{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":113846915,"defaultBranch":"dev","name":"pal","ownerLogin":"GPUOpen-Drivers","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2017-12-11T10:45:47.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/29980581?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1723795880.0","currentOid":""},"activityList":{"items":[{"before":"d3bfe509cd335b06d1276a5bb3ceb8dce622ca34","after":"eca6b999c2bec3b04f0c52a8a64f3c1d03db8059","ref":"refs/heads/master","pushedAt":"2024-08-29T07:23:57.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit d543c4af\n\nFix some CodeQL alerts\nDo not try query value for empty components\nAdd Gfxip Major.Minor.Stepping to DbgOverlay\nAdd instruction timing flag to SqttData chunk\nUpdate submodule address-lib\nRefresh chip headers\nRestore the hard-coded AsicRevision values\nRemove the IsVirtual CP DMA checks\nBegin QueueTimings in the PAL Trace preparation phase\nFix SetcutoffSeverityLevel/SetOriginationTypeMask\nFix the random core dump of dEQP-VK.wsi.wayland.swapchain.render.10swapchains2 when enabling zwp_linux_dmabuf_v1_interface\nRemove UsesUavs check for DepthOnlyOpt\nGpuProfiler forces RequestDebugVMID\nDeprecate some unused fields in InheritedStateParams\nTweak handling of PAL Private Pm4Opt key\nSupport dynamic lds from DynamicSharedPointer\nFix null-pointer access issue\nCheck for nullptr in case MALLOC fails\nUpdate FileArchiveCacheLayer to fix some race conditions, corruption issues, and inefficiencies\nAdvertise untypedBufferView separately in srdSizes\nOptimize CmdFillMemory\nUse Max(...) function instead of a recursive call\nOmit the DevDriver overlay if a RGP trace is actively running or if a PALTrace trace is actively running\nDisable OREO mode\ntypedBufferView is bytes, not DWs\nMove ShouldUploadUsingDma to Device and Check for Valid Dma Upload Ring\nUpdate submodule devdriver\nEnable displayDcc by default except strix1\nUpdate numWorkGroupsReg on a Per-Dispatch Basis\nFix potential memleak in CrashAnalyzer\nAdd strix1 support","shortMessageHtmlLink":"Update pal from commit d543c4af"}},{"before":"d3bfe509cd335b06d1276a5bb3ceb8dce622ca34","after":"eca6b999c2bec3b04f0c52a8a64f3c1d03db8059","ref":"refs/heads/dev","pushedAt":"2024-08-28T10:37:12.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit d543c4af\n\nFix some CodeQL alerts\nDo not try query value for empty components\nAdd Gfxip Major.Minor.Stepping to DbgOverlay\nAdd instruction timing flag to SqttData chunk\nUpdate submodule address-lib\nRefresh chip headers\nRestore the hard-coded AsicRevision values\nRemove the IsVirtual CP DMA checks\nBegin QueueTimings in the PAL Trace preparation phase\nFix SetcutoffSeverityLevel/SetOriginationTypeMask\nFix the random core dump of dEQP-VK.wsi.wayland.swapchain.render.10swapchains2 when enabling zwp_linux_dmabuf_v1_interface\nRemove UsesUavs check for DepthOnlyOpt\nGpuProfiler forces RequestDebugVMID\nDeprecate some unused fields in InheritedStateParams\nTweak handling of PAL Private Pm4Opt key\nSupport dynamic lds from DynamicSharedPointer\nFix null-pointer access issue\nCheck for nullptr in case MALLOC fails\nUpdate FileArchiveCacheLayer to fix some race conditions, corruption issues, and inefficiencies\nAdvertise untypedBufferView separately in srdSizes\nOptimize CmdFillMemory\nUse Max(...) function instead of a recursive call\nOmit the DevDriver overlay if a RGP trace is actively running or if a PALTrace trace is actively running\nDisable OREO mode\ntypedBufferView is bytes, not DWs\nMove ShouldUploadUsingDma to Device and Check for Valid Dma Upload Ring\nUpdate submodule devdriver\nEnable displayDcc by default except strix1\nUpdate numWorkGroupsReg on a Per-Dispatch Basis\nFix potential memleak in CrashAnalyzer\nAdd strix1 support","shortMessageHtmlLink":"Update pal from commit d543c4af"}},{"before":null,"after":"6b7e7222da60937d373cca8777ba55dbf5e7bd86","ref":"refs/heads/dev_promote","pushedAt":"2024-08-16T08:11:20.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Improve the refresh rate precision. (AMDVLK #268)","shortMessageHtmlLink":"Improve the refresh rate precision. (AMDVLK #268)"}},{"before":"695337aa8131913d883ff15787cd75f6fcb11229","after":"d3bfe509cd335b06d1276a5bb3ceb8dce622ca34","ref":"refs/heads/master","pushedAt":"2024-08-13T05:31:22.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit a33b9bbb\n\nFix SDMA ODR warnings\nMRT Use Minimum Resolution\nAdd 2 new TraceSources needed for pull UserMarker into RRA\nMemory Leak in Execute Indirect Shader\nUpdate submodule devdriver\nExpose the flag 'enablePops' to PAL pipeline user\nFix the regression issue for the dynamicDualSourceBlend in fastlinklibrary mode\nMinor fix for split barrier\nBresenham lines with adjacency failing\nCleanup gfx query pool code.\nDon't read over PCIe when using SDMA predication\nImplement workaround WaCwsrThreadgroupTrap\nUpdate spill table register in rpm GenerateCmdDisptachTaskMesh on gfx10/11\nHandle Directory Error in Command Buffer Dumping\nUpdate submodule address-lib\nMoves supportSplitReleaseAcquire from gfxipProperrties to queue queueProperties\nRemove redundant Inits within ElfProcessor and PipelineProcessor\nFix PAL_CLIENT_INTERFACE_VERSION\nFix SaschaWillems perf regression due to OREO_MODE Workaround\nDefine Release PAL_ANALYSIS_ASSUME to void(0)\nUpdate PM4 headers\nAlways log out layout transition info in interface logger (for barrier log)\nAdd alignment asserts for several gfx9 packets\nResolve address alignment\nSet reg for PA_SU_POLY_OFFSET_DB_FMT_CNTL with no depth buffer bound.\nFix Vector Erase and define a new FastErase() function\nChange definition of release token used for CmdRelease and CmdAcquire\nOptimize split barrier: defer CpDma wait from Release to Acquire time\nImplement Compute InheritStateFromCmdBuf\nAdd PERF_COUNTER_WINDOW Support\nConvert enum opcodes for comparison\nFix spill threshold so that driver uses minimum spill threshold\nAdd SPM buffer wrap warning\nOptimize split barrier to not acquire retired release token again\nChange SubresRange member data types to save size\nFix VKD3D Execute Indirect Test Failures\nUse double when converting float to fixed point number\nRemove outdated ADDRLIB_VERSION checks\nRenderOp trace controller\nAllow mipmaps for T2T SDMA copies\nUpdate ASIC Info chunk def\nUse pairwise tree reduction in msaaFmaskResolve8x\nRemove hard-coded NullGpuId and AsicRevision numbers\nAdd RGP instrumentation marker to amp/task ACE dispatch\nCorrect the matrixRowOrder in plane 2 of the P208 format\nRelax blit operations\nPlace an upper limit on internalMemMgr pool size growth\nRename PipelineUploader to CodeObjectUploader and move it to its own files\nCorrect the sampleLocX of the packed format\nNavi2 VRS data cannot be read from miptail\nChanges to ExecuteIndirectV2 path on Gfx9\nController config update and active controller\nDisable use of PERF_COUNTER_WINDOW\nAdd treatSrcAsSrgb flag into ScaledCopyInfo\nAdding L2 cache values for individual ASICs\nRenderOp Controller updates\nCustomize Offset in Cmd Generator Creation\nFix interface log for ExternalImageOpenInfo to match tip definition\nUpdate COVERAGE_TO_SHADER_SELECT bit when underestimation is enabled\nRemove legacy settings\nAdd new enum for FsrOvr-ApArea\nFix nullptr deref in SPM Trace processing","shortMessageHtmlLink":"Update pal from commit a33b9bbb"}},{"before":"695337aa8131913d883ff15787cd75f6fcb11229","after":"d3bfe509cd335b06d1276a5bb3ceb8dce622ca34","ref":"refs/heads/dev","pushedAt":"2024-08-08T09:09:36.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit a33b9bbb\n\nFix SDMA ODR warnings\nMRT Use Minimum Resolution\nAdd 2 new TraceSources needed for pull UserMarker into RRA\nMemory Leak in Execute Indirect Shader\nUpdate submodule devdriver\nExpose the flag 'enablePops' to PAL pipeline user\nFix the regression issue for the dynamicDualSourceBlend in fastlinklibrary mode\nMinor fix for split barrier\nBresenham lines with adjacency failing\nCleanup gfx query pool code.\nDon't read over PCIe when using SDMA predication\nImplement workaround WaCwsrThreadgroupTrap\nUpdate spill table register in rpm GenerateCmdDisptachTaskMesh on gfx10/11\nHandle Directory Error in Command Buffer Dumping\nUpdate submodule address-lib\nMoves supportSplitReleaseAcquire from gfxipProperrties to queue queueProperties\nRemove redundant Inits within ElfProcessor and PipelineProcessor\nFix PAL_CLIENT_INTERFACE_VERSION\nFix SaschaWillems perf regression due to OREO_MODE Workaround\nDefine Release PAL_ANALYSIS_ASSUME to void(0)\nUpdate PM4 headers\nAlways log out layout transition info in interface logger (for barrier log)\nAdd alignment asserts for several gfx9 packets\nResolve address alignment\nSet reg for PA_SU_POLY_OFFSET_DB_FMT_CNTL with no depth buffer bound.\nFix Vector Erase and define a new FastErase() function\nChange definition of release token used for CmdRelease and CmdAcquire\nOptimize split barrier: defer CpDma wait from Release to Acquire time\nImplement Compute InheritStateFromCmdBuf\nAdd PERF_COUNTER_WINDOW Support\nConvert enum opcodes for comparison\nFix spill threshold so that driver uses minimum spill threshold\nAdd SPM buffer wrap warning\nOptimize split barrier to not acquire retired release token again\nChange SubresRange member data types to save size\nFix VKD3D Execute Indirect Test Failures\nUse double when converting float to fixed point number\nRemove outdated ADDRLIB_VERSION checks\nRenderOp trace controller\nAllow mipmaps for T2T SDMA copies\nUpdate ASIC Info chunk def\nUse pairwise tree reduction in msaaFmaskResolve8x\nRemove hard-coded NullGpuId and AsicRevision numbers\nAdd RGP instrumentation marker to amp/task ACE dispatch\nCorrect the matrixRowOrder in plane 2 of the P208 format\nRelax blit operations\nPlace an upper limit on internalMemMgr pool size growth\nRename PipelineUploader to CodeObjectUploader and move it to its own files\nCorrect the sampleLocX of the packed format\nNavi2 VRS data cannot be read from miptail\nChanges to ExecuteIndirectV2 path on Gfx9\nController config update and active controller\nDisable use of PERF_COUNTER_WINDOW\nAdd treatSrcAsSrgb flag into ScaledCopyInfo\nAdding L2 cache values for individual ASICs\nRenderOp Controller updates\nCustomize Offset in Cmd Generator Creation\nFix interface log for ExternalImageOpenInfo to match tip definition\nUpdate COVERAGE_TO_SHADER_SELECT bit when underestimation is enabled\nRemove legacy settings\nAdd new enum for FsrOvr-ApArea\nFix nullptr deref in SPM Trace processing","shortMessageHtmlLink":"Update pal from commit a33b9bbb"}},{"before":"2076667db1fa57f03d1624d1c59c3d024afb42f6","after":"695337aa8131913d883ff15787cd75f6fcb11229","ref":"refs/heads/master","pushedAt":"2024-06-26T08:24:38.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 025d0bdb\n\nAdd the missed PFP version to check\nChange WsiPlatform to scoped enum\nEnable the support for the cooperativeMatrix on gfx10\nExpand queue filtering for GpuProfiler\nFilter redundant elements for CS and \"middle\" user data entries on gfx11+\nImprove interface logger to generate barrier log compatible JSON files\nIncrease the MaxUserDataEntries from 128 to 160\nInterfaceLogger callback support\nMake null pipeline signatures device class members\nRemove an unused private panel setting-CsMaxWavesPerCu\nUpdate address-lib\nUpdate devdriver\nUpdate RPM shaders\nFix map table entries in the interface logging layer\nFix the topology for barycentric","shortMessageHtmlLink":"Update pal from commit 025d0bdb"}},{"before":"2076667db1fa57f03d1624d1c59c3d024afb42f6","after":"695337aa8131913d883ff15787cd75f6fcb11229","ref":"refs/heads/dev","pushedAt":"2024-06-26T05:46:22.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 025d0bdb\n\nAdd the missed PFP version to check\nChange WsiPlatform to scoped enum\nEnable the support for the cooperativeMatrix on gfx10\nExpand queue filtering for GpuProfiler\nFilter redundant elements for CS and \"middle\" user data entries on gfx11+\nImprove interface logger to generate barrier log compatible JSON files\nIncrease the MaxUserDataEntries from 128 to 160\nInterfaceLogger callback support\nMake null pipeline signatures device class members\nRemove an unused private panel setting-CsMaxWavesPerCu\nUpdate address-lib\nUpdate devdriver\nUpdate RPM shaders\nFix map table entries in the interface logging layer\nFix the topology for barycentric","shortMessageHtmlLink":"Update pal from commit 025d0bdb"}},{"before":"c789abc769d264f73e3dad3f21238b876918f594","after":"2076667db1fa57f03d1624d1c59c3d024afb42f6","ref":"refs/heads/master","pushedAt":"2024-06-05T03:29:59.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit e72181b7\n\n[CMAKE] Support precompiled headers to improve build times\n[PAL CodingStandards] Clarify standard on default values\nAdd 3 new rpc calls to DriverUtilsService\nAdd 32Bit SPM Support\nAdd gl2Uncached flag for OpenExternalSharedGpuMemory changing MType\nAdd metadata and PipelineInfo flags for CPS\nAdd more enum operators\nAdd override for client compression on resource clears\nAdd pixel scale for DbgOverlay\nAdd static_assert no min/max\nAdd support for EndOfRecreation and VirtualPresent which are passed to SetStencilRef but may not make it to PAL as a stencilRef\nAdd support for hsa abi AMD_KERNEL_CODE_PROPERTIES_ENABLE_SGPR_DISPATCH_ID\nAdd unused_imm_const to PAL metadata.\nAdd variable initialization to fix a bad memory access in developer mode debug driver build\nAdd workaround that disables the use of OREO's Opaque mode\nAllow for compute library metadata with no .shader_functions section\nClean-up VCN code\nCleanup old interface versions\nComplete the Device::InitOutPutPaths\nConsolidate some sdma functions into base class\nDon't assert on HiddenNone\nEliminate unnecessary null DMA packets in command stream by setting appropriate bit in release mem command\nFix a POPS issue on GFX11\nFix barrier operation report\nFix layout of prints in CmdBufferLogger.\nFix memory leak when RGP is connected\nFix overlay time graph GetScaledGpuTime\nFix palTime missing in some files\nFix SQWGP SPM Muxsel\nFix typo in logs\nForce SE Broadcast for SQ Perfcounter Select\ngpaSession: Allow multi-queue timed submits\nInvestigate ability to signal/wait on semaphores without GPU submission\nMove chrono helpers to palTime.h\nRefactor perfcounter block instance count names\nRefactor PerfCounter defines\nRefine DCC_CONTROL reg setting\nRemove GFX9 from the gfx9 merged headers\nRemove meta eq debug functionality\nRemove unused member box from ImgBarrier and memory from MemBarrier\nSkip layout transition when both old and new usage are LayoutUninitializedTarget\nSpeed up OcclusionQueryPool::End\nSupport create cs pipeline from hsa ELF binary with multi kernels\nSupport new SRD structure D3D12_VERTEX_BUFFER_VIEW\nUpdate address-lib\nUpdate DevDriver to v24.05.01\nUpdate GFX9 HWL SDMA PerfCtr Select\nUse LargeEmbeddedData for new chunks to generate command data in UseExecuteIndirectShaders path","shortMessageHtmlLink":"Update pal from commit e72181b7"}},{"before":"c789abc769d264f73e3dad3f21238b876918f594","after":"2076667db1fa57f03d1624d1c59c3d024afb42f6","ref":"refs/heads/dev","pushedAt":"2024-06-04T03:29:53.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit e72181b7\n\n[CMAKE] Support precompiled headers to improve build times\n[PAL CodingStandards] Clarify standard on default values\nAdd 3 new rpc calls to DriverUtilsService\nAdd 32Bit SPM Support\nAdd gl2Uncached flag for OpenExternalSharedGpuMemory changing MType\nAdd metadata and PipelineInfo flags for CPS\nAdd more enum operators\nAdd override for client compression on resource clears\nAdd pixel scale for DbgOverlay\nAdd static_assert no min/max\nAdd support for EndOfRecreation and VirtualPresent which are passed to SetStencilRef but may not make it to PAL as a stencilRef\nAdd support for hsa abi AMD_KERNEL_CODE_PROPERTIES_ENABLE_SGPR_DISPATCH_ID\nAdd unused_imm_const to PAL metadata.\nAdd variable initialization to fix a bad memory access in developer mode debug driver build\nAdd workaround that disables the use of OREO's Opaque mode\nAllow for compute library metadata with no .shader_functions section\nClean-up VCN code\nCleanup old interface versions\nComplete the Device::InitOutPutPaths\nConsolidate some sdma functions into base class\nDon't assert on HiddenNone\nEliminate unnecessary null DMA packets in command stream by setting appropriate bit in release mem command\nFix a POPS issue on GFX11\nFix barrier operation report\nFix layout of prints in CmdBufferLogger.\nFix memory leak when RGP is connected\nFix overlay time graph GetScaledGpuTime\nFix palTime missing in some files\nFix SQWGP SPM Muxsel\nFix typo in logs\nForce SE Broadcast for SQ Perfcounter Select\ngpaSession: Allow multi-queue timed submits\nInvestigate ability to signal/wait on semaphores without GPU submission\nMove chrono helpers to palTime.h\nRefactor perfcounter block instance count names\nRefactor PerfCounter defines\nRefine DCC_CONTROL reg setting\nRemove GFX9 from the gfx9 merged headers\nRemove meta eq debug functionality\nRemove unused member box from ImgBarrier and memory from MemBarrier\nSkip layout transition when both old and new usage are LayoutUninitializedTarget\nSpeed up OcclusionQueryPool::End\nSupport create cs pipeline from hsa ELF binary with multi kernels\nSupport new SRD structure D3D12_VERTEX_BUFFER_VIEW\nUpdate address-lib\nUpdate DevDriver to v24.05.01\nUpdate GFX9 HWL SDMA PerfCtr Select\nUse LargeEmbeddedData for new chunks to generate command data in UseExecuteIndirectShaders path","shortMessageHtmlLink":"Update pal from commit e72181b7"}},{"before":"a4cbf005fc8e6cfb72fb5d39dd09f45272c0f184","after":null,"ref":"refs/heads/dev_promote","pushedAt":"2024-05-15T07:39:51.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"}},{"before":"42e29f4d3d528c6a59c2db77fb6d70127fb9ffab","after":"c789abc769d264f73e3dad3f21238b876918f594","ref":"refs/heads/master","pushedAt":"2024-05-15T07:39:00.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit acd9848c\n\nAdd an explicit log file for GpuProfiler\nAdd ETW event collection to QueueTimings\nAdd IsGfx9Hwl to cleanup device logic\nAdd missing overrides and fix more build warnings\nAdd PalToHResult() and HResultToPal()\nAdd PC Perfcounter support for Gfx11\nAdd PHOENIX2 support\nAdd support for special/UTF8 characters in cache/debug file paths\nAdd TdrController for PAL-Trace\nAdd zwp_linux_dmabuf_v1 support to Wayland WS\nAllow std::bit_cast\nAlways issue required async cache flush/inv event for gfx layout blt\nChange experiments setting to enable only\nCorrect programing of depth for poly offset\nDeterministic CmdBuffer Log Names\nExpose file headers for archives\nFactor out SlowColorClear1-7\nMake CP DMA blt buffer only and optimize BLT flags for mage only barrier\nMake GpuProfiler aware of PerformanceData when fastlink\nMake GpuProfiler aware of PerformanceData when fastlink\nNull check for active controller\nOffscreen GPU Profiling frame support\nOptimize CoherClear to CoherColorTarget or CoherDepthStencilTarget transition\nOptimize misaligned metadata WA to reduce GL2 sync for BLT access flags\nPrevent DbgLogMgr destructor use of loggers\nRemove assert in ~DbgLogMgr\nRemove dead legacy sdma perfctr handling\nRemove IsGfx9 and IsGfx10Plus\nRemove more gfx9 stuff\nRemove the default disabled legacy barrier path for pre/post clear sync\nRemoveOldestFilesOfDirUntilSize() shouldn't return an error on an empty dir\nReorder Perfctr Init after queue info is known\nRevise the UcodeVersion check to enable task shaders\nSet up PBB dirty tracking so that it PBB is written\nSqtt Data: Add Trace Buffer Size and split chunk into Header and Data\nSubmodule address-lib update\nSubmodule devdriver update to v24.04.03\nSubmodule SWD update\nUpdate and create DevDriver Logger\nUpdate CacheLayerBase\nUpdate GpuPerfExp Trace Source\nUse pSymbolName under PAL major 827\nUse std::chrono\nFix 32Bit UMDs errors when using the callback mechanism\nFix a typo for \"streamout_vertex_strides\"\nFix dbgLogger file path gen\nFix one typo about PAL version check\nFix pMsaaState warning\nFix trace markers being missing from UberTrace SQTT","shortMessageHtmlLink":"Update pal from commit acd9848c"}},{"before":"2682a9122ca4b6b0d5875e41ea347b6377922e47","after":"c789abc769d264f73e3dad3f21238b876918f594","ref":"refs/heads/dev","pushedAt":"2024-05-14T09:29:42.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit acd9848c\n\nAdd an explicit log file for GpuProfiler\nAdd ETW event collection to QueueTimings\nAdd IsGfx9Hwl to cleanup device logic\nAdd missing overrides and fix more build warnings\nAdd PalToHResult() and HResultToPal()\nAdd PC Perfcounter support for Gfx11\nAdd PHOENIX2 support\nAdd support for special/UTF8 characters in cache/debug file paths\nAdd TdrController for PAL-Trace\nAdd zwp_linux_dmabuf_v1 support to Wayland WS\nAllow std::bit_cast\nAlways issue required async cache flush/inv event for gfx layout blt\nChange experiments setting to enable only\nCorrect programing of depth for poly offset\nDeterministic CmdBuffer Log Names\nExpose file headers for archives\nFactor out SlowColorClear1-7\nMake CP DMA blt buffer only and optimize BLT flags for mage only barrier\nMake GpuProfiler aware of PerformanceData when fastlink\nMake GpuProfiler aware of PerformanceData when fastlink\nNull check for active controller\nOffscreen GPU Profiling frame support\nOptimize CoherClear to CoherColorTarget or CoherDepthStencilTarget transition\nOptimize misaligned metadata WA to reduce GL2 sync for BLT access flags\nPrevent DbgLogMgr destructor use of loggers\nRemove assert in ~DbgLogMgr\nRemove dead legacy sdma perfctr handling\nRemove IsGfx9 and IsGfx10Plus\nRemove more gfx9 stuff\nRemove the default disabled legacy barrier path for pre/post clear sync\nRemoveOldestFilesOfDirUntilSize() shouldn't return an error on an empty dir\nReorder Perfctr Init after queue info is known\nRevise the UcodeVersion check to enable task shaders\nSet up PBB dirty tracking so that it PBB is written\nSqtt Data: Add Trace Buffer Size and split chunk into Header and Data\nSubmodule address-lib update\nSubmodule devdriver update to v24.04.03\nSubmodule SWD update\nUpdate and create DevDriver Logger\nUpdate CacheLayerBase\nUpdate GpuPerfExp Trace Source\nUse pSymbolName under PAL major 827\nUse std::chrono\nFix 32Bit UMDs errors when using the callback mechanism\nFix a typo for \"streamout_vertex_strides\"\nFix dbgLogger file path gen\nFix one typo about PAL version check\nFix pMsaaState warning\nFix trace markers being missing from UberTrace SQTT","shortMessageHtmlLink":"Update pal from commit acd9848c"}},{"before":null,"after":"a4cbf005fc8e6cfb72fb5d39dd09f45272c0f184","ref":"refs/heads/dev_promote","pushedAt":"2024-04-10T09:41:13.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Complete the ndDevice::InitOutPutPaths","shortMessageHtmlLink":"Complete the ndDevice::InitOutPutPaths"}},{"before":"2682a9122ca4b6b0d5875e41ea347b6377922e47","after":null,"ref":"refs/heads/dev_promote","pushedAt":"2024-04-10T08:03:38.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"}},{"before":"42e29f4d3d528c6a59c2db77fb6d70127fb9ffab","after":"2682a9122ca4b6b0d5875e41ea347b6377922e47","ref":"refs/heads/dev","pushedAt":"2024-04-10T08:03:21.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"[AMDVLK-357] Fix the warning passing NULL to non-pointer","shortMessageHtmlLink":"[AMDVLK-357] Fix the warning passing NULL to non-pointer"}},{"before":null,"after":"2682a9122ca4b6b0d5875e41ea347b6377922e47","ref":"refs/heads/dev_promote","pushedAt":"2024-04-10T07:05:16.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"[AMDVLK-357] Fix the warning passing NULL to non-pointer","shortMessageHtmlLink":"[AMDVLK-357] Fix the warning passing NULL to non-pointer"}},{"before":"abb22ae22308954cd9ce76b526c2f805e2dec1ba","after":"42e29f4d3d528c6a59c2db77fb6d70127fb9ffab","ref":"refs/heads/master","pushedAt":"2024-03-26T04:22:14.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 664d85b7\n\nAdd \"Preparing\" stage to TraceSession\nAdd `IsTimingInProgress` function to QueueTimings Trace Source\nAdd CsSeSaEnableMask to limit the cs shader to run on specified SE/SA, together with CsCuEnLimitMask we can get a specified CU's LDS content\nAdd dynamic topology and merge user data\nAdd Experiments Loader\nAdd GpuProfiler Command Buffer range\nAdd GpuProfiler setting to disable session reuse\nAdd IB2 MEC to pm4 dumps for gfx9+\nAdd Init function to QueueTimingsTraceSource\nAdd 'pciId' field to various PAL Trace chunks\nAdd predication for thread trace marker event writes\nAdd QueueTimings TraceSource\nAdd settings file manager to null device\nChange CmdWaitMemoryValue interface to take GPU VA\nChange GetShaderFunctionCode to accept a StringView.\nCompact ColorBlendStateCreateInfo structure\nCompact MsaaStateCreateInfo\nCompact TriangleRasterStateParams\nConsider metadata VGPR/SGPR/LDS usage when linking\nDeprecate CreateInfo Param to StateObject Size Interfaces\nExpose 'IPlatform::GetPciId'\ngfx9/CmdBarrier: Do not allow read-only DSV for early phase depth/stencil transitions.\nHave LocalHeap as backup for the ExecuteIndirectV2 buffer when InvisibleHeap does not exist\nHook up ZSurfWorkaround\nImprove ExecuteIndirectV2 implementation.\nOptimize copy in UpdateUserDataTableCpu\nRefactor acquire release cache operation determination logic\nRemove AMDGPU_IS_NAVI() in ndDevice.cpp\nRemove assert in GfxDevice::VertsPerPrimitive()\nRemove dead gfx9 code\nRemove the code that forces binningMaxAllocCountNggOnChip on Navi31 to 255 to fix the fps drop in Sashawillems - computeparticles\nSome code cleanup\nSupport stdout ucode logging\nUpdate `ApiInfo` chunk in PAL Trace\nUpdate CodeObject & QueueTimings trace sources\nUpdate experiment defaults\nUpdate static assert in RegPairHandler for UConfig\nUpdate devdriver to v24.02.21\nUpdate submodule address-lib\nUpdate submodule SWD\narFile fixes\nFix gpuProfiler for NPRT archive pipeline\nFix missing overrides\nFix potential barrier bugs\nFix streamout query comment\nFix the issue that crashAnalysis::CmdDrawIndexedIndirectMulti forwards to the wrong function in the next layer\nFix unused spillTable with userDataMask","shortMessageHtmlLink":"Update pal from commit 664d85b7"}},{"before":"abb22ae22308954cd9ce76b526c2f805e2dec1ba","after":"42e29f4d3d528c6a59c2db77fb6d70127fb9ffab","ref":"refs/heads/dev","pushedAt":"2024-03-21T10:19:30.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 664d85b7\n\nAdd \"Preparing\" stage to TraceSession\nAdd `IsTimingInProgress` function to QueueTimings Trace Source\nAdd CsSeSaEnableMask to limit the cs shader to run on specified SE/SA, together with CsCuEnLimitMask we can get a specified CU's LDS content\nAdd dynamic topology and merge user data\nAdd Experiments Loader\nAdd GpuProfiler Command Buffer range\nAdd GpuProfiler setting to disable session reuse\nAdd IB2 MEC to pm4 dumps for gfx9+\nAdd Init function to QueueTimingsTraceSource\nAdd 'pciId' field to various PAL Trace chunks\nAdd predication for thread trace marker event writes\nAdd QueueTimings TraceSource\nAdd settings file manager to null device\nChange CmdWaitMemoryValue interface to take GPU VA\nChange GetShaderFunctionCode to accept a StringView.\nCompact ColorBlendStateCreateInfo structure\nCompact MsaaStateCreateInfo\nCompact TriangleRasterStateParams\nConsider metadata VGPR/SGPR/LDS usage when linking\nDeprecate CreateInfo Param to StateObject Size Interfaces\nExpose 'IPlatform::GetPciId'\ngfx9/CmdBarrier: Do not allow read-only DSV for early phase depth/stencil transitions.\nHave LocalHeap as backup for the ExecuteIndirectV2 buffer when InvisibleHeap does not exist\nHook up ZSurfWorkaround\nImprove ExecuteIndirectV2 implementation.\nOptimize copy in UpdateUserDataTableCpu\nRefactor acquire release cache operation determination logic\nRemove AMDGPU_IS_NAVI() in ndDevice.cpp\nRemove assert in GfxDevice::VertsPerPrimitive()\nRemove dead gfx9 code\nRemove the code that forces binningMaxAllocCountNggOnChip on Navi31 to 255 to fix the fps drop in Sashawillems - computeparticles\nSome code cleanup\nSupport stdout ucode logging\nUpdate `ApiInfo` chunk in PAL Trace\nUpdate CodeObject & QueueTimings trace sources\nUpdate experiment defaults\nUpdate static assert in RegPairHandler for UConfig\nUpdate devdriver to v24.02.21\nUpdate submodule address-lib\nUpdate submodule SWD\narFile fixes\nFix gpuProfiler for NPRT archive pipeline\nFix missing overrides\nFix potential barrier bugs\nFix streamout query comment\nFix the issue that crashAnalysis::CmdDrawIndexedIndirectMulti forwards to the wrong function in the next layer\nFix unused spillTable with userDataMask","shortMessageHtmlLink":"Update pal from commit 664d85b7"}},{"before":"42ac98760ce5d5ae50299d1ffdc2cfaee573d2fa","after":null,"ref":"refs/heads/qiaojbao-patch-1","pushedAt":"2024-02-23T10:25:06.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"}},{"before":"6eaadddc5f56f7e388ace7d0b113684d36df51dd","after":"abb22ae22308954cd9ce76b526c2f805e2dec1ba","ref":"refs/heads/master","pushedAt":"2024-02-22T10:41:19.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 6f6b4e19\n\nUpdate devdriver to v24.01.18\nAdd smaller SampleLocations struct\nAdd support for the new Driver Overlay path\nAdd TestAnyFlagSet64 and TestAllFlagsSet64\nBarrier optimization: allow pipelined decompress in acquire release path\nCollect dispatch dims for DispatchMesh\nCompact InputAssemblyStateParams\nDisable ExecuteIndirectV2 by default on Navi3x\nDrop remaining static_asserts in offline compile\nDrop static assert for LogicOp/DepthRange/DepthClampMode bitstride\nDrop StencilRefMaskParams static assert from RPM generation\nExpose Swizzle mode at Pal::SubresLayout\nInclude more barrier-related calls in CrashAnalysisCmdBuffer\nRemove a redundant operation from ConstructExecuteIndirectPacket()\nRemove Navi3x/Phx macros\nReset m_flags.hasHybridPipeline in CmdBuffer::ResetState() to let Speedway continue\nFix nested bug with wireframe toss point","shortMessageHtmlLink":"Update pal from commit 6f6b4e19"}},{"before":"6eaadddc5f56f7e388ace7d0b113684d36df51dd","after":"abb22ae22308954cd9ce76b526c2f805e2dec1ba","ref":"refs/heads/dev","pushedAt":"2024-02-22T08:35:23.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 6f6b4e19\n\nUpdate devdriver to v24.01.18\nAdd smaller SampleLocations struct\nAdd support for the new Driver Overlay path\nAdd TestAnyFlagSet64 and TestAllFlagsSet64\nBarrier optimization: allow pipelined decompress in acquire release path\nCollect dispatch dims for DispatchMesh\nCompact InputAssemblyStateParams\nDisable ExecuteIndirectV2 by default on Navi3x\nDrop remaining static_asserts in offline compile\nDrop static assert for LogicOp/DepthRange/DepthClampMode bitstride\nDrop StencilRefMaskParams static assert from RPM generation\nExpose Swizzle mode at Pal::SubresLayout\nInclude more barrier-related calls in CrashAnalysisCmdBuffer\nRemove a redundant operation from ConstructExecuteIndirectPacket()\nRemove Navi3x/Phx macros\nReset m_flags.hasHybridPipeline in CmdBuffer::ResetState() to let Speedway continue\nFix nested bug with wireframe toss point","shortMessageHtmlLink":"Update pal from commit 6f6b4e19"}},{"before":"6353b182d28fe85f28b40d1809291b515fcddd3e","after":"6eaadddc5f56f7e388ace7d0b113684d36df51dd","ref":"refs/heads/dev","pushedAt":"2024-01-26T04:06:02.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 12101ea6\n\nUpdate copyright to 2024\nUpdate devdriver to v24.01.08.\nAdd DescribeCallback for ComputeCmdBuffer\nAdd ImageResolveSrcAsNorm flag for ResolveImageCompute\nAdd linear for NonBCViews to pass linear 3D BCn image on vulkan cts\nAdd new PM4 ExecuteIndirectV2 as a substitute to the ExecuteIndirect PM4 (V1)\nAdd offset to trap handler memory base address\nAdd SQTT EXCLUDE_NONDETAIL_SHADERDATA option for TT 3.2 and beyond\nAdd standardized use for extern functions\nAdd TraceConfig trace source\nAdding info to RGD execution marker\nAdjust settings targets to not trigger unconditional rebuild of all downstream sources.\nArchive loader: Global memory for continuations ray sorting\nArchivePipeline::QueryAllocationInfo full implementation\nBug fix and enhancement to GpuDebug\nCleanup PAL interface changes <= 792\nClear cached GPUVA in pm4CmdBuffer::Reset\nCorrect scratch size for color export shader\nDistributed compression support for displayable images\nDrop some asserts related to pipeline bind\nDump depth/stencil view info\nELF prefix change in archive loader\nEnable gang-submit on Linux\nEnable Thread Trace Marker bits on Indirect Multi PM4s as per Spec.\nExpose FixupQueryForNoGangedAce() to fix cts stuck\nForce ext32perfcntr use for secure ranges\nImplement automatically generating gpu hang report for Linux\nImplement command buffer allocation and synchronization\nMake CmdBuffer::CmdNop public\nMisc fix on crashAnalysisCmdBuffer\nMove GfxImage object creation to constructor.\nNew GpuProfiler option to filter ThreadTrace dumping queue ID\nNow linear is valid for 3D image as 2D for gfx10\nRefactor & fix SPM global timestamp counter programming\nRegenerate all rpm shaders with newer PalTools version\nRemove an unnecessary assert of EsGsLdsSize metadata\nRemove Navi1x macros\nRemove PAL gfx10 macros\nReplace outdated assert and comment with alert\nRestore static_assert\nRework Pipeline Dynamic State\nSupport Abs32Lo and Abs32Hi in archive loader\nSupport for loading archive pipelines\nFix a typo in RsrcProcMgr::InitDepthClearMetaData()\nFix AMDGPU(Linux) platform Device enumeration failure if GFX9 device present\nFix few Vulkan applications' hang and crash with Kernel:6.6.rc1\nFix LogExecutionMarkerInfo function\nFix the incorrect calculation in ConvertColorToX9Y9Z9E5()\nFix Xorg crash while running dEQP-VK.mesh_shader* on upstream stack","shortMessageHtmlLink":"Update pal from commit 12101ea6"}},{"before":"6353b182d28fe85f28b40d1809291b515fcddd3e","after":"6eaadddc5f56f7e388ace7d0b113684d36df51dd","ref":"refs/heads/master","pushedAt":"2024-01-26T04:04:52.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 12101ea6\n\nUpdate copyright to 2024\nUpdate devdriver to v24.01.08.\nAdd DescribeCallback for ComputeCmdBuffer\nAdd ImageResolveSrcAsNorm flag for ResolveImageCompute\nAdd linear for NonBCViews to pass linear 3D BCn image on vulkan cts\nAdd new PM4 ExecuteIndirectV2 as a substitute to the ExecuteIndirect PM4 (V1)\nAdd offset to trap handler memory base address\nAdd SQTT EXCLUDE_NONDETAIL_SHADERDATA option for TT 3.2 and beyond\nAdd standardized use for extern functions\nAdd TraceConfig trace source\nAdding info to RGD execution marker\nAdjust settings targets to not trigger unconditional rebuild of all downstream sources.\nArchive loader: Global memory for continuations ray sorting\nArchivePipeline::QueryAllocationInfo full implementation\nBug fix and enhancement to GpuDebug\nCleanup PAL interface changes <= 792\nClear cached GPUVA in pm4CmdBuffer::Reset\nCorrect scratch size for color export shader\nDistributed compression support for displayable images\nDrop some asserts related to pipeline bind\nDump depth/stencil view info\nELF prefix change in archive loader\nEnable gang-submit on Linux\nEnable Thread Trace Marker bits on Indirect Multi PM4s as per Spec.\nExpose FixupQueryForNoGangedAce() to fix cts stuck\nForce ext32perfcntr use for secure ranges\nImplement automatically generating gpu hang report for Linux\nImplement command buffer allocation and synchronization\nMake CmdBuffer::CmdNop public\nMisc fix on crashAnalysisCmdBuffer\nMove GfxImage object creation to constructor.\nNew GpuProfiler option to filter ThreadTrace dumping queue ID\nNow linear is valid for 3D image as 2D for gfx10\nRefactor & fix SPM global timestamp counter programming\nRegenerate all rpm shaders with newer PalTools version\nRemove an unnecessary assert of EsGsLdsSize metadata\nRemove Navi1x macros\nRemove PAL gfx10 macros\nReplace outdated assert and comment with alert\nRestore static_assert\nRework Pipeline Dynamic State\nSupport Abs32Lo and Abs32Hi in archive loader\nSupport for loading archive pipelines\nFix a typo in RsrcProcMgr::InitDepthClearMetaData()\nFix AMDGPU(Linux) platform Device enumeration failure if GFX9 device present\nFix few Vulkan applications' hang and crash with Kernel:6.6.rc1\nFix LogExecutionMarkerInfo function\nFix the incorrect calculation in ConvertColorToX9Y9Z9E5()\nFix Xorg crash while running dEQP-VK.mesh_shader* on upstream stack","shortMessageHtmlLink":"Update pal from commit 12101ea6"}},{"before":"6bdc69412b3c3936bcc6e23edf9e338d66b69d5f","after":"6353b182d28fe85f28b40d1809291b515fcddd3e","ref":"refs/heads/dev","pushedAt":"2024-01-03T03:52:29.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update pal from commit 48ad0c6e\n\n* Update RDF\n* Update DevDriver to v23.11.22\n* Update TraceSession::CollectTrace Result codes based on trace state\n* Update platform settings to devdriver settings\n* Update core settings to devdriver settings\n* Update amdgpu_drm.h\n* Make targets for settings generation.\n* Correct CreateGraphicsPipeline in DeviceDecorator derived classes\n* Relax the PipelineAbiProcessor and handle unnamed symbols\n* Do not cache symbols with type 'section'\n* Initialize GpuPerfExpTraceSource\n* Add setting to control DCC MIN_COMPRESSED_BLOCK_SIZE\n* Add ConvertWcharStringToUtf8\n* Add a new function to Util::Thread class to assign a name to a thread\n* Add the support of setting VGT_STRMOUT_CONFIG.EN_PRIMS_NEEDED_CNT\n* Add missing L2 flush after early phase gfx blit.\n* CodeGen: Add default value for string size\n* Support for SQTT & SPM UberTrace chunks\n* Move settings source files from json to yaml\n* Prepare change for later WaitSync change\n* Totally remove the gfx9 P2P BAR workaround\n* Dump PM4 in Exe named directory\n* Conditional TraceSource activation\n* Suppress warning from [[unlikely]] on Clang 10/11 and gcc\n* VK_EXT_device_fault - Driver Implementation - Linux\n* Couple of barrier code cleanup\n* Generic Gang-Submit Cleanups\n* Change indirect dispatch/draw interfaces to take GPUVAs\n* Remove some deprecated references to addr1 defines that are no longer needed\n* Remove code referencing pre-Navi AddrLib enums\n* Remove CMake overrides for Gfx9 address-lib\n* Remove Validation v nonValidation Distinction from the independent layer\n* Remove Gfx9 classes and formats.\n* Remove CmdDispatchDynamic support\n* Navi3x: set sxblendopt to 0 for alphaToCoverage case\n* SX Blend optimization fix with alpha to coverage state\n* Fix double-free causing test stop during extended 3dmark testing\n* Fix hash size comments\n* Fix PAL decorator class missing variables like stackSizeInBytes\n* Fix various aspects of app profile-related string handling\n* Fix compiler warnings [NFC]\n* Fix hangs and crashes in few Vulkan applications\n* Fix AddrLib version check\n* Fix incorrect application of ppPbbPBBBreakBatchDifferenceWithPrimLimit_FpovLimit_DeallocLimit_A_ WAR\n* Fix incorrect usage of sizeof()\n* Fix typo PAL_CLIENT_INTERFACE_MAJOR_VERSION guard for CmdDrawIndexedIndirectMulti.\n* Fix the failures for dynamicDualSourceBlend also include fast link library cases\n* Fix incorrect function name in UberTraceService\n* Fix for missing SQTT detailed instruction tokes for some Navi3 GPUs\n* Fix Navi3 board power throttle during long time aging","shortMessageHtmlLink":"Update pal from commit 48ad0c6e"}},{"before":"6bdc69412b3c3936bcc6e23edf9e338d66b69d5f","after":"6353b182d28fe85f28b40d1809291b515fcddd3e","ref":"refs/heads/master","pushedAt":"2023-12-28T02:48:23.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"chuang13","name":"Chen Huang","path":"/chuang13","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/129034108?s=80&v=4"},"commit":{"message":"Update pal from commit 48ad0c6e\n\n* Update RDF\n* Update DevDriver to v23.11.22\n* Update TraceSession::CollectTrace Result codes based on trace state\n* Update platform settings to devdriver settings\n* Update core settings to devdriver settings\n* Update amdgpu_drm.h\n* Make targets for settings generation.\n* Correct CreateGraphicsPipeline in DeviceDecorator derived classes\n* Relax the PipelineAbiProcessor and handle unnamed symbols\n* Do not cache symbols with type 'section'\n* Initialize GpuPerfExpTraceSource\n* Add setting to control DCC MIN_COMPRESSED_BLOCK_SIZE\n* Add ConvertWcharStringToUtf8\n* Add a new function to Util::Thread class to assign a name to a thread\n* Add the support of setting VGT_STRMOUT_CONFIG.EN_PRIMS_NEEDED_CNT\n* Add missing L2 flush after early phase gfx blit.\n* CodeGen: Add default value for string size\n* Support for SQTT & SPM UberTrace chunks\n* Move settings source files from json to yaml\n* Prepare change for later WaitSync change\n* Totally remove the gfx9 P2P BAR workaround\n* Dump PM4 in Exe named directory\n* Conditional TraceSource activation\n* Suppress warning from [[unlikely]] on Clang 10/11 and gcc\n* VK_EXT_device_fault - Driver Implementation - Linux\n* Couple of barrier code cleanup\n* Generic Gang-Submit Cleanups\n* Change indirect dispatch/draw interfaces to take GPUVAs\n* Remove some deprecated references to addr1 defines that are no longer needed\n* Remove code referencing pre-Navi AddrLib enums\n* Remove CMake overrides for Gfx9 address-lib\n* Remove Validation v nonValidation Distinction from the independent layer\n* Remove Gfx9 classes and formats.\n* Remove CmdDispatchDynamic support\n* Navi3x: set sxblendopt to 0 for alphaToCoverage case\n* SX Blend optimization fix with alpha to coverage state\n* Fix double-free causing test stop during extended 3dmark testing\n* Fix hash size comments\n* Fix PAL decorator class missing variables like stackSizeInBytes\n* Fix various aspects of app profile-related string handling\n* Fix compiler warnings [NFC]\n* Fix hangs and crashes in few Vulkan applications\n* Fix AddrLib version check\n* Fix incorrect application of ppPbbPBBBreakBatchDifferenceWithPrimLimit_FpovLimit_DeallocLimit_A_ WAR\n* Fix incorrect usage of sizeof()\n* Fix typo PAL_CLIENT_INTERFACE_MAJOR_VERSION guard for CmdDrawIndexedIndirectMulti.\n* Fix the failures for dynamicDualSourceBlend also include fast link library cases\n* Fix incorrect function name in UberTraceService\n* Fix for missing SQTT detailed instruction tokes for some Navi3 GPUs\n* Fix Navi3 board power throttle during long time aging","shortMessageHtmlLink":"Update pal from commit 48ad0c6e"}},{"before":"b86c83013a299c4ea265ab5382cb6e8ff57e14a3","after":"6bdc69412b3c3936bcc6e23edf9e338d66b69d5f","ref":"refs/heads/dev","pushedAt":"2023-11-22T07:57:12.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"chuang13","name":"Chen Huang","path":"/chuang13","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/129034108?s=80&v=4"},"commit":{"message":"Update pal from commit 22607d14\n\n* Fix log running out of GPU memory error for gpu profiler SQTT dumping\n* Only expand once and retile once for VpBlt on VPE\n* PipelineStageFlag Improvements\n* Handle misaligned offset in BestFitAllocator\n* Add CodeObject TraceSource\n* Remove AddrMgr1\n* Add setting to control DCC MIN_COMPRESSED_BLOCK_SIZE\n* Remove p2pCopyToInvisibleHeapIllegal\n* Disallow acquire point PreShader for PWS barrier\n* Revert L1 invalidation in COMPUTE_DISPATCH_INITIATOR\n* Correct graphics shader library in debug overlay\n* Speed up slow clears with VRS\n* Fix dEQP-VK.ray_tracing_pipeline.misc.* tests failing\n* Update address-lib\n* Remove video support for pre-Navi GPUs.\n* Enable DEBUG for asserts\n* Remove Duplicate Setting of nggCullingDataAddr\n* GetGenericGpuSymbol usage cleanup\n* Fix [[unlikely]] [[unlikely]]\n* Implement new BindPipelineValidation callback\n* Report 32-bit SPM data out to RGP.\n* Require clients to default-init structs.\n* HIP changes for secondary trap handler\n* Update the panel description to indicate TT buffer size may change internally","shortMessageHtmlLink":"Update pal from commit 22607d14"}},{"before":"b86c83013a299c4ea265ab5382cb6e8ff57e14a3","after":"6bdc69412b3c3936bcc6e23edf9e338d66b69d5f","ref":"refs/heads/master","pushedAt":"2023-11-21T08:07:44.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"chuang13","name":"Chen Huang","path":"/chuang13","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/129034108?s=80&v=4"},"commit":{"message":"Update pal from commit 22607d14\n\n* Fix log running out of GPU memory error for gpu profiler SQTT dumping\n* Only expand once and retile once for VpBlt on VPE\n* PipelineStageFlag Improvements\n* Handle misaligned offset in BestFitAllocator\n* Add CodeObject TraceSource\n* Remove AddrMgr1\n* Add setting to control DCC MIN_COMPRESSED_BLOCK_SIZE\n* Remove p2pCopyToInvisibleHeapIllegal\n* Disallow acquire point PreShader for PWS barrier\n* Revert L1 invalidation in COMPUTE_DISPATCH_INITIATOR\n* Correct graphics shader library in debug overlay\n* Speed up slow clears with VRS\n* Fix dEQP-VK.ray_tracing_pipeline.misc.* tests failing\n* Update address-lib\n* Remove video support for pre-Navi GPUs.\n* Enable DEBUG for asserts\n* Remove Duplicate Setting of nggCullingDataAddr\n* GetGenericGpuSymbol usage cleanup\n* Fix [[unlikely]] [[unlikely]]\n* Implement new BindPipelineValidation callback\n* Report 32-bit SPM data out to RGP.\n* Require clients to default-init structs.\n* HIP changes for secondary trap handler\n* Update the panel description to indicate TT buffer size may change internally","shortMessageHtmlLink":"Update pal from commit 22607d14"}},{"before":"2de164b431f8a27652e63513ae73338dc512e5bf","after":"b86c83013a299c4ea265ab5382cb6e8ff57e14a3","ref":"refs/heads/dev","pushedAt":"2023-10-23T08:38:46.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"WenqingLiAMD","name":"Wenqing Li","path":"/WenqingLiAMD","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/65225636?s=80&v=4"},"commit":{"message":"Update pal from commit e1d28c51\n\n* Drop support for gfx8-9\n* 128BPP fastclear support\n* Fix graphics pipeline library regressions\n* Sample coordinates of Fmask surface should be the same as origin image's\n* Fix test fail about dEQP-VK.query_pool.statistics_query.*_cq\n* Fixup for likely/unlikely, redundant\n* Add Local heap gpu memory management for GpaSession\n* Use ShaderLibrary owned code object data\n* Split ListDir into CountFilesInDir and GetFileNamesInDir\n* Adds TraceError RDF chunk\n* Force configures the Display DCC offset as displayDccOffset[MaxNumPlanes] when sharedMetadata is used\n* Fixup for add SQTT token mask, redundant\n* Add a setting to change engine preference for image copies\n* Optimize MSAA clears.\n* Refactor GetGenericGpuSymbol\n* GFX10: AC01 clears can use R9G9B9E5 format\n* Discard usage of GpuHeapGartCacheable as a backup heap\n* Bump CMake minimum version to 3.21\n* Update devdriver\n* Refactor ComputeShaderLibrary function list creation\n* Add Util::AtomicCompareExchangePointer\n* Fix and refactor Continuation stack computation.\n* Update Navi31 numSdpInterfaces to 24\n* Fix GpuHeapAccess does not handle TMZ heaps properly\n* Fix perfcounter failures when measuring a large number of UMCCH counters\n* InterpolatorCount is invalid for prerasterOutputSemantic\n* Fix warnings that clang 16 complained about\n* Fix memory growth in GpaSession\n* Remove SQ_NON_EVENT from nested postamble\n* Use `[[likely]]` and `[[unlikely]]` instead of `__builtin_expect()`\n* Use RingSizes from cmdBuffer.\n* VK_EXT_primitives_generated_query - Driver Implementation\n* Make some asserts debug only\n* Add options SurfaceCaptureFrameStart/FrameCount/FilenameHashType for surface capture\n* Add panel setting to control T#/Target.COLOR_TRANSFORM\n* Update Coding Standard doc to explicitly allow attributes\n* Add GetGpuInfo lookup helpers for IDs and info\n* [VK_EXT_nested_command_buffer] Check that the command buffer is not nested when setting allowIb2Launch\n* Added build option PAL_BUILD_RDF with default ON\n* Update the comment that offset between drawRing and controBuffer must be 256 from pm4 spec\n* Add new SQTT token mask setting","shortMessageHtmlLink":"Update pal from commit e1d28c51"}},{"before":"2de164b431f8a27652e63513ae73338dc512e5bf","after":"b86c83013a299c4ea265ab5382cb6e8ff57e14a3","ref":"refs/heads/master","pushedAt":"2023-10-20T16:31:55.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"chuang13","name":"Chen Huang","path":"/chuang13","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/129034108?s=80&v=4"},"commit":{"message":"Update pal from commit e1d28c51\n\n* Drop support for gfx8-9\n* 128BPP fastclear support\n* Fix graphics pipeline library regressions\n* Sample coordinates of Fmask surface should be the same as origin image's\n* Fix test fail about dEQP-VK.query_pool.statistics_query.*_cq\n* Fixup for likely/unlikely, redundant\n* Add Local heap gpu memory management for GpaSession\n* Use ShaderLibrary owned code object data\n* Split ListDir into CountFilesInDir and GetFileNamesInDir\n* Adds TraceError RDF chunk\n* Force configures the Display DCC offset as displayDccOffset[MaxNumPlanes] when sharedMetadata is used\n* Fixup for add SQTT token mask, redundant\n* Add a setting to change engine preference for image copies\n* Optimize MSAA clears.\n* Refactor GetGenericGpuSymbol\n* GFX10: AC01 clears can use R9G9B9E5 format\n* Discard usage of GpuHeapGartCacheable as a backup heap\n* Bump CMake minimum version to 3.21\n* Update devdriver\n* Refactor ComputeShaderLibrary function list creation\n* Add Util::AtomicCompareExchangePointer\n* Fix and refactor Continuation stack computation.\n* Update Navi31 numSdpInterfaces to 24\n* Fix GpuHeapAccess does not handle TMZ heaps properly\n* Fix perfcounter failures when measuring a large number of UMCCH counters\n* InterpolatorCount is invalid for prerasterOutputSemantic\n* Fix warnings that clang 16 complained about\n* Fix memory growth in GpaSession\n* Remove SQ_NON_EVENT from nested postamble\n* Use `[[likely]]` and `[[unlikely]]` instead of `__builtin_expect()`\n* Use RingSizes from cmdBuffer.\n* VK_EXT_primitives_generated_query - Driver Implementation\n* Make some asserts debug only\n* Add options SurfaceCaptureFrameStart/FrameCount/FilenameHashType for surface capture\n* Add panel setting to control T#/Target.COLOR_TRANSFORM\n* Update Coding Standard doc to explicitly allow attributes\n* Add GetGpuInfo lookup helpers for IDs and info\n* [VK_EXT_nested_command_buffer] Check that the command buffer is not nested when setting allowIb2Launch\n* Added build option PAL_BUILD_RDF with default ON\n* Update the comment that offset between drawRing and controBuffer must be 256 from pm4 spec\n* Add new SQTT token mask setting","shortMessageHtmlLink":"Update pal from commit e1d28c51"}},{"before":"36490c70f37dea830f4b8fe6327a6af5fc001793","after":"2de164b431f8a27652e63513ae73338dc512e5bf","ref":"refs/heads/master","pushedAt":"2023-09-29T06:21:20.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"WenqingLiAMD","name":"Wenqing Li","path":"/WenqingLiAMD","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/65225636?s=80&v=4"},"commit":{"message":"Update pal from commit 89573f19\n\n* Add navi32 support\n* Add new SQTT token mask setting\n* Fix RGP setting clock modes in MGPU systems\n* Faster cache layer hashing\n* Minor tweaks for graphics pipeline shader library fast link\n* Scissor in scaled copy compute\n* Add logging of the GpuMemoryDesc struct for each creat…\n* Allow for library with no funcs in LinkWithLibraries\n* Read frontend/backend stack size from .shader_functions\n* Add a 'default chip' for RGA offline compiles\n* Queue: ensure the first element of internal submit info is initialized\n* Change SQTT token mask setting to supress instruction tokens\n* Add a setting to control if issuing marker event is allowed for SQTT\n* Handle ScissorRect before CopyImageCompute\n* Enable 128BPP DCC fast clear support\n* Fix hangs caused by command allocator auto-trim\n* Optimize rpm blt active flags\n* Fix negative 32-bit SPM counter values.\n* Remove cp dma sync in nested postamble\n* Respect elf alignment for memory allocation\n* Fix SlowClearCompute path for color masked\n* Set imageVaLocked=1 unconditionally for all RPM CTV and DSV create\n* Look up color export shader symbol from shader_functions section\n* Add guardband check in restore graphics state\n* Bumps to C++20\n* Prefer MSAA slow clears on graphics in some cases for performance\n* Minor barrier fix for clear since PAL supports read-modify-write clears\n* Minor barrier fix with CmdClearDepthStencil\n* Patch SPI_SHADER_Z_FORMAT from color export shader\n* Added forceShaderRingToVMem as a performance-tuning setting","shortMessageHtmlLink":"Update pal from commit 89573f19"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOC0yOVQwNzoyMzo1Ny4wMDAwMDBazwAAAASnTJXS","startCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOC0yOVQwNzoyMzo1Ny4wMDAwMDBazwAAAASnTJXS","endCursor":"Y3Vyc29yOnYyOpK7MjAyMy0wOS0yOVQwNjoyMToyMC4wMDAwMDBazwAAAAOLkDT6"}},"title":"Activity · GPUOpen-Drivers/pal"}