Skip to content

Commit b5c44bd

Browse files
[BACKPORT] Bump LLVM to pick fixes for Gfx12 Hazards (#1824)
* Squashed 'external/llvm-project/' changes from 656552edc693..a3a3e6997bd7 a3a3e6997bd7 [AMDGPU] Rewrite GFX12 SGPR hazard handling to dedicated pass (#118750) 59613ac2371b Revert "[asan] Enable wait4 test on Android" (#125011) a8d4335ee08b [flang][AIX] Handle more trig functions with complex argument to have consistent results in folding (#124203) 00c096e604ad [clang] StmtPrinter: Handle DeclRefExpr to a Decomposition (#125001) 3c6aa04cf4de [CodeGenPrepare] Replace deleted ext instr with the promoted value. (#71058) 69c24684f66d [BOLT] Fix test. NFC (#124851) b32e55df246e workflows/release-binaries: Stop using ccache (#124415) cdc09a118a71 [mlir][IntRangeInference] Infer values for {memref,tensor}.dim (#122945) de7438e47283 [NVPTX] Auto-Upgrade some nvvm.annotations to attributes (#119261) f0d05b099daf [asan][test] Attempt to fix suppressions-alloc-dealloc-mismatch.cpp on Darwin (#124987) e094c0fa6778 [SandboxVec][Legality] Don't vectorize when instructions repeat (#124479) 62f6d637c015 [libc++] Add clang-21 to failing tests on Windows (#124955) fdd4e9f1012c [clang] UEFI handle unsupported triples. (#124824) 774b12c4a06b [memprof] Initialize AllocInfoIter and CallSitesIter (NFC) (#124972) 8a86e6aefead [MemProf] Constify a couple of methods used during cloning (#124994) 5921295dcaa1 Revert "[SLP] getSpillCost - fully populate IntrinsicCostAttributes to improve cost analysis." (#124962) 4479a2273a12 [DA] add testcase (#116631) 956c0707d909 [mlir][tosa] Change the start and size of slice to tosa shape type (#124209) 46f9cddfd7e4 [DA] enable update_analyze_test_checks.py (#123435) 61ea63baafb5 [Hexagon] Add support for decoding PLT symbols (#123425) ab1ee912be95 [flang][cuda] Remove the need of special compile definition for CUFInit (#124965) 15ab7be2e049 [flang][OpenMP] Parse WHEN, OTHERWISE, MATCH clauses plus METADIRECTIVE (#121817) abc8812df025 [Clang][P1061] Add stuctured binding packs (#121417) 608012ace43b [libc++] Simplify the implementation of iostream.cpp (#124103) 07a0e2be86f3 [clang] Track function template instantiation from definition (#112241) bcf306e0ebb5 [libc] Update include directory for libcMPCWrapper target when LIBC_MPC_INSTALL_PATH is set. (#124810) a368402d6330 [ExtractAPI] merge anon declarators even if they're array types (#120801) bac95752748a [flang] Reset all extents to zero for empty hlfir.elemental loops. (#124867) b8708753c89d [flang] Allow non-index length parameter on exprs fed into hlfir.get_length. (#124827) 8a334af417ff [clang] Remove the deprecated flag `-frelaxed-template-template-args`. (#111894) d3161defd6b7 [PassBuilder] VectorizerEnd Extension Points (#123494) 1822462e2a0c [InstCombine][VectorCombine][NFC] Move a test from InstCombine to (#124948) 57f17319796a [KernelInfo] Remove unused include. 953354c7afcc [KernelInfo] Fix layering violation, Analysis cannot depend on Passes ae6d5dd58bca [MemProf] Prune unneeded non-cold contexts (#124823) bda19768de03 [gn build] Port 18f8106f310e 5dae05f68452 Fix MSVC signed/unsigned mismatch warning. NFC. c8593239a3b5 [flang][OpenMP] Make parsing of trait properties more context-sensitive (#122900) aa2952165cd1 Fix typo "tranpose" (#124929) 27e01d1d74bf [X86] Use new Flags argument to storeRegToStackSlot to simplify code. NFC (#124658) 7fff2527f8a3 [PowerPC] Use SelectionDAG::makeEquivalentMemoryOrdering(). NFC (#124889) 18f8106f310e [KernelInfo] Implement new LLVM IR pass for GPU code analysis (#102944) 15412d735a4f [clang-tidy] Refactor: remove typos in 'AllowedTypes' option in various checks (#122957) 0d4efa27252c [MLIR][Linalg] Introduce linalg.contract (#123618) 88e00141f81c [PhaseOrdering][X86] Add additional hadd/hsub test coverage d444558f51a8 [bazel] Port 25ae1a266d50f24a8fffc57152d7f3c3fcb65517 9adc99bcc564 [AMDGPU] Always emit SI_KILL_I1_PSEUDO for uniform floating point branches. (#124028) e0054e984cac [MLIR][OpenMP] Emit nullary check for mapped pointer members and appropriate size select based on results (#124604) 67752f619179 [libc++] Simplify vector<bool>::__construct_at_end (#119632) 8a43d0e8736c [Attributor] Check correct IRPosition in AANoCapture::isImpliedByIR() 35defdf4707d Revert "[ReachingDefAnalysis][NFC] Use at instead of lookup for DenseMap access" 3ce97e4aa98a [ReachingDefAnalysis][NFC] Use at instead of lookup for DenseMap access 6900768719ff [mlir][Conversion] Fix typos in MemRef descriptor comments (#124923) 4fc514af516f [MLIR][LLVM] Fix import of dso_local attribute (#124822) 382d3599c203 [flang][cuda] Propagate the data attribute on the converted calls (#124877) 29441e4f5fa5 [IR] Convert from nocapture to captures(none) (#123181) 3c3c850a45c8 [ReachingDefAnalysis] Extend the analysis to stack objects. (#118097) 8baa0d9d545f [CodeGen] Avoid repeated hash lookups (NFC) (#124885) 8e4c5cb0063e [Serialization] Migrate away from PointerUnion::dyn_cast (NFC) (#124884) 7ab8f286b8e4 [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124883) c583df46d404 [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124882) a34a087fc597 [clang][bytecode] Handle non-primitive vector element types (#124926) 0c63ec5347d2 [NFC][SIWholeQuadMode] Remove redundant arguments (#124930) 349eab186986 [asan] Enable wait4 test on Android (#124879) 540402c05bad [sanitizer_common] Remove workarounds for older Android (#124881) bfabd5be5359 [libc++] Remove some private symbols from the ABI (#121497) a01097faca35 [mlir][spirv] Add definition for VectorTimesMatrixOp (#124571) f226cabbb1b9 [ValueTracking] Handle nonnull attributes at callsite (#124908) 39a72be5e59c [Clang] [NFC] Introduce `ConstDynamicRecursiveASTVisitor` (reland) (#124821) e9c2e0acd747 [AArch64] Match GCC behaviour for zero-size structs (#124760) 25ae1a266d50 [mlir][spirv] Make ConvertToSPIRVPass into a test pass (non-public) 5a4945fa4d51 [RISCV][SIFIVE] Fix latencies for vector integer arithmetic long latency (#124855) 2e43f3922377 [NFC][SIWholeQuadMode] Perform less lookups (#124927) ce96c26cd612 Revert "[llvm][DebugInfo] Attach object-pointer to DISubprogram declarations (#122742)" (#124853) 4573c857da88 Revert "[mlir][spirv] Make ConvertToSPIRVPass into a test pass (non-public) (#124301)" 4a1a697427bb [SLP][NFC]Unify ScalarToTreeEntries and MultiNodeScalars, NFC 7441e87fe053 [libclc] Move several integer functions to CLC library (#116786) cf37ae5caeaf [InstCombine] Add one-use check when folding fabs over selects (#122270) 058d183980a2 [mlir][spirv] Make ConvertToSPIRVPass into a test pass (non-public) (#124301) 86e20b00c313 [libcxx] Use _ftelli64/_fseeki64 on Windows (#123128) d412fe531dce [Offload] Enable mlir and flang in bot build (#124915) 008e16242b86 [mlir][x86vector] Restrict BF16 dot test to x86 (#124916) 3a29dfe37c58 [LLVM][AMDGPU] Add Intrinsic and Builtin for ds_bpermute_fi_b32 (#124616) 978e0839ae2f [OpenMP] Allow OMP6.0 features. (#122108) 66e0498dafbf [GlobalISel] Do not run verifier after ResetMachineFunctionPass (#124799) 983562d8c521 [AMDGPU][NFC] Simplify t16/fake16 TableGen definitions. (#122693) 8d6b24167b7b [mlir] Make `TypedStrAttr` actually enforce the string type. (#124770) 2f401458266c [RISCV][TTI]Use processShuffleMasks for cost estimations/actual per-register shuffles 87782b216fd3 [mlir][x86vector] AVX512-BF16 Dot op (#124800) 9534d27e3321 [X86] vector-idiv-sdiv-512.ll - regenerate VPTERNLOG comments 12cdf4330d32 [libclc] Move (add|sub)_sat to CLC; optimize (#124903) ecc71de53f87 [flang] Implement IERRNO intrinsic (#124281) 2b55ef187cb6 [VPlan] Add helper to run VPlan passes, verify after run (NFC). (#123640) 36b3c43524c8 [AArch64] PAUTH_PROLOGUE should not be duplicated with PAuthLR (#124775) 5e43418e0edb [ARM] Forbid use of TLS with execute-only (#124806) f20b8e35b3bc [MLIR][Linalg] Fixes for Winograd decomposition and for tiling (#123675) 690f251063d6 [LoopInterchange] Handle LE and GE correctly (#124901) 89ca3e72ca03 [CostModel][X86] Reduce worst case v8i16/v16i8 SSE2 shuffle costs (#124789) e80d934aa6c1 [MCJIT][test] Move remaining MCJIT interpreter tests to Interpreter/ subdirectory (#124744) 89e767f1277b [LoopIdiom] Move up atomic checks for memcpy/memmove (NFC) (#124535) db567eaca071 [lldb][NFC] Format part of ScriptInterpreterPython.cpp 9ea64dd87813 [lldb] Make Python >= 3.8 required for LLDB 21 (#124735) 776ef9d1bec6 [LoopVectorize][NFC] Regenerate some early exit test CHECK lines (#124900) 2a1f79582f93 [MLIR] Fix import of invokes with mismatched variadic types (#124828) 1ac3665e66c7 [clang] Restrict the use of scalar types in vector builtins (#119423) e902cf2df171 [LLD][COFF] Write both native and EC export symbols to the import library on ARM64X (#124833) 5a34e6fdceac [flang] Implement CHDIR intrinsic (#124280) e811cb00e533 [flang] Implement !DIR$ UNROLL [N] (#123331) 51c7338cc671 [clang][bytecode] Fix dummy handling for p2280r4 (#124396) 9326633abd0e [lldb] Remove PATH workaround for Android (#124682) 5d3ae5161210 Reapply "[mlir][python] allow DenseIntElementsAttr for index type (#118947)" (#124804) c836b8956d39 [LoopVectorize][NFC] Disable output for tests that don't need it (#124747) 07efe2c18a63 [SCEV] Check correct value for UB (#124302) a3aa452a21f9 [CodeGen] RegisterCoalescer: Remove unused AliasAnalysis dependency (#124773) 71edfd623026 [AMDGPU][NewPM] Sketch out a AMDGPUPassRegistry skeleton (#124785) 267e293510ad [bazel] Introduce MAX_CLANG_ABI_COMPAT_VERSION (for #123998) d1997322cb89 [TableGen] Use MCRegister::id() to avoid an implicit conversion from MCRegister to unsigned. NFC d34462409757 MLIR: asan: Fix python tests under asan on Linux (#123303) 2ec27848c00c [MLIR] normalize-memrefs: Normalize memref.alloca (#123293) 1c3ea59daac2 [gn build] Port 79cbad188afd 3feb72449623 [AsmPrinter][ELF] Support profile-guided section prefix for jump tables' (read-only) data sections (#122215) 77e44e59665e [gn] port 60927405f561 b0924ed64e0d [AggressiveInstCombine] Add tests for memchr inline threshold (NFC) (#121711) d4159e2a1d1d [MLIR][NVVM] Add support for griddepcontrol Ops (#124603) ab6d41eae12b Add llvm-driver to the list of toolchain tools (#123083) b28e879a8323 [clang-format] Simplify ConfigParseTest for int/unsigned options (#124704) 9ec4f474c0f4 [GlobPattern] Fix doxygen docs (#124361) 654b76321a60 [flang][cuda] Allow to set the stack limit size (#124859) 9052b37ab1aa [ORC][LLI] Remove redundant eh-frame registration plugin construction from lli. 8675cd3facc0 [RISCV][VLOPT] Compute demanded VLs up front (#124530) cdced8e5bc42 [MLIR] Implement emulation of static indexing subbyte type vector stores (#115922) 4c7aa6f98368 [msan] Fix -Wunused-variable in non-assertion builds after #124421 23763a1200ed [compiler-rt][rtsan] mremap for Linux interception. (#124234) 3a3377579f13 [mlir][math]Update `convertPowfOp` `ExpandPatterns.cpp` (#124402) 3bd3e06f3fe4 Bump version to 21.0.0git (#124870) 8c2574832ed2 Reland: [clang] improve print / dump of anonymous declarations (#124858) fdadef9be33f [msan] Handle x86_avx512_(min|max)_p[sd]_512 intrinsics (#124421) ff271d04a2fe [RISCV][VLOPT] Fix assertion failure across blocks (#124734) 60927405f561 Don't use CLANG_VERSION_MAJOR to check that the value passed to -fclang-abi-compat= is valid (#123998) d0052ebbe2e2 [ORC][LLJIT] Install eh-frame registration plugin during platform setup. 79cbad188afd [SandboxVec] Clear Context's state within runOnFunction() (#124842) 6b654a09c2b6 [sanitizer] Support "alloc_dealloc_mismatch" suppressions (#124197) 98d6dd398873 [LLVM][LangRef][noalias] Remove Redundant Line and Improve Wording (#124685) 617278e7b0c9 [lld][WebAssembly] Fix for shared library symbols WRT replacing lazy symbols (#124619) 28507ac62928 [MLIR] Fix thread safety of the deleter in PyDenseResourceElementsAttribute (#124832) b8cdc5ea2741 Reapply "[msan] Add avx512-intrinsics.ll and avx512-intrinsics-upgrade.ll test case (#123980)" (#124500) 8aae191cb6ad [BPF] Remove 'may_goto 0' instructions (#123482) e7f02241ad4d [nfc][llvm] Clean up isUEFI checks (#124845) 437040fe3df4 [CMake] Use shm_open, not clock_gettime, to detect librt (#120727) 38902153fe66 [PassBuilder] Add RelLookupTableConverterPass to LTO (#124053) 381218950e9c [clang-cl]: generate debug info when `novtable` is specified (#124643) 4a426079d60a [msan] Use horizontal add to compute shadow for horizontal sub (#124835) 548ecde42886 Add extra explicit triple to fix errors from #110102 3a1e157454ec [bazel] More fixes for 35df525fd00c2037ef144189ee818b7d612241ff 6bb6c305752a [clang] NFC: cleanup check template argument (#124668) 4eb7c349a616 [bazel] Fix build after 35df525fd00c2037ef144189ee818b7d612241ff 1f7eb6f403bc [lldb] Make SBProgress move-only (#124843) c2fba023475f [ValueTracking] Fix bug of using wrong condition for deducing KnownBits (#124481) 35df525fd00c [mlir][Vector] Add support for poison indices to `Extract/IndexOp` (#123488) a06c89387621 [libc++abi] Make once_flag constinit in cxa_exception_storage (#124627) e89e7c4685aa [lldb] Fix lldb-dotest.in to use args determined by CMake (#124811) 9d7999885a6a [lldb] Update API headers for SBProgress (#124836) 26f4e2a701f7 [libc][docs] Fix the RPC documentation leaking ports 2cbf2798ae55 [clang] fix nondeduced mismatch with nullptr template arguments (#124498) ed199c8d7668 [libc] Workaround for gcc complaining about implicit conversions with the ternary ?: operator. (#124820) 6338bde5681c [VPlan] Use cast<VPRecipeBase> in verifier (NFC). c47a57393c2b [libc] Fix conversion warnings for float16 tests. (#124830) 9f660625bee6 [libc++] Disable CFI in __libcpp_allocate (#124805) 9d8dc45d1708 [flang] IEEE underflow control for Arm (#124807) 7bd9c780e363 [msan][NFCI] Generalize handleIntrinsicByApplyingToShadow to allow alternative intrinsic for shadows (#124831) bfefa15cc18f [mlir][bufferization] Use original type when convert arg for users (#124826) b1972689a431 [clang] fix template argument conversion (#124386) a255da0c5373 [libc] clean up duplicated includes in generated headers (#124524) 5dab5bf0851f [Clang] handle [[warn_unused]] attribute for unused private fields (#120734) 310f55875f2f [CUDA] Make target intrinsics work with ptx 8.7 (#124818) 964565c868c4 [NFC][docs][HIP] Update HIP docs around `hipstdpar` and SPIR-V (#124803) de9b0ddedc43 Add explicit triple to fix errors from #110102 589bef333eb3 [emacs][lsp][tblgen] add tblgen-lsp-server support for emacs lsp-mode (#76337) ef232a7e3486 [BOLT][AArch64] Remove nops in functions with defined control flow (#124705) 1d5ce614a7cd [CodeGen] Avoid repeated hash lookups (NFC) (#124677) 69c9bed21b90 [Serialization] Migrate away from PointerUnion::dyn_cast (NFC) (#124676) 2ef497e5b026 [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124675) 986581f6bcef [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124674) e29c085812e2 [clang] disallow narrowing when matching template template parameters (#124313) eb10e94180dd NFC: convert clang/test/AST/HLSL/StructuredBuffers-AST.hlsl to unix line endings b108fbe6ea42 [bazel] Fixes for 7f37b34, 0165e33 and 3c64f86 a4d17c44f149 [clang][Sema] Generalize HeuristicResolver::resolveTypeToRecordDecl() to resolveTypeToTagDecl() (#124699) 8578b816fa90 [Driver] Fix a build error 822f74a91106 [Clang] Cleanup docs and comments relating to -fextend-variable-liveness (#124767) cfdd937b30ac [llvm] Android has sched_getaffinity() now (#124689) 063db51cd446 Reapply "[msan] Add handlers for AVX masked load/store intrinsics (#123857)" 17d1523207c6 [Clang] Minor fixes to 'gpuintrin.h' header bb9533598277 [libclc][NFC] Clang-format includes 0b80491cd5e7 [flang] Support non-index shape/shift/slice for CG box operations. (#124625) 1b729c3d70ce Revert "[mlir][python] allow DenseIntElementsAttr for index type (#118947)" 9dd762e8b105 [mlir][python] allow DenseIntElementsAttr for index type (#118947) 2f077dfbdf76 [flang] Support non-index shape in hlfir.get_extent codegen. (#124622) 947d8ebbf337 [SLP]Unify getNumberOfParts use 1bc5fe669f54 [mlir][python] implement GenericOp bindings (#124496) 822954b4a977 [TySan] Add initial documentation for Type Sanitizer (#123595) a080498c92e4 [libc] Update include directory for libcMPCommon target when LLVM_LIBC_MPFR_INSTALL_PATH is set. (#124765) d76ea250c8b9 Reland [InstCombine] Teach foldSelectOpOp about samesign (#124320) ef92e6b99fc2 [BoundsChecking] Update ubsantrap to use GuardKind (#124613) f388ca3d9d9a [NFC] Remove redundant `sys/file.h` header. (#124684) 648912582ccb [libc] Add `FILENO` related macros to `unistd.h`. (#124688) 56a0a7f6d188 [flang][cuda] Adding support for more atomic calls (#124671) df122fc734ce [SPIR-V] Change a way SPIR-V Backend API works with user facing options (#124745) a58e774fba42 [mlir][tosa] Make TOSA MUL's Shift an Input (#121953) 5a8fe9e9a777 [libc] Revise the definition of `{get, set}rlimit`. (#124701) a8c82d5fde1d [libclc] Optimize isfpclass-like CLC builtins (#124145) 8ce0d05b88f3 [libc] Revise the definition of `posix_spawn`. (#124686) e38f4f6904b7 Revert "[clang] improve print / dump of anonymous declarations (#124605)" 48df9480dab5 [NFC] Suppress spurious deprecation warning with MSVC (#124764) 75622e3f8d9d [MLIR] Define `getArgument()` for Toy tutorial passes d459784cbea3 [IR][SPIR-V] Replace of PointerType::get(Type) with opaque version (NFC) (#124755) a0b049055df5 [RISC-V] Fix incorrect epilogue_begin setting in debug line table (#120623) 2abde54aabc4 [gn build] Port de4bbbfdccb6 7d172f96ff2c [CostModel][X86] getShuffleCosts - convert all shuffle cost tables to be CostKind compatible. NFC. (#124753) 68d90cff580f [AMDGPU][GlobalISel] Fix assert on APInt creation. (#124608) 9d8d538e40ef libclc: clspv: add missing clc_isnan.cl dependency (#124614) 6cb71d7e85d5 [flang][NFC] Restrict -funroll-loops tests to known working targets (#124594) ee1c6a6bc150 MachineVerifier: Move test into AMDGPU directory 3007f31e7458 [LoopUnroll] Add AArch64 tests for multi-exit loop unrolling. 820c6ac7f5e4 [Clang] call HandleImmediateInvocation before checking for immediate escacalating expressions (reland) (#124708) 78b5bb702fe9 [libclc][NFC] Move key math headers to CLC (#124739) 3a975d697f11 [gn] port 7e22180c20f (BuiltinsHexagon.td) 4bd2307a2085 [clang][ExprConst] Don't diagnose a non-existent init as not constant (#124575) 79499f010d2b [NFC][DebugInfo] Deprecate iterator-taking moveBefore and getFirstNonPHI (#124290) a1ab5b4c8725 [SLP]Check the MainOp matches the requirements for the instructions 22687aa97bda [CodeGen] Correctly handle non-standard cases in RemoveLoadsIntoFakeUses (#111551) 37b595cb3715 [gn] port b968fd950266 (BuiltinsNVPTX.td) 11db7fb09b36 [GlobalISel] Catching inconsistencies in load memory, result, and range metadata type (#121247) 0cf6714279d4 [lldb][AArch64] Fix GCS register field detection 1d5fbe83c3bc [SLP]Adjust NumberOfParts value for adjusted number of buildvector scalars e84f6b6a88c1 [mlir] Fix conflict of user defined reserved functions with internal prototypes (#123378) 13dcc95dcd49 [Offload] Rework offloading entry type to be more generic (#124018) 015aed18ee35 [SelectionDAG] WidenVecOp_INSERT_SUBVECTOR - Replace `INSERT_SUBVECTOR` with series of `INSERT_VECTOR_ELT` (#124420) 606cf887416b [ExecutionEngine][test][RISCV] Don't mark RISCV as unsupported (#124464) 8ad9e1ecb7e5 [Clang] Fix use of deprecated method and missing triple 62c16d82194f [MCJIT][test] Move MCJIT non-interpreter tests to MCJIT subdirectory (#124463) afa4681ce443 [flang][debug] Add support for common blocks. (#112398) e1b5826c2b9d [LLVM][CMake][MSVC] Wrap linker options for ICX with LLVM_BUILD_INSTUMENTED (#124573) c0861e9cbbc3 Revert "[flang] IEEE underflow control for Arm (#124617)" 4424c44c8c4e [Clang] Add fake use emission to Clang with -fextend-lifetimes (#110102) 500a1834d92d [RISCV][VLOPT] Fix some typos in vl-opt-op-info.mir test. NFC 8ea018ce1de0 [DAGISel] Fix MMRA Handling in copyExtraInfo (#124730) 8353aa2a53b3 [llvm][Docs] Add LLDB AArch64 GCS Release note a7f4044bd019 [clang][SME] Emit error for OpenMP captured regions in SME functions (#124590) 431024506c6f [bazel] Remove DebugInfo files covered by more specific targets (#124138) 3a51466caf93 [LLD][COFF] Add support for delay-load imports on ARM64X (#124600) c5840cc609a3 [lldb][AArch64] Add register fields for Guarded Control Stack registers (#124295) 83433d936195 [OpenMP][IRBuilder] Handle `target ... nowait` when codegen targets host (#124720) 1b551e76a2fe [LLD][COFF] Call setLocation on DelayAddressChunk when inserting into the addresses vector (NFC) (#124736) 6d0dd3d5c41e [lldb][Docs] Add Guarded Control Stack to AArch64 Linux page (#117860) 304a99091c84 [NFC][DebugInfo] Use iterators for insertion at some final callsites 7f845cba2ccc [libc++] Update the CI to Clang-20 and drop Clang-17 support (#117429) 75aa5a35568b [lldb][AArch64] Add Guarded Control Stack support for Linux core files (#124293) 65f81df47390 [Docs][DebugInfo] Summarise what people need to do for RemoveDIs now (#124725) 8017ca1d0056 Reapply "[AArch64] Combine and and lsl into ubfiz" (#123356) (#124576) 97aa56ada5d2 [LLD][COFF] Move delayLoadHelper and tailMergeUnwindInfoChunk to SymbolTable (NFC) (#124729) db6fa74dfea3 [AArch64] Implement FP8 Neon reinterpret intrinsics (#120476) 2c934dc5e1a3 [clang][bytecode] Always compile most recent function decl (#124722) cdea38f91afc Reland "[LoopVectorizer] Add support for chaining partial reductions #120272" (#124282) b29bf3de051b [llvm][Docs] Re-order the LLDB release notes 0165e3346fdb [llvm][Object] Add missing const qualifier for value_type in content_iterator (#124106) 7cd6f8557814 [CodeGen][NFC] Format RegisterCoalescer sources (#124697) 71ab44a8193c [Clang] Add "extend lifetime" flags and release note (#110000) 8fe7860610e3 LAA/test: cover invariant stores with unit stride (#124586) d69b78573355 [bazel] Add Builtins for #122873(NVPTX) and #123460(Hexagon) de4bbbfdccb6 [Build Attributes] Standardize names according to convention. (#124556) 9c92824dd638 [libc++] Remove Android header no longer in use (#124691) cfc8ef0ad8f7 [libclc] Move copysign to CLC library; fix & optimize (#124598) 4a00c84fbb74 [AArch64] Allow register offset addressing mode for prefetch (#124534) 0f61558b97aa [LoopVectorize][NFC] Remove unused variable in addUsersInExitBlocks (#124553) 1295aa2e814d [Clang] Add -fwrapv-pointer flag (#122486) 458542f454cd [mlir][linalg] Relax structured op region filler check (#123741) f76f534e8c81 [NVPTX][Script] Update lit.cfg.py (#124395) f4de28a63c81 [StrTable] Switch intrinsics to StringTable and work around MSVC (#123548) cd57c9530b91 [NFC][AMDGPU] Autogenerating test cases (#124507) 8e97f50eed71 [clang][ASTImporter] Fix possible crash at import of function template (#124273) c8d3ccfa165d [RISCV] Use llvm::reverse instead of make_range(rbegin, rend). NFC 7e22180c20fa [StrTable] Mechanically convert Hexagon builtins to use TableGen (#123460) 0865ecc5150b [clang] Extend diagnose_if to accept more detailed warning information, take 2 (#119712) aab25f20f6c0 [HLSL][SPIRV][DXIL] Implement `WaveActiveMax` intrinsic (#123428) 00f692b94f9a Reland "[MLGO] Count LR Evictions Rather than Relying on Cascade (#124440)" 0cb7636a462a [RISCV] Add MIPS extensions (#121394) b968fd950266 [StrTable] Mechanically convert NVPTX builtins to use TableGen (#122873) ea9993a9a350 [RISCV] Add P550 scheduler model. (#124639) 085f7fb560ee [ELF] Remove redundant isExported computation 5e43dd5bde99 [test] Add missing -triple=x86_64 8f8a640e9ab5 [ELF,test] Test static-pie __global_pointer$ b593110d89ae [compiler-rt] Deprecate LLVM_ENABLE_PROJECTS in favor of LLVM_ENABLE_RUNTIMES (#124016) 4167ea2cb082 Revert "[libcxx] Use alias for detecting overriden function" (#124431) 952685a43d04 [ELF,test] Add static-pie test related to demoted lazy symbol d4af658323c6 [RISCV] Support multiple memory operands in expandRV32ZdinxStore. f10441ad0032 [ELF] Refine includeInDynsym condition 6a9d0e53ae04 [llvm-cov] Prevent assertion failure in sumMCDCPairs 42432ada8e0b [AMDGPU][NFC] Sort AMDGPUPassRegistry entries alphabetically (#124544) 839cdb58e71f [libc][complex] remove -ffreestanding flag from MPFRUtils compile options (#124702) d606f68a2e37 [bazel] Fix build after e0c7f081f1582d49f81ec4c6cdbf5d6ef13c58ba d50ebd47ae57 [clang-format] Add style option `PenaltyBreakBeforeMemberAccess` (#118409) 3a439e2caf0b [mlir][dataflow] disallow outside use of propagateIfChanged for DataFlowSolver (#120885) 3c64f86314fb [mlir] Add OpAsmTypeInterface for pretty-print (#121187) 7f37b34d3191 [libc][complex] Testing infra for MPC (#121261) 7109f521975e [ELF,test] Don't rely on --export-dynamic --gc-sections behavior for non-pie static linking 2d0688797cc3 [RISCV] Renaming muladdi to muliadd as per v0.5 spec. (#124237) bd8578bf2e8f [JITLink] Add 'REQUIRES: asserts' to MachO_ptrauth_dependencies.s. cc97653d534e AMDGPU: Custom lower 32-bit element shuffles (#123711) 334a1cdbfaaf [SandboxIR] createFunction() should always create a function (#124665) fa9ac62d02fd [ubsan] Parse and use <cutoffs[0,1,2]=70000;cutoffs[5,6,8]=90000> in LowerAllowCheckPass (#124211) 5ab43c3e7d9b [JITLink] Use continue to reduce loop indentation and improve readability. NFC. 08d14e10ca4b [SLP] Fix CommonMask will be transformed into an incorrect mask if createShuffle is called multiple times. (#124244) 79685b59b788 [JITLink][aarch64] Fix dependence tracking for Pointer64Authenticated edges. 8b29c5ccddb8 [JITLink] Move debugging output helper array into LLVM_DEBUG. NFC. 6e4105574ebb [NFC][AMDGPU] Improve code introduced in #124607 (#124672) 9c1d2f84e91f [JITLink] Add "Pointer64Authenticated" string to aarch64::getEdgeKindName. ba789c6f82a9 [Flang] Add semantics checks for CrayPointer usage in DSA list (#123171) 0bbfd96a3a8e [gn build] Port e0c7f081f158 e0c7f081f158 [lld-macho] Refactor BPSectionOrderer with CRTP. NFC c4891089125d [flang] Added hlfir.reshape definition/lowering/codegen. (#124226) 5ece348f77e9 [lldb] Android 9 has added the spawn.h header (#124452) 641694729df1 [lldb] Clean up Socket headers for Android (#124453) ad697b28f479 ReleaseNotes: add lld/ELF notes 760a786d1580 [Clang] Prevent `mlink-builtin-bitcode` from internalizing the RPC client (#118661) 7873d3b50b44 Revert "[Clang] [NFC] Introduce `ConstDynamicRecursiveASTVisitor`" (#124667) 8900c09ebfd7 [mlir][nvgpu] Fix crash when handling 0D memref in OptimizeSharedMemoryPass (#124517) 740e6aeccdb2 Revert "[AMDGPU] Use the AMDGPUToolChain when targeting C/C++ directly (#99687)" eeefa72f7c72 [Clang] [NFC] Introduce `ConstDynamicRecursiveASTVisitor` (#122991) d839e765f03c [TargetLowering] Inline the only caller of one of the forceExpandWideMUL functions. NFC 194c74e01668 [clang][Sema] Null-check type in resolveMemberExpr() before checking for auto type (#124628) f75860f89522 [AArch64] Implement NEON FP8 intrinsics for fused multiply-add (#123615) f949f876daed [clang] improve print / dump of anonymous declarations (#124605) aa65f93b71de Revert "[MLGO] Count LR Evictions Rather than Relying on Cascade (#124440)" 3cfda4f11842 Don't use -z,defs linker flag when building shared libraries with emscripten (#123396) f94c481543bd [clang] Track source deduction guide for alias template deduction guides (#123875) a34159f85ee0 [TSan][Apple] Fix interceptor build error (#124351) 934532d8b18a remove unused var after refactoring fa6e976602a4 [llvm-exegesis] Use TestBase for TargetTest (#121895) e98b2028c758 [NFCI]Refactor AsmPrinter around jump table emission (#124645) 8b1edc0f0ceb Revert "[ELF] Remove redundant isExported computation" 8cc83b66e20e [MLGO] Count LR Evictions Rather than Relying on Cascade (#124440) 7fd58339b4c7 [clang] Add __nullptr as a keyword to C (#123119) 0e372c3ea31d Revert "[Clang] call HandleImmediateInvocation before checking for immediate escacalating expressions" (#124646) 89c5576ff903 OpenMP: Fix Python 3 SyntaxErrors (#123940) a7a4c16c672b [mlir][Vector] Support efficient shape cast lowering for n-D vectors (#123497) 3b2b7ec07d97 [AMDGPU] Handle invariant marks in `AMDGPUPromoteAllocaPass` (#124607) 5a81a559d69f [GISel] Explicitly disable BF16 tablegen patterns. (#124113) c310b4e7bd2a Revert "[Clang] __has_builtin should return false for aux triple builtins (#121839) (#124626) 5d1c596ab47b [AMDGPU][True16][MC] true16 for minimummaximum/max/min/max3/min3 (#124184) b707d52fc557 [compiler-rt][Mips] Properly guard references to _ABIN32 (#124492) aca08a8515d4 [TableGen] Add assert to validate `Objects` list for `HwModeSelect` (#123794) 6578790076fe Reapply "[Fuchsia][cmake] Allow using FatLTO when building runtimes" (#119252) (#121820) 713482fccf82 [VPlan] Use State.get to extract lane mask for BranchOnMask. 5815a311050a [Clang] call HandleImmediateInvocation before checking for immediate escacalating expressions (#124414) ad9da92cf6f7 [LoopUnroll] Add RuntimeUnrollMultiExit to loop unroll options (NFC) (#124462) 804b81d39f2d [AArch64] Add FP8 Neon intrinsics for dot-product (#123613) 3bf8e671dbc8 [libc++][format] Add tests for flat_(|multi)map formatting (#124418) c24e5f982e27 [GlobalMerge] Fix inaccurate debug print. (#124377) 0cbb1d5673dd [GlobalMerge] Use constructor to set all bits in BitVector. NFC (#124375) e734f01bffb8 [clang] Prevent duplicated instantiation of enumerators of unscoped member enumerations (#124407) bd5d361c0598 [mlir][vector] add support for linearizing vector.bitcast in VectorLinearize (#123110) 285009f202ca [NFC][DebugInfo] Rewrite more call-sites to insert with iterators (#124288) a85b2dc45a5f [Clang] only inherit the parent eval context inside of lambdas (#124426) c4c76eabb88f [flang] IEEE underflow control for Arm (#124617) 08a18efc664a [libc++][doc] Update the release notes for LLVM 20. (#124403) d732c86c9282 [flang] Don't take corank from actual intrinsic argument (#124029) 1e9b60cfa431 [flang] Recognize and check EVENT_QUERY (#123429) 2e5a5237daf8 [flang][debug] Avoid redundant debug data generation for derived types. (#124473) aa34a6ab2990 [RISCV] Add register allocation hints for lui/auipc+addi fusion. (#123860) 539b2e06542f [WebAssembly] Fix catch block type in wasm64 (#124381) 1bcf44be4c6a [MLOpt] Add tf_xla_runtime to LLVMDevelopmentExport (#124610) e77d428e46d9 [AMDGPU] Do not remat instructions with PhysReg uses (#124366) d1139b32d251 [AMDGPU][True16][CodeGen] true16 codegen pats for v_mad_u16 (#124000) 4e812756745e [tsan] Add support for linux/riscv64 in lib/tsan/go/buildgo.sh (#124557) 817e777296a5 [CodeGen] Avoid repeated hash lookups (NFC) (#124506) e0c5a8553d62 [memprof] Migrate away from PointerUnion::dyn_cast (NFC) (#124505) 5d2434166787 [lld] Migrate away from PointerUnion::dyn_cast (NFC) (#124504) 4075915ebdfc [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124503) dc6411d3e1ab [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124502) 5d6d982df61d [AArch64] Generate zeroing forms of certain SVE2.2 instructions (11/11) (#116837) 754b94638e89 [lld] Support RUN_LLD_MAIN_TWICE for the ELF port (#124441) 1782168c527b [X86] Fix a warning 610e33a54775 [Polly] Ensure i1 preload condition 7b1becd940cb [libc] Add CMake cache file for the GPU build (#124589) 88cca8ea209b [libc++] Add more missing bits to the locale base API (#122531) 658f8500c84f [bazel] Remove obsolete mlir-cpu-runner alias 99bd2e3f123b [AArch64] Add Neon FP8 conversion intrinsics (#123612) 19f052443df0 [Clang] fix test on 32 bits target after 561132e (#124593) 1e89355dadce [clang-format] Treat `f<N | M>(a)` as template function call (#124438) 44c9e46fce12 [InstrRef] Fix mismatch between LiveDebugValues and salvageCopySSA (#124233) e7de60369836 [X86] combineCMov - pull out repeated getValueType calls. NFC. b80965efc1ea [Support] Report OOM from `allocate_buffer` (#85449) 08c364280a79 [flang] Improve error message on bad complex literal. (#124331) e252c402104b [flang] Fix spurious error due to bad expression shape calculation (#124323) fee393e4ea2b [flang][runtime] Don't crash on ASYNCHRONOUS='NO' in child I/O (#124208) 873426bea3dd [lldb-dap] Ensure the IO forwarding threads are managed by the DAP object lifecycle. (#122783) b0fab14e9ca2 [flang] Fix spurious error in character sequence association (#124204) c596aae47ad8 [flang] Catch assumed-length interoperability error (#124179) d1d952c206ef [flang] Interpret 'Q' exponent letter as kind=16 even on x86 (#124158) ec6b2c63d93d [flang] Fold character array constructor with unknown length (#123983) f5ddb1012de1 [flang] Fix crash in module file generation (#123859) 3ac00784ac3c [flang] Fix crash on erroneous program (#123843) b16c98969720 [flang] Fix check for coarray actual passed to implicit interface (#123836) 210e675cfd7b [flang] Accept CHANGE TEAM/END TEAM as branch target (#123822) 73db9ee1e87b [clang][Sema][FMV] Add a note to the 'cannot become multiversioned' diagnostic (#124364) 73f9034036c9 [flang] Fix failure to fold character array (#123418) 4e231014c198 [AArch64] Generate zeroing forms of certain SVE2.2 instructions (10/11) (#116836) 2625510ef809 [flang] Refine EVENT_TYPE/LOCK_TYPE usage checks (#123244) 512b44d5e153 [flang] Define ATOMIC_ADD as an intrinsic procedure (#122993) 34b139594aa2 [NFC][DebugInfo] Switch more call-sites to using iterator-insertion (#124283) 038b42ba5b47 [flang] Safer hermetic module file reading (#121002) c3a0fcc98206 [libclc] Optimize CLC vector any/all builtins (#124568) 38b3f45a8112 [Offload] Fix offload-info interface 6ff86f2c0a5b [AMDGPU] Use the AMDGPUToolChain when targeting C/C++ directly (#99687) 81d18ad86419 [NFC][DebugInfo] Make some block-start-position methods return iterators (#124287) 5f5cdf40382f [lldb][TypeSystemClang] CreateParameterDeclarations: don't specify SmallVector size 09a29fcc8dbb [VPlan] Don't collect live-ins in collectUsersInExitBlocks. (NFC) (#123819) 749443a307e8 [NFC][DebugInfo] Mop up final instruction-insertion call sites (#124289) 1bb784a7489e [LowerMatrixIntrinsics] multiply-minimal.ll - use -passes="..." to allow DOS to correctly evaluate the RUN command ad2b2aa50bc8 [PhaseOrdering] vector-trunc.ll - use -passes="default<O2>" to allow DOS to correctly evaluate the RUN command 7f24b9acd189 [CI] Support multiple jobs in metrics container (#124457) 16e9601e193e [Flang] Adjust the trampoline size for AArch64 and PPC (#118678) 178f47143a3b [CostModel][X86] getShuffleCost - shuffles with only one defined element are always cheap (#124412) cb6f021af235 [RISCV][VLOPT] Remove unnecessary passthru restriction (#124549) f95f10c7e65b [AArch64] Generate zeroing forms of certain SVE2.2 instructions (9/11) (#116835) e492083f55d9 [OpenACC] Add AutomaticAllocationScope to recipe ops (#124337) 1e2d5f7943d0 [NFC][mlir][OpenMP] Remove mentions of `target` from generic `loop` rewrite (#124528) 3322ba493ad7 Revert "[flang] IEEE underflow control for Arm" (#124570) 20f72d19fc58 Revert "[flang] arm build fix" (#124569) 1eb4e9f88b82 [flang] arm build fix (#124562) 559287575b5b [GlobalMerge][NFC] Reland "Skip sorting by profitability when it is not needed" 212f344b84b4 [InstCombine] Handle constant expression result in tryFactorization() e14962a39cc6 [NFC][DebugInfo] Use iterators for instruction insertion in more places (#124291) 5aafc6d58f34 [Polly] Fix typos discovered by codespell (#124545) 62340ff8d844 [AMDGPU][True16][MC] true16 for v_cmpx_xx_f16 (#123419) 092372da15e5 [mlir][Tensor] Rework `ReifyRankedShapedTypeInterface` implementation for `tensor.expand_shape` op. (#113501) 27c917307563 [Clang] Remove unnecessary Decl transform & profiles for SizeOfPackExpr (#124533) 1f5335c1db5d Make index computation used divsi/remsi (#124390) 5c5bbffe75ca [clang][ASTImporter] Import source location of explicit object parameter instead of copying it (#124305) 081723b9db84 [lldb][TypeSystem] Ensure that ParmVarDecls have the correct DeclContext (#124279) 561132e71b29 [Clang] Fix immediate escalation of template function specializations. (#124404) eaa5897534cb [libclc] Optimize CLC vector is(un)ordered builtins (#124546) f1d5e70a00fb [SLP][NFC]Do not check poison values for corresponding vectorized entries f95a8bde3425 [AArch64] Refactor implementation of FP8 types (NFC) (#123604) 54928a10c8db [clang] __STDC_NO_THREADS__ is no longer necessary for VS 2022 1939 and above (#117149) f07505849c8e [Offload] Fix server thread from being shut down if unused 3a4376b8f906 LAA: handle 0 return from getPtrStride correctly (#124539) 3684ec425904 [flang] IEEE underflow control for Arm (#124170) 86705eb6242b [X86] huge-stack-offset.ll - add gnux32 test coverage e7592d83e0ac [Offload][NFC] Make sure the thread is not running already d7e561b913d2 [flang][OpenMP] Support `bind` clause code-gen for standalone `loop`s (#122674) e9e06bea8661 [lldb][AArch64][NFC] Move a comment in GCS tests 347fb208c1e3 [libclc] Optimize CLC vector relational builtins (#124537) ef54e0bbfbef [AArch64] Avoid generating LDAPUR on certain cores (#124274) b31e9747d086 [lldb][AArch64] Fix expression evaluation with Guarded Control Stacks (#123918) d8ad1eef8ffe [AArch64] Generate zeroing forms of certain SVE2.2 instructions (7/11) (#116833) e21b80464a44 [compiler-rt][rtsan] socketpair interception. (#124107) cfdd7d736a94 [compiler-rt][rtsan] sched cpu affinity for linux interception. (#124194) 6087c3049656 [lldb] Simplify preprocessor conditional (#124522) 98e52db4a5e5 Revert "[flang][NFC] Restrict -funroll-loops tests to known working targets" (#124536) b8d921003d1f [flang][NFC] Restrict -funroll-loops tests to known working targets (#123939) b7286dbef9dc Reland "[LoopVectorize] Add support for reverse loops in isDereferenceableAndAlignedInLoop #96752" (#123616) ddbfe6f7d207 [Sema] Fix __array_rank instantiation (#124491) ac87d6b03642 [mlir][arith] Fold `arith.cmpi eq, %val, %one : i1` -> `%val` and `arith.cmpi ne, %val, %zero : i1 -> %val` (#124436) 43a50deb6345 [MLIR][ROCDL] Add GFX940 SMFMAC (2:4 sparsity) instructions to the ROCDL dialect (#124435) 14ffff384740 [clang] Add dump() support for lvalue APValues (#124476) 8f17f51deb12 [mlir][tosa] Fix comments format(NFC) (#124520) 87103a016fbf [AArch64] Implement NEON FP8 vectors as VectorType (#123603) 3b5e9eed2f67 [NVPTX] Add float to tf32 conversion intrinsics (#124316) 7211bf48a62b [flang][driver] add negative from of -fsave-main-program (#124110) bbf377060adc [win/asan] GetInstructionSize: Support some more 7 or 8 byte instructions. (#124011) 351ee30529c0 [win/asan] GetInstructionSize: Support some more 6 byte instructions. (#124006) bd38c4993aa4 [AArch64] Generate zeroing forms of certain SVE2.2 instructions (8/11) (#116834) a01e1d4e044e [clang][Sema] Handle dependent qualifier in HeuristicResolver::resolveDeclRefExpr() (#124515) 9452ee4f750a [clang-format] Treat uppercase identifiers after struct as macros (#124397) 6805d7e8aa5f [test] Convert remove-note.test from \r\n to \n after #118739 b9efbed468ec Revert "Move HIP fatbin sections farther away from .text" 1a4d6de1b532 [ELF] Remove redundant isExported computation 2a26292388fc [ELF] Make isExported accurate early 0e6b58202ca9 [ELF] Improve parseSymbolVersion tests in for compileBitcodeFiles 7107f55d82f8 [clang] NFC: remove redundant dyn_cast b2647ffbf797 Revert "[msan] Add avx512-intrinsics.ll and avx512-intrinsics-upgrade.ll test case (#123980)" b6eeec586fa6 Revert "[msan] Fix-forward avx512-intrinsics-upgrade.ll (#124495)" b9d301cc7e4f Revert "[msan] Add handlers for AVX masked load/store intrinsics (#123857)" a6044a05cd16 [msan] Fix-forward avx512-intrinsics-upgrade.ll (#124495) 84af3ee5124d [ELF] Replace Fatal with Err 980e86f130ee [msan] Add avx512-intrinsics.ll and avx512-intrinsics-upgrade.ll test case (#123980) db79fb2a91df [msan] Add handlers for AVX masked load/store intrinsics (#123857) bfa7de0df5d8 X86: Support FCANONICALIZE on f64/f80 for i686 with SSE2 or AVX (#123917) c9637afec7ed [Clang] Fix createConstexprUnknownAPValues to use zero offset when ceating APValue (#124478) e278e1b6ece0 [NFC][CodeGen] Fix typos in code comments. (#124382) 80ab237c1187 [LLD][COFF] Add REQUIRE x86 to arm64x-import.test (NFC) fb01a289038c [LLD][COFF] Implement support for hybrid IAT on ARM64X (#124189) 1c4341d17649 [SandboxVec][DAG] Fix interval check without Node ccc066e8d5a7 [TableGen] Avoid repeated map lookups (NFC) (#124448) e8e75e08c921 [lld-macho] Remove unneeded functions from BPSectionOrderer. NFC f8ab91f74f15 [LVI][CVP] Add test for trunc bittest. (NFC) 2af819fa3d80 [MemProf] Add test for hot hints (#124394) 33ad474c45e6 [Clang] Add predefined macros for integer constants (#123514) 0c784851c50b [MathExtras] Favor using the hexadecimal FP constants (#123180) e4514293f999 [Clang] Correctly determine constexprness of dependent lambdas. (#124468) dec47b76f406 [CostModel][X86] Update baseline CTTZ/CTLZ costs for x86_64 (#124312) 81d38da65e33 [LV] Add more tests for narrowing interleave groups for AArch64. 8035d38daab0 [Flang][OpenMP]Add parsing support for DISPATCH construct (#121982) 850852e9a45f [CodeGen] Avoid repeated hash lookups (NFC) (#124455) f09a6f632584 [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124447) ab895ad2bfb6 [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124446) f46eb1430992 [AMDGPU] Replace unsigned with Register in SIMachineScheduler. NFC 37fdde6025c8 [CodeGen] Remove implict conversions from Register to unsigned from MachineOperand. NFC 753028bc81c1 [Xtensa] Move XtensaUtils to MCTargetDesc 44b85743498a [AArch64] Fix movk parsing with an .equ operand (#124428) 6bb70a94da1b workflows/release-binaries: Enable builds on Linux/AArch64 (#120786) 9b6990ff2531 [Github][CI] Add Windows Premerge Job for Testing (#122661) c1ec5beb4ab3 [clang-format] Fix a TableGen crash on comment after l_paren (#124380) c1f10ef0a5c1 [ELF] SHF_LINK_ORDER: replace Fatal with ErrAlways 18335f4800ae [ELF] Ignore --time-trace if disableOutput f359c1f524bf [ELF] Disable error handling script if disableOutput 988978f964fb [ELF,test] Add env LLD_IN_TEST=1 to make some tests work if RUN_LLD_MAIN_TWICE f21c35d54f8f [ELF] Replace some Fatal with Err b7195e8e040d [ELF,test] Add env LLD_IN_TEST=1 to make some tests work if RUN_LLD_MAIN_TWICE a9e92beb253d [ELF] openAuxiliaryFile: open /dev/null if disableOutput and filename is "-" 6b87f01aaaa9 [ELF] MergeInputSection: replace Fatal with Err 7db789b57027 [ELF] Replace a few Fatal with Err c7579bfba596 [ELF] -o -: suppress output if disableOutput 4f4804817165 [ELF] SHF_MERGE: avoid Fatal d2c7cabe0453 [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124434) 0f3c2884f3cc [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124433) 19a6ac18ef3e [ELF] EhFrame: replace failOn with errOn 04d5608057f7 [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124430) 563c7c5539f0 [clang] Migrate away from PointerUnion::dyn_cast (NFC) (#124425) 1395cd015f2e [VPlan] Support multi-exit loops in HCFG builder. 2131115be5b9 [InstCombine] Drop Range attribute when simplifying 'fshl' based on demanded bits (#124429) 77c325b64630 [LLD][COFF] Keep hasData true in NullChunk constructor (#124368) 89f2fee9f806 [InstCombine] Add test for incorrect retention of Range attribute in fshl e5b0132d157a SCEV: add samesign tests for exit-limit computation (#124304) 5e65f430414d [SLP][NFC]Add a test, producing serie of extrtactelements, building non-extendable tree 4bcd8184a093 [TargetLowering] Pull similar code out of the forceExpandWideMUL into a helper. NFC (#124371) 2655ae54db6d [mlir] Fix deprecated pointer union casts in toy example (#124422) d578d0bb135c [gn build] Port def50f701f6a def50f701f6a [libc++] implement `std::flat_multimap` (#113835) aba0476f23fc [lldb] Delete lldbutil.PrintableRegex (NFC) b178c2d63e07 [SandboxVec][DAG] Fix trim schedule 8b6211472793 [lldb] Delete unused lldbutil.print_registers (NFC) 14b44179cb61 [libc++][format][3/3] Improves formatting performance. (#108990) 485b1ac8a265 [SandboxIR][Docs] C++ highlighting for code block 5cb2db3b51c2 [SandboxVec][Scheduler] Forbid crossing BBs (#124369) 21f04b1458c5 Hold a queue of iterator ranges (not operations) in wouldOpBeTriviallyDead (#123642) 9325a61aa096 Revert "[GlobalMerge][NFC] Skip sorting by profitability when it is not needed" (#124411) 6383a12e3b43 [VPlan] Refactor HCFG builder to preserve original vector latch (NFC). de5ff8ad07ae [libc++][test] Improves C++ Standard filtering. (#89499) 1a53d4baeb02 [clang][cmake] Apply bolt optimizations as part of the clang target (#119896) 52bffdf9f5bb [IPSCCP][FuncSpec] Protect against metadata access from call args. (#124284) 2696e4fb9567 [libc++] Reduce std::conjunction overhead (#124259) 7974f12b1e36 [HLSL] Suppress a warning in #122820 [-Wunused-but-set-variable] 84d4037488f5 Reapply "[libc++] Fix tests for clang::no_specializations for C++17 and C++20" 72918fd11dd8 [GlobalISel] Avoid repeated hash lookups (NFC) (#124393) 0cc74a894188 [CodeGen] Avoid repeated hash lookups (NFC) (#124392) 62bd217b5a1c [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124391) 186d6546d9c5 [Index] Migrate away from PointerUnion::dyn_cast (NFC) (#124389) c216081e981e [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124388) f607e3fd23ef [Clang][Sema] Reject declaring an alias template with the same name as its template parameter. (#123533) 1f26ac10ca1b [X86] Better handling of impossibly large stack frames (#124217) 8a6b44bf4cfe Revert "[libc++] Fix tests for clang::no_specializations for C++17 and C++20" ac1ba1f9dd70 [CodeGen] Introduce a VirtRegOrUnit class to hold virtual reg or physical reg unit. NFC (#123768) 3b35b4c7f914 [mlir] Allow fallback from file line col range to loc (#124321) 8e31050bc2e0 [clang-format] Fix a bug in annotating overloaded co_await decl (#124240) 4ea44eb1e292 [WebAssembly] Fix EH feature flags when compiling multiple files (#124374) 07ed8187acc3 [OpenMP] Replace nvvm.annotation usage with kernel calling conventions (#122320) d92bac8a3ebb [HLSL] Introduce address space `hlsl_constant(2)` for constant buffer declarations (#123411) 1b4bd4e1a512 [BOLT][AArch64] Remove assertions from jump table heuristic (#124372) db1ee18eda63 NFC: Typo fix c725a95e088d [MemProf] Convert Hot contexts to NotCold early (#124219) 48657bf29b01 [flang][cuda] Handle launch of cooperative kernel (#124362) 6409799bdcd8 [SandboxVec][Legality] Pack from different BBs (#124363) 280c7d719834 [CI] Increase Configurability of Monolithic Windows Build (#124328) d789915f35a9 [Github] Bump Runner Version in Containers (#124324) 241e5d8c5c42 [AMDGPU][True16][MC] true16 for v_cmpx_eq_f16 (#124038) e2005d146194 [LLDB] Reapply #123873 SBSaveCore Docstrings (#124355) 4df9c17e5f43 [libc++] Fix tests for clang::no_specializations for C++17 and C++20 ac75d3228014 [SandboxVec][VecUtils] Filter out instructions not in BB in VecUtils:getLowest() (#124360) d910fbcbd10c [RISCV][NFC] cR Constraint Release Note cff7ad56babc [SandboxVec][Utils] Implement Utils::verifyFunction() (#124356) 05fd4d5775e2 [flang][cuda] Perform inlined assignment when field is c_devptr (#124322) 4b209c5d87c8 [SandboxIR][Region] Add cost modeling to the region (#124354) 34c6c5e72f48 [BOLT][AArch64] Fix PLT optimization (#124192) 77c23fd0aa15 [AMDGPU] Update AMDGPUUsage.rst to document two intrinsics (#123816) 425d25f5df4c [AArch64][WinCFI] Fix a crash due to missing seh directives (#123993) b41987beaeda [SandboxVec][DAG] Fix MemDGNode chain maintenance when move destination is non-mem (#124227) 73b462321c29 [libc] Include size_t type header in strings.h (#124352) df9b31f1e0cd [clang][Sema] Handle undeduced auto types in HeuristicResolver (#124236) ee054404dfde [flang][cuda] Carry over the cuf.proc_attr attribute to gpu.launch_func (#124325) ae8b560899c0 [MemProf] Disable hot hints by default (#124338) 074a25fb2678 [RISCV][MC] Create an AsmOperand for carry-in vmask (#124317) d398c0c97aa0 [libc][cpio] Add cpio.h header. (#123798) 1b1270f30bbd [FMV][GlobalOpt] Enable static resolution of non-FMV callers. (#124314) 95d993a83886 [MLIR] Fix import of calls with mismatched variadic types (#124286) 3b30f20c60d0 [libc++][TZDB] Fixes CI. 83df39c649fe [mlir][inline] Fix Issue#82401: Infinite loop in MLIR inliner for indirect recursive call. (#124026) ab976a171213 PreISelIntrinsicLowering: Lower llvm.exp/llvm.exp2 to a loop if scalable vec arg (#117568) 3861b9db882d [gn build] Port 0cd794d4860e 825e712959d4 [HLSL] cbuffer: create host layout structs (#122820) e4009ed3d68b [clang][docs] Update bytecode interpreter docs (#124252) d9b8120259a5 [lld/COFF] Fix -start-lib / -end-lib more after reviews.llvm.org/D116434 (#124294) 544a3cb65b6b [flang][cuda] Handle variable with initialization in device global pass (#124307) a9ad601f7c54 [RISCV] Use vrsub for select of add and sub of the same operands (#123400) 7293455cf292 [lldb] Add SBThread.selected_frame property (#123981) 12f82fbe0723 [compiler-rt] Fix Windows test after profile summary change (#124318) 0cd794d4860e [libc++][chrono] implements UTC clock. (#90393) bd8a8181288c [Offload] Add cuLaunchHostFunc to dynamic cuda 134401deea5e [Offload] Move RPC server handling to a dedicated thread (#112988) c025b96ef9bb [ELF] Symbol::extract : remove unneeded file->lazy check 5daecd4a3b9c [Support] Fix namespace after #123990 474f5d2aefb4 [FMV][AArch64] Remove features predres and ls64. (#124266) d88293d8a200 [mlir][vector] Disable `BreakDownVectorBitCast` for scalable vectors (#122725) ba6774f997ee [mlir][xegpu] Fix verifier diagnostic recursion (#124148) a976036a100b [libc][NFC] Remove extra ; in exhaustive_test.h. (#124216) 3da7de34a2bc [flang][runtime] Disable optimization for traceback related functions. (#124172) cccb55491223 [lldb] Remove unused posix_openpt function definition for Android (#124257) 7842374103b2 [NFC][TableGen] Emit nested namespaces in InstrInfoEmitter (#124210) 256f40d0e6b2 [libc] Use the NVIDIA device allocator for GPU malloc (#124277) 37bf0a10fb4c [SCEV] Add test for #123550 (NFC) 5d2393a222c7 [InstCombine] Avoid repeated hash lookups (NFC) (#124243) ec66c4af0926 [AMDGPU][True16][CodeGen] true16 codegen pattern for f16 canonicalize (#122000) a94226f9e6f5 [llvm-ml] Remove unsafe getCurrentSegmentOnly() call (#123355) 6c11b7e689c8 [CodeGen] NFC: Change order of checks in MachineInstr->isDead() (#124207) a12d7e4b611f [SLP] getVectorCallCosts - don't provide scalar argument data for vector IntrinsicCostAttributes (#124254) c546b5317c51 [ValueTracking] Pass changed predicate `SignedLPred` to `isImpliedByMatchingCmp` (#124271) 9cf52fe1f94f [flang][OpenMP][NFC] test the current private dealloc runtime calls (#124017) 2068b1ba031e [X86] Fix ABI for passing after i128 (#124134) f2b253b9613a [SelectionDAG] Fix an incorrect DebugLoc on a COPY (#122963) acde3f722ff3 [mlir:python] Compute get_op_result_or_value in PyOpView's constructor. (#123953) 77c780d64b95 [bazel] Port eb206e9ea84eff0a0596fed2de8316d924f946d1 e5e55c04d6af [GlobalMerge][NFC] Skip sorting by profitability when it is not needed (#124146) 970094d50b08 [llvm-opt-report] Show scalable vectorization factors (#123367) 11b040192640 [AMDGPU] Restore SP from saved-FP or saved-BP (#124007) 02c6002d1cd2 [lldb][AArch64] Add Guarded Control Stack registers (#123720) 6292a808b352 [NFC][DebugInfo] Use iterator-flavour getFirstNonPHI at many call-sites (#123737) a5cc897cdedf [gn build] Port 0ee037b861f9 eb206e9ea84e [mlir] Rename mlir-cpu-runner to mlir-runner (#123776) b4e81fd1eb22 [gn] port 4018317407006b2c63 d6e0798a2a7c [Mips] Add the missing judgment when processing function handleMFLOSlot (#121463) acc13dbe4dcb [libc][workflow] improve ci coverage with windows-2025 and arm ubuntu (#123745) 57b48987f6c2 [lldb] Use the first address range as the function address (#122440) b60c118f53e6 MachineUniformityAnalysis: Improve isConstantOrUndefValuePhi (#112866) 5ce349589a4a [Coro] Prebuild a module-level debug info set and share it between all coroutine clones (#118628) 4831fa8632f6 AMDGPU/GlobalISel: RegBankLegalize rules for load (#112882) 8e6d6a55108c [clang-tidy][NFC] improve performance misc-unused-using-decls (#123454) 46a08ce83262 [clang-tidy][NFC] simplify `TimerGroup` in `ClangTidyProfiling` (#123958) 88136f964529 [mlir][vector] Canonicalize gathers/scatters with trivial offsets (#117939) 0ee037b861f9 AMDGPU/GlobalISel: AMDGPURegBankLegalize (#112864) 965ff7fa309d [NVPTX] Add SM versions for 101 and 120 (#124155) 33fc477e6e26 [clang][NFC] Update `cxx_dr_status.html` dac49e8ddd5d [Arm] Fix generating code with UB in NeonEmitter (#121802) 8e7027350903 [NFC][DebugInfo] Use iterator moveBefore at many call-sites (#123583) e6030d389571 [clang][bytecode] Use std::allocator calls for Descriptor source (#123900) 9bb3c62948d0 [Flang][Driver] Deprecate Ofast (#101701) 625e0a40f1a9 [SLP][X86] Add missing SSE2/SSE4 checks from vector rotate tests f8a56df36e39 AMDGPU/GlobalISel: AMDGPURegBankSelect (#112863) 3208801aec18 [AArch64][GlobalISel] Add some bf16 GISel test coverage. NFC aff1242b8ee8 [LV] Align debug location of the widen-phi to the original phi. (#120338) 77465967130a [SLP][X86] Add VBMI2 coverage for funnel shift tests 865104a1042e [AArch64][SME] Change output class of FORM_TRANSPOSED_REG_TUPLE pseudos (#123755) 0510d4ea59a1 [bazel]Fix bazel build after 631a6e0004e57ca85569b99ea411418627925697 4b6fc4934685 [llvm][Docs] Clarify the process for requesting a merge on your behalf (#124154) 1c0af8dced4a [TySan] Added tests for methods of ignoring instrumentation (#124125) 148da0662850 [Sema] Migrate away from PointerUnion::dyn_cast (NFC) (#124229) b4ef11d0e20b [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124228) 1fa56038f6ea [CostModel][X86] getIntrinsicInstrCost - lrint/llrint costs can use getCastInstrCost without argument data 990837f91de3 [mlir][arith][tensor] Disable index type for bitcast (#121455) df3bc54eff3b [flang] Avoid repeated hash lookups (NFC) (#124230) 196f7c2a4f47 [Utils] Identity map module-level debug info on first use in CloneFunction* (#118627) 24e70e393072 [libc++] Switch experimental library macros to 0/1 macros (#124030) 97df7411fd99 [llvm][Docs] Make it clear where lit test files live (#124121) eda16991adeb [lldb] Enable the use of dladdr() on Android (#124187) a6cfde62bb89 [lldb] Check Android API for existence of getgrgid_r() introduced in 24 (#124182) bfd9bc274586 [AMDGPU] SIPeepholeSDWA: Disable on existing SDWA instructions (#124131) e289cb545ada [lldb] Remove more workrounds for Android that have been fixed upstream (#124176) ee2722fc882e [X86][AVX10.2-BF16] Remove [NE]P from intrinsic and instruction name (#123335) 886adf8fb3ba [clang][ExprConst] Let diagnostics point to std::allocator calls (#123744) ddd2f57b2966 [X86] Use NSW/NUW flags on ISD::TRUNCATE nodes to improve X86 PACKSS/PACKUS lowering (#123956) f6253f8fe9ff [JITLink][LoongArch] Add label addition and subtraction relocations (#122262) b84b717f093b [CostModel] getTypeBasedIntrinsicInstrCost - add default cost approximations for funnel shifts (#124175) 8ef171ee831f [msan] Handle horizontal add/subtract intrinsic by applying to shadow (#124159) 45d83ae7df65 [mlir] [math] Fix the precision issue of expand math (#120865) 212cdc9a377a Revert "[ORC] Enable JIT support for the compact-unwind frame info format..." 9dd5aed4edf0 [JITLink][LoongArch] Support R_LARCH_ALIGN relaxation (#122259) a9c61e0d7655 [NewPM] LiveIntervals: Check dependencies for invalidation (#123563) 3c79a04cc231 [AMDGPU] Add amdgpu-sw-lower-lds pass to NPM codegen addIRPasses. (#124102) 6330f1e052db [clang-format] Fix a regression in `PointerAlignment: Left` (#124085) 2f39d138dc38 [DirectX] Handle dx.RawBuffer in DXILResourceAccess (#121725) 6db73fa481be [SandboxVec][Scheduler] Fix clear() to clear all state (#124214) f3d2e75ead4e [compiler-rt][rtsan] inotify api for Linux interception. (#124177) 02a30049926b [compiler-rt][rtsan] preadv(64)/pwritev(64) interception. (#124115) fd174f0ff3e7 [clang-repl] Try to XFAIL testcase on arm32 without affecting arm64 darwin. 435609b70c8b [LLVM][NVPTX] Add support for griddepcontrol instruction (#123511) c9bc242e387f [clang-repl] The simple-exception test now passes on arm64-darwin. a001cc0e6cdc [ORC] Destroy defunct MaterializationUnits outside the session lock. 9fecb4f90717 [CodeGen] Fix a warning 6735d527f994 [MIPS] [MSA] Widen v2i8, v216 and v2i32 vectors (#123040) 378dcf61014b [C++20] [Modules] Fix may-be incorrect ADL for module local entities (#123931) 0ef39a882bb3 MachineCSE: Remove check for subreg on a def operand (#124095) e06b7030303b [RISCV][NFC] Remove Redundant Inline Asm Logic (#124202) b11529bfa26f [libc] Use -fno-math-errno to for __builtin_fma* to generate fma instructions. (#124200) 4405f7288193 [llvm][NFC] Remove CV-qualified base class in PassManagerInternal.h (#124193) 001326437442 [clang][test] Add .cuh as a recognized extension for lit test files (#124080) acb7859f075f [MachineSink] Extend loop sinking capability (#117247) 24f177df61f6 [X86][AVX10.2-BF16] Update VCOMISBF16 intrinsics and instructions (#123307) 969eb4ec4c22 [msan][NFC] Correct and clarify comment for getShadowPtrOffset() d2234ca16310 [SandboxVec][BottomUpVec] Fix packing when PHIs are present (#124206) 33c44074714d [RISCV] Support cR Inline Asm Constraint (#124174) 788318484d96 Revert "[InstCombine] Teach foldSelectOpOp about samesign" (#124123) 26fc07d5d887 [ORC] Add explicit narrowing casts to fix build errors. 67a8857989ed [flang][cuda] Handle pointer allocation with double descriptors (#124183) 13dae34819e5 [DSE] Enable the initializes improvement in DSE (#124058) c676104875f3 [lld-macho] Implement symbol string deduplication (#123874) 28ad8978ee20 Reland: [clang] unified CWG2398 and P0522 changes; finishes implementation of P3310 (#124137) bc74a1edbe5e [IA] Generalize the support for power-of-two (de)interleave intrinsics (#123863) 10772807ab72 Reapply "[llvm-objcopy][ELF] Add an option to remove notes (#118739)" 3d59e30cbcfe [flang][cuda] Allow DO CONCURRENT in cuf kernel (#124190) 4065d985ab0f [flang][cuda] Allow complex type in cuf kernel reduce (#124185) c7053ac202de [SandboxVec][BottomUpVec] Disable crossing BBs (#124039) 621e5cd82045 Revert "[llvm-objcopy][ELF] Add an option to remove notes (#118739)" 4f0325873fac [ORC] Enable JIT support for the compact-unwind frame info format on Darwin. 25825d4ac968 [WebAssembly] Enable Wasm EH features only once (#124042) 42043c423ff0 Reapply "Verifier: Add check for DICompositeType elements being null" 9324e6a7a5c5 [llvm-objcopy][ELF] Add an option to remove notes (#118739) f5bd623d0600 [RISCV][VLOPT] Rename vx to vf where appropriate in test case 0e213834df11 Revert "[LoopVectorizer] Add support for chaining partial reductions (#120272)" (#124198) fc952b2a6933 [AArch64] Add pre-index store patterns for bf16. bf258dbd57ee [RISCV][VLOPT] support fp sign injection instructions (#124195) f402e06e7d81 [RISCV][VLOPT] Add vector fp min/max instructions to isSupportedInstr (#124196) d87441a24204 [llvm-profdata] Fix detailed summary format on Windows (#124169) 631a6e0004e5 [libc][wchar] implement wcslen (#124150) 3ed28bbf195d [bazel] Generate CXX11AttributeInfo.inc for 4018317407006b2c632fbb75729de624a2426439 1688c8719fca s/requires/REQUIRES to fix the test on release build f2942b907786 [CodeGen] NFC: Move isDead to MachineInstr (#123531) 76ed4b1cec57 [TableGen] Add StringInit pools to RecordKeeperImpl::dumpAllocationStats. (#124164) e30a4fc3e20b [TargetLowering] Improve one signature of forceExpandWideMUL. (#123991) e19261faf5c7 [TableGen] Add a SmallPtrSet to track WriteRes that are referenced by some ReadAdvance. NFC (#124160) 0b7cbd23a043 [Darwin][Sanitizers][CrashReporter] Adopt initializer for Crash Reporter Annotations struct (#123978) 7a831eb924e3 [VPlan] Remove unused VPLane::getNumCachedLanes. (NFC) c4ca87ee78fa [yaml2obj] Don't use uninitialized Type (#123274) e10d551aa482 [mlir][PDLL] Allow (and ignore) `-D` tablegen macros. (#124166) 24b137365004 [libc][docs] Add Unistd header's page to the status of implementations doc (#123068) c118864223c6 [MLIR][ROCDL]Add MFMA_*_F8F6F4 instructions to the ROCDL dialect (#123830) cb981cc540ba [libc] added btowc to wchar.h generated header (#124168) d17e4ca7e02b [bazel][NFC] Add td_library for downstream use (#124156) 401831740700 [Clang] restrict use of attribute names reserved by the C++ standard (#106036) c3ecbe6792bb Disable the test again. * https://lab.llvm.org/buildbot/#/builders/127/builds/2148/steps/7/logs/stdio shows a failure. c9b7303b9b18 Add [[maybe_unused]] to a variable used only in assert in VPlan.h (#124173) 775d0f36f748 [GVN] Handle scalable vectors with the same size in VNCoercion (#123984) 4cf1fe240589 [lldb] Add missing operations to GetOpcodeDataSize (#120163) a6211a64dc22 [gn build] Port ff17a4136ded 1c28b9237382 [Clang] __has_builtin should return false for aux triple builtins (#121839) 0d0190815d8f [TailDup] Allow large number of predecessors/successors without phis. (#116072) 6d5697f7cb4e [SystemZ] Fix ICE with i128->i64 uaddo carry chain 3dec24d2a284 Stats are sorted before they are printed. Try fixing test failure by checking stats in its print order. bec4c7f5f7fb [InstCombine] Unpack scalable struct loads/stores. (#123986) 2f6b0b4a8522 [RISCV] Add SiFive sf.vqmacc tests to vmv-copy.mir. NFC (#124075) db6b7a84e6e4 [libc][NFC] Strip all training whitespace and missing newlines (#124163) 029069316544 [llvm-profdata] XFAIL broken test on windows (#124165) 604514601415 [libc] change return type of pthread_setspecific to int in generated header (#124072) e0cd57decb3a [compiler-rt] Remove support and workarounds for Android 4 and older (#124056) 5a7d92f7a09d [NFC] Remove invalid features from test and autogenerate checks. (#124130) ff17a4136ded [lldb] Remove support and workarounds for Android 4 and older (#124047) 66e49e38aeed [YAML] Don't validate `Fill::Size` after error (#123280) c7e6ca76cb4b [SLP][NFC]Add dump() method for ScheduleData struct type for better debugging 6a97897d5c15 [NFC][modules] Create objects on the stack (#124034) ed512710a5e8 [Offload] Make MemoryManager threshold ENV var size_t type. (#124063) e06222459675 [test] Remove misleading '' 3d72619d7519 [InstCombine] Add a test for splitting scalable structs. NFC 6d4e72abb85a [GVN] Add extra vscale tests with different types. NFC e00f1f843610 [ELF] Error for executable .note.GNU-stack unless -z execstack or -r a2453097e3b4 [llvm-profdata] Add block percent to detailed summary (#105915) caf0540b91b0 [LoopVectorizer] Add support for chaining partial reductions (#120272) 8c138bee6e5a [flang][cuda] Handle pointer allocation with source (#124070) 1a8f49fdda5b [mlir][python][cmake] Allow skipping nanobind compile options changes. (#123997) 082b148041ec [TableGen] Pass CodeGenProcModel reference instead of index to addWriteRes/addReadAdvance. NFC 2f76e2b27d9d [Driver] -fno-plt: warn for unsupported targets d8cd8d56ea98 [SLP] getSpillCost - fully populate IntrinsicCostAttributes to improve cost analysis. (#124129) 7ddeea359811 [LLVM][AMDGPU] MC support for ds_bpermute_fi_b32 (#124108) bda39a606783 [ExecutionEngine] Include <map> (#124083) bb019dd165ce [CodeGen] Avoid repeated hash lookups (NFC) (#124078) d05008363d4e [lldb] Avoid repeated map lookups (NFC) (#124077) 113e1fdc8c7f [CodeGen] Migrate away from PointerUnion::dyn_cast (NFC) (#124076) e622468f164f [AST] Migrate away from PointerUnion::dyn_cast (NFC) (#124074) fa299294c068 [SLP][NFC]Modernize code base in several places 7db4ba3916d3 [GlobalMerge][NFC] Fix inaccurate comments (#124136) bca6dbd3a241 [X86] Add additional i128 abi test (NFC) df299958e64c [mlir] Fix warnings 1937a36209bc [RISCV] Add @earlyclobber to SiFive custom matrix multiply instruction. (#124060) 4bd0440bd2a6 [RISCV] Add RVVConstraint to SiFive custom matrix multiply instructions. (#124055) cb426b18c2f6 [NFC][DirectX] Clean-up of `DXIL.td` (#124005) 96410edd4748 mark test as unsupported as I investigate test failure on certain environments e1aa1e43decf [WPD]Provide branch weight for checking mode. (#124084) 6fe0fc60341b [CallingConv] Return ArrayRef from AllocateRegBlock() (NFC) (#124120) f61d93ffc456 [Flang] Generate math.acos op for non-precise acos intrinsic calls (#123641) c3b40c7ea215 [X86] Regenerate test checks (NFC) 6206f5444fc0 [AMDGPU] Occupancy w.r.t. workgroup size range is also a range (#123748) 26b61e143b7e [LoopVectorizer] Propagate underlying instruction to the cloned instances of VPPartialReductionRecipes (#123638) 1f0964f81e5a [llvm][Docs] Clarify finding maintainers 25653e558c29 [AVX10.2] Update convert chapter intrinsic and mnemonics names (#123656) 4d3a5309248e [bazel]Fix(2) bazel build past 2e6cc79f816d942ab09d6a310cd925c1da148aa9 (#124118) 99d450e9f516 Revert "[AMDGPU] SIPeepholeSDWA: Disable on existing SDWA instructions (#123942)" 0c71fdd1575b [NVPTX] Fix ctor / dtor lowering when NVPTX target is not enabled (#124116) e28e93550a74 AMDGPU: Make vector_shuffle legal for v2i32 with v_pk_mov_b32 (#123684) 92b839e9c824 [gn] fix mistake in ee99c4d4845db ff55c9bc63dd [llvm][amdgpu] Handle indirect refs to LDS GVs during LDS lowering (#124089) d8eb4ac41d88 [Support] Remove ciso646 include (#123578) 6fdaaafd89d7 [AMDGPU] SIPeepholeSDWA: Disable on existing SDWA instructions (#123942) fb3fa41aee47 MachineRegisterInfo: Use variable for TRI d3d605b7cdee [FileCheck] Use move semantics instead of std::swap. NFC. (#123304) 590e5e20b12f [M68k] Fix llc pass test after 3630d9ef65b30af7e4ca78e668649bbc48b5be66 0c66644270ab [bazel]Fix bazel build past 2e6cc79f816d942ab09d6a310cd925c1da148aa9 (#124112) 90e9895a9373 [X86] Handle BSF/BSR "zero-input pass through" behaviour (#123623) 0e944a30954e [SCFToGPU] Convert scf.parallel+scf.reduce to gpu.all_reduce (#122782) e069518f82bc SCEV: cover a codepath in isImpliedCondBalancedTypes (#123070) 9705500582b9 [libclc] Move nextafter to the CLC library (#12…
1 parent c2f864d commit b5c44bd

File tree

10,470 files changed

+1409733
-451017
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

10,470 files changed

+1409733
-451017
lines changed

README.md

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -60,18 +60,18 @@ In general (with all invocations given from the build directory)
6060

6161

6262
The result of this pipeline should, most simply, be passed to the `rocm-run`
63-
script in `mlir/utils/widgets//rocm-run`, which calls `mlir-cpu-runner` with
63+
script in `mlir/utils/widgets//rocm-run`, which calls `mlir-runner` with
6464
the appropriate flags and infers the pathnames for libraries correctly.
6565

6666
In more detail, the result of the above pipeline can be passed to
67-
`./external/llvm-project/llvm/bin/mlir-cpu-runner` .
67+
`./external/llvm-project/llvm/bin/mlir-runner` .
6868

69-
`mlir-cpu-runner` needs to link the generated host code against libraries that
69+
`mlir-runner` needs to link the generated host code against libraries that
7070
map from MLIR operations to the HIP runtime.
7171
The required command-line arguments (if running from `build/`) are
7272

7373
```sh
74-
./external/llvm-project/llvm/bin/mlir-cpu-runner --shared-libs=./external/llvm-project/llvm/lib/libmlir_rocm_runtime.so,./lib/libconv-validation-wrappers.so,./external/llvm-project/llvm/lib/libmlir_runner_utils.so --entry-point-result=void
74+
./external/llvm-project/llvm/bin/mlir-runner --shared-libs=./external/llvm-project/llvm/lib/libmlir_rocm_runtime.so,./lib/libconv-validation-wrappers.so,./external/llvm-project/llvm/lib/libmlir_runner_utils.so --entry-point-result=void
7575
```
7676

7777
Adding `--debug-only=serialize-to-blob` to the `rocmlir-driver` invocation

external/llvm-project/.azuredevops/rocm-ci.yml

Lines changed: 0 additions & 27 deletions
This file was deleted.

external/llvm-project/.ci/generate-buildkite-pipeline-premerge

Lines changed: 2 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -128,9 +128,8 @@ if [[ "${windows_projects}" != "" ]]; then
128128
limit: 2
129129
timeout_in_minutes: 150
130130
env:
131-
CC: 'cl'
132-
CXX: 'cl'
133-
LD: 'link'
131+
MAX_PARALLEL_COMPILE_JOBS: '16'
132+
MAX_PARALLEL_LINK_JOBS: '4'
134133
commands:
135134
- 'C:\\BuildTools\\Common7\\Tools\\VsDevCmd.bat -arch=amd64 -host_arch=amd64'
136135
- 'bash .ci/monolithic-windows.sh "$(echo ${windows_projects} | tr ' ' ';')" "$(echo ${windows_check_targets})"'

external/llvm-project/.ci/generate_test_report.py

Lines changed: 88 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -19,12 +19,13 @@ def junit_from_xml(xml):
1919

2020
class TestReports(unittest.TestCase):
2121
def test_title_only(self):
22-
self.assertEqual(_generate_report("Foo", []), ("", "success"))
22+
self.assertEqual(_generate_report("Foo", 0, []), ("", "success"))
2323

2424
def test_no_tests_in_testsuite(self):
2525
self.assertEqual(
2626
_generate_report(
2727
"Foo",
28+
1,
2829
[
2930
junit_from_xml(
3031
dedent(
@@ -45,6 +46,7 @@ def test_no_failures(self):
4546
self.assertEqual(
4647
_generate_report(
4748
"Foo",
49+
0,
4850
[
4951
junit_from_xml(
5052
dedent(
@@ -70,10 +72,51 @@ def test_no_failures(self):
7072
),
7173
)
7274

75+
def test_no_failures_build_failed(self):
76+
self.assertEqual(
77+
_generate_report(
78+
"Foo",
79+
1,
80+
[
81+
junit_from_xml(
82+
dedent(
83+
"""\
84+
<?xml version="1.0" encoding="UTF-8"?>
85+
<testsuites time="0.00">
86+
<testsuite name="Passed" tests="1" failures="0" skipped="0" time="0.00">
87+
<testcase classname="Bar/test_1" name="test_1" time="0.00"/>
88+
</testsuite>
89+
</testsuites>"""
90+
)
91+
)
92+
],
93+
buildkite_info={
94+
"BUILDKITE_ORGANIZATION_SLUG": "organization_slug",
95+
"BUILDKITE_PIPELINE_SLUG": "pipeline_slug",
96+
"BUILDKITE_BUILD_NUMBER": "build_number",
97+
"BUILDKITE_JOB_ID": "job_id",
98+
},
99+
),
100+
(
101+
dedent(
102+
"""\
103+
# Foo
104+
105+
* 1 test passed
106+
107+
All tests passed but another part of the build **failed**.
108+
109+
[Download](https://buildkite.com/organizations/organization_slug/pipelines/pipeline_slug/builds/build_number/jobs/job_id/download.txt) the build's log file to see the details."""
110+
),
111+
"error",
112+
),
113+
)
114+
73115
def test_report_single_file_single_testsuite(self):
74116
self.assertEqual(
75117
_generate_report(
76118
"Foo",
119+
1,
77120
[
78121
junit_from_xml(
79122
dedent(
@@ -166,6 +209,7 @@ def test_report_single_file_multiple_testsuites(self):
166209
self.assertEqual(
167210
_generate_report(
168211
"ABC and DEF",
212+
1,
169213
[
170214
junit_from_xml(
171215
dedent(
@@ -198,6 +242,7 @@ def test_report_multiple_files_multiple_testsuites(self):
198242
self.assertEqual(
199243
_generate_report(
200244
"ABC and DEF",
245+
1,
201246
[
202247
junit_from_xml(
203248
dedent(
@@ -238,6 +283,7 @@ def test_report_dont_list_failures(self):
238283
self.assertEqual(
239284
_generate_report(
240285
"Foo",
286+
1,
241287
[
242288
junit_from_xml(
243289
dedent(
@@ -272,6 +318,7 @@ def test_report_dont_list_failures_link_to_log(self):
272318
self.assertEqual(
273319
_generate_report(
274320
"Foo",
321+
1,
275322
[
276323
junit_from_xml(
277324
dedent(
@@ -312,6 +359,7 @@ def test_report_size_limit(self):
312359
self.assertEqual(
313360
_generate_report(
314361
"Foo",
362+
1,
315363
[
316364
junit_from_xml(
317365
dedent(
@@ -351,12 +399,18 @@ def test_report_size_limit(self):
351399
# and output will not be.
352400
def _generate_report(
353401
title,
402+
return_code,
354403
junit_objects,
355404
size_limit=1024 * 1024,
356405
list_failures=True,
357406
buildkite_info=None,
358407
):
359408
if not junit_objects:
409+
# Note that we do not post an empty report, therefore we can ignore a
410+
# non-zero return code in situations like this.
411+
#
412+
# If we were going to post a report, then yes, it would be misleading
413+
# to say we succeeded when the final return code was non-zero.
360414
return ("", "success")
361415

362416
failures = {}
@@ -385,7 +439,11 @@ def _generate_report(
385439
if not tests_run:
386440
return ("", None)
387441

388-
style = "error" if tests_failed else "success"
442+
style = "success"
443+
# Either tests failed, or all tests passed but something failed to build.
444+
if tests_failed or return_code != 0:
445+
style = "error"
446+
389447
report = [f"# {title}", ""]
390448

391449
tests_passed = tests_run - tests_skipped - tests_failed
@@ -400,17 +458,17 @@ def plural(num_tests):
400458
if tests_failed:
401459
report.append(f"* {tests_failed} {plural(tests_failed)} failed")
402460

403-
if not list_failures:
404-
if buildkite_info is not None:
405-
log_url = (
406-
"https://buildkite.com/organizations/{BUILDKITE_ORGANIZATION_SLUG}/"
407-
"pipelines/{BUILDKITE_PIPELINE_SLUG}/builds/{BUILDKITE_BUILD_NUMBER}/"
408-
"jobs/{BUILDKITE_JOB_ID}/download.txt".format(**buildkite_info)
409-
)
410-
download_text = f"[Download]({log_url})"
411-
else:
412-
download_text = "Download"
461+
if buildkite_info is not None:
462+
log_url = (
463+
"https://buildkite.com/organizations/{BUILDKITE_ORGANIZATION_SLUG}/"
464+
"pipelines/{BUILDKITE_PIPELINE_SLUG}/builds/{BUILDKITE_BUILD_NUMBER}/"
465+
"jobs/{BUILDKITE_JOB_ID}/download.txt".format(**buildkite_info)
466+
)
467+
download_text = f"[Download]({log_url})"
468+
else:
469+
download_text = "Download"
413470

471+
if not list_failures:
414472
report.extend(
415473
[
416474
"",
@@ -435,11 +493,23 @@ def plural(num_tests):
435493
"</details>",
436494
]
437495
)
496+
elif return_code != 0:
497+
# No tests failed but the build was in a failed state. Bring this to the user's
498+
# attention.
499+
report.extend(
500+
[
501+
"",
502+
"All tests passed but another part of the build **failed**.",
503+
"",
504+
f"{download_text} the build's log file to see the details.",
505+
]
506+
)
438507

439508
report = "\n".join(report)
440509
if len(report.encode("utf-8")) > size_limit:
441510
return _generate_report(
442511
title,
512+
return_code,
443513
junit_objects,
444514
size_limit,
445515
list_failures=False,
@@ -449,9 +519,10 @@ def plural(num_tests):
449519
return report, style
450520

451521

452-
def generate_report(title, junit_files, buildkite_info):
522+
def generate_report(title, return_code, junit_files, buildkite_info):
453523
return _generate_report(
454524
title,
525+
return_code,
455526
[JUnitXml.fromfile(p) for p in junit_files],
456527
buildkite_info=buildkite_info,
457528
)
@@ -463,6 +534,7 @@ def generate_report(title, junit_files, buildkite_info):
463534
"title", help="Title of the test report, without Markdown formatting."
464535
)
465536
parser.add_argument("context", help="Annotation context to write to.")
537+
parser.add_argument("return_code", help="The build's return code.", type=int)
466538
parser.add_argument("junit_files", help="Paths to JUnit report files.", nargs="*")
467539
args = parser.parse_args()
468540

@@ -477,7 +549,9 @@ def generate_report(title, junit_files, buildkite_info):
477549
if len(buildkite_info) != len(env_var_names):
478550
buildkite_info = None
479551

480-
report, style = generate_report(args.title, args.junit_files, buildkite_info)
552+
report, style = generate_report(
553+
args.title, args.return_code, args.junit_files, buildkite_info
554+
)
481555

482556
if report:
483557
p = subprocess.Popen(

0 commit comments

Comments
 (0)