind_after_ies*\n\nAffects:\ndEQP-VK.dgc.ext.compute.misc.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 6962\n\nChange-Id: I5d4530ce5d9f324561e9c06ea8219c8a34871dde\n" }, { "commit": "8ce9fe01f9705ce6e90476678f718020100d1ba5", "tree": "07236eb28636fbdbfab419ab00210a472fb549e9", "parents": [ "8e56810dff425a2e3699d38fb39b317f6acfb6de" ], "author": { "name": "Cristina Murillo", "email": "cmurillo@igalia.com", "time": "Thu Aug 27 23:12:00 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 17:21:53 2026" }, "message": "Video: Add H.265 Main10 video encode tests\n\nAdd 10-bit (Main10 profile) variants of the H.265 encode tests. The\npicture format, session/decode profiles, SPS bit_depth fields and\nprofile tier level are derived from the clip profile bit depth, so the\nsame code path drives both 8- and 10-bit content. A test-time format\ncheck reports NotSupported on encoders without the 10-bit picture\nformat. The 10-bit clip is fetched as yuv/720x480_420_10le.yuv.\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 6716\n\nNew tests:\ndEQP-VK.video.encode.h265.i_10bit_*\ndEQP-VK.video.encode.h265.i_p_10bit_*\ndEQP-VK.video.encode.h265.i_p_b_13_10bit_*\n\nCo-authored-by: Stéphane Cerveau \u003cscerveau@igalia.com\u003e\nChange-Id: Ia616df2739e2a29c6fc5537a68cc31af53f7dd6e\n" }, { "commit": "8e56810dff425a2e3699d38fb39b317f6acfb6de", "tree": "8ff650bb49655ee00487a9249825dd01b5199dfc", "parents": [ "68c8abbdd31ef6b50be41717cb59ce927aca0b9c", "fc08192de01d0d3b629d2e1a5a162b8d9a5abf07" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:55:49 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:56:06 2026" }, "message": "Merge vk-gl-cts/opengl-cts-4.6.8-es-3.2.14 into vk-gl-cts/main\n\nChange-Id: I897798916383293af3be335d6e2641b4626fca76\n" }, { "commit": "fc08192de01d0d3b629d2e1a5a162b8d9a5abf07", "tree": "2d5e94fc522771e46e29ed95c9c9cb990cab73c5", "parents": [ "527db26174a17c2e6c9181c7348936cedef82cae" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Mon Oct 19 11:26:41 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:49:23 2026" }, "message": "Gate the DSA transform feedback query in transform_feedback2_states\n\nThe desktop GL path of transform_feedback2_states queried the transform feedback buffer start and size with glGetTransformFeedbacki64_v, which needs GL 4.5 or ARB_direct_state_access, while the test only requires GL 4.0 or ARB_transform_feedback2. On a GL 4.0 implementation without direct state access the entry point is null and the test crashes. Without the DSA query, use glGetInteger64i_v on the bound transform feedback object, as the ES path does.\n\nAffects:\nKHR-GL*.transform_feedback2.transform_feedback2_states\n\nComponents: OpenGL\n\nVK-GL-CTS public issue: 632\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\nClaude-Session: https://claude.ai/code/session_01DuoE3DwqRkRisLXAbHL1Gb\nChange-Id: If7f826b9d7b0543ab9f1ba9ae97afc26165a5ce8\n" }, { "commit": "527db26174a17c2e6c9181c7348936cedef82cae", "tree": "ac542cf2ab9890981dc643a2ecbda548708dc908", "parents": [ "fb79a340535701abaded3ec0f185db6114f2a1a2" ], "author": { "name": "shihwang", "email": "shihwang@amd.com", "time": "Mon Aug 31 16:28:53 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:49:20 2026" }, "message": "Fix GL object leaks in mesh shader tests\n\nFix GL object leaks (buffers, textures, framebuffers, and shader\nprograms) in the Mesh Shader API, Property, and Smoke test groups by\nadding deinit() implementations that release all resources created\nduring iterate().\n\nAffects:\nKHR-Single-GL45.meshShader.apiTests.*\nKHR-Single-GL45.meshShader.propertyTests.*\nKHR-Single-GL45.meshShader.smokeTests.*\nKHR-Single-GL46.meshShader.apiTests.*\nKHR-Single-GL46.meshShader.propertyTests.*\nKHR-Single-GL46.meshShader.smokeTests.*\n\nVK-GL-CTS issue: 7055\nComponents: OpenGL\n\nChange-Id: Ib1bb3ef7385e12c0d95ba2245f375be3cd08b758\n" }, { "commit": "fb79a340535701abaded3ec0f185db6114f2a1a2", "tree": "da479fc685181a9cd53b4758623f0a39a2779c8e", "parents": [ "e936d2d15ee39bab878d6a9b6dd903a3d64ff88e", "a107d9d70e5640c98c40638b90e95231e7959a07" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:16:10 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:16:10 2026" }, "message": "Merge vk-gl-cts/opengl-es-cts-3.2.13 into vk-gl-cts/opengl-cts-4.6.8-es-3.2.14\n\nChange-Id: I913181dfa85870424a4ec6f18d6ac36876b143d4\n" }, { "commit": "a107d9d70e5640c98c40638b90e95231e7959a07", "tree": "5b6081528f5e53946db913f147e63583705d92a8", "parents": [ "6011abd2bc485342f97a338fb0a241597b22f642" ], "author": { "name": "Tapani Pälli", "email": "tapani.palli@intel.com", "time": "Tue Oct 20 06:10:19 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 16:16:03 2026" }, "message": "Fix blend equation tests with 1010102 window visual\n\nSome tests fail on Wayland with 1010102 EGL visuals, there were\n3 different kind of issues found causing failures:\n\n* dithering causes problems (vendor specific)\n* rounding errors that exposed only with 1010102\n* hardcoded for 8888 type\n\nAffects:\ndEQP-GLES2.functional.fragment_ops.blend.equation_*\n\nComponents: OpenGL ES\nVK-GL-CTS issue: 7123\n\nSigned-off-by: Tapani Pälli \u003ctapani.palli@intel.com\u003e\nChange-Id: I0d03183bf1319d556e889d4cd2ecacc02146d28e\n" }, { "commit": "6011abd2bc485342f97a338fb0a241597b22f642", "tree": "a663fa7a7c3ea5be382d04813d1e980fbe7c336f", "parents": [ "b36500edf8a07a958a8b830c001913935fc202ff", "1e136a8666ea8590398970b948b2054daa06f96a" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 15:45:47 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 15:46:06 2026" }, "message": "Merge vk-gl-cts/opengl-es-cts-3.2.12 into vk-gl-cts/opengl-es-cts-3.2.13\n\nChange-Id: Id87b489db91161f60c6c144088f1265485c245fc\n" }, { "commit": "1e136a8666ea8590398970b948b2054daa06f96a", "tree": "d3073a3786751c17e3bc4c6a1aba3f5e9a39aa63", "parents": [ "6be4d9f88c132eddc250b5f90082796eacb67234" ], "author": { "name": "Tingwei Guo", "email": "tingwei.guo@arm.com", "time": "Tue Aug 04 07:41:10 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 15:45:37 2026" }, "message": "Remove exact-once VS execution assumptions\n\nFix two tests that assume exact-once execution of\nvertex-stage side effects.\n\n- AdvancedMatrixVSFS now uses an idempotent assignment instead of +\u003d.\n- Atomic counter indexing validation now tolerates spec-allowed\n re-execution by checking lower bounds and observed ranges.\n\nComponents: OpenGL\n\nVK-GL-CTS issue: 6799\n\nAffects:\nKHR-GLES31.core.shader_storage_buffer_object.advanced-matrix-vsfs\ndEQP-GLES31.functional.shaders.opaque_type_indexing.atomic_counter.*_vertex\ndEQP-GLES31.functional.shaders.opaque_type_indexing.atomic_counter.*_tessellation_evaluation\n\nChange-Id: I2383b1a9a97158edf798872a5e85e3f13f55fbe3\n" }, { "commit": "68c8abbdd31ef6b50be41717cb59ce927aca0b9c", "tree": "2e59aa85c97b9b139a80c7c495d06584077602ac", "parents": [ "1d90adaac847c02a760d334b51361db6ada6c0e1" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 15:16:39 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 15:16:39 2026" }, "message": "Keep renamed descriptor indexing tests in their Android levels\n\nThe descriptor indexing tests were renamed without updating the Android test lists, so the renamed tests moved to the 2026-03-01 level, and the 2020, 2021 and 2024 list files left behind made check_all fail. Rename them in the Android test lists too.\n\nAffects: none\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7032\n\nChange-Id: I6f2782ecbd5cd859ecfdbbca8de9de2670c75052\n" }, { "commit": "1d90adaac847c02a760d334b51361db6ada6c0e1", "tree": "6b994b70eb144ca99a527647ce84e58dd6833e21", "parents": [ "23ecd86038e915de68704cbb28053ba8097f9cb6" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Mon Nov 02 10:59:47 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:27:27 2026" }, "message": "Keep the test duration in compact logs\n\nWith --deqp-log-compact, TestLog skips the additional data of each case, including the TestDuration that the session executor writes after every case. Keep TestDuration, so that runs with compact logs can still report the time of each case.\n\nAffects: none\n\nComponents: Framework\n\nVK-GL-CTS issue: 6467\n\nChange-Id: I57b801e9b65c6281eac34e9757611756c6526743\n" }, { "commit": "23ecd86038e915de68704cbb28053ba8097f9cb6", "tree": "a14f5cdd033c28d6fc828ba1f3fdee382a007595", "parents": [ "b62f70f03ae4e6b632421474706b5fcb0a460ca2" ], "author": { "name": "Ryan Zhang", "email": "ryan.zhang@nxp.com", "time": "Thu Oct 15 02:51:22 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:27:25 2026" }, "message": "Fix CreateContextNoConfigCase to skip APIs not advertised by any config\n\nWithout checking the renderable type bits advertised across all configs,\na conformant implementation that supports EGL_KHR_no_config_context but\nnot, say, OpenGL ES 3 would be required to create a context for an\nAPI it never claimed to support.\n\nPublic PR: https://github.com/KhronosGroup/VK-GL-CTS/pull/643\n\nAffects:\ndEQP-EGL.functional.create_context.no_config\n\nComponents: EGL\n\nSigned-off-by: Ryan Zhang \u003cryan.zhang@nxp.com\u003e\nChange-Id: I956480689da5de34361d0c9c07e6234b99731ba1\n" }, { "commit": "b62f70f03ae4e6b632421474706b5fcb0a460ca2", "tree": "1e4ccc467dad6de9220c618cbd3fccc49cf3c6ed", "parents": [ "a030a8edfa201b630146968a6cbbdc3a4f3de5e8" ], "author": { "name": "Tapani Pälli", "email": "tapani.palli@intel.com", "time": "Tue Oct 20 04:53:26 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:27:22 2026" }, "message": "Fix immutable texture level check in geometry shader test\n\nSimilar to commit 009e30a13e we need separate error handling for\nimmutable texture level tests.\n\nAffects:\nKHR-GLES31.core.geometry_shader.layered_fbo.fb_texture_invalid_level_number\n\nComponents: OpenGL ES\nVK-GL-CTS issue: 7122\n\nSigned-off-by: Tapani Pälli \u003ctapani.palli@intel.com\u003e\nChange-Id: Ib6d3c18a63a98eb34605d1d35c598e9f1e9f94e3\n" }, { "commit": "a030a8edfa201b630146968a6cbbdc3a4f3de5e8", "tree": "92d46abc2896c15600357c348f56af0ecc2ab006", "parents": [ "b9f3978ecea6c9d84f4360ece5c881a70301197b" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Tue Oct 13 10:58:41 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:27:14 2026" }, "message": "Test using large shared memory across multiple entry points\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7024\n\nNew tests:\ndEQP-VK.spirv_assembly.instruction.compute.multiple_shaders_extended.four_entry_points_large_shared_memory_spirv1p4\ndEQP-VKSC.spirv_assembly.instruction.compute.multiple_shaders_extended.four_entry_points_large_shared_memory_spirv1p4\n\nChange-Id: I0f50d60b7ffab312756cd1cd0ba294be8cb27fac\n" }, { "commit": "b9f3978ecea6c9d84f4360ece5c881a70301197b", "tree": "ff013341c9ace9c8d00bc75f1fc0fc5ac9975f67", "parents": [ "952f89821211f547007acefc2eb4557f18e12204" ], "author": { "name": "Sanjeevi Suresh", "email": "sanjeevisuresh@google.com", "time": "Tue Oct 13 08:22:49 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:27:11 2026" }, "message": "Add unexecuted_oob_uniform_index non robust buffer access test\n\nAdds a new Amber test case `unexecuted_oob_uniform_index` under the\n`robustness.non_robust_buffer_access` group to verify that an\nimplementation does not execute a storage buffer load that is guarded by\na branch which is never taken at runtime, when the load address is\ndynamically uniform.\n\nTo validate the behavior, the fragment shader reads its load index from\na push constant, so the address is dynamically uniform but not a\ncompile-time constant, and guards `values[pc.index]` behind\n`if (pc.index \u003c\u003d 15u)`. The push constant is set to 0x0F000000, far\noutside the 16-element storage buffer, so the branch is never taken.\nWith robustBufferAccess disabled, a correct implementation never performs\nthe load and every fragment writes green. An implementation that\nspeculatively executes or hoists the uniform load out of the branch\naccesses an unmapped address, which can cause a GPU page fault and\nVK_ERROR_DEVICE_LOST, or a framebuffer mismatch.\n\nNew tests:\ndEQP-VK.robustness.non_robust_buffer_access.unexecuted_oob_uniform_index\n\nComponents: Vulkan\nVK-GL-CTS Issue: 7027\n\nChange-Id: I2ba9290eac566d2938ed4ff4121af952b5d1ffa5\n" }, { "commit": "952f89821211f547007acefc2eb4557f18e12204", "tree": "4b57200a000e377ffa4299142bdac7883b8f92c1", "parents": [ "8a2b8cbb29f9afcac19e86dae22ea463e4f9d905" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Fri Oct 23 09:38:27 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:27:07 2026" }, "message": "Check descriptor set layout pNext is not used after creation\n\nIn addition, this commit reorganizes descriptor set indexing tests to\ndivide them into smaller subgroups, to ease addition of the new tests\nwithout creating large subtrees.\n\nRenames:\ndEQP-VK.descriptor_indexing.misc_* \u003d\u003e dEQP-VK.descriptor_indexing.misc.misc_*\ndEQP-VK.descriptor_indexing.*minNonUniform \u003d\u003e dEQP-VK.descriptor_indexing.min_non_uniform.*minNonUniform\ndEQP-VK.descriptor_indexing.*_no_runtime_array \u003d\u003e dEQP-VK.descriptor_indexing.min_non_uniform.*_no_runtime_array\nOther dEQP-VK.descriptor_indexing.* \u003d\u003e dEQP-VK.descriptor_indexing.after_bind_and_loop.*\n\nNew tests:\ndEQP-VK.descriptor_indexing.pnext.*\n\nVK-GL-CTS issue: 7032\nComponents: Vulkan\n\nChange-Id: I34b93b6496dbe37a8029ef70623a4bf67833b434\n" }, { "commit": "8a2b8cbb29f9afcac19e86dae22ea463e4f9d905", "tree": "5984e1f14f385850477da5b2f1305b784f1d5de2", "parents": [ "fc8a5fefe9716f04cb6f62f39b668b7469f3a595" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Fri Oct 16 14:44:46 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:56 2026" }, "message": "Query queue family properties with count 0 and non-NULL array\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7068\n\nNew tests:\ndEQP-VK.api.info.get_physical_device_properties2.queue_family_properties_zero_count.basic\n\nChange-Id: Ic34541a525c85eaae6176bb26ed581b7ff9f9ffb\n" }, { "commit": "fc8a5fefe9716f04cb6f62f39b668b7469f3a595", "tree": "609c5d3a92e1c8d808a2a7f327e95b8d150c6feb", "parents": [ "daf8295a3a96695e7251145c22ad4b2658c2165e" ], "author": { "name": "Marcin Zając", "email": "marcin.zajac@cognizant.com", "time": "Fri Oct 16 15:25:47 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:50 2026" }, "message": "Test tanh(-/+Inf) with round towards zero rounding mode\n\nRun tanh(Inf) and tanh(-Inf) in fp16 and fp32 shaders that declare the RoundingModeRTZ execution mode and check that the results are exactly 1.0 and -1.0\n\nVK-GL-CTS issue: 6926\n\nComponents: Vulkan\n\nNew tests:\ndEQP-VK.spirv_assembly.instruction.*.float_controls2.*.input_args.tanh_testedWithout_NotInf_*_rtz_*\n\nChange-Id: Ic2b2b0f2b372c6227790d3b81ad6c33e26710be5\n" }, { "commit": "daf8295a3a96695e7251145c22ad4b2658c2165e", "tree": "03e599f846f2dc07a5ac4482a1cbae5fece27a84", "parents": [ "62e3a58961cb12903bc62e7952210f02ca4bfda6" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Mon Oct 19 14:32:19 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:47 2026" }, "message": "Add test mixing mesh and regular draws with gl_DrawID\n\nThis seems to be problematic for some drivers.\n\nNew tests:\ndEQP-VK.mesh_shader.ext.misc.draw_id_mix*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7023\n\nChange-Id: Idd9fb168c33edb439b1003b98c8fbbf94407f5ff\n" }, { "commit": "62e3a58961cb12903bc62e7952210f02ca4bfda6", "tree": "6ba5d500b3dad6f7ee934a650bcc8511f5b194b2", "parents": [ "0eb88bfb777fa30cc6cac7c8200ee3ddfae5a324" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Mon Oct 19 18:55:42 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:43 2026" }, "message": "Test OpSelect of bool vectors\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7043\n\nNew tests:\ndEQP-VK.spirv_assembly.instruction.spirv1p4.opselect.bool_vector_*\n\nChange-Id: I8bf2f80c06abf9b76d52feb9533646585e80a4cf\n" }, { "commit": "0eb88bfb777fa30cc6cac7c8200ee3ddfae5a324", "tree": "a25c0f58dbdf1e92eaee3bd40440ec7a212fdbcf", "parents": [ "0f544e75276c744eebb25faadba605168baa40f5" ], "author": { "name": "ziga-lunarg", "email": "ziga@lunarg.com", "time": "Mon Oct 19 13:52:58 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:36 2026" }, "message": "Test 2d array to 3d image copies with different formats\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7041\n\nNew tests:\ndEQP-VK.api.copy_and_blit.*image_to_image*.different_format*\ndEQP-VKSC.api.copy_and_blit.*image_to_image*.different_format*\n\nAffected tests:\ndEQP-VK.api.copy_and_blit.*image_to_image*\ndEQP-VKSC.api.copy_and_blit.*image_to_image*\n\nChange-Id: I37b9b8a7cf6966ae91e778ff0281e461bd22d395\n" }, { "commit": "0f544e75276c744eebb25faadba605168baa40f5", "tree": "28b135a058763a4bff75fbdb88296e877ce7df25", "parents": [ "03de694e0759ef0528644b77e8359445865c39c1" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Tue Oct 20 12:16:39 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:32 2026" }, "message": "Make the depth and stencil copy areas different for 2D DS copy tests\n\nAffects:\ndEQP-VK.api.copy_and_blit.*.image_to_image.*.depth_stencil.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7129\n\nChange-Id: Idbcaf67fdcb081ae910e7ee7bf2efaaede4ffe9f\n" }, { "commit": "03de694e0759ef0528644b77e8359445865c39c1", "tree": "8606f2edaf57cb12bdd351d3c3d0b2fb57a04fec", "parents": [ "cdc920bbe380d16ced9d2b0fd618283a691a0892" ], "author": { "name": "Madeeha Javed", "email": "javed@igalia.com", "time": "Tue Oct 20 07:52:17 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:25 2026" }, "message": "Add tests for descriptor set over allocation\n\nA descriptor set could be allocated past the end of the descriptor\npool\u0027s storage. On some implementations, when the storage capacity\ncrosses page boundaries, the descriptor can end up outside the buffer.\nThe new tests verify for various desciptor types and set counts if\nover-allocation fails properly and that the allocated sets are updated\nand used without failures.\n\nNew tests:\ndEQP-VK.api.descriptor_pool.out_of_bounds_set_count_*\ndEQP-VKSC.api.descriptor_pool.out_of_bounds_set_count_*\n\nAffects:\nN/A\n\nComponents: Vulkan\nVK-GL-CTS issue: 7028\n\nChange-Id: If6b7d7cc2e22f16dd8c63e5efc0d99ffb1f3cb9b\n" }, { "commit": "cdc920bbe380d16ced9d2b0fd618283a691a0892", "tree": "76838c1404a54ec88098fe0cb66ddc5cd533b2e8", "parents": [ "0eea0a99678cefa3bc66ad0fabb3139ca27df130" ], "author": { "name": "Mateusz Bahyrycz", "email": "mateusz.bahyrycz@cognizant.com", "time": "Wed Oct 14 15:03:17 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:20 2026" }, "message": "Upload compute variant textures on the compute queue\n\nShaderRenderCaseInstance prepared all textures on the universal queue\nwith graphics stages in the barriers, while the compute variants sample\nthem on the dedicated compute queue without a queue family ownership\ntransfer. The depth compute variants are removed.\n\nAffects:\ndEQP-VK.glsl.texture_functions.*_compute\ndEQP-VK.glsl.texture_gather.compute.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 6988\n\nChange-Id: I1bfa55897ccafdd95e1c70ff3b99290ca475fc92\n" }, { "commit": "0eea0a99678cefa3bc66ad0fabb3139ca27df130", "tree": "98b17eadad0adcb8088012c7d0e96764c7bb5077", "parents": [ "e8eed57b99357b838b0a6160dd3222d493fed481" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Oct 08 17:20:44 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:14 2026" }, "message": "Initialize feature and property structs before querying them\n\nSome checkSupport functions passed structs with uninitialized members to vkGetPhysicalDeviceFeatures2 or vkGetPhysicalDeviceProperties2, so members the driver does not write were read from the stack.\n\nAffects:\ndEQP-VK.pipeline.*.blend_operation_advanced.*\ndEQP-VK.spirv_assembly.instruction.*.float_controls.*\ndEQP-VK.spirv_assembly.instruction.*.64bit_compare.*\ndEQP-VK.subgroups.subgroup_uniform_control_flow.*\ndEQP-VK.texture.mipmap.*.image_view_min_lod.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7039\nChange-Id: I69f6127560feb3d579823d9710f60a838b8f24e6\n" }, { "commit": "e8eed57b99357b838b0a6160dd3222d493fed481", "tree": "a12dfced1486083f2bfbd32731e4fd81f886c1d6", "parents": [ "c2c05e95f97f8b73582a5630c995867ba56f648f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Oct 15 19:28:33 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:11 2026" }, "message": "Subtract offsets from mesh indirect count address ranges\n\nThe device address variants of the mesh shader indirect count draws passed the whole buffer size as the size of ranges that start at an offset, so the ranges ran past the end of the buffers. The indirect draw path already subtracts the offset.\n\nAffects:\ndEQP-VK.mesh_shader.ext.api.draw_indirect_count.*_device_address\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7081\n\nChange-Id: I8a9f4cf6e8a3bd7d928f9b491848320617ac090c\n\n\n" }, { "commit": "c2c05e95f97f8b73582a5630c995867ba56f648f", "tree": "912e3653446d7f6702ff45c86eb89b5d1069e7a8", "parents": [ "8e504597f513072273290d5e8377e34d07389f87" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Thu Oct 15 19:28:30 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:08 2026" }, "message": "Compute device address copy ranges from the addressing rules\n\nThe device address variants of the buffer and image copy tests sized each address range from the buffer row length and image height alone, which overstates 1D array layers so the range runs past the buffer and ignores the layer count in image to buffer copies so the range is too small. Size each range from the first to the last byte the copy accesses, as Buffer and Image Addressing defines it.\n\nAffects:\ndEQP-VK.api.copy_and_blit.device_address.buffer_to_image.*\ndEQP-VK.api.copy_and_blit.device_address.image_to_buffer.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7080\n\nChange-Id: Ib061090e229ce4d400945c57ca08bacf8236686f\n" }, { "commit": "8e504597f513072273290d5e8377e34d07389f87", "tree": "e9568032c1a97996b6abfc20d1c3327559cc7ae1", "parents": [ "6262ccb6083d42ed35dd48aa4ae3d5bc717cf406" ], "author": { "name": "Mateusz Bahyrycz", "email": "mateusz.bahyrycz@cognizant.com", "time": "Mon Oct 19 14:54:44 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:05 2026" }, "message": "Skip depth/stencil copies on a universal queue without graphics\n\nOn a device without graphics the universal queue is a compute queue.\nSeveral tests copy depth/stencil aspects between buffers and images on\nit, either as the tested copy or to upload and read back their images,\nwhich needs VK_QUEUE_GRAPHICS_BIT unless VK_KHR_maintenance10 is used.\nThese cases now report NotSupported. The check is moved from the host\nimage copy tests to vktTestCase so that all modules can share it. The\ncompute and transfer queue variants, which test VK_KHR_maintenance10,\nare unchanged.\n\nAffects:\ndEQP-VK.api.copy_and_blit.*depth_stencil*\ndEQP-VK.api.copy_and_blit.*depthstencil*\ndEQP-VK.api.copy_and_blit.*.image_to_image*\ndEQP-VK.api.ds_color_copy.*\ndEQP-VK.image.host_image_copy.*\ndEQP-VK.image.load_store.without_any_format.*\ndEQP-VK.image.store.without_format.*\ndEQP-VK.synchronization*.op.single_queue.*\ndEQP-VK.synchronization*.signal_order.shared_*\ndEQP-VK.synchronization*.timeline_semaphore.*\n\nComponents: Framework, Vulkan\n\nVK-GL-CTS issue: 6992\n\nChange-Id: Id67c24d0104fd2c324d9933e19c6c4132d43cde0\n" }, { "commit": "6262ccb6083d42ed35dd48aa4ae3d5bc717cf406", "tree": "9ca1741973edab46ab8106b7bf961433963ec523", "parents": [ "ee7c08b48cd06132bc8d2d489a90696cdd2c64c2" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 09 02:50:59 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:26:02 2026" }, "message": "Include the alignment of optimal matrices in cooperative vector weight sizes\n\nEach weight matrix is stored as written by the test, then converted to the optimal layout at the next 64-byte boundary, but the size of the weights left out that padding. When the converted size is not a multiple of 64, which the spec allows, the last converted matrix is written past the end of the buffer. Size the weights from the end of the converted matrix.\n\nAffects:\ndEQP-VK.cooperative_vector.matmul.*Optimal*\ndEQP-VK.cooperative_vector.training.*Optimal*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7013\n\nChange-Id: I6f6868f3fbe50867a34b6955581a60e932152b4d\n" }, { "commit": "ee7c08b48cd06132bc8d2d489a90696cdd2c64c2", "tree": "c8c3a2e12dd9c3a6b27a79163e13b7f9b81bd401", "parents": [ "1092584f5941c87cb9bf35cdcd5a226bede1fa8f" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Wed Oct 14 10:52:26 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:25:58 2026" }, "message": "Fix matrix conversion parameters in cooperative vector outer product tests\n\nThe float32 outer product tests queried the weight matrix size with a source stride computed from the float16 input type for a float32 matrix, smaller than a row, and read back each result matrix with srcSize set to the whole buffer although the source starts at that matrix\u0027s offset. Compute the stride from the matrix type and pass the size of one result matrix.\n\nAffects:\ndEQP-VK.cooperative_vector.training.*outerproduct*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7038\nChange-Id: Ia057621996c4f88cd7087ed2c7c940ae0fa10a0e\n" }, { "commit": "1092584f5941c87cb9bf35cdcd5a226bede1fa8f", "tree": "c11da1db2f19c894bd4917cc951af159dfd964bb", "parents": [ "6d17e24db24efb0fd35768185b4ba9f4dc4f36ba" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Wed Oct 14 07:38:18 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:25:55 2026" }, "message": "Use valid workgroup sizes in workgroup scope cooperative matrix tests\n\nThe workgroup scope tests queried VkCooperativeMatrixProperties2EXT for 8 to 256 invocations, which need not be multiples of the subgroup size or within cooperativeMatrixWorkgroupScopeMaxWorkgroupSize. They now use the power of two multiples of the subgroup size up to that limit.\n\nAffects:\ndEQP-VK.compute.*.cooperative_matrix.*.workgroupscope.*\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7035\nChange-Id: Ifca372238716db8c35585b514b0ba760eb043828\n" }, { "commit": "6d17e24db24efb0fd35768185b4ba9f4dc4f36ba", "tree": "fa2db0bcfcc3ba01321c36a362641fa6902924e6", "parents": [ "98bf2f38f265c5604287b20c71a08addc5000d2d" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 06:16:34 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:25:52 2026" }, "message": "Destroy the surface when swapchain_counter setup throws\n\nThe test destroyed its display surface in the test instance destructor, which does not run when the constructor throws, such as the NotSupported thrown when a windowing system owns the display, so the instance was destroyed with the surface alive (VUID-vkDestroyInstance-instance-00629). Hold the surface in a Unique.\n\nAffects:\ndEQP-VK.wsi.display_control.swapchain_counter\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 7085\n\nChange-Id: I104b7cb3c66b5f39d3e4a39bb6ef08de0cf8c527\n" }, { "commit": "98bf2f38f265c5604287b20c71a08addc5000d2d", "tree": "91b35c7067dbc91ca97d7edbaeb8b9031a933578", "parents": [ "4117e6d0f67ed1898529f75b22619c3a208864bb" ], "author": { "name": "Ria Singhla", "email": "rsinghla@google.com", "time": "Wed Aug 19 14:57:37 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:25:48 2026" }, "message": "Adding test for Private memory access in loop compiler bug\n\nAdds a new SPIR-V assembly Amber compute test (`private_memory_in_loop.amber`)\nto reproduce a compiler bug where accessing `Private` (internal global) memory\nwithin a loop causes incorrect code generation and data corruption.\n\nThe test requires 16-bit storage (`storageBuffer16BitAccess`) and\n16-bit float (`shaderFloat16`) features, and has been integrated into the\ndEQP framework under the `dEQP-VK.compute.pipeline.basic` test group.\n\nNew Test: dEQP-VK.compute.pipeline.basic.private_memory_in_loop\n\nComponents: Vulkan\nVK-GL-CTS issue: https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/6830\n\nChange-Id: I7c1ec728e02371c6145551772c8fbefdd0616de7\n" }, { "commit": "4117e6d0f67ed1898529f75b22619c3a208864bb", "tree": "040047c9ee267785f61e2e3b191121cfd7d7fd48", "parents": [ "63eada55fb16e58d2197535467bc149cd5aba2a6" ], "author": { "name": "Mikołaj Bednarek", "email": "mmbednarek@google.com", "time": "Thu Oct 15 14:11:27 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:25:40 2026" }, "message": "Add alpha to coverage constant alpha tests\n\nAdd new tests to verify that alpha to coverage works correctly\nwhen the alpha value is known at compile time.\nDrivers optimizations related to constant alpha values might\nnot take alpha to coverage into account.\nThese tests verify that the driver behaves correctly.\n\nAlso enable depth testing with an occluded quad in opaque alpha to\ncoverage tests to verify that covered samples write depth values.\n\nComponents: Vulkan\n\nNew tests:\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_invisible_constant\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_invisible_constant_check_depth\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_invisible_constant_sparse\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_invisible_constant_sparse_check_depth\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_opaque_constant\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_opaque_constant_sparse\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_translucent_constant\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_translucent_constant_sparse\n\nAffects:\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_opaque\ndEQP-VK.pipeline.*.multisample*.alpha_to_coverage.samples_*.alpha_opaque_sparse\n\nChange-Id: I3081063742cbd6b1d3d02e773cfc25af9d144e48\n" }, { "commit": "63eada55fb16e58d2197535467bc149cd5aba2a6", "tree": "be690c4de411bf844db6eadd58ba72eca99004e7", "parents": [ "dc51d716a70671743b3d339ca283cc66d0e33ec4", "1b527b7e90e87b0cd2d19f20d229232d2ffe65ce" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:01:17 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 14:01:17 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.7 into vk-gl-cts/main\n\nChange-Id: I2651cf265508d8295e42a854e9e18ab6d0cb79d8\n" }, { "commit": "1b527b7e90e87b0cd2d19f20d229232d2ffe65ce", "tree": "a4bad3f066fcdf06c51738f742c2ffe723e51b77", "parents": [ "7bf654019994085962b721e3050bbfc3540f4ab4" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 23 15:53:52 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 12:20:33 2026" }, "message": "Assert the descriptor resource in the independent sets utilities (clang-analyzer-core.CallAndMessage)\n\nclang-tidy reports calls through a null resourceInfo when writing descriptor sets: the resource lookup may fail, which is expected only for samplers. An assertion after the lookup states that every other descriptor type has a resource; it helps the analyzer once the deAssertFail change (CL 21435, vulkan-cts-1.4.1) is in the same tree.\n\nNo behaviour changes.\n\nAffects: none\n\nComponents: Vulkan\n\nVK-GL-CTS issue: 4780\n\nChange-Id: Ifcab7b31da6c90aff38727e667c83d1fe421441b\n" }, { "commit": "7bf654019994085962b721e3050bbfc3540f4ab4", "tree": "842757552bd14b367d58a3a53d64e465e7aa6e7e", "parents": [ "bad32071a6f3923c92f117c8661997478c66ea46" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 14:14:20 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 12:20:29 2026" }, "message": "Chain initialized flags2 structs in video format support queries\n\nThe extended flags video format support queries filled the image create and usage flags2 structs with push_back after vectors already sized to the format count, so each VkVideoFormatPropertiesKHR chained two zero-initialized structs with sType 0, which VUID-VkVideoFormatPropertiesKHR-pNext-pNext forbids.\n\nAffects: dEQP-VK.video.capabilities.extended_flags.*_video_format_support_query\nComponents: Vulkan\nVK-GL-CTS issue: 7094\nChange-Id: Iccecceb917962df603f42e1b2701f6cbf47cb080\n" }, { "commit": "bad32071a6f3923c92f117c8661997478c66ea46", "tree": "6f0e567325309f4be1da261a733fb520fb27689b", "parents": [ "8cf1134ddec1414c4cc3bbaaa94a84508840956a", "4d4bb716d016b554455274b6a810caeedd8368c8" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 12:13:01 2026" }, "committer": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 12:13:01 2026" }, "message": "Merge vk-gl-cts/vulkan-cts-1.4.6 into vk-gl-cts/vulkan-cts-1.4.7\n\nChange-Id: I9ebf964a7149af206df33fff444d06f6b84c8c04\n" }, { "commit": "4d4bb716d016b554455274b6a810caeedd8368c8", "tree": "1fa7eb389a129dd15f84bdc38145978096d1f43f", "parents": [ "b7d22cf6ee67b1c92c6640e40ffc70c600097e10" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Mon Oct 19 10:25:23 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 09:27:33 2026" }, "message": "Take sparseAddressSpaceSize into account in buffer mem. req. tests\n\nAffects:\ndEQP-VK.api.buffer_memory_requirements.create_sparse_binding_sparse_residency.ext_mem_flags_excluded.method1.size_req_other_usage_bits\ndEQP-VK.api.buffer_memory_requirements.create_sparse_binding_sparse_residency.ext_mem_flags_excluded.method2.size_req_other_usage_bits\ndEQP-VK.api.buffer_memory_requirements.create_sparse_binding_sparse_residency.ext_mem_flags_excluded.method2.size_req_storage_usage_bits\ndEQP-VK.api.buffer_memory_requirements.create_sparse_binding_sparse_residency.ext_mem_flags_included.method1.size_req_storage_usage_bits\ndEQP-VK.api.buffer_memory_requirements.create_sparse_binding_sparse_residency.ext_mem_flags_included.method1.size_req_transfer_usage_bits\ndEQP-VK.api.buffer_memory_requirements.create_sparse_binding_sparse_residency.ext_mem_flags_included.method2.size_req_transfer_usage_bits\n\nComponents: Vulkan\nVK-GL-CTS issue: 7059\n\nChange-Id: I2507e5b31ba7c8395fc6be8d7f09dbaf0e65d50b\n" }, { "commit": "b7d22cf6ee67b1c92c6640e40ffc70c600097e10", "tree": "c1b7412c7e2ed5665dcd8b8d4d35686dc396bc18", "parents": [ "888961d18a2a17063dede6a8d3aea51d46c0986b" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Oct 16 18:01:20 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 09:27:30 2026" }, "message": "Do not generate conditional indirect copy cases on transfer queues\n\nConditional rendering can only be recorded in command buffers from a queue family with graphics or compute capabilities (VUID-vkCmdBeginConditionalRenderingEXT-commandBuffer-cmdpool), so the transfer queue variants of the indirect memory to image copy cases are invalid on every implementation.\n\nRemoved tests:\ndEQP-VK.api.copy_and_blit.core.memory_to_image_indirect_transfer_queue.2d_images.conditional_off\ndEQP-VK.api.copy_and_blit.core.memory_to_image_indirect_transfer_queue.2d_images.conditional_on\ndEQP-VK.api.copy_and_blit.dedicated_allocation.memory_to_image_indirect_transfer_queue.2d_images.conditional_off\ndEQP-VK.api.copy_and_blit.dedicated_allocation.memory_to_image_indirect_transfer_queue.2d_images.conditional_on\n\nComponents: Vulkan\nVK-GL-CTS issue: 7102\nChange-Id: Idd2873bdc1435183c90869cab1af00ec6180bb7a\n" }, { "commit": "888961d18a2a17063dede6a8d3aea51d46c0986b", "tree": "af68309f208e1c6a4bf092270611f9d118abedd0", "parents": [ "3ac7da17c031b21e5212877e40574ed2262075a3" ], "author": { "name": "Khronos: Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Sun Oct 18 12:19:43 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 09:27:27 2026" }, "message": "Create the host image copy comparison image without host transfer usage\n\nWhen identicalMemoryTypeRequirements is set, the preinitialized cases create a second image to check that host transfer usage does not change the memory types, but they clear the usage bit from the create flags instead of the usage, so both images are the same. The second image is also bound to memory without the device address flags that the capture replay flag of capture_replay.heap needs (VUID-vkBindImageMemory-image-08113, VUID-vkBindImageMemory-image-09202). Clear the usage bit and create the second image without memory, since only its memory requirements are queried.\n\nAffects:\ndEQP-VK.image.host_image_copy.preinitialized.*\ndEQP-VK.image.host_image_copy.capture_replay.heap\nComponents: Vulkan\nVK-GL-CTS issue: 7112\nChange-Id: I51a465b3a0ac9d301d001e985084d3f3f4afe7e8\n" }, { "commit": "3ac7da17c031b21e5212877e40574ed2262075a3", "tree": "caa7c38eeb1b218262ebb30e1227b721e539eddb", "parents": [ "a731d98a79b6aeda521177b26010d39ecacb15f2" ], "author": { "name": "Ricardo Garcia", "email": "rgarcia@igalia.com", "time": "Wed Oct 21 10:53:27 2026" }, "committer": { "name": "Lorenzo Dal Col", "email": "lorenzo@khronosgroup.org", "time": "Fri Nov 06 09:27:23 2026" }, "message": "Add missing geometryShader support check in maintenance10 resolve tests\n\nIn addition, make the vertex shader use SPIR-V 1.5 instead of 1.4, which\nwas likely a typo. SPIR-V 1.5 is needed so the shader uses the\nShaderLayer capability instead of ShaderViewportIndexLayerEXT. This\nmatches the support check that the implementation must use Vulkan 1.2+.\n\nAffects:\ndEQP-VK.pipeline.*.multisample.m10_resolve.*.full_multilayer*.*\ndEQP-VK.pipeline.*.multisample.m10_resolve.*.regions_multilayer*.*\ndEQP-VK.pipeline.*.multisample.m10_resolve.*.full_3d.*\n\nComponents: Vulkan\nVK-GL-CTS issue: 7131\n\nChange-Id: Ib2ce3a66cba1bf1fad48b0c6840b507528d01b62\n" } ], "next": "a731d98a79b6aeda521177b26010d39ecacb15f2" }