kc3-lang/angle/src/libANGLE/renderer

src/libANGLE/renderer

Log

Author Commit Date CI Message
Tim Van Patten 317f81db 2025-05-01T14:17:47 Fix EGL_RENDER_BUFFER query if EGL_SINGLE_BUFFER unsupported eglCreateWindowSurface() can have the EGL attribute EGL_RENDER_BUFFER specified: EGL_RENDER_BUFFER specifies which buffer should be used by default for client API rendering to the window, as described in section 2.2.2. If its value is EGL_SINGLE_BUFFER, then client APIs should render directly into the visible window. If its value is EGL_BACK_BUFFER, then all client APIs should render into the back buffer. The default value of EGL_RENDER_BUFFER is EGL_BACK_BUFFER. Client APIs may not be able to respect the requested rendering buffer. To determine the actual buffer that a context will render to by default, call eglQueryContext with attribute EGL_RENDER_BUFFER (see section 3.7.4). To support EGL_SINGLE_BUFFER, Vulkan surfaces must support the VkPresentModeKHR value VK_PRESENT_MODE_SHARED_DEMAND_REFRESH_KHR: VK_PRESENT_MODE_SHARED_DEMAND_REFRESH_KHR specifies that the presentation engine and application have concurrent access to a single image, which is referred to as a shared presentable image. The presentation engine is only required to update the current image after a new presentation request is received. Therefore the application must make a presentation request whenever an update is required. However, the presentation engine may update the current image at any point, meaning this mode may result in visible tearing. However, this is only available on Vulkan devices that support the extension VK_KHR_shared_presentable_image. Add checking in Surface::initialize() to update Surface::mRenderBuffer to EGL_BACK_BUFFER if the backend implementation does not support EGL_SINGLE_BUFFER. This includes adding supportsSingleRenderBuffer() to query the backend if it supports single render buffer mode, which defaults to False. The Vulkan backend overrides this and the result is based on support for VK_PRESENT_MODE_SHARED_DEMAND_REFRESH_KHR. Bug: b/412446258 Test: EGLLockSurface3Test.WindowMsaaSurfaceReadTest/ES2_Vulkan_NoFixture Test: EGLSingleBufferTest.VerifyMutableRenderBufferKHR/* Change-Id: I4e6d56f01a895a5bd887580e6ffa34d574c87fad Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6506764 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Sunny Sachanandani 8ea69d84 2025-05-07T12:26:48 metal: Synchronize imported IOSurfaces with command buffer We want to skip the various Metal waitUntilScheduled calls in Chromium on single GPU systems and use only shared events for synchronization. However, this doesn't work correctly today because ANGLE can do CPU readbacks from the IOSurface without synchronizing with the command buffer which contains the shared event wait. This is because ANGLE doesn't know about the dependency between the IOSurface texture and the shared event wait in the command buffer. This CL makes BindTexImage mark the IOSurface texture as used so that we wait for command buffer execution if we do a CPU readback. It reuses the setWriteDependency mechanism on mtl::CommandBuffer. Bug: chromium:40273077 Change-Id: I5d480f2c28abbd6d36cf1e1045461c84b3e4a1f0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6514098 Reviewed-by: Quyen Le <lehoangquyen@chromium.org> Auto-Submit: Sunny Sachanandani <sunnyps@chromium.org> Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
Amirali Abdolrashidi 5c441a53 2025-05-06T13:00:32 Vulkan: Prefer MSRTSS for QCOM devices The feature flag `preferMSRTSSFlagByDefault` was originally added to avoid allocating extra memory if the images are created with the MSRTSS flag by default. If it is disabled, the image will be recreated with the said flag only if it bound to an MSRTT framebuffer. * preferMSRTSSFlagByDefault is now enabled for QCOM driver versions beyond 0777.0. Bug: b/329911999 Change-Id: Iea72f864eac5ecfa97b5c99de5bd66ef09c54c0e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6516037 Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu fcfc9760 2025-05-06T11:26:40 Reformat VVL error message The test VulkanPerformanceCounterTest_ES31.TextureSampleByDrawDispatchDraw no longer runs into the VVL error it used to have. Remove the VVL error from the skipped VVL error list. Bug: angleproject:42265504 Change-Id: I8eff929524b3e1bd015d96db2cc45b2c8f331170 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6514411 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi 7447e8ac 2025-05-06T10:33:47 Vulkan: Driver workarounds on Qualcomm Dynamic rendering is disabled on 512.800 due to crashes on some devices. Imageless framebuffers are disabled on 512.805, suspecting cause of crash on some other devices. Bug: chromium:415738891 Bug: chromium:415968761 Change-Id: I71f533c7954b628ce897ed1dac478aed119658e4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6512663 Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Yuxin Hu c6e0448b 2025-05-05T10:49:32 Reformat VVL error message The SYNC-HAZARD-WRITE-AFTER-READ VVL error is no longer observed on TraceTest.special_forces_group_2, removed the error message from skipped VVL error list. The TraceTest.special_forces_group_2 now runs into a different SYNC-HAZARD-READ-AFTER-WRITE VVL error, on Linux AMD GPU. An existing entry in the skipped VVL error list already covers this error message. Reformat that entry by adding the extraProperties, and changing the messageContents1/messageContents2 fields to nullptr for ease of clean up later. Bug: angleproject:42264123 Bug: angleproject:397775556 Change-Id: I1681cd7f76dd2090e78584b08c25224d5b05d963 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6514289 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi 2ca90dab 2025-05-02T22:38:38 Vulkan: Don't calculate vertex input state when dynamic When vertex input state is dynamic, VkGraphicsPipelineCreateInfo::pVertexInputeState can be nullptr. This change makes sure time is not wasted calculating this state at pipeline creation time. The change also ensures drivers don't rely on info provided at compile time by mistake. Bug: angleproject:42265637 Change-Id: I54da8be7769f0dd0767df6d7774baa956475bf8d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6510082 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Matthew Denton c60d98ce 2025-05-03T01:06:02 WGPU: Ship driver uniforms in the bind groups to the shader This writes driver uniforms to a WGPU buffer and ships that buffer to the shader. Right now this creates a new buffer every time a driver uniform changes. Bug: angleproject:389145696 Change-Id: Icfbce4d995ef593e703a597bb96e72173ec29cda Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6477136 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Yuxin Hu ce795fa8 2025-04-28T15:31:25 Reformat VVL error message Set messageContents1 and messageContents2 fields to nullptr, for ease of clean-up in follow-up changes. Make the extra properties fields more specific. Bug: angleproject:42265363 Bug: angleproject:415382790 Bug: angleproject:415383266 Change-Id: I032bc82c6aeb16281c868778db827318fa833005 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6496650 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Amirali Abdolrashidi 18ec8824 2025-05-02T12:42:26 Vulkan: Exclude RADV from global cache preference * Added exception for AMD/RADV machines from the following flag: preferGlobalPipelineCache * (Due to no information regarding the effect of this flag on those platforms) Bug: b/411442610 Change-Id: Ifd9e2fe02e488d51cbb9e070106e86ac21e8533d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6509821 Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Auto-Submit: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 16e0dfd9 2025-03-20T10:59:26 Vulkan: Add feature flag to recycle VkEvent This CL adds a feature flag to enable/disable the VkEvent recycling. Given ARM prefers taking CPU overhead of creating and destroying VkEvent when last reference count to VkEvent goes away instead of the extra GPU overhead comes with vkCmdResetEvents, VkEvent recycling is disabled on ARM. Bug: angleproject:409441956 Change-Id: I288be6f7cfcbc4dd9c4701d20fba44349f25befc Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6376784 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Junji Watanabe f6e1d4d7 2025-05-02T12:02:09 Fix include header case mismatches This CL fixes the include header names to be aligned with the actual file names. It's problematic for cross compile on Linux, where file system is case sensitive. Context: Goma/Reclient's deps scanner was able to handle those case mismatches. But, Siso's builtin remote execution doesn't resolve them automatically. Bug: b/401959048 Change-Id: I98282899e31ea801058f82d4cad770413634f4f3 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6508684 Reviewed-by: Charlie Lao <cclao@google.com> Auto-Submit: Junji Watanabe <jwata@google.com> Reviewed-by: Yuly Novikov <ynovikov@chromium.org> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Amirali Abdolrashidi a3b20796 2025-04-25T13:12:08 Vulkan: Add flag to prefer renderer pipeline cache * Added the following feature: preferGlobalPipelineCache * The renderer's pipeline cache will be used to create pipelines, including for warm-up. * mPipelineCache in the program is not saved when this flag is enabled. * Currently enabled for NVIDIA and AMD. * Impact on captured trace (w/ compiling inactive shaders) * Decrease in peak system memory usage on Windows/NVIDIA: * ~7600 MB -> ~3900 MB * Decrease in peak RES memory usage on Linux/NVIDIA: * ~7100 MB -> ~3600 MB Bug: b/411442610 Change-Id: I04929569f0f8d59a77c52505072faa0244ef1393 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6495155 Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com>
Vasiliy Telezhnikov c52183bd 2025-04-11T11:17:10 Reset sample coverage after fbo change on Adreno Adreno drivers cache some internal value based number of samples in currently bound FBO and glSampleCoverage value, but doesn't update it on FBO change. So if we set sample coverage to 1 when FBO with 1 samples bound and then switch to FBO with 4 samples it behaves like coverage is 0.25 instead of 1. To avoid this, reset sample coverage after each FBO change. Bug: chromium:408364831 Change-Id: I7d5e36926e8568f4c652049b43ecdbe382ddc293 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6447878 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Vasiliy Telezhnikov <vasilyt@chromium.org>
Shahbaz Youssefi 9bcfb38f 2025-04-30T23:12:19 Vulkan: Disable MSRTSS on buggy qualcomm drivers Bug: chromium:413427770 Change-Id: I62499c3aa04f3786a0deb6afcaafb7976f2a6446 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6501621 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Matthew Denton 4b69ba93 2025-04-28T16:45:56 WGPU: upload texture bind groups and configured samplers GLSL samplers are split into separate sampler/texture variables in WGSL. Before this CL, the WGSL translator generated shaders that look like: @group(1) @binding(@@@@@@) var ANGLE_sampler_samp2D : sampler; @group(1) @binding(@@@@@@) var ANGLE_texture_samp2D : texture_2d<f32>; @group(1) @binding(@@@@@@) var ANGLE_sampler_sampCube : sampler; @group(1) @binding(@@@@@@) var ANGLE_texture_sampCube : texture_cube<f32>; This CL replaces those with actual binding numbers @group(1) @binding(0) var ANGLE_sampler_samp2D : sampler; @group(1) @binding(1) var ANGLE_texture_samp2D : texture_2d<f32>; ... Such that @binding(n*2) is the WGSL sampler variable corresponding to the n-th GLSL sampler and @binding(n*2+1) is the WGSL texture variable corresponding to the n-th GLSL sampler. This CL then generates binding group layouts matching the above, and uploads textures and configured samplers in bind groups. This makes some of the deqp_gles2 tests 2d texture tests pass, though some fail because they need a flipped y coordinate. Not yet supported: 1. arrays of samplers 2. shadow samplers 3. cube textures Bug: angleproject:389145696 Change-Id: I2ab18ae5ebb4d1289101266bd9451576aa04ce2a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6382272 Reviewed-by: Liza Burakova <liza@chromium.org> Auto-Submit: Matthew Denton <mpdenton@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi c884460a 2025-04-24T16:43:04 Vulkan: Fix deferred clears vs noop multidraw Bug: chromium:407828338 Change-Id: I5da22aeb72605bb7943fa5ae079ae297d00888f7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6488794 Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Cody Northrop <cnorthrop@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Neil Zhang e5680ba6 2025-04-14T17:43:07 Fix mInFlightCommands.empty() assert in CommandQueue destroy When initialize a higher priority shared Context, ANGLE internal call updateContextsPriority to submit previously flushed commands. This submission is not tracked by any Context. Add Renderer::mSubmittedResourceUse to record submitted queue serials not belonging to any Context to fix the assertion. Bug: angleproject:410996556 Change-Id: I7a91e47c0233832b8806637f295dffd1ff54f729 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6458623 Reviewed-by: Igor Nazarov <i.nazarov@samsung.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Neil Zhang <Neil.Zhang@arm.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi d392eeae 2025-04-22T10:54:13 Vulkan: Force host-image-copy for LUMA textures on ARM Bug: angleproject:42266771 Change-Id: I79a8fe14ba389195d5846261752be847bc13bcf4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6479371 Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Xin Yuan cdcf6a63 2025-04-22T10:19:33 Add R10X6G10X6B10X6A10X6 support on angle Bug: angleproject:409355677 Change-Id: I8115ca73a038360f673552532e9d9b87ff846867 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6438114 Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Peng Huang d698500d 2025-04-08T09:05:42 Fix an ASSERT() crash in ExternalFence SyncWaitFd(). https://registry.khronos.org/vulkan/specs/latest/man/html/vkGetFenceFdKHR.html If pGetFdInfo->handleType is VK_EXTERNAL_FENCE_HANDLE_TYPE_SYNC_FD_BIT and the fence is signaled at the time vkGetFenceFdKHR is called, pFd may return the value -1 instead of a valid file descriptor. So when above scenario happens, ANGLE will get a -1 fd from vkGetFenceFdKHR(), and then Angle will call SyncWaitFd() with the invalid fd, it will cause an ASSERT() crash. Bug:angleproject:409342980 Change-Id: I48d4d21d531e40f8ba275918454632ae14fec04b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6439303 Auto-Submit: Peng Huang <penghuang@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Igor Nazarov 691140da 2025-04-22T21:05:44 Vulkan: Do not skip VUID-vkQueueSubmit-pSignalSemaphores-00067 ... when "VK_EXT_swapchain_maintenance1" is enabled. Original problem is a VVL bug, introduced in: 62e9314e2eb9c58af42f6f1b613d200d5362598d layers: Track semaphore is in use by swapchain It was later partially fixed in: 88934ae6acf0c8b8ae169df9607e779ac4a70239 layers: Fix waiting on present fence did not have effect Bug is only fixed when using present fences (when "VK_EXT_swapchain_maintenance1" is enabled). Without present fences, VUID error still prints the following: Swapchain image 0 was presented but not re-acquired, so VkSemaphore 0x41c700000041c7 may still be in use and cannot be safely reused with image index 0. a) Use a separate semaphore per swapchain image. Index these semaphores using the index of the acquired image. Since problem only happens in a shared present mode, re-acquiring the image causing "VUID-vkAcquireNextImageKHR-surface-07783". It is also unclear from the Vulkan spec if the image may be re-acquired in shared present mode, but it is currently not necessary. When not using present fences, ANGLE tracks present semaphores by attaching submit serial to the previous present operation when same image is presented. This guarantees to use more than one present semaphore (in a shared present mode) and not to reuse semaphore with signal operation pending. This is more strict synchronization than what is suggested by the VUID error in the item (a). However, it is debatable if this is ANGLE's or VVL error. If it is not a VVL error, it at least should provide correct message of how to fix the problem in case of a shared present mode. Theoretically, ANGLE may fix this error by only using present semaphore for the initial present, and skip it for all subsequent present operations, since image is already shared with the presentation engine and present operation is only used for notification purposes. Bug: angleproject:408190758 Change-Id: Id0dc3c0efcba72e844ba6b2c241d0076f70fcd51 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6480175 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 6f9f343a 2025-04-16T14:45:33 Reland "Vulkan: Prune superseded updates during host-image-copy" This is a reland of commit f8a8adf63c5834a725d1bd752acc6d63c3f45c05 Original change's description: > Vulkan: Prune superseded updates during host-image-copy > > Bug: angleproject:42266771 > Change-Id: I7aa8e85720a54537389df615b575ddfa9951770d > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6462475 > Reviewed-by: Yuxin Hu <yuxinhu@google.com> > Reviewed-by: Cody Northrop <cnorthrop@google.com> > Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> > Commit-Queue: Cody Northrop <cnorthrop@google.com> Bug: angleproject:42266771 Change-Id: I95e6f1d2421327a6394e627f03ef5b4cddf4a1e1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6480011 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Tingwei Guo 3f88a505 2025-04-03T17:29:20 GenerateMipmap encode and decode sRGB when GL_SKIP_DECODE_EXT is set According to the spec, if TEXTURE_SRGB_DECODE_EXT is set before GenerateMipmap for a texture with format sRGB, mipmap generation should skips the encode and decode steps during mipmap generation, matching the mipmap generation for a non-sRGB texture. Bug: angleproject:406829410 Change-Id: Icea8c27a1ab2f20e9fd6c8d79187c0c5cfe8e789 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6428506 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Tingwei Guo <tingwei.guo@arm.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: mohan maiya <m.maiya@samsung.com>
Shahbaz Youssefi 50a4ad19 2025-04-22T09:29:17 Revert "Vulkan: Prune superseded updates during host-image-copy" This reverts commit f8a8adf63c5834a725d1bd752acc6d63c3f45c05. Reason for revert: Has a bug in the calculation of image layers whose updates should be dropped. Bug: angleproject:42266771 Original change's description: > Vulkan: Prune superseded updates during host-image-copy > > Bug: angleproject:42266771 > Change-Id: I7aa8e85720a54537389df615b575ddfa9951770d > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6462475 > Reviewed-by: Yuxin Hu <yuxinhu@google.com> > Reviewed-by: Cody Northrop <cnorthrop@google.com> > Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> > Commit-Queue: Cody Northrop <cnorthrop@google.com> Bug: angleproject:42266771 Change-Id: I2b992c8598f58c40386338ade3cf59ffdb03abd5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6480413 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Shahbaz Youssefi f355e2b3 2025-04-15T18:58:25 Vulkan: Remove preferDriverUniformOverSpecConst This was practically true for every vendor on Android (where rotation matters). For Qualcomm, it was also true due to a bug in version checking and didn't seem to be causing any concerns. Where pre-rotation is supported, it is better to enable this feature to avoid excessive pipeline creation. This change removes the feature and makes sure ANGLE always uses uniforms for rotation instead of spec consts. While technically this may have an adverse effect on platforms that never need pre-rotation, the ability is retained for all vendors since pre-rotation is finding its way into more platforms and would likely eventually be needed everywhere anyway. Bug: angleproject:42265878 Bug: angleproject:42262166 Change-Id: I4b64c04da46db08cfdd44b60789b66d93d8e8b17 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6459025 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: mohan maiya <m.maiya@samsung.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Cody Northrop <cnorthrop@google.com> Reviewed-by: Cody Northrop <cnorthrop@google.com>
Le Hoang Quyen 07eeca12 2025-04-17T20:58:19 Metal: prevent padding in state cache's structs. Fixed: angleproject:411845677 Change-Id: I8a9635634e66f86c132205f75825a916481f7ad0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6468239 Commit-Queue: Quyen Le <lehoangquyen@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org>
Le Hoang Quyen f6a350de 2025-04-17T20:43:08 Metal: fix rasterSamplecount not used in comparing render passes. RenderPassDesc's rasterSamplecount was not used to comparing two render passes' descriptors. Note: no reported bug found related to this missing comparison yet. This bug most likely affects render passes without any attachment (PLS?). Because if the 2 render passes have the same attachments, their rasterSamplecount values must also be equal. There was also a missing rasterSamplecount comparison in RenderPipelineOutputDesc::operator==. However, this function is not used. When comparing two render pipelines, RenderPipelineDesc::operator== is used instead, which is currently implemented using memcmp. So removing the former to avoid confusions. Bug: angleproject:411845677 Change-Id: Ic9b1867226561d096c1fb8ca2194b657d29daa18 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6468238 Reviewed-by: Kenneth Russell <kbr@chromium.org> Commit-Queue: Quyen Le <lehoangquyen@chromium.org>
Shahbaz Youssefi f8a8adf6 2025-04-16T14:45:33 Vulkan: Prune superseded updates during host-image-copy Bug: angleproject:42266771 Change-Id: I7aa8e85720a54537389df615b575ddfa9951770d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6462475 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Cody Northrop <cnorthrop@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Cody Northrop <cnorthrop@google.com>
Amirali Abdolrashidi a9f2f87f 2025-04-15T13:46:34 Vulkan: Update version macros * Updated the deprecated version macros * VK_VERSION_MAJOR -> VK_API_VERSION_MAJOR * VK_VERSION_MINOR -> VK_API_VERSION_MINOR * VK_VERSION_PATCH -> VK_API_VERSION_PATCH * VK_MAKE_VERSION -> VK_MAKE_API_VERSION * Added a separate case for QCOM drivers in getVersionString(). * The major version is now OR-ed with 512. * (New drivers seems to have a major version of 512, which falls not within the major field, but the variant field, when using the VK_API_VERSION_x macros.) Bug: angleproject:42267006 Change-Id: Ic9b761d182ff49c3ede43c7bd14f366f945fbf36 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6460682 Reviewed-by: Cody Northrop <cnorthrop@google.com> Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Alexey Knyazev 4b9525b2 2025-04-04T00:00:00 Remove all getClientMajorVersion helpers Client version checks should use proper struct compare operations. Bug: angleproject:408843436 Change-Id: I92cd91d1e7e6daa761f79060835aa534c5671264 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6467460 Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya a6a7de5a 2025-04-16T14:28:13 Vulkan: Sync perf counters in finishImpl(...) .. instead of finish(...) so other code paths like onDestroy(...) have a chance to update perf counters before decoupling from the sharegroup Bug: angleproject:386749841 Change-Id: Ib0f847dda2274686ed9e5d3c8211ee5a364e4b65 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6463366 Commit-Queue: mohan maiya <m.maiya@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Alexey Knyazev 408f5853 2025-04-04T00:00:00 Make gl::Version data members private * Further simplified gl::Version implementation by making data members private and merging them. * Used proper version struct comparisons instead of accessing individual version components. * Moved known version constants to Version.h for broader availability. * Removed no longer used helpers: * PrivateState::getClientMinorVersion() * State::getClientMinorVersion() * Context::getClientMinorVersion() Bug: angleproject:408843436 Change-Id: I3ae8f495269d649253fa2381ecbfc018a184fa20 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6460787 Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Shahbaz Youssefi b64d380e 2025-04-15T00:54:22 Vulkan: Ghost texture's image backing if overwritten If the texture's image is in use by the GPU but is overwritten completely, this change releases the old image and creates a fresh one. If the texture was used in a render pass, this avoids breaking the render pass. Otherwise, it allows the new image to be initialized with VK_EXT_host_image_copy functionality. In the very least, an unnecessary barrier is avoided. As a targeted optimization, this functionality is limited to non-array 2D color textures which are known to benefit from this ghosting in applications. Other texture types can be easily added, but need lots of tests to be added. Bug: angleproject:42265356 Change-Id: I18a7587599e36f9f70109264ddc1003b24b8b2df Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6456345 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 0ed51ec7 2025-04-10T22:41:22 Vulkan: Disable color attachment usage for LUMA formats Bug: angleproject:42265356 Change-Id: If97fd59e85febfdf345a36097901a2015aa4183b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6447236 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Amirali Abdolrashidi 49d0a332 2025-04-08T16:07:17 Vulkan: Remove ring buffer allocators * Removed the ring buffer allocator functionality from ANGLE: angle::RingBufferAllocator * Also removed the related common files. * (Pool allocators will be used at all times.) * Removed the placeholder functions from the pool allocator. * Removed the following BUILD flag: angle_enable_vulkan_shared_ring_buffer_cmd_alloc * Removed redundant line from ContextVk. Bug: b/410036490 Change-Id: I368fb93a66ddfd192018b09f65004a32339abd5a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6442640 Reviewed-by: Igor Nazarov <i.nazarov@samsung.com> Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Alexey Knyazev 33e035f9 2025-04-04T00:00:00 Enforce correct usage of gl::Version In preparation for the implementation update, ensured that gl::Version struct is always initialized with 8-bit values. Bug: angleproject:408843436 Change-Id: Idcc925b49cadc7fdc4f05ff2acaed8020e72a32e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6458758 Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Mavis Deng 4be3143c 2025-04-10T15:18:52 Vulkan: Vertex attribute hole crash workaround When the start vertex index of glDrawArrays is not 0, this hole case will crash again. * Added a condition to bypass the attribute update if the attribute pointer is null. * Added a glDrawArrays(GL_TRIANGLES, 3, 3) in unit test VertexAttribPointerCopyBufferFromInvalidAddress Bug: angleproject:359729255 Change-Id: I1d7f50dc56080ef0b4e48c4c3c983189a26171c1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6445172 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com> Auto-Submit: Mavis Deng <mavis.deng@arm.com> Commit-Queue: Cody Northrop <cnorthrop@google.com>
Shelly Wang ac8607a9 2025-03-27T16:46:21 Vulkan: Disable graphics pipeline on ARM due to performance loss ARM is investigating on Graphics Pipeline Library(GPL) on contents whose performance loss is unacceptable. Therefore, it's recommended to disable GPL until the problematic area gets fixed and the final performance data is acceptable. Bug: angleproject:404581992 Change-Id: I2bcb5f8bca3d64be0e947d7262a615ef867532ef Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6400174 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Cody Northrop <cnorthrop@google.com> Reviewed-by: Cody Northrop <cnorthrop@google.com>
Shahbaz Youssefi a3f2545f 2025-04-11T14:03:08 Tests that exercise overwriting textures while in use The first three tests are credit Cheryl Wei <@arm.com>, adopted from https://chromium-review.googlesource.com/c/angle/angle/+/3366193 The change also exposed the fact that the force_fallback_format with ASTC leads to assertion failure if ASTC emulation is not built. This change adjusts the format table to fix that as well. Bug: angleproject:42265356 Change-Id: Ib68355f317472d8cc3f035b492b273ac452a8217 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6447202 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Igor Nazarov d32c6c03 2025-04-11T15:12:24 Vulkan: Treat present SURFACE_LOST_KHR as OUT_OF_DATE ... when in shared present mode. Handle `VK_ERROR_SURFACE_LOST_KHR` the same way as `VK_ERROR_OUT_OF_DATE_KHR`, because on some platforms (observed on Android) swapchain recreate still succeeds making this error behave the same as OUT_OF_DATE. In case of a real surface lost, following swapchain recreate will also fail, effectively deferring the failure. This will avoid generating swap buffers failure unnecessarily, which may cause abortion of the application. Bug: angleproject:409809766 Change-Id: Ie8abd6b10f160254c94abc0f799195d7414681e3 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6450580 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Yuly Novikov 5dd05578 2025-04-14T07:34:57 Revert "GL: Allow untranslated shaders to pass through on GLES" This reverts commit 4e77552b86a89b449ada6d6c18f84285f5812b1d. Reason for revert: breaks ChromeOS and fuzzers Bug: angleproject:398857482 Original change's description: > GL: Allow untranslated shaders to pass through on GLES > > Add an EGL extension EGL_ANGLE_create_context_passthrough_shaders which > uses the NULL translator and passes the original shader to the driver. > The parser is still used for shader reflection. > > Bug: angleproject:398857482 > Change-Id: I7c5fcc318c7e11931f78c08dcbf4764bf77d397d > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6297527 > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Commit-Queue: Geoff Lang <geofflang@chromium.org> Bug: angleproject:398857482, angleproject:410423936 Bug: chromium:410114655, chromium:410100607, chromium:410121218 Bug: chromium:410052365, chromium:410290507, chromium:410178288 No-Presubmit: true Change-Id: I45b01960637a1cda05d21a7df6d07465f6a8f5e9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6448984 Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Auto-Submit: Yuly Novikov <ynovikov@chromium.org> Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Geoff Lang 4e77552b 2025-02-24T18:04:32 GL: Allow untranslated shaders to pass through on GLES Add an EGL extension EGL_ANGLE_create_context_passthrough_shaders which uses the NULL translator and passes the original shader to the driver. The parser is still used for shader reflection. Bug: angleproject:398857482 Change-Id: I7c5fcc318c7e11931f78c08dcbf4764bf77d397d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6297527 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Kimmo Kinnunen 78048112 2025-01-15T15:13:14 Fix IndexRange::vertexIndexCount calculation Use only one loop. Avoid redundant primitive restart index parameter. Avoid calling GetPrimitiveRestartIndexFromType() because the algorithm relies on the value being numeric_limits<T>::max(). Fixes a bug where primitive restart case would process the value after first primitive restart twice, once in both for loops. This would result in incorrect vertexIndexCount. Fix by removing IndexRange::vertexIndexCount, and instead using IndexRange::mCount == 0 to signify empty range. Bug: angleproject:401284933 Change-Id: Ifaeb9949f2e852fb7c5ef80bc47f72bfabba21a6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6333541 Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com>
Yuxin Hu 1c0bcb06 2025-04-08T18:42:43 Reformat VVL error angleproject:42265363: dead_by_daylight hits the VVL error covered by another entry in the skip list. Move the dead_by_daylight comment above the other entry. angleproject:42265427: VVL error added to the skip list in this bug still occurs on blade_and_soul_revolution and FramebufferFetch* tests. Change the VVL error message to use the new formats: remove the messageContents1 and messageContents2, and add more extraProperties. Bug: angleproject:42265363 Bug: angleproject:42265427 Change-Id: Ie913a7345b9f16580507894edd4a2befa29e9418 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6442305 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Liza Burakova 3dcf109d 2025-04-10T17:19:09 WebGPU: Refactor syncClientArrays. This change adds a few helper methods for syncClientArrays to improve readability Bug: angleproject:401499251 Change-Id: Ida427f80e07f469b83ff12850169b2110e65743c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6441946 Reviewed-by: Matthew Denton <mpdenton@chromium.org> Commit-Queue: Liza Burakova <liza@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Alexey Knyazev 595d79e4 2025-04-04T00:00:00 Do not expose GL_ARB_sync That extension does not exist in OpenGL ES. Fixed: angleproject:409717088 Change-Id: I82234f0051c2c646ef15c22e1a0da1704ef3a35b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6448233 Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Igor Nazarov ce394c49 2025-04-10T10:33:57 Vulkan: Fix undefined behavior accessing empty vertex buffer This fixes undefined behavior detected by VVL: [ SYNC-HAZARD-READ-AFTER-WRITE ] vkCmdDraw(): READ_AFTER_WRITE hazard detected. vkCmdDraw reads vertex VkBuffer 0xe000000000e, which was previously written by vkCmdBeginRenderingKHR. No sufficient synchronization is present to ensure that a read (VK_ACCESS_2_VERTEX_ATTRIBUTE_READ_BIT) at VK_PIPELINE_STAGE_2_VERTEX_ATTRIBUTE_INPUT_BIT does not conflict with a prior write (VK_ACCESS_2_COLOR_ATTACHMENT_WRITE_BIT) at VK_PIPELINE_STAGE_2_COLOR_ATTACHMENT_OUTPUT_BIT. This is regression from the old CL: https://crrev.com/c/angle/angle/+/3297026 Vulkan: Don't attempt to convert vertices in empty buffers Above change made `if (bufferVk->getSize() == 0)` branch a dead code. In the current code this branch is `if (numVertices == 0)` and is no longer a dead code. It sets zero stride to prevent reading beyond the empty buffer (which is only 16 bytes). This fix also sets `mCurrentArrayBufferStrides[attribIndex]` to zero when it is not a streaming attribute (in case of zero sized buffer). Bug: chromium:1271671 Change-Id: I3460201ae7366b8de1089efef6fbefe359d28460 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6440439 Commit-Queue: Igor Nazarov <i.nazarov@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton f4ff9241 2025-04-08T14:37:08 WGPU: add utils for converting between GLSL and WGSL types ...specifically those related to textures and samplers. Bug: angleproject:389145696 Change-Id: If3bec56b70c04443928ef0d4e1e1a40b5c578cc0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6441818 Reviewed-by: Liza Burakova <liza@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Shahbaz Youssefi 02ae054c 2025-04-07T13:48:29 Translator: Fix init of inactive output variables Bug: chromium:398401939 Change-Id: I0df494b945b8d0e805a62cf7645d06bf233f36ca Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6438495 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Jiaqi Li a6f44806 2025-04-01T17:30:53 Wrong condition to call copyBufferDataToImage In function copyBufferDataToImage, it call vulkan vkCopyBufferToImage directly. To call copyBufferDataToImage directly, it needs that there's no overlap when copy. If mImage is not valid, it will assert in mImage->getActualFormatID(). Need to check its validation at beginning. Bug: angleproject:406566447 Change-Id: I875c2a97bec52251b9c5b3d196466e8bf5ce662d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6400313 Auto-Submit: Jiaqi Li <Jiaqi.Li@arm.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Yuxin Hu a0a5c565 2025-04-07T15:05:11 Vulkan: Add assert to check only clear is staged on swapchain image The swap after clear optimizations on swapchain image multisampled image https://chromium-review.googlesource.com/c/angle/angle/+/6072416 works with the assumption that the only staged updates on swapchain multisampled image is clear. This change adds an assert to validate the assumption holds true before applying the optimization. If the assert is ever triggered, we will know the assumption is violated, and we need to look into how to tweak the optimization or add other optimizations to solve the violated case. Bug: angleproject:382006939 Change-Id: I7ce23dd7767b7730f1fa3ec89cb6616689339602 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6437217 Reviewed-by: mohan maiya <m.maiya@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Cody Northrop d27a1fd0 2025-04-03T19:59:20 Compiler: Allow denorm float values in the lexer This adds an option to preserve denorm values in the lexer, skipping the explicit zero conversions for below range floats. There are applications in the wild that expect to be able to use denorm float values. They are typically immediately converted to integer values, not used in floating point operations. The option is only enabled for Vulkan backends. Test: FloatLexTest, DenormFloatsToIntValues, app traces Bug: b/406827038 Change-Id: Iab5a1a69a540b78ccbce8ea90b532d2d4976e29e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6432237 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Cody Northrop <cnorthrop@google.com>
Geoff Lang 0e28c030 2025-04-02T17:26:28 ScalarizeVecAndMatConstructorArgs on WGSL shaders. WGSL does not support mixing types in the constructors of vectors and matrices. Use the ScalarizeVecAndMatConstructorArgs traverser to scalarize these constructors and update it to cast to the destination type if the inputs are not in that type already. Bug: angleproject:395659799 Change-Id: I4c270f2b18e79393e3c187916f33ad719bc88a63 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6427067 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Igor Nazarov f798b0d2 2025-04-04T08:17:57 Vulkan: Remove enablePreRotateSurfaces feature Removed to simplify the code and to avoid the problem for which `presentSubOptimalReturnedOnTransformChange` feature was added. Platforms without the per-rotation support always have `VK_SURFACE_TRANSFORM_IDENTITY_BIT_KHR` as the `currentTransform`, so the code will perform the same as when the feature was disabled. Add `warmUpPreRotatePipelineVariations` to explicitly control per-rotation pipeline warm up. Bug: angleproject:42262166 Change-Id: I44f6c221c11105f01f62f62622987b1955bc58aa Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6433586 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Igor Nazarov b77f8af7 2025-04-02T19:01:48 Vulkan: Ignore SUBOPTIMAL on Android without pre-rotate Android returns `VK_SUBOPTIMAL_KHR` when window orientation changes, while ignores size changes. When "enablePreRotateSurfaces" is disabled, each present will return `VK_SUBOPTIMAL_KHR` if `currentTransform` is not `VK_SURFACE_TRANSFORM_IDENTITY_BIT_KHR`, causing swapchain recreate on each swap. This behavior makes disabling "enablePreRotateSurfaces" on Android worse than suboptimal. New "presentSubOptimalReturnedOnTransformChange" feature is currently enabled only on Android for the reason described above. It may be enabled for other platforms in the future if required. Change also improves error handling, by moving `deferAcquireNextImage()` call right after the `queuePresent()`, before any other call that may fail. This will avoid repeated present of the same image if something fails after successful `queuePresent()` call. The `checkForOutOfDateSwapchain()` and `computePresentOutOfDate()` was combined into a single method to improve error handling and to simplify the code. Bug: angleproject:397848903 Bug: angleproject:42266082 Bug: angleproject:42262166 Change-Id: I7ba2389ad2ddafc0c77c3068cd85ac9b9251ccdd Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6424753 Commit-Queue: Igor Nazarov <i.nazarov@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Yuxin Hu 127aec7e 2025-04-03T13:54:17 Vulkan: Remove VVL error skips that have been fixed Test: --gtest_filter='MemoryBarrierBufferTest.TransformFeedbackBitWriteThenCapture*' Bug: angleproject:42266506 Change-Id: I0261a26157f6e44d25e6631faec96a55e470ea66 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6431825 Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Yuxin Hu b8151923 2025-04-03T11:34:18 Vulkan: Remove VVL error skips that have been fixed Test: --gtest_filter='MemoryBarrierBufferTest.TransformFeedbackBitWriteThenCapture*' Bug: angleproject:42266506 Change-Id: I314ee10d9ee67f5871312d730f35c2232a9372b7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6430466 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
angle-autoroll 92cb669b 2025-04-03T15:26:42 Manual roll vulkan-deps from 12ef946decd5 to 48f50085fcda (21 revisions) Suppress VUID-vkQueueSubmit-pSignalSemaphores-00067 due to https://github.com/KhronosGroup/Vulkan-ValidationLayers/pull/9840 Manual roll requested by ynovikov@google.com https://chromium.googlesource.com/vulkan-deps.git/+log/12ef946decd5..48f50085fcda Also rolling transitive DEPS: https://chromium.googlesource.com/external/github.com/KhronosGroup/glslang/+log/e57f993cff981c8c3ffd38967e030f04d13781a9..c3d39de93955f884e443c39e9ffecf86e4aac883 https://chromium.googlesource.com/external/github.com/LunarG/VulkanTools/+log/27e60d9cd9d68cfcb6ca597956a4e57ac330413b..9e702e25aa472c8528b49c874a3c03d514df1ede https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Headers/+log/8c88e0c4c94a21de825efccba5f99a862b049825..8e82b7cfeca98baae9a01a53511483da7194f854 https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Tools/+log/056e3ed2c7ca7b30aa803685f806de7a2402aa85..4bd1536ed79003a5194a4bd8c9aa2fa17a84c15b https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-ValidationLayers/+log/73d7d74bc979c8a16c823c4eae4ee881153e000a..b2fee806c378105400a1ed4bc4b63ee3c08ddc4e If this roll has caused a breakage, revert this CL and stop the roller using the controls here: https://autoroll.skia.org/r/vulkan-deps-angle-autoroll Please CC angle-team@google.com,romanl@google.com,ynovikov@google.com on the revert to ensure that a human is aware of the problem. To file a bug in ANGLE: https://bugs.chromium.org/p/angleproject/issues/entry To report a problem with the AutoRoller itself, please file a bug: https://issues.skia.org/issues/new?component=1389291&template=1850622 Documentation for the AutoRoller is here: https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md Bug: angleproject:408190758 Change-Id: I835d826a71590c788c29cecd7958951e5b648df8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6430461 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Geoff Lang 4d478713 2025-04-03T08:10:40 Revert "Metal: Make StateCache descriptors hash consistent" This reverts commit 4334125be506e0c2c5dde4cfc29b31a7ff2bdec6. Reason for revert: Pipeline creation failures. Bug: chromium:407310859 Original change's description: > Metal: Make StateCache descriptors hash consistent > > Various descriptors would compare equality and do hashing based on the > underlying storage. The storage contains alignment padding and > bitfields. The contents of these are not defined, even when the > constructors tried to memset and memcpy. > > The various hash and compare functions also seemed to be missing some > elements, like rasterSampleCount. > > Missing == on fields and hashing unused memory may lead to inconsistency > where a==b is true but hash(a) != hash(b). > > Fix by: > Remove the memset/memcpy and write out the operator== and > hash() consistently. > > Initialize the members to their default values. > > Use uint32_t : 1 instead of bool if the struct tries to optimize for > size, since mixing types stops the packing. > > Use uint32_t to get the alignment to uint32_t, which is mostly what > is expected. > > Use uint32 someVar : N where N is payload bits + padding bits for > the last member. This way the assignment clears the padding bits. > The goal os to get the operator== to compare the padding bits, which > lets the compiler elide the memberwise compares and use word-wise > compares where it makes sense. > > This is an attempt to fix rare crashes related to map inserts > and lookups. std::unordered_map will crash if operator== is not > consistent with hash. > > Bug: angleproject:403372465 > Change-Id: I0a3a2ff327ac9f65e7d41bc9585cda54ce93ab86 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6355828 > Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com> > Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> > Reviewed-by: Alexey Knyazev <lexa.knyazev@gmail.com> > Reviewed-by: Geoff Lang <geofflang@chromium.org> Bug: angleproject:403372465 Change-Id: Id111c0728b603776fd8085fd395819bb5aa14722 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6428466 Commit-Queue: Quyen Le <lehoangquyen@chromium.org> Reviewed-by: Quyen Le <lehoangquyen@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Roman Lavrov a7e8c957 2025-04-01T17:08:01 Split SetUniform into inline and noinline path Makes a single smaller function for the common cases, and setting uniforms with a stride or boolean conversion is extracted to separate function calls. This results in the main SetUniform being about half in assembly size and yields a small but noticeable performance improvement in my tests on an Android device. Impact on the binary size is negligible. Bug: b/383305597 Change-Id: I630ff41cfeea6ce9e24e30cdebfbb58382e67de7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6422893 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Le Hoang Quyen 0f9feb49 2025-03-15T23:51:04 Metal: Use render passes for B2T PBO copies. This would make it possible to avoid setting ShaderWrite usage to all textures in future. Bug: chromium:392521084 Change-Id: I25145f3578589c71ee469df08162d3a3f17c8cbe Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6356129 Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Quyen Le <lehoangquyen@chromium.org>
Yuxin Hu 9ee2c7e7 2025-04-02T16:01:45 Vulkan: Remove VVL error skips that have been fixed Bug: angleproject:42265220 Change-Id: I828f1d96a19efd0de9220b2168cfa9c3aab80912 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6427323 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Alexey Knyazev f56c8e02 2025-03-28T00:00:00 Fix & cleanup advanced blend equations support * Made advanced blend equation extensions enableable to facilitate better testing. * Made initial BLEND_ADVANCED_COHERENT state to be unconditionally true regardless of the extension support. If the coherent extension is unsupported, that state must not be observable anyway. * Ensured that advanced blend equation functionality is supported on unextended OpenGL ES 3.2 contexts. * Removed redundant ES 2.0+ context version checks because these extensions must never be enabled on ES 1.x contexts. * Fixed ValidateBlendBarrierKHR to set the correct error code and properly mark invalid calls. * Added more negative test cases and enabled tests on OpenGL ES 3.0 contexts. Bug: angleproject:42262258 Bug: angleproject:406922380 Change-Id: I709376ad40b09972a31a691ca3813e87dd2c8165 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6421857 Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Yuxin Hu a625043f 2025-04-01T16:09:14 Vulkan: Remove VVL error skips that have been fixed Bug: angleproject:42265220 Change-Id: Ia2048817c6238031c83aaa237c6c88b0e9a9ebf9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6423550 Auto-Submit: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Roman Lavrov 23c1e11e 2025-04-01T11:27:02 Disable platform trace events by default, except Chromium build Observed overhead in perf builds is too significant, some workloads generate thousands of events per second. There does not seem to be a simple way to get around the overhead - it could be somewhat reduced but the event rate is just too high. ANGLE_TRACE_EVENT*(...) etc become a no-op by default. Can be re-enabled with angle_enable_platform_trace_events=true gn arg Bug: angleproject:404542398 Change-Id: I8d7363ee28d22226f1860e24f2399bb4129cfb84 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6419308 Commit-Queue: Roman Lavrov <romanl@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Feels: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Amirali Abdolrashidi bbf92d12 2025-03-21T12:43:14 Vulkan: Add workaround flag for extra submit fence This change is a workaround for some devices exhibiting a device-lost error related to the fence usage in the command batches when working with EGL sync objects, such as a cycle similar to below: - <Some GL work> - eglCreateSyncKHR() - glFlush() - eglDupNativeFenceFDANDROID() - eglDestroySyncKHR() - <More GL work> (The exact point of error and the reason for the device-lost remain unknown.) This change is meant to mitigate such errors on those devices if needed. However, for the rest, the feature will remain disabled. * Added a new feature flag: enableExtraSubmitFence * Modified the usage of mFence in CommandBatch * It is also used when enableExtraSubmitFence is enabled, even if there is an external fence. * When there is an external fence, this object is used to create an empty submission after the primary commands are submitted using the external fence. * Added the config to enable EnableExtraSubmitFence for the following suite: EGLSyncTest Bug: b/384477641 Change-Id: I05f8bdcc804967c4984416af802326d22afd7a46 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6378778 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao ecc378cc 2025-03-03T16:43:33 Reland "Dont use Subject/Observer for SwapchainImageChanged" This is reland of the following CL https://chromium-review.googlesource.com/c/angle/angle/+/6319893 Because we do deferred ANI (VkAcquireNextImage) call until image is needed, we need a way to force Context to go through FramebufferVk::syncState call (FramebufferVk::syncState calls WindowSurfaceVk::getAttachmentRenderTarget, which end up calling ANI. Right now we uses subject/observer mechanism, by sending angle::SubjectMessage::SwapchainImageChanged to all observers of WindowSurfaceVk. In this case it is egl::Surface. Then eglSurface redirects this message to its observers, which are all gl::Framebuffer's attachments: color, depth, stencil. Even though only color attachment needs to be notified, but because we don't have a separate list of observers, depth/stencil attachment also receive the notification and they early out. Then gl::Framebuffer sets DIRTY_BIT_COLOR_BUFFER_CONTENTS_0 dirty bit and send the angle::SubjectMessage::DirtyBitsFlagged to Context, which dirty DrawFBO and ReadFBO and dirty cached state. Note that this is specific for swap image changed case, there is no surface property change (surface property change will still trigger the subject/observer message with SubjectMessage::SubjectChanged message, but this occurs rarely). This gets worse for apps that uses multiple contexts, for the example pokemon_masters_ex has three contexts, each context has its own default frame buffer that attach to the same surface, and we never remove non-current context from the observer list. This end up with egl::Surface has 12 observers and for every frame, it loop over the list of 12 observers and send message (virtual function call) to each of them. Color attachment also ends up sending two messages to Context, one for Read FBO and another for Draw FBO. There are total 21 virtual function calls. Even for single context usage, you have 6 virtual function calls, for every frame. EGL spec says "an EGLSurface must be current on only one thread at a time", any other context must call EGLMakeCurrent in order to use this surface, which will add all necessary dirty bits at that time. So we really only need to notify current context. In this CL, SwapchainImageChanged no longer uses subject/observer mechanism, so this message is removed. This CL still uses subject/observer mechanism to send DirtyBitsFlagged from Framebuffer back to context. We could call setDrawFramebufferDirty and setReadFramebufferDirty directly, but that will require to remove the "const" decoration out of gl::Context which generates too much code diff, so onStateChange(angle::SubjectMessage::DirtyBitsFlagged) is still used. Bug: angleproject:400711938 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6319893 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Change-Id: I017b0e8934b5194a520828fa5c4af1d6e3ef9aac Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6404621
Yuxin Hu eb8deecd 2025-03-28T14:03:50 Remove VVL error skips that have been fixed Bug: angleproject:42265220 Change-Id: I1588db6ec2846faf92eef9042ba483547c0f5127 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6410511 Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Kimmo Kinnunen 4334125b 2025-03-14T16:12:32 Metal: Make StateCache descriptors hash consistent Various descriptors would compare equality and do hashing based on the underlying storage. The storage contains alignment padding and bitfields. The contents of these are not defined, even when the constructors tried to memset and memcpy. The various hash and compare functions also seemed to be missing some elements, like rasterSampleCount. Missing == on fields and hashing unused memory may lead to inconsistency where a==b is true but hash(a) != hash(b). Fix by: Remove the memset/memcpy and write out the operator== and hash() consistently. Initialize the members to their default values. Use uint32_t : 1 instead of bool if the struct tries to optimize for size, since mixing types stops the packing. Use uint32_t to get the alignment to uint32_t, which is mostly what is expected. Use uint32 someVar : N where N is payload bits + padding bits for the last member. This way the assignment clears the padding bits. The goal os to get the operator== to compare the padding bits, which lets the compiler elide the memberwise compares and use word-wise compares where it makes sense. This is an attempt to fix rare crashes related to map inserts and lookups. std::unordered_map will crash if operator== is not consistent with hash. Bug: angleproject:403372465 Change-Id: I0a3a2ff327ac9f65e7d41bc9585cda54ce93ab86 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6355828 Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com> Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Robic Sun 7fb1ac40 2025-03-10T20:28:56 Support EGLImageTargetTexture2DOES + GenerateMipmap According to the spec, If an application adds mipmap levels to a texture originally created with EGLImageTargetTexture2D, OpenGL will allocate new memory for the updated texture, copying existing data, including from the EGLImage source. So, relevant support is added. Bug: angleproject:405840226 Change-Id: I4495df2d8ee1e18b4482808a290dcd861ef85b77 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6335305 Auto-Submit: Robic Sun <Robic.Sun@arm.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Charlie Lao <cclao@google.com>
Igor Nazarov 8cf89716 2025-03-14T20:17:07 Vulkan: Remove perFrameWindowSizeQuery feature Feature was enabled for all platforms in order for surface to be resized before acquire next image (not only after swap). Remove it, as if it's always enabled to simplify the code. Bug: angleproject:397848903 Bug: angleproject:42262287 Bug: angleproject:42262286 Bug: angleproject:40096601 Change-Id: I768772e30f5f38f68992e5b82c84430732aa77d9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6354166 Commit-Queue: Igor Nazarov <i.nazarov@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu e4bfa483 2025-03-27T10:54:19 Remove VVL error skips that have been fixed Bug: angleproject:42265220 Change-Id: I9d6cb7ee4ace27cad5ef3db84d83ded6ca9095f8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6404613 Auto-Submit: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
Igor Nazarov 081934f6 2025-03-12T15:22:31 Vulkan: Add avoidInvisibleWindowSwapchainRecreate feature There is a bug on some NVIDIA drivers on Linux X11 when `vkAcquireNextImageKHR()` constantly returns `VK_ERROR_OUT_OF_DATE_KHR` if recreate the swapchain while window is not visible. If not recreate the swapchain after window resize - nothing happens, acquire and present works without errors. But if recreate the swpahchain to the `VkSurfaceCapabilitiesKHR::currentExtent` the next `vkAcquireNextImageKHR()` will return `VK_ERROR_OUT_OF_DATE_KHR`. ANGLE will retry to recreate the swapchain one more time and fail the call. Enabling the "avoidInvisibleWindowSwapchainRecreate" feature will avoid swapchain recreation when window size changes while it is not visible. Test: angle_end2end_tests --gtest_filter=EGLSurfaceTest.ResizeInvisibleWindow/* Bug: angleproject:397848903 Bug: angleproject:42264022 Bug: angleproject:42263074 Change-Id: I48588bf467d15c0e84b923092e06a42c22084dcc Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6348739 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Igor Nazarov caf14998 2025-02-25T17:28:05 Vulkan: Fix window surface resize and size query Added `mSizeState` enum member. When it is not `Resolved`, surface size is not yet resolved and may change any time in response to native window resize. Otherwise, surface size is resolved and will not change until next swap. Updated `getUserWidth()`/`getUserHeight()` methods. They return current window size when surface size is not resolved, or current surface size, otherwise. Window size is queried either by `getCurrentWindowSize()`, when surface is sized by swapchain, or by querying vulkan surface capabilities. The "perFrameWindowSizeQuery" feature controls when `mSizeState` is updated. Initially it is not resolved. Marked as not resolved in `invalidateSwapchain()`, and in `deferAcquireNextImage()` when the feature is enabled. It is marked as resolved after swapchain is created when feature is disabled, or in `AcquireNextImageUnlocked()` function after acquire next image, otherwise. Enabled some tests that were previously failing. Replaced old resize tests with new versions. Test: angle_end2end_tests --gtest_filter=EGLSurfaceTest.Resize*/* Bug: angleproject:397848903 Bug: angleproject:373659619 Bug: angleproject:153329980 Bug: angleproject:42266013 Bug: angleproject:42265843 Bug: angleproject:42265529 Bug: angleproject:42264022 Bug: angleproject:42263074 Bug: angleproject:42261800 Bug: angleproject:40096826 Change-Id: I3ad836960a68229fab6c94624022f1a0aaf2c3e5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6300645 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com> Reviewed-by: Charlie Lao <cclao@google.com>
Igor Nazarov e44357c6 2025-03-25T15:50:15 Vulkan: Do not defer ANI in shared present mode Calling ANI in shared present mode is not necessary after the initial acquire. Skip calling `deferAcquireNextImage()` in that mode. For code simplicity and as a preparation for (anglebug.com/400711938), the `invalidateSwapchain()` method is now also defers ANI and notifies `SubjectMessage::SurfaceChanged` to the parent class. `Framebuffer` message handling was updated to also reset color buffer dirty bit and to notify `SubjectMessage::DirtyBitsFlagged` if needed, as in the `SubjectMessage::SwapchainImageChanged` message case (which will be removed in future CL). This way, single `SubjectMessage::SurfaceChanged` message is enough to handle swapchain recreate. The `VK_SUBOPTIMAL_KHR` is no longer treaded as OUT_OF_DATE when in shared present mode. Added for consistency (since we are already skipping "perFrameWindowSizeQuery" checks), to preserve content, and to match the Android native GLES driver behavior. Call `invalidateSwapchain()` when swapchain operations fail to avoid repeated swapchain use and to be able to recover from the error. The `checkForOutOfDateSwapchain()` was split into two methods: - checkForOutOfDateSwapchain(): - Called only after present. - Checks present out of date result and present mode compatibility. - Invalidates the swapchain and updates the present mode if the above checks fails. - prepareSwapchainForAcquireNextImage(): - Calls `queryAndAdjustSurfaceCaps()` and `recreateSwapchain()` if swapchain is invalid. - Calls `queryAndAdjustSurfaceCaps()` and checks surface properties when "perFrameWindowSizeQuery" is enabled. Then calls `recreateSwapchain()` if something changed. Other changes: - The `prepareForAcquireNextSwapchainImage()` method was replaced with `prepareSwapchainForAcquireNextImage()`. - Removed `doDeferredAcquireNextImageWithUsableSwapchain()` and `postProcessUnlockedAcquire()` methods because of redundancy. - Move image invalidation code into `acquireNextSwapchainImage()` to make `doDeferredAcquireNextImage()` simpler. - Convert `resizeSwapchainImages()` into `createSwapchainImages()` for simplicity. Updated old and added new tests. Test: angle_end2end_tests --gtest_filter=EGLSurfaceTest.ReadFramebufferBindingSyncState/* Test: angle_end2end_tests --gtest_filter=EGLSingleBufferTest.OnCreateWindowSurface/* Test: angle_end2end_tests --gtest_filter=EGLSingleBufferTest.OnSetSurfaceAttrib/* Test: angle_end2end_tests --gtest_filter=EGLSingleBufferTest.WindowResize/* Test: angle_end2end_tests --gtest_filter=EGLSingleBufferTest.WindowRotation/* Bug: angleproject:400711938 Bug: angleproject:397848903 Bug: angleproject:42262606 Change-Id: I2247417aa8b7b5afc10a8420083aeb845895aec9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6387920 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Charlie Lao e339f91c 2025-03-21T10:22:59 Vulkan: Split asyncCommandBufferResetAndGarbageCleanup into two Right now this one feature flag controls garbage clean up and command buffer reset. If this is enabled, we are seeing command buffer reset some times runs on small core and some times gets blocked by mutex lock inside vulkan driver. This could take quite long while main rendering thread is blocked by ANGLE's CommandPoolAccess lock. This CL splits this feature flag into two separate feature flag: asyncGarbageCleanup controls garbage clean up in the async thread or not. asyncCommandBufferReset controls commandBuffer.reset in the async thread or not. This CL also disables commandBuffer.reset in async thread only on ARM given there is no data shows other GPUs suffer form the same problem. Bug: angleproject:378718508 Change-Id: Ice87b5b91568a0a95e0064da2b70243516ff6753 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6381893 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Roman Lavrov <romanl@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu f841d12a 2025-03-26T14:22:16 Remove VVL error skips that have been fixed Bug: angleproject:42265220 Change-Id: Ieb0638e3c28dd4dacaeebdcc45de381c994f4bdb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6397943 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Kimmo Kinnunen c2679dbc 2025-03-10T12:15:08 Key vertex array index range cache on restart Key vertex array index range cache on primitive restart enablement. Add the test to WebGLCompatibilityTest as otherwise the VertexArray::getIndexRange() is not testable. Currently WebGL compatibility contexts allow changing primitive restart. Bug: angleproject:401284933 Change-Id: I48a53770d7dcb2276b89fd743f4834c53c8d8d1e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6333538 Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org>
Liza Burakova 5024ccef 2025-03-25T11:37:19 WebGPU: Emulate line loops with primitive restarts Bug: angleproject:383356846 Change-Id: Iab1d658cb3c5e32f88241f6757d1c2e5a84bf8ec Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6297524 Commit-Queue: Liza Burakova <liza@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Aurora Zhang 34db1c20 2025-03-20T12:56:42 Vulkan: glReadPixels should nearest sample the YUV render target Vulkan backend uses texture except texelFetch to obtain the YUV image value for glReadPixels. According to the chroma filter set in the image, it may apply LINEAR sampling. * Change this filter to NEAREST inside glReadPixels to get the correct pixel values. * Modify the fragment shader used in glReadPixels to make sure sampling from the center of the pixel. * Add a new end2end test. Bug: angleproject:404394628 Change-Id: I0e574a464639e11671f0a09eee3e9f38b5abd919 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6366299 Auto-Submit: Aurora Zhang <Aurora.Zhang@arm.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Yuly Novikov 3cfc0ce2 2025-03-24T07:41:33 Revert "Vulkan:Dont use Subject/Observer for SwapchainImageChanged" This reverts commit 48103cb2f2b292cb50cc5a29546b358b2e47fd29. Reason for revert: assert fails https://ci.chromium.org/ui/p/angle/builders/ci/android-arm64-exp-test/7085/overview I 22:27:33.697 77.533s _RunTestsOnDevice(17221FDF6000A4) [ RUN ] EGLAndroidAutoRefreshTest.SwapCPUThrottling/ES3_Vulkan_NoFixture INFO:root:ERR: SurfaceVk.cpp:3165 (getCurrentFramebuffer): ! Assert failed in getCurrentFramebuffer (../../src/libANGLE/renderer/vulkan/SurfaceVk.cpp:3165): mAcquireOperation.state == ImageAcquireState::Ready Original change's description: > Vulkan:Dont use Subject/Observer for SwapchainImageChanged > > Because we do deferred ANI (VkAcquireNextImage) call until image is > needed, we need a way to force Context to go through > FramebufferVk::syncState call (FramebufferVk::syncState calls > WindowSurfaceVk::getAttachmentRenderTarget, which end up calling ANI. > Right now we uses subject/observer mechanism, by sending > angle::SubjectMessage::SwapchainImageChanged to all observers of > WindowSurfaceVk. In this case it is egl::Surface. Then eglSurface > redirects this message to its observers, which are all gl::Framebuffer's > attachments: color, depth, stencil. Even though only color attachment > needs to be notified, but because we don't have a separate list of > observers, depth/stencil attachment also receive the notification and > they early out. Then gl::Framebuffer sets > DIRTY_BIT_COLOR_BUFFER_CONTENTS_0 dirty bit and send the > angle::SubjectMessage::DirtyBitsFlagged to Context, which dirty DrawFBO > and ReadFBO and dirty cached state. Note that this is specific for swap > image changed case, there is no surface property change (surface > property change will still trigger the subject/observer message with > SubjectMessage::SubjectChanged message, but this occurs rarely). This > gets worse for apps that uses multiple contexts, for the example > pokemon_masters_ex has three contexts, each context has its own default > frame buffer that attach to the same surface, and we never remove > non-current context from the observer list. This end up with > egl::Surface has 12 observers and for every frame, it loop over the list > of 12 observers and send message (virtual function call) to each of > them. Color attachment also ends up sending two messages to Context, one > for Read FBO and another for Draw FBO. There are total 21 virtual > function calls. Even for single context usage, you have 6 virtual > function calls, for every frame. > > EGL spec says "an EGLSurface must be current on only one thread at a > time", any other context must call EGLMakeCurrent in order to use this > surface, which will add all necessary dirty bits at that time. So we > really only need to notify current context. In this CL, > SwapchainImageChanged no longer uses subject/observer mechanism, so this > message is removed. > > This CL still uses subject/observer mechanism to send DirtyBitsFlagged > from Framebuffer back to context. We could call setDrawFramebufferDirty > and setReadFramebufferDirty directly, but that will require to remove > the "const" decoration out of gl::Context which generates too much code > diff, so onStateChange(angle::SubjectMessage::DirtyBitsFlagged) is still > used. > > Bug: angleproject:400711938 > Change-Id: I61354516fd0aa307714b7abd30c6b6e45ff7b496 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6319893 > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Yuxin Hu <yuxinhu@google.com> Bug: angleproject:400711938 Change-Id: Ib7899d1ac63a1f86af0953a1d25922578c470fc9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6387755 Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Jiaqi Li cbe26d0c 2025-03-04T16:56:26 Wrong texture coordinates sent to VULKAN for blit In ANGLE's Vulkan backend, complex blit operations are implemented using draw calls that utilize samplers and shaders, with texture coordinates passed via push constants. In this test case, texture coordinates are calculated based on the effective framebuffer size, which is determined by the smallest attachment dimensions. However, the draw call is executed on an attachment with its original size, leading to unexpected texture coordinates and incorrect sampling results. Use ReadImageExtent size instead of framebuffer effective size to calculate the sample texture coordinate. Bug: angleproject:400584608 Change-Id: I06f0a1e48463f36aef7189f772515561d2e752d1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6321348 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Jiaqi Li <Jiaqi.Li@arm.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Igor Nazarov 1166eec4 2025-03-20T02:02:50 Move TRACE_EVENT inside acquireNextSwapchainImage() The `acquireNextSwapchainImage()` is no longer called from surface initialize, which allows to move the trance event inside the function. Change also uses loop to handle repeated ANI retry instead of call duplication. This is done to further reduce code duplication in the next CL. Additionally, fixed assert that was missed in recently merged CL: Vulkan: Remove recreateSwapchain ContextVk dependency https://crrev.com/c/angle/angle/+/6298733 Bug: angleproject:397848903 Bug: angleproject:42261625 Change-Id: Iae2ce35018d4fe105430271b56804af8c74152c2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6372565 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 48103cb2 2025-03-03T16:43:33 Vulkan:Dont use Subject/Observer for SwapchainImageChanged Because we do deferred ANI (VkAcquireNextImage) call until image is needed, we need a way to force Context to go through FramebufferVk::syncState call (FramebufferVk::syncState calls WindowSurfaceVk::getAttachmentRenderTarget, which end up calling ANI. Right now we uses subject/observer mechanism, by sending angle::SubjectMessage::SwapchainImageChanged to all observers of WindowSurfaceVk. In this case it is egl::Surface. Then eglSurface redirects this message to its observers, which are all gl::Framebuffer's attachments: color, depth, stencil. Even though only color attachment needs to be notified, but because we don't have a separate list of observers, depth/stencil attachment also receive the notification and they early out. Then gl::Framebuffer sets DIRTY_BIT_COLOR_BUFFER_CONTENTS_0 dirty bit and send the angle::SubjectMessage::DirtyBitsFlagged to Context, which dirty DrawFBO and ReadFBO and dirty cached state. Note that this is specific for swap image changed case, there is no surface property change (surface property change will still trigger the subject/observer message with SubjectMessage::SubjectChanged message, but this occurs rarely). This gets worse for apps that uses multiple contexts, for the example pokemon_masters_ex has three contexts, each context has its own default frame buffer that attach to the same surface, and we never remove non-current context from the observer list. This end up with egl::Surface has 12 observers and for every frame, it loop over the list of 12 observers and send message (virtual function call) to each of them. Color attachment also ends up sending two messages to Context, one for Read FBO and another for Draw FBO. There are total 21 virtual function calls. Even for single context usage, you have 6 virtual function calls, for every frame. EGL spec says "an EGLSurface must be current on only one thread at a time", any other context must call EGLMakeCurrent in order to use this surface, which will add all necessary dirty bits at that time. So we really only need to notify current context. In this CL, SwapchainImageChanged no longer uses subject/observer mechanism, so this message is removed. This CL still uses subject/observer mechanism to send DirtyBitsFlagged from Framebuffer back to context. We could call setDrawFramebufferDirty and setReadFramebufferDirty directly, but that will require to remove the "const" decoration out of gl::Context which generates too much code diff, so onStateChange(angle::SubjectMessage::DirtyBitsFlagged) is still used. Bug: angleproject:400711938 Change-Id: I61354516fd0aa307714b7abd30c6b6e45ff7b496 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6319893 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Geoff Lang b9567391 2025-03-20T14:10:21 Prefer Device type requests over adapter LUID in D3D11. When the user requests both a non-hardware device type and a specific DXGI adapter LUID, prefer to honor the device type request. Chrome always requests LUIDs of the first available GPU to make sure that it prefers integrated GPUs over discreet. It only requests non-hardware device types when software rendering is desired. This behaviour change satisfies Chrome's intentions in both cases. Bug: chromium:402163834 Change-Id: Idc714498260591cafc37d810c83de08a364a72f6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6377165 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Yuxin Hu ac242c1a 2025-03-20T11:18:45 Vulkan: Fix the feature flag typo Bug: b/333987915 Change-Id: Iea21592532352d95c69c0cf8dfcad96b58563ec5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6377544 Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Charlie Lao <cclao@google.com> Auto-Submit: Yuxin Hu <yuxinhu@google.com>
Kimmo Kinnunen c36d0cf0 2025-03-14T14:09:41 Metal: Improve ColorBlitUtils shader key hash Did not hash transformLinearToSrgb. Bug: angleproject:403372465 Change-Id: Ie06cf3535775f753e0468505e8b31fd513130b98 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6356891 Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Alexey Knyazev <lexa.knyazev@gmail.com> Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com>
Yuxin Hu 6bd7a518 2025-03-19T13:35:48 Vulkan: Allow pipelineStatisticsQuery substitute PGO Some devices are missing VK_EXT_primitives_generated_query extension and primitivesGeneratedQuery (PGO) feature, which are required for GLES3.2 extensions geometryShaderEXT and tessellationShaderEXT. On such devices, when app requests to create a GLES 3.2 context, the eglCreateContext call will fail due to not able to support GLES 3.2. We add a new feature flag allowPipelineStaticsForPrimitivesGeneratedQuery, and enable it on devices where we can fallback to using pipelineStatisticsQuery (PSQ) as an approximation for VK_EXT_primitives_generated_query, to broaden support for GLES 3.2 contexts. This is an opt-in feature for vendors that can functionally substitute PGQ with PSQ. Expose geometryShaderEXT and tessellationShaderEXT extensions if allowPipelineStaticsForPrimitivesGeneratedQuery feature is enabled. Bug: b/333987915 Change-Id: I9f0affbc3cc383ed6c50bcfc5d4d6f7f8a708aaa Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6362627 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: mohan maiya <m.maiya@samsung.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Mohan Maiya 05c491e1 2025-03-15T11:56:07 Vulkan: Optimize GraphicsDriverUniforms update Unless RP is closed there is no need to dirty GraphicsDriverUniforms when the program executable changes. Bug: angleproject:386749841 Test: VulkanPerformanceCounterTest.NoUpdatesToGraphicsDriverUniformsOnProgramChange* Change-Id: Id02e8a17de93e2b73103666fc6cc62ce3cdd8f43 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6358315 Commit-Queue: mohan maiya <m.maiya@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu 5b5a5e21 2025-03-17T17:59:18 Reformat VVL error message part1 Remove the string entries from skipped VVL error message. TraceTest.manhattan_31 with SwiftShader no longer runs into VVL error, remove the corresponding comments. Bug: angleproject:316337308 Bug: angleproject:42265196 Change-Id: I113d2e7953b3f14ed9811df91204d046a5fe4fb4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6364335 Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Mavis Deng d668a13c 2025-03-14T17:18:18 Fix sync issue between XFB output and texture buffer input For the following scenario, where the first draw writes to the transform feedback buffer and the second draw reads from the same buffer as a texture buffer, it is necessary to end the render pass between the two draws and add a pipeline barrier. // xfb write to tex_buffer glBindBufferBase(GL_TRANSFORM_FEEDBACK_BUFFER, 0, tex_buffer); glBeginTransformFeedback(); glDrawArrays(); glEndTransformFeedback(); // Draw with texture buffer tex_buffer glBindBuffer(GL_TEXTURE_BUFFER_EXT, tex_buffer); glTexBufferEXT(GL_TEXTURE_BUFFER_EXT, ..., tex_buffer); glDraw(); Bug: angleproject:403319685 Change-Id: I9381a336ce61dea696c93158bb617a41afcfc583 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6356070 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Mavis Deng <mavis.deng@arm.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Charlie Lao <cclao@google.com>
Chris Dalton 292bb458 2025-03-14T12:17:49 Cut MAX_COLOR_ATTACHMENTS_WITH_ACTIVE_PIXEL_LOCAL_STORAGE_ANGLE Since the number of real devices that support framebuffer fetch, but do not support draw_buffers_indexed, is effectively zero, MAX_COLOR_ATTACHMENTS_WITH_ACTIVE_PIXEL_LOCAL_STORAGE_ANGLE is an unnecessary complication. These hypothetical devices can fall back on shader images or just not support pixel local storage at all. This closes the door for WEBGL_shader_pixel_local_storage to ever be implemented on top of the ES2 extension that inspired it all, EXT_shader_pixel_local_storage, but WebGL can just use framebuffer fetch instead, or maybe even EXT_shader_pixel_local_storage2 some day. This dramatically simplifies the WEBGL_shader_pixel_local_storage extension. Bug: angleproject:40096838 Change-Id: I9c51f2ce000620f883cb7917d4e8618c8e9ee803 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6354657 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Commit-Queue: Chris Dalton <chris@rive.app>
Shahbaz Youssefi fca3fed3 2025-03-18T11:06:10 Manual roll vulkan-deps from 552ac332629b to bf9998679aee (25 revisions) https://chromium.googlesource.com/vulkan-deps.git/+log/552ac332629b..bf9998679aee Also rolling transitive DEPS: https://chromium.googlesource.com/external/github.com/KhronosGroup/glslang/+log/0051b92b6fff2fd3dbc99b7dae316fa51a7f4767..8842cf92e3de290f275c46d55cbfe42b7d0775a6 https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Tools/+log/9e470ef0f95d6923fa19dc5c5dc48c1442eb5a8b..ba1359d203e544bd458373ba249a47c44e97b071 https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-Loader/+log/8491f31d36a8c642e588220b89729129721fec31..54cbefd25dbcaeb2bb03da207afce6cad7fb5dd1 https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-ValidationLayers/+log/74a6b7dc828078cf2f1524d3947f65a29a1562e9..b2c8bd421aa5c1502fd56c3c07fb2831a8ceb882 If this roll has caused a breakage, revert this CL and stop the roller using the controls here: https://autoroll.skia.org/r/vulkan-deps-angle-autoroll Please CC angle-team@google.com,syoussefi@google.com on the revert to ensure that a human is aware of the problem. To file a bug in ANGLE: https://bugs.chromium.org/p/angleproject/issues/entry To report a problem with the AutoRoller itself, please file a bug: https://issues.skia.org/issues/new?component=1389291&template=1850622 Documentation for the AutoRoller is here: https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md Bug: angleproject:404418769 Change-Id: I359e903e2d061b7bbde6b18d234d8927d2cca56c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6368274 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Roman Lavrov <romanl@google.com>
Shahbaz Youssefi dae3c851 2025-03-14T11:44:53 Vulkan: Bake non-shader state into linked pipeline When using VK_EXT_graphics_pipeline_library, previously ANGLE would create three pipelines libraries: * The Shaders library was created based on the GL program's shaders + a few static states. This typically hit the program's own pipeline's cache that was warmed up during link. * The VertexInput and FragmentOutput libraries were created at draw time, which used the global pipeline cache At draw time, immediately after creating the non-Shaders libraries, the three libraries were linked into the final pipeline to be used by the draw call. This caused an inefficiency; because the non-Shaders libraries were created independently from the Shaders library, they had to be compiled pessimistically, for example because they could not be optimized to take into account the precision of the fragment shader's outputs or whether any value is const (typically alpha being set to one). Given the creation of VertexInput and FragmentOutput libraries is typically quite fast (the former being no-op and dynamic state anyway), this change removes the need for creating those libraries, and directly specifies the vertex input and fragment output state when creating the final pipeline out of the Shaders library. In this way, the same fragment output state can be tailored to the exact shaders it is being used with and incur a smaller overhead. In this change, the linked pipeline is cached in the GL program's pipeline cache, which is never synced to the blob cache as producing it is assumed to be fast already. Bug: angleproject:42265839 Change-Id: I8496ea37771555522bdc9de94043a1b56fa5967e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6354205 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: mohan maiya <m.maiya@samsung.com> Reviewed-by: Charlie Lao <cclao@google.com>
Le Hoang Quyen f64a9547 2025-03-15T23:45:33 Metal: remove "Compute" from EnsureComputeShaderInitialized() name Previously this function was named EnsureComputePipelineInitialized() because it would create a compute pipeline. However at some point, the pipeline's creation was moved to a pipeline cache. Thus this function now only creates a shader function. This can be used to create both compute & graphics shader functions so we should remove the "Compute" part from its name. Bug: None Change-Id: I73d101142ec1fc72d150579bc409b1dea3196940 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6357511 Reviewed-by: Alexey Knyazev <lexa.knyazev@gmail.com> Commit-Queue: Quyen Le <lehoangquyen@chromium.org>
Igor Nazarov c935e06d 2025-02-24T20:13:46 Revert "Vulkan: Fix Wayland surface size queries" This reverts commit a277f1bce563999296620c29d83c760c0fb9763d. Reason for revert: No longer required, because `mIsSurfaceSizedBySwapchain` is always true on Wayland which will provide same behavior as the reverted override methods. It is also a preparation for the follow up change. Bug: angleproject:397848903 Bug: angleproject:40096815 Change-Id: If83ef133b422494d73565068e06e2d6a737f6315 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6298737 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Igor Nazarov daaae2bd 2025-02-24T19:36:39 Vulkan: Remove surface caps query from createSwapChain This is mainly a refactoring with minor bug fixes. The `createSwapChain()` is called at initialize and at swapchain recreate. Before recreate there is always `queryAndAdjustSurfaceCaps()` call, which queries current surface sizes and min image count. However, surface capabilities may be queried again from `createSwapChain()` to get compatible present modes and the min image count. Problem is that surface caps `currentExtent` may change after the second query in the `createSwapChain()`, while the `extents` parameter is already calculated. Such situation will violate the specification. To eliminate the possibility of the above problem and to simplify the code, query of compatible present modes is moved to the `queryAndAdjustSurfaceCaps()`. To save performance, `mCompatiblePresentModes` is only updated if swapchain is already invalid (recreate is imminent). Function is now also handles `mEmulatedPreTransform` to further simplify the code. The `mSurfaceCaps` member was replaced with local variable. The `mIsSurfaceSizedBySwapchain` boolean is added instead to check it in the `getUserWidth()`/`getUserHeight()` methods. This also fixes a possible bug (anglebug.com/168327817) on Fuchsia when this expression could be false: mSurfaceCaps.currentExtent.width == kSurfaceSizedBySwapchain This is because previously `queryAndAdjustSurfaceCaps()` updated `mSurfaceCaps` with the values from `getCurrentWindowSize()`. It seems that the bug never happened on practice, otherwise the ASSERT in the same function would fail. The `prepareForAcquireNextSwapchainImage()` is used at initialize instead of `createSwapChain()` to remove code duplication. Fixed (anglebug.com/168327817) checking `mPreTransform` without enabled "enablePreRotateSurfaces" feature. Bug: angleproject:397848903 Bug: angleproject:168327817 Change-Id: I02e8e3b815b30350e12476853d8dcffed3fe7c38 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6298736 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Igor Nazarov ac251fa8 2025-02-24T17:31:58 Vulkan: Remove doDeferredAcquireNextImage ContextVk dependency This is the continuation of the previous CL. After this change `lockSurface()` calls `doDeferredAcquireNextImage()` as the rest of the code do. The `ImageHelper::invalidateSubresourceContentImpl()` required `ContextVk` pointer before this change. The `ContextVk` is only used to print the performance warning when `layerIndex` exceed the maximum count, which is not possible in case of a `WindowSurfaceVk`. Added `layerLimitReachedOut` pointer instead of writing the warning. It is processed in `invalidateSubresource[Stencil]Content()`. Added `invalidateEntireLevel[Stencil]Content()` which do not require `ContextVk` to use in the `WindowSurfaceVk`. Bug: angleproject:397848903 Bug: angleproject:42264593 Change-Id: I58451a4818ad56fa196c3c971df3a5f7793f2bfe Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6298735 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Igor Nazarov 2f5a91ba 2025-02-24T17:25:10 Vulkan: Remove checkForOutOfDateSwapchain ContextVk dependency This is the continuation of the previous CL. After this CL it is now possible to call `prepareForAcquireNextSwapchainImage()` from `lockSurface()` that will also do necessary swapchain recreation in case of window resize. These methods required `ContextVk` pointer before this change: - RenderTargetVk::releaseImageAndViews() - replaced with `releaseSwapchainImage()` that does not release framebuffer because Window Surface framebuffers are not managed by the cache. - added `release()` method that does release the framebuffer. - WindowSurfaceVk::releaseSwapchainImages() - use `Renderer::collectGarbage()` instead of `ContextVk::addGarbage()`. - use `ImageHelper::releaseImage()` instead of `ImageHelper::releaseImageFromShareContexts()`. The `finalizeImageLayoutInShareContexts()` was not required since renderpass must be already ended, because swapchain recreate is only possible after present. Removal of `addToPendingImageGarbage()` is not going to cause OOM problems, because repeated swapchain recreate calls are not possible without swap (submissions). Bug: angleproject:397848903 Bug: angleproject:42264593 Change-Id: Iacfa3a144aa980659569b7100be25a44ebb9f0a0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6298734 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Igor Nazarov <i.nazarov@samsung.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
kmx.io     mail     discord kmxgit v0.4.0